22 PAGE BLENDER PAGE BLENDER SPECIAL Practical inspiration inspiration for the the 3D community www.3DArtistonline.com
107 10 7
16GB
YOUR QUESTIONS ANSWERED NSWERED
BLENDER
OF FREE VIDEOS
ISSUES
& MORE
SOLVED ENHANCED HAIR & FUR MASTER PBR PBR TEXTURES PLAY WITH LIGHTING
HOUDINI
TRICKS
Discover new techniques with the best Discover artists around and the SideFX team
CREATURE
VFX EXPLORED Framestore takes you behind the scenes of Guardians Of The Galaxy Vol.. 2
MAYA CONCEPT
MODELLING
INSIDE THE THE AMAZING WORLD WOR LD OF MEDIAMONKS ISSUE 107
Digital Edition GreatDigitalMags.com GreatDigital Mags.com
CREATE STUNNING GAME PROPS FOR UNREAL UNREAL
Blender problems solved Page 40
Nazar Noschenko bit.ly/2oGxsMj Software Blender, Marvelous
Designer, Photoshop
3
Future Publishing Ltd Richmond House, 33 Richmond Hill Bournemouth, Dorset, BH2 6EZ +44 (0) 1202 586200 Web: www.3dartistonline.com www.greatdigitalmags.com www.futureplc.com
Editorial
Editor Steve Holmes
[email protected] 01202 586248
Features Editor Carrie Mok Production Editor Rachel Terzian Art Editor Newton Ribeiro de Oliveira Editor in Chief Amy Hennessey Senior Art Editor Will Shum Photographer James Sheppard Contributors Orestis Bastounis, Gurmukh Bhasin, James Brady, Paul Champion, Ian Failes, Ara Kermanikian, Daniel Pedersen, Davide Pellino, Mike Rutter Advertising Digital or printed media packs are available on request. Commercial Sales Director Clare Dove Advertising Manager Mike Pyatt 01225 687538
[email protected] Account Director George Lucas
[email protected] Advertising Sales Executive Chris Mitchell
[email protected] International 3D Artist is available for licensing. Contact the International department to discuss partnership opportunities. Head of International Licensing Matt Ellis
[email protected] Subscriptions For all subscription enquiries:
[email protected] 0844 249 0472 Overseas +44 (0) 1795 592951 www.imaginesubs.co.uk Head of Subscriptions Sharon Todd FileSilo.co.uk Support: Support: Circulation Circulation Director Darren Pearce 01202 586200
Your Blender problems solved Page 40
Production Head of Production US & UK Mark Constance Production Project Manager Clare Scott Advertising Production Manager Joanne Crosby Digital Editions Controller Jason Hudson Production Manager Frances Twentyman Management Finance & Operations Director Marco Peroni Creative Director Aaron Asadi Art & Design Director Ross Andrews Printing & Distribution William Gibbons & Sons Ltd, 26 Planetary Road, Willenhall, West Midlands, WV13 3XT Distributed in the UK, Eire & the Rest of the World by Marketforce, 5 Churchill Place, Canar y Wharf, London, E14 5HU 0203 787 9060 www.marketforce.co.uk
I
t struck me the other day that one of the best things about computer graphics is that artists are really willing to help each other out creatively, more often than not. There’s no such thing as a stupid question in our field, as the nature of the multitude of tools available to us means that there’s rarely just one ‘correct’ way to do things – there are often several. This rings true whether you’re a 20-year veteran or a complete beginner, really. There’s always something new to learn, and a great way to do so is by absorbing advice from the artists around you. One community
that certainly feels like it’s at the forefront of collaborative learning is Blender’s. With free sharing of information at its core, it’s no wonder that the scene is driven by the ebb and flow of questions and answers, so we thought we’d do the right thing and highlight this in the magazine this month. Head over to our phenomenal cover feature to find solutions to some of your most interesting queries – these questions were asked by you and have been answered by some of the most knowledgeable Blender gurus out there. Also, don’t miss our incredible behind-the-scenes access on Guardians Of The Galaxy Vol. 2, a magical collection of Houdini techniques and much more!
Steve Holmes, Editor
Sign up, share your art and chat to other artists at www.3dartistonline.com Get in touch...
4
[email protected]
@3DArtist
Facebook.com/3DArtistMagazine
Distributed in Australia by Gordon & Gotch Australia Pty Ltd, 26 Rodborough Road, Frenchs Forest, New South Wales 2086 + 61 2 9972 8800 www.gordongotch.com.au Disclaimer The publisher cannot accept responsibility for any unsolicited material lost or damaged in the post. All text and layout is the copyright of Future Publishing Ltd. to ensure all information is correct at time of print, prices and availability may companies mentioned herein. If you submit material to Future Publishing via post, email, social network or any other means, you automatically grant Future Publishing an irrevocable, perpetual, royalty-free licence to use the material across its entire portfolio, in print, online and digital, and to deliver the material to existing and future clients, including but not limited to international licensees for reproduction in international, licensed editions of Future Publishing products. Any material you submit is sent at your risk and, although every care is taken, neither Future Publishing nor its employees, agents or subcontractors shall be liable for the loss or damage.
© 2017 Future Publishing Ltd ISSN 1759-9636
This issue’s team of pro artists… …
NAZAR Z NOSCHENKO N SCHENK
DANIEL PEDERSEN
DAVIDE PELLINO
artstation.com/artist/nazarnoschenko Nazar is responsible for our beautiful cover image this month, and over on p40 he’s joined by some incredible talent to help solve some of your most pressing problems in Blender. 3DArtist username Nazar username Nazar
oceanthebard.artstation.com Continuing our Blender special is Daniel, who reveals his workflow for creating an awesome character using some of the latest functionality in version 2.78, plus some handy posing tips. 3DArtist username Oceanthebard username Oceanthebard
behance.net/DavidePellino Davide wraps up our blockbuster Blender content with an amazing tutorial looking at how to add cartoon flair to your renders. You can find his unmissable guide over on p56. 3DArtist username Capodoglio username Capodoglio
GURMUKH BHASIN
JAMES BRADY
MIKE RUTTER
gurmukhbhasin.com Quite a few of you will have heard of Gurmukh before now due to his amazing concept art – particularly his sci-fi work. On p64 he’s jumped on board to show off his modelling workflow for a pirate ship. 3DArtist username Gurmukh username Gurmukh
artstation.com/artist/jamesbrady We can’t get enough of Unreal Engine at the moment, so games professional James has returned to the magazine with a masterclass in asset creation and PBR texturing on p72. 3DArtist username Sabertooth_00 username Sabertooth_00
thedigimonsters.com Mike joins us from The Digi Monsters, an exciting new game art and VFX company featuring some really talented artists. Over on p80, he deconstructs how to create a futuristic motion graphics ident. 3DArtist username username DigiMonsterBoss
ARA KERMANIKIAN
ORESTIS BASTOUNIS ASTOUNIS
IAN FAILES
kermaco.com
twitter.com/mrbastounis twitter.com/mr twitter. com/mr astounis After we gave him a little rest last month, before he knew it another workstation was knocking on Orestis’s door. This time it’s from Scan and features the recent NVIDIA Quadro P4000. 3DArtist username n/a username n/a
vfxblog.com
Drawing on many years of experience in computer graphics, Ara has taken a critical look at the Wacom MobileStudio Pro 16 on p86 – a tablet with more or less workstation architecture. 3DArtist username kermaco username kermaco
6
By the time you read this, Guardians Of The Galaxy Vol. 2 will 2 will be absolutely huge. On p22, Ian has spoken to Framestore and Marvel about bringing the latest MCU adventure to life. 3DArtist username n/a username n/a
What’s in the magazine and where
News, reviews & features 12 The Gallery A hand-picked collection of incredible artwork to inspire you
22 Galaxy Quest Ian Failes journeys deep into space to discover the new heights of VFX work on Guardians Of The Galaxy Vol. 2
With the Comb brush I adjust the hairs into the desired shape Nazar Noschenko and a selection of pros solve your pressing Blender problems Page 41
28 43 Houdini Tricks, Tips and Techniques We delve into the magic of Houdini with essential techniques from pro users and the SideFX team
34 Inside the MediaMonks Monastery The Dutch media production giant takes us behind the scenes of creating innovative 3D experiences
40 Your Blender Issues Solved Our Blender experts answer your questions from social media
84 Subscribe Today! Save money and never miss an issue by snapping up a subscription
86 Review: Wacom MobileStudio Pro 16 Ara Kermanikian puts the high-end tablet through its paces 40
88 Review: Scan 3XS Classic 3D Will this entry-level workstation impress Orestis Bastounis?
98 Technique Focus: Scandinavian bedroom
43 Houdini Tricks, Tips and Techniques
Create a motion graphics ident
Artem Glazov discusses his lighting techniques for arch vis with Corona
28
Save up to 20% Model & texture an asset for Unreal Engine 4 8
72
SUBSCRIBE TODAY
80
Turn to page 84 for details ls
Render a pirate ship
64
34
Inside the MediaMonks Monastery
The Pipeline 48 Step By Step: Create a character with Blender 2.78 Daniel Pedersen gives us his stepby-step workflow for modelling and rigging a character
For the groom, in particular, particu ar, we really tried for that animalistic feel
56 Step By Step: Achieve a fantasy illustration style in Blender Davide Pellino reveals how to create a 3D render in a cartoon style
Christopher Townsend reveals the secrets behind the visual effects of Guardians Of The Galaxy Vol.2 Page 25
64 Step By Step: Render a pirate ship Gurmukh Basin sets sail with Maya and discusses manipulating UVs and library parts
72 Step By Step: Model & texture an asset for Unreal Engine 4 James Brady explains his triple-A industry-standard workflow for texturing a game asset
80 Pipeline Techniques: Create a motion graphics ident 22
Achieve a fantasy illustration style in Blender
Mike Rutter of The Digi Monsters explains how he uses Maya to create a channel ident
The Hub 92 Community News Director Alexander Heringer brings us in the loop with the tools that helped create film festival favourite, Circle
94 Industry News Nuke 11 is previewed at NAB and Arnold replaces mental ray in 3ds Max
96 Readers’ gallery The very best images of the month from our online community
Visit the 3D Artist online shop at
Create a character with Blender 2.78
48
56
for back issues, books and merchandise
9
Free with your magazine Blender videos from Pluralsight
3 premium CGAxis HDRIs
25 textures from 3DTotal.com
Learn to animate a character for games
A collection of images to use in your work
Download plenty of high-quality textures
Plus, all of this is yours too… mountains of high-resolution screenshots to accompany our Blender character tutorial scene files to help you develop your own Unreal Engine game assets cartoon flair to your renders accompany our tutorials
Log in to www.filesilo.co.uk/3DArtist Register to get instant access to this pack of must-have creative resources, how-to videos and tutorial assets
Free for digital readers too! Read on your tablet, download on your computer
The hom e of great dow nloa ds –exclusive to your favo uri te m ag azines from Fu ture!
Secure and safe online access, from anywhere Free access for every reader, print and digital Download only the files you want, when you want All your gifts, from all your issues, in one place
Get started Everything you need to know about accessing your FileSilo account
01
Follow the instructions on screen to create an account with our secure FileSilo system. Log in and unlock the issue by answering a simple question about the magazine.
Unlock every issue
gifts from more than 40 issues Access our entire library of resources with a money-saving
02
You can access FileSilo on any computer, tablet or smartphone device using any popular browser. However, we recommend that you use a computer to download content, as you may not be able to download files to other devices.
Over 50 hours of video guides
More than 1,000 textures
Hundreds of 3D models
The very best walkthroughs around
Brought to you by quality vendors
Vehicles, foliage, furniture… it's all there
Head to page 84 to subscribe now 03
If you have any problems with accessing content on FileSilo take a look at the FAQs online or email our team at the address below
[email protected]
r
Already a print subscriber? ID – the eight-digit alphanumeric code that is printed above your address details on the mailing label of your subscription copies. It can also be found on any renewal letters.
t
r
i
More added every issue
Have an image you feel passionate about? Get your artwork featured in these pages
Create your gallery today at www.3dartistonline.com
12
This is an image for all parents. Remember when you came home from a night out and your little boy has left the house in an absolute state? He might have had a life-defining moment. Best of luck dude, Mars 2030 will be here soon. This project was done over a few weeks of spare time and is an interpretation of a line drawing by Andy Estrada
Zeno Pelgrims zenopelgrims.com Zeno is a lighting artist for animation, probably waiting for uncompleted render buckets!
Software Maya, ZBrush, Mari, Arnold, Nuke
Work in progress…
Zeno Pelgrims, Mars 2030, 2017
13
I’ve always been fascinated by mythical beings from folklore, forest gods and monsters, and I wanted to create a benign forest creature that evokes a sense of deep history. The project was created for a tutorial series on the Blender Cloud (cloud.blender.org). All the original Blender files are published under a Creative Commons licence, as well as a full making-of video with commentary Andy Goralczyk, Waking the Forest, 2017
Andy Goralczyk artificial3d.com Andy is currently working at the Blender Institute in Amsterdam on the Agent 327 open-source feature film Software Blender, Krita
Work in progress…
14
I’ve always found human characters to be diffcult to create, especially human faces. This work was my personal challenge and an attempt to achieve a more or less convincing look. I was thinking about who this girl might be and how to express it through visual details. It was also a good exercise to improve my understanding of V-Ray hair material Liudmila Kirdiashkina, Portrait of a Girl, 2017
Liudmila Kirdiashkina artstation.com/artist/liu_k Liudmila is a freelance 3D artist. Currently she mainly works on hard-surface objects Software ZBrush, Maya, Mari, V-Ray for Maya
Work in progress…
15
Marcos de M. Sampaio notan.com.br Based in São Paulo, Marcos is a 3D generalist and is a founder of Notan Studio
Software ZBrush, Modo, Photoshop
Work in progress…
I’ve been wanting to create a character in a kind of Zootopia style for a while, and I also wanted it to tell a funny story too. So recently I had this idea of a happy and naive hedgehog who’s excited about bringing more love to the world, but he’s also totally unaware of his spines Marcos de Moraes Sampaio, Free Hugs, 2017 16
This robot is a personal work made for fun and practice. I did this robot to try some new designs, colours, shapes and to practise procedural textures in KeyShot Rafael Gomes Amarante, 23-v2, 2017
Rafael G. Amarante rafaelamarante.com Rafael is a freelance concept designer and 3D artist for videogame cinematics Software Cinema 4D, KeyShot, Photoshop
Work in progress…
17
In depth
Arthur Gatineau arthurgatineau.com Arthur is a freelance 3D artist specialised in modelling, texturing and shading Software 3ds Max, ZBrush, Mari, V-Ray
Work in progress…
I decided to do this project to improve my modelling, texturing and shading skills. As you may have already seen in my portfolio, I like to create dirty and dusty textures and materials instead of brand-new shiny ones as I find them more interesting to develop Arthur Gatineau, The Man Who Can’t Breathe, 2017
18
19
I came across images of a character from the film Insidious and I immediately saw the potential of doing it. [There are] interesting textures on his face with a lot of imperfections and veins, and some complex materials such as the dirty plastic mask or the cloth. My goal was to make this image as fast as possible in my free time Arthur Gatineau,
The Man Who Can’t Breathe, 2017
MODELLING THE SHIRT RIGHT Thanks to Marvelous
Designer I was able to quickly create the cloth piece by doing very simple T-shirt patterns. As I always consider my MD meshes as bases, I imported it in ZBrush to keep refining the shapes and add details.
TEXTURING THE FACE RIGHT The skin texture was so
much fun to develop. I ended up using a tweaked marble texture as a base to create the veins. No rush: layer after layer it’s taking shape. 20
THE GALLERY
LIGHTING THE FINAL SHOT ABOVE I wanted to create a
very dramatic mood. The lighting setup is very simple: an HDRI with a low intensity as a fill light, a strong light on top of the character as a key light, and a rim light.
TESTING THE MATERIALS LEFT I constantly go back
and forth between different lighting conditions during the look development. You really need to make sure the values you’re putting in for the specular or roughness are holding up in all situations. 21
XXXXXX QUEST GALAXY XXXXXX XXXXXXXXX
GALAXY QUEST The VFX team behind behind Guardians Of The Galaxy Vol. 2 tells us how it reached d astronomically ly new heights – and shot counts – in CGI GI
O ) d e d u l c n i y a r n u S ( . m a e T O R E 4 M S e l b s b o u i d H u e t S h t l d e n v r a a , A M S 7 E 1 , 0 2 A S A © t N h : g e i r g y a p i o m c d s n e u g o a r g m k i l c l a A B
22 2
n 18 February 2017, 7, director James Gunn noted on social media ed a that t at he e had ad just ust finalised a visual e effects ects shot s ot for o r Guardians uar ans t iteration. terat on. But ut that t at Of The Galaxy Vol. 2 in its 749th record did not last long. Later ater in March, Gunn followed up with advice that hat a final VFX shot shot had had reached 859 versions. It’s a stunning indication on of o the t e work wor that t at has as gone into this sequel to the he 2014 hit, which which was was considered a risky venture re in the Marvel Cinematic Universe. All in, there are are more more than than 2,000 visual effects shots ts inn the t e film, m, including ncud ngaa host of computer-generated ated characters c aracters such suc as as Rocket, Baby Groot and a particularly mean tentacled alien known ass the Abilisk, plus several complicated 3D environments nments and space scenes. finds out from rom vvisual sua eeffects ects 3D Artist supervisor Christopher Townsend ownsend – already a ready aa veteran of several Marvel vel adventures – and from Framestore and Weta Digital, just two of the studios behind the work, r , how ow the t e characters c aracters and worlds of Guardians Of The he Galaxy aaxy Vol. o . 2 were brought to the screen..
Guardians Of The Galaxy Vol.2 is the sequel to the 2014 hit based on the Marvel Comics superhero team
23
XXXXXX QUEST GALAXY XXXXXX XXXXXXXXX
THE RETURN OF ROCKET RACCOON The genetically engineered Rocket, voiced by Bradley Cooper and acted on set by Sean Gunn, was a crowd favourite in the first film. His wise-cracking comments, stunningly photoreal fur and clothing and his large but fragile presence on the screen were things Townsend wanted to replicate. But he also wanted to take Rocket to new places. “I wanted to go in more of an animalistic way, and to make Rocket slightly more of a believable character so you appreciate that, when he talks, it’s coming out of him,” the visual effects supervisor told 3D Artist. “We did a lot of work with Bradley in getting him to be more physical when he performed and relate that more to what Rocket was doing.” For shooting, Sean Gunn (James’ brother) played Rocket, often dressed in a green-screen onesie. A furred stuffie of Rocket and one in costume were also utilised during production for lighting reference. Townsend pushed on set for Sean’s performance to be the guide for Rocket’s animators, not just for eyelines or to have someone in the scene for the other actors. “He’s an incredibly flexible guy, Sean,” outlines Townsend. “He was able to crouch down and get to the right eyeline. So we would do takes with the full cast on set crouched down at Rocket’s
ROCKETING HAIR COUNTS
HEAD
WHISKERS
LIMBS
TAIL
400,323
156
161,586
66,881
hairs
hairs
hairs TOTAL: 24
hairs
628,946 hairs
(That’s approximately 50GB of data from work-in-progress grooms and textures)
Rocket returns in Guardians Of The Galaxy Vol. 2, along with Baby Groot
eyeline height and performing and saying the lines. And then we would shoot, and then we’d take Sean out and we’d shoot it again without him, with just him saying it off camera, but we would use the visuals of that first performance as our guide.” From a CG perspective, Framestore took the reins of Rocket once again in developing the character’s skeletal structure, muscle system and groom. “For the groom, in particular,” says Townsend, “we really tried for that animalistic feel, and there are moments in the film where he reverts back to being more animal, so it just fit the moments.” Perhaps the most major change with Rocket, however, was that a total of four visual effects studios – Framestore, Weta Digital, Method Studios and Trixter – would each contribute final Rocket shots. That was simply based on the scope of the work required, but at times, notes Townsend, “we are literally inter-cutting from one studio’s Rocket to another.”
BRINGING UP BABY GROOT Equally engaging from the first film was Groot, voiced ever so sparingly by Vin Diesel. His rebirthed character in Vol. 2 is of course a smaller and cuter version, but still completely
synthetic. Interestingly, from an animation perspective, the visual effects teams set out not to ‘overdo’ Baby Groot’s cuteness. “James Gunn would often talk about keeping him very Buster Keaton, just playing him very flat,” recalls Townsend. “He’s this little toddler of a character, who is oblivious to what’s going on around him at times. He’s joyful and happy about so many things. He’ll see this big, bright explosion and, rather than being terrified, he’ll be really happy about it. “Groot’s often just in the background,” adds Townsend, “and we were very careful to keep him alive at all times. There were moments where some massive thing is happening in the foreground and if you just look over at Groot he’s in the background staring nonchalantly off the screen, or he’s looking down at his feet or twiddling with the leaves on his arm or something. So he’s doing these things totally not in the moment because that’s one of the aspects of a child that we were channelling.” In order to find that right Baby Groot design and feel, Townsend did embark on some motion capture tests with a toddler. “It was fantastic and it was cute but it looked exactly like a little child dressed up as Baby Groot. We didn’t actually want that – we wanted to always maintain a slightly alien feel to him. So we ended up doing things like broadening his waist to make him more sort of linear. We flattened the face a bit more, we reduced any cuteness in his cute little butt as much as we could.”
REALISING ROCKET Having worked on Rocket previously for Guardians Of The Galaxy , Framestore “knew the character inside and out”, as animation supervisor Arslan Elver explains in this outline of how the raccoon model was rebuilt and animated from scratch.
1
Rocket is slightly thinner and slightly more upright for this film. Framestore also worked on more realistic mouth shapes – they had learnt from the previous outing that a character with a long muzzle presented issues with lip sync.
2
The studio’s hair toolset, fcHairFilters, working in conjunction with Arnold for rendering, had been improved upon since the first film and was used to groom Rocket this time around as well.
3
Rocket’s performance was entirely keyframed, but ultimately came from several direct sources (in addition to studying animal references). One source was the voice acting by Bradley Cooper who sometimes wore a head-mounted camera; animators looked at the actor’s mouth and jaw for lip sync reference.
4
Another source was Sean Gunn on set, in which the actor would sometimes kneel down to Rocket’s height to provide appropriate eyelines either for final shots or takes.
5
Yet another source of reference were pieces of footage shot by the animators of themselves. “We managed to keep a consistent performance from Rocket when we used ourselves as reference, because the people who shot the reference were people who knew the character quite well,” says Elver.
s o i d u t S l e v r a M 7 1 0 2 © t h g i r y p o c s e g a m i l l A
25
XXXXXXQUEST GALAXY XXXXXX XXXXXXXXX
ALIEN EFFECTS Find out how Framestore went about the visual effects for the tentacled Abilisk
s o i d u t S l e v r a M 7 1 0 2 © t h g i r y p o c s e g a m i l l A
26
In addition to the central CG characters of Rocket and Groot, there are many other CG creatures seen in Guardians Of The Galaxy Vol. 2. Among them is the Abilisk, a hellish tentacled alien the heroes encounter early on in the film. Framestore handled the visual effects, taking reference from a combination of real-life animals. Christopher Townsend says, “We wanted it somewhat like an octopus but to also have the lack of grace of an elephant seal on land. And all that time we wanted to create something which had the colour of a little pinkie mouse, like a newborn mouse, and make it really translucent, with all these gross hairs on it.” For the Abilisk’s tentacles, Framestore had to balance giving the limbs enough weight and jiggle in animation, versus what could be achieved via simulation in the creature FX department. “In the end,” says Framestore’s Arslan Elver, “we gave the tentacles enough velocity to hit the floor but we didn’t have any jiggle in the animation. All the jiggle in the bottom part of the tentacle was done in CFX.” The Abilisk proved to be representative of a common theme in Vol. 2 – that of juxtaposition, as Townsend explains. “It’s got that pinkie mouse colour but it’s the size of a massive monster that is incredibly dangerous, and then spews out My Little Pony -like pink and blue and rainbow-coloured bubbles that then explode in these massive dangerous ways.”
Look closely and you’ll see the adorable tiny Baby Groot on the shoulder of Yondu
Peter Quill/ Star-Lord (Chris Pratt) readies for battle
CRACKING A FRACTAL CRISIS
Baby Groot gets ready to dance
Behind the scenes with our heroes inside the spacecraft Milano
As well as an array of digital characters, Guardians Of The Galaxy Vol. 2 features some incredibly distinctive and fantastical new worlds. One of those was for an area dubbed the ‘planet hollow’ inside the planet Ego, which Weta Digital crafted as a hugely complex digital asset for the film’s final act. The inspiration for the look and feel of the ten kilometre-wide planet hollow were fractals and Mandelbrot designs. The hollow was a large asset built to work as a single set with a few moving parts. “This meant that every location is present in almost every shot,” explains Weta Digital visual effects supervisor Guy Williams. “This resulted in a large number of assets needing to be built to a high spec. Add to that the challenge of building geometry that was almost unbuildable, thanks to the large Mandelbulb tower structures and the fractal walls.” In fact, the forms seen in the planet hollow were so complex that Weta Digital’s original attempt with the program Mandelbulb, which production had used to look dev some of the scenes, could not generate enough resolution. “Since most of the environment is fractaline,” says fellow Weta Digital visual effects supervisor, Kevin Smith, “it has repeating structures at all scales, and Mandelbulb was running out of memory and crashing on our high-memory renderwall slots before we got anywhere close to the resolution we would need for the shots.” There were also issues in constructing fractals that, even though implicit functions could render very fast, were not as art direction-friendly as they needed to be. “We went to R&D with our problem and came up with a way to explicitly generate some of the same structures that we were seeing in the reference art,” details Smith. “It was great until we realised that the sphere-packing methods that were being used to generate the geo were still kind of implicit. “We had to run these massive Boolean operations on the renderwall and again we were back to processes that took hundreds of gigabytes to run and crashed as often as not, and still weren’t giving us quite the resolution that we wanted,” adds Smith. “The geometry was also extremely messy UV-wise and was going to be very time-consuming to retopologise for texturing. We did use the tools that R&D developed to generate generic background shapes and helped geo for the models department.” But a solution was found when one of Weta Digital’s modellers realised that, although Mandelbulb could not write out geometry very fast (but could render geometry really fast), places in the parameter space that had
The interior of Ego was generally rock and rock-like structures, so shading was generally done procedurally structures that were appealing for on screen could be rendered out like turntable passes at 8K. “We took those renders,” says Smith, “and fed them to our photogrammetry software and bam, we had high-res bits and pieces that modellers could use to build the hero geometry that we needed. Luckily for us the interior of Ego was generally rock and rock-like structures, so shading was generally done procedurally with a lot of hand-painted helper passes.” Extra detail came from a couple dozen tiny fractal pieces that were called ‘plaque’. “The plaque was about six families of coral-like dressing that we derived from some simple fractal forms,” says Williams. “Each family had up to 20 variants. We then ‘grew’ this dressing all over the towers and other Mandelbulb shapes in the hollow. This created an amazing sense of age and texture and resulted in some huge poly counts.” Artists used Weta Digital’s procedural geometry painting software Genesis to spray hundreds of millions of the plaque dressing as instances, resulting in upwards of 600 billion polygons in some scenes. Final rendering of the rather surreal – yet believable – imagery was carried out in the studio’s proprietary path tracer, Manuka.
GUARDIANS OF THE VF X Like so many of Marvel’s recent films in its Cinematic Universe, the visual effects for Guardians Of The Galaxy Vol. 2 were realised by an enormous group of artists around the world. The VFX studios involved included Weta Digital, Framestore, Method Studios, Trixter, Animal Logic, Scanline VFX, Lola VFX, Luma Pictures, Cantina Creative, Technicolor VFX and an in-house team. Christopher Townsend, in contemplating just how the complicated worklo ad gets finished, gives a lot of credit to James Gunn for calling out the efforts of the VFX crew, including in the director’s social media tweets about shot iterations. “I think him just saying that he acknowledges that there were so many people that go into these individual shots is so great. We have nearly two-and-a-half thousand shots, and there’s a lot of people working for a long time creating them.” 27
43 33 HOUDINI TRICKS, IC , TIPS AND TECHNIQUES I
CONTRIBUTORS Gustavo Åhlén gustavoahlen.com Founder, Svelthe
Hermes Crespo zerovfx.com FX artist, Zero VFX
Ian Farnsworth imageworks.com FX lead, Sony Pictures Imageworks
Upgrade your knowledge and speed up your workflow with this magical insight from expert Houdini users
A
s any aspiring or professional VFX and games artist will know, there’s often a new technique or workflow hack out there to discover that will increase your productivity and make a big difference to how you work on a day-to-day basis. Generally, these nuggets of wisdom are discovered through the necessity of meeting production deadlines, but here, we’ve collected 12 industry experts and invited them to share their hard-earned advice from years of experience in their field. Over the next few pages you’ll find a trove of Houdini knowledge guaranteed to enhance your workflow, covering a broad assortment of topics including modelling, lighting, dynamics, fluids, hair and fur,
28
general workflow and more. There is something for everyone regardless of your experience or specialisation, courtesy of professional artists working at Criterion Games, DICE, MPC, Red Ring Entertainment, SideFX, Sony, Svelthe and Zero VFX. Accompanying the tips, Red Ring Entertainment also explains some of the benefits it discovered while working with Houdini 16 to create a fearsome werewolf for the Houdini 16 Amarok launch video in partnership with SideFX. Criterion Games is working on a recently announced triple-A games franchise and will be looking for technical artists with Houdini knowledge. Keep an eye on their careers section at www.criteriongames.com.
David Hipp moving-picture.com Senior FX TD, MPC
Björn Henriksson dice.se Technical artist, DICE
Joshua Matthews joshua-matthews.com Artist & intern, SideFX
Joan Panis moving-picture.com Head of FX, MPC
NicholasPliatsikas criteriongames.com Senior technical artist/pipeline lead, Criterion Games
Kalin Stoyanov redringent.com Co-founder/lead animator, Red Ring Entertainment
Charles Trippe zerovfx.com FX lead at Zero VFX
Fianna Wong sidefx.com Technical marketing lead, SideFX
Tomas Zaveckas zerovfx.com FX lead, Zero VFX
CONTROL FLUID BOUNCES
01 Focusing on the FLIP Tank shelf tool, use the Axis option to control fluid boundaries. This feature is really important in order to effectively control which axes fluids will and won’t bounce on. Uncheck +X and +Y to avoid boundaries over these axes, so the fluid on the shelf will be cut out.
Gustavo Åhlén
02 Attribute blur mesh point positions, then CHEAP CURVATURE MAPS
subtract the original position from the blurred to compute the delta (difference) with an Attribute Wrangle. In the same wrangle use a dot product of the normals with the delta to compute the concavity and convexity values. Values are -1 concave, 0 flat and 1 convex. Remap values as required and normalise delta and normal values in the dot product.
Nicholas Pliatsikas
02 control time, downsample, Depth Falloff and other parameters.
11
03 Unfortunately, in previous versions of
Gustavo Åhlén
Houdini the Wavelet Field Generation is broken. If you are going to add Wavelet Turbulence into your pyro sims, you have to add the Wavelet Fields manually in order for it to work. You can break most of what you need for this out of the upres solver.
07 Grabbing a bunch of nodes and
You can bend your smoke trail work slightly to one side with a volume VOP and a curve. Give it a rest position and a new position as well as a new volume with new bounds to put stuff in the old volume. Look up the position in the volume b ased on the deforming curve to offset the volume.
ADD WAVELET TURBULENCE
David Hipp & Ian Farnsworth
AUTO-WIRING NODES
drag-and-dropping them onto Merge, Boolean, Copy SOPs and others, you can blindly hook them up. Change the input’s order with Shift+R to cycle through until you get what you want. Remember to drop the outputs into the first input of the node.
04 Generate points in a Voronoi pattern
Fianna Wong
using the Voronoi node in a point VOP, or Voronoi VEX function in a point Wrangle. Subtract dist1 from dist2 and plug it in a colour ramp that goes into the colour of the points. This helps visualise the Voronoi pattern for tweaking the frequency. Delete points lower than a cer tain colour value.
08 The Performance Monitor helps identify
VORONOI POINT CLOUD
Joan Panis
LOOK FOR BOTTLENECKS
performance bottlenecks in your scene. Red is where the highest action is. Nodes highlighted in orange and green are moderate action. You can scrub the Timeline at the bottom of the Performance Monitor pane to see the specific frame that you want to query.
Fianna Wong INTERSECTION ANALYSIS
05 Intersections are a common problem,
BAKING WORKFLOW
09 The Bake Texture ROP has a big
especially in procedural modelling. Having the ability to query their exact location and primitives involved is a good start to solve the problem. Intersection analysis does just this. From here you can, for example, find and clean up the shortest path outputs, as these naturally take shortcuts through geometry.
upgrade and you can get good output with much less hassle. It also ships with premade setups for common channels like curvature and occlusion; just activate the channels on the Bake Texture ROP node and remember to drop a Bake Exports node in your Shader.
Björn Henriksson
Björn Henriksson
USE VELOCITY INPUT
SMART CLUSTERING MESHES
ARTISTIC VOLUME MANIPULATION
David Hipp
12
POLYPATH AND REWIRE VERTICES
Here are two wrapper nodes that solve common low-level problems when constructing topology: PolyPath simplifies a series of connected polylines or curves to a single one . Rewire Vertices makes it possible to change vertex order using an attribute, enabling you to insert vertices to an existing primitive without re-creating.
Björn Henriksson FLIP TANK COLLISIONS
13 Collisions can be generated with the Geometry node into the Root p ane (scene). Add a mesh and select Collisions Tab>Static Object. In the Root pane, go into AutoDopNetwork, select the Static Object node and enable Use Deforming Geometry. If the SOP contains animated geometry, the static object’s geometry will also animate.
Gustavo Åhlén
14
PRESERVE COMPONENT ORDERING
06 Linking the third FLIP solver input to
10 Smart generate points in meshes with
the Source Volume node allows FLIP Tank velocity control. To begin, create a Geo node in the OBJ path (root). Next, add a Grid n ode to originate the plane, and an Ocean Spectrum node to create volumes with wind and tidal pull on ocean waves information for animating ocean surface geometry. Finally, connect both to
the Cluster Points node. Set the mode to average points and higher iterations for improved clustering. You can cluster any attribute, for example you can use P for position-based attributes. Also, the Attribute weight on this node can consolidate clusters for values above 1.
Deleting, merging, packing or unpacking geometry may cause point/primitive order issues. Keep original point/prim numbers by creating an ID attribute for your points/primitives before modifying geometry. Then you can reorder based on the original ID attribute post modification. The Sort node based on your stored ID attribute lets you restore order/relative.
Nicholas Pliatsikas
Nicholas Pliatsikas 29
43 33 HOUDINI TRICKS, TIPS TIPS AND TECHNIQUES
15
CONTROL SIM RESOLUTION
For higher-resolution fluid, go to the FLIP Object node in the DOP network and control sim resolution by decreasing the Particle Separation parameter. Default parameter values work well for prototypes. It creates very fast simulations. To increase your final sim resolution, decrease this value but avoid values lower than 0.02 to preserve computer stability.
Gustavo Åhlén
16 ATTRIBUTE TYPES
Even though colours, UVs, normals and positions are represented with three floating point numbers, Houdini will treat these very differently depending on how these attributes are labelled under the hood. You can set an existing attribute type using an Attribute Create SOP or the setattribtypeinfo() VEX function in the Wrangle.
Charles Trippe
17 UVS TO WORLD SPACE
To use UVs in World Space start with a Fuse node to make unique points (Fuse also unfuses points). Utilise the Attribute Promote node to promote the vertex UV attribute to a point Attribute, then use a point Wrangle to assign the UVs to position (@P=@uv;). Once done, points can be fused back together with the Fuse node.
Nicholas Pliatsikas
18 FORGET ‘$’, EMBRACE THE @
In Version 16 you will find that local variables like $TX has, in many places, been phased out in favour of the VEX approach of @P.x. Classic nodes like Point have been replaced by Attribute Expression. If you haven’t already, now is the time to read up on VEX expressions.
Björn Henriksson
19
BULLET RBD SIMS
When you are working on creating a bullet simulation, it’s often a better idea to shatter complex collision geometry and bring that in as a series of convex hulls rather than to try it as a concave hull. This is because it will run faster and also become much more stable to sim.
cloud. Radius and Max Point in the Point Cloud Open are key to a good result. The while loop is to repeat the operation multiple times. This gives a stringy look but can be slow to compute.
20
Setting Particle Separation below 0.05 may cause computer instability. Download Intel SDK for OpenCL Applications and install the runtime driver. The advantage of using OpenCL on the CPU is the ability to use all available RAM on the system, unlike the GPU which is limited to RAM on the video card. This works for Smoke and Pyro solvers, as well as fluids.
Gustavo Åhlén
21 TIGHTEN POINT CLOUDS
Using Point Cloud Open and Point Cloud Filter inside a while loop, we can tighten a point
30
Gustavo Åhlén
Joan Panis
22
MAKE FASTER TOOLS
Do your heaviest processing upfront in the network or in the non-interactive sections of your tools, for example a render-time process or click-button solution to store data for use in the more interactive parts later down the pipe. This way paint nodes, or remapping values with ramps, can feel more interactive when tweaking.
28 COMPILE BLOCKS
A major performance boost comes in the form of these nodes. Encapsulate your for-each networks and Houdini will thread the iterations. A minute-long cook suddenly cut down to a few seconds. Be aware that not all nodes are compatible with this workflow yet (or verbified as SideFX call it), but most are.
Björn Henriksson
Nicholas Pliatsikas
23
MAKE DEEPS LIGHTER
For smaller file sizes when rendering with Deeps in Mantra, use options hidden by default. Edit the rendering parameter on your Mantra node and filter by the word ‘deep’ to display all attributes of deeps. Bring Compression to a lower level and increase Z-Bias to reduce the file size, making it manageable for compositing.
29 CACHE MANAGER
Houdini is resource hungry and can take up all your machine’s RAM quickly. Houdini tries to release memory as it goes, however from time to time it’s good to flush the cache manually instead of closing Houdini and reopening it. You can access the cache manager through Window>Cache Manager.
Joan Panis
Joan Panis
24
RBD TRANSFORMS ON POINTS FROM SIMULATION
For non-deforming RBD simulations, the only data needed is rigid transformations for simulated pieces, which are later used to transform the original static pieces (packed disk primitives). Using points with transform data to alter packed disk primitives will save massive amounts of disk space and be faster to iterate as well as preview.
Tomas Zaveckas ENVIRONMENT 25 SIMPLIFY GEOMETRY Scatter points on Lidar Geometry, then save the rest position as a new attribute on the points. Make these points zero on the y axis and use Triangulate 2D. Use the rest attribute on these points to replace the current point position to the rest position and create a remeshed Lidar. The complexity is controlled by the number of points.
Joan Panis
David Hipp UNLOCK SYSTEM RAM
Null node output, select the Camera node and check Keep Position When Parenting, or the moving object won’t be tracked by the camera.
26 GENERAL CACHE WORKFLOW
By default Houdini caches simulated information in memory for quick playback. Always check to make sure Cache Simulation stays on when using a DOP simulation, because DOPs can behave unpredictably without caching. You can also set how much memory to use for cached simulation frames.
David Hipp
27 MOTION PATH TRACKING
To track the motion path of an object, first add a Null node (null1), then connect to the output a Camera node (cam1) and the geometry (geo1) to track. After connecting both nodes to
HIDDEN 30 EXPLORE PARAMETERS There are a ton of hidden parameters to explore to improve your renders. A quick example would be to add Geometry Filter Width to the Geometry node to help render small curves and points without increasing pixel samples.
Hermes Crespo
31 MAYA ENGINE
The Maya plugin has new attributes to let your tools control the Maya output. For example, you can specify Shaders with maya_ shading_group or set with maya_selection_set. There’s now a debug function enabling you to open a Houdini scene with your setup reconstructed. Very useful for understanding how attributes behave.
Björn Henriksson
32 SEARCH LARGE NETWORKS
You can search large networks easily by hitting Ctrl+F in the Network Editor pane. In the Find Node window that appears type the node you’re looking for. You can filter by name, or browse them one by one using Previous or Next to find the node you want.
Fianna Wong LIGHTING 33 ENVIRONMENT WITH PYRO To light up an environment using a pyro object, instead of using a light source we can tweak Mantra for something similar. Turn the Diffuse Limit up in Mantra under the Limit tab. You might have to add more Pixel Samples or up the Diffuse Quality to get rid of the noise.
Joan Panis
PRODUCE LIKE THE PROS Red Ring Entertainment reveals how Houdini 16 was used to create the stunning Amarok wolf with Kalin Stoyanov
LIGHTING SETUP We designed a special lighting rig for the eyes in order for it to look like an animal in the night that’s lit by a flashlight. For the rest of the body we chose to use a standard three-point lighting setup. There is no global illumination in the scene.
FAST BONE PLACEMENT
SETTING UP FUR
The entire rig and all of the animation was produced with Houdini 16. The new fast bone placement with medial axis support is the perfect tool for placing bones inside the geometry. You can control interior snapping using the Placement menu. It successfully managed to find the centre of the mesh almost flawlessly.
We built eight separate fur systems for the different body parts to control the specific look of each area. For faster render times we lowered the density in some places. The new Operators give endless options to control your fur in two stages; first manipulate the guides and then refine with the Generate Fur nodes.
OPTIMISED ANIMATION PLAYBACK
BIHARMONIC SKINNING ALGORITHM
In Houdini 16 the rig is much more responsive and animation is done in a much more fluent way. Compared to Houdini 15 you can play an animation at almost real-time. The evaluation of the nodes has been tremendously optimised. The little enhancements to the Animation Editor proved very useful in production.
The new biharmonic skinning algorithm was so good that there was almost no need for additional weight painting. In the oftenproblematic places like shoulders and knees it was able to create great results. We were shocked that, after capturing, the character moved as if it had muscles.
31
43 33 HOUDINI TRICKS, TIPS AND TECHNIQUES
Joshua Matthews worked on the splash screen and other hair/ fur examples for Houdini 16
HAIR AND FUR TIPS Discover how to achieve the best results when creating, grooming and cooking hair and fur in Houdini 16 with Joshua Matthews
34
RESAMPLE THE GUIDES
For curly hair, add a Guide Process frizz to your guide groom, then add a Resample node. Be sure to turn on Maximum Segments and turn off Maximum Segment Length. Change Treat Polygons As to Subdivision Curves. In your hairgen, add a Hair Clump node and turn on Curling. Increasing Amplitude gives your hair smooth, tight curls.
35
USE NOISE MASKS
For hair variation, place a Guide Process with the Set Length operation, then add a Noise Mask. It works as a multiplier to your original value. If length is 3, it randomises between 0 and 3. You can further tweak by changing Frequency, Amplitude, Bend, Frizz, Direction, Wave, Smooth and Straighten.
36
SPEED UP COOKING
Creating VDB for preventing hair intersecting geometry can be slow. For quicker cooking, place a Guide Collide with VDB at the end of the Node Tree. Merge in skin geometry
32
40 and use a VDB from polygons to make a manual collision skin. Guidecollidevdb lets you control how much the guides avoid the skin.
37
CONTROL HAIR LENGTH
Use VEX for a procedurally faded haircut. Connect a point Wrangle with the input and output of the skin. Add f@hairLength = fit(chr amp(‘Ramp’,relbbox(0,@P.y)),0,1,0,1);. This sets the hair length of each point equal to the relative y position of the first input. Chramp creates a ramp parameter for more control and ‘fit’ keeps values between zero and one. In the Set Length node, set Length to be controlled by the skin attribute. It must match the name of the attribute created in the Wrangle.
38
MORE ACCURATE
39
PLACE GUIDE PROCESSES
SIMULATIONS
To have guides collide with each other accurately, select your Guide Simulate node and click Create External DOP Network to place a dopnet that’s linked to all guide sim parameters for greater control. Select Wire Object>Guides and turn on Width Visualization to see each guide’s collision radius. In your guide sim, turn on self collide.
Think about cooking time when choosing where to put processes. It’s preferable to wait for hairgen to cook than guides, as usually the guides need to be simulated. If you build your fur system mostly in the hairgen, you will find that your simulation times will be much more manageable.
40
TRANSMISSION VS OPACITY
Transmission controls how much light passes through a surface, making beautiful effects on hair when backlit. Opacity doesn’t affect light passing through. Opacity with falloff can help emulate softness in fur, specifically on the ends. Here I chose transmission, as with opacity falloff you may lose integrity of the strands at the ends.
41
CACHING OUT HAIRGEN
The act of caching out curves bypasses vm procedural rendering, so you no longer need to visualise your guides and hairgen at the same time. This also means you don’t need to recook the guides every time you change something in the Shader settings. Caching out the hairgen is great for still renders of fur, especially turntables.
42
USE GOOD UVS
43
USE POINT COLOUR
UVs aren’t necessary for hair and fur, but can make a huge difference in the flow and direction of your curves. Inside the Guide Initialize node, there are options called UV Blend and UV Rotation. This will group the guides into sections by UV patch, and change the direction and rotation of each patch individually.
To create different coloured hair patches, use a simple paint SOP on the skin and an Attribute Transfer in the hairgen to overwrite the Cd attribute. Be sure to transfer from points to primitives. In the Shader turn on Tint with Point Color so the Shader is affected by the Cd attribute. Remember, Cd value is multiplied with the base colour rather than replacing.
INSIDE THE MEDIAMONKS MONASTERY
ANIMATION
MOBILE
A CAMPAIGNS
C RICH MEDIA
VFX & POST
R The headquarters for the Dutch digital agency are located in Amsterdam
34
INSIDE THE
MEDIA MONKS MONASTERY
FILMS
F
Pierre Nelwan reveals the secret disciplines of the creatives taking the world by storm
O
ne of the first things you see adorning the website of global powerhouse MediaMonks’ website is the slogan, “Crafted with care, coded by coffee, celebrated with champagne.” It’s no surprise, then, that a blazing passion for creating exciting, lively work is complemented by hundreds of awards; as of writing, MediaMonks has 97 Cannes Lions and 186 FWAs to its name, and there’s seemingly no end to its genius. Co-founded in 2001 by Wesley ter Haar, Gin Roberscheuten and Terrence Koeman, with CEO Victor Knaap joining soon af ter, MediaMonks has its headquarters in Amsterdam, but the studio is ever-growing. With ten offices across four continents and 600 monks (as the employees call themselves), MediaMonks is perhap s one of the biggest creative production companies on the planet. The studio’s work takes a multitude of forms as it strives to create innovative campaigns across film, games, VFX, animation and more with some of the world’s leading brands. Adverts, websites, apps, VR and AR, physical installations – nothing is amiss here. Take Toyota Prius Prime: The Impossible Quest, a story-driven VR game using 3D p ainting as a narrative device. MediaMonks developed a 3D painting program that integrates the user’s artwork with the story, enabling them to experience the car while drawing their focus to its features within the context of the story. “By
35
INSIDE THE MEDIAMONKS MONASTERY
turning a demo into a game, we managed to get people queuing for a car ad,” says Pierre Nelwan, global head of animation. “The Impossible Quest is quite possibly our most thrilling 4D installation yet. Debuted at TechCrunch Disrupt, it puts you in the middle of a VR car chase designed by legendary Hollywood concept artist Syd Mead. The real-time 3D experience starts with the creation of a 3D artwork using our custom VR painting software. You’re then put behind the wheel of the new Toyota Prius Prime to race it to safety. The experience is directed by our very own Jason Zada and enhanced by various 4D effects that literally blow you away.” Nelwan’s story at MediaMonks began ten years ago when he was headhunted by ter Haar, the company’s COO. “He found my portfolio on a well-known Dutch Flash forum and invited me for a cup of tea at the office. We hit it off right from the start since we liked the same things: making awesome things while listening to hip-hop. So I started the day after.” He leads a 30-monk (and counting) animation army spread across all of the offices worldwide, and it’s an incredible feat to be able to keep the work momentum flowing cross-continent, no matter the time or project. “All of our offices are connected to each other as we share the same projects and resources. We offer our clients a 24-hour production service, so our Amsterdam team might end their day with a handover to our LA team, resulting in a constant stream of production.” MediaMonks reached its ten-office milestone in 2016, adding four locations and even more worldwide talent to the already burgeoning creatives. “With so many different people, talents and cultures there’s obviously differences between each office. For example, the team in São Paulo are amazing at experiential productions, while the monks in Stockholm are the masters of VR. In terms of animation and VFX, a lot of our output is based at our HQ office, but it’s not exclusive to one place. Generally, we try to make sure all offices are formed by talented individuals in multiple disciplines, so we can cater to b oth the local and international markets.” Each discipline and medium requires a different animation-production approach, explains Nelwan, and each one is more complex than the last, “but always interesting. For example, making an animation in a real-time Unity environment for VR
All offices are formed by talented individuals in multiple disciplines, so we can cater to local and international markets 36
5
DISCIPLINES OF A TOP MEDIAMONK
Pierre Nelwan gives us his five tips for prospective artists after a job at MediaMonks
1
SHOW PASSION AND MOTIVATION IN WHAT YOU DO (ESPECIALLY WHEN YOU’RE A VETERAN)
2
SHOW CONFIDENCE IN YOUR WORK
3
ALWAYS AIM FOR MORE. NEVER SETTLE FOR ‘OK’
4
DON’T BE AVERAGE. SURPRISE US AND STAND OUT. INDIVIDUALITY IS WHAT MAKES US
5
SPELL MEDIAMONKS PROPERLY (HONESTLY, YOU’D BE SURPRISED BY THE NUMBERS OF TYPOS…)
Deep Thoughts brings together “not-so like-minded” animators, getting them to escape their comfort zones
Intel: History Comes Alive is an interactive installation that lasts around 12 minutes
Each episode of Deep Thoughts features around 15 monks from animation and sound design
MediaMonks opened its tenth office in 2016
GAMES
G
The biggest challenge was to realistically map detailed 3D models onto the historical characters completely differs to creating an animation for broadcast television.” A good MediaMonks project, reveals Nelwan, is one that has “a strong idea, a smooth process and a flawless execution”. The animation series Deep Thoughts is one of Nelwan’s favourite projects. Released bi-monthly, episodes are themed around a different word, such as pets, cereal, money and fear. Each one features lightning-fast animations injected with bright colours that interpret the topic in just under two minutes. The themes carry their own colour palettes; take cereal for example, which has salmon pinks and dodger blues, whereas pets features pastels, mint greens and bright purples. Using the specified colour palette, the team then creates a short 2D or 3D animation without any other restrictions. “It’s fascinating to see different signatures come together in one animation,” says Nelwan. “What makes it even more awesome is that the project is done during our downtime.” On the complete opposite end of the style spectrum is History Comes Alive, an interactive installation done with JWT. History Comes Alive uses Intel’s RealSense and produces realistic 3D models of a participant’s facial features. These are then mapped onto historical characters in a series of 3D, fully animated stories. “The biggest challenge was to realistically map detailed 3D models onto the historical characters, smoothly blending them into animations. We achieved this by matching the character’s skin to that of the user, but also by reinterpreting the heads of the historical characters ever so slightly to create a better match with the user’s face. “Another challenge was to create a pipeline that could render the new frames and integrate them into a personalised animation starring the users as a historical character. Our hosting engineers helped to create a fabulous cloud-based render farm that could do all of this and deliver the animation as a video through WeChat – just minutes after users had scanned their face.” Some of the tools used by MediaMonks for this type of high-end animation work are Maya and V-Ray, which have been used since the start of the animation department. The Adobe Creative Suite is favoured generally, as are Cinema 4D and After Effects for motion graphics, and Toon Boom and Flash for cell animation. Wacom or Cintiq tablets accompany powerful PCs, and pencils and paper are always waiting in the background for quick sketches and doodles while waiting for renders.
37
INSIDE THE MEDIAMONKS MONASTERY
The Deep Thoughts web series has featured topics including money, pets, cereal and fear
MAKING HISTORY How MediaMonks and JWT worked to make ordinary people a part of history in Intel: History Comes Alive History Comes Alive brings art and technology together in a live experiential event that connects people in China to their remarkable heritage. To feature ordinary people in an extraordinary way, MediaMonks produced four original films and an installation to make it personal. In total, 18,000 frames of original animation work made up the films. The campaign demonstrates Intel’s RealSense technology by producing realistic 3D models of people’s facial features. By scanning their face, participants are featured in a series of fully animated stories in which their faces are mapped onto historical characters. The project was first brought to life on the ancient walls of Xi’an and now lives in live events around China. Around 40 concept artists and animators created 100 scenes of epic moments featuring 22 heroes and villains from Chinese history.
Pierre Nelwan has been at MediaMonks since 2006
EXPERIMENTAL
E 38
Love and show passion in what you do, since that mindset will bring you far in everything
Toyota Prius Prime: The Impossible Quest experiencebegins with 3D art created by the user
MediaMonks created a 3D painting tool for the Toyota VR project that integrates the user’s artwork with the VR story
“We’ve added some additional software packages for specific tasks over the last few years, like ZBrush and Houdini for instance. But Maya and V-Ray are still our main focus for the moment,” says Nelwan. “We build a lot of our own tools to maximise the workflow between software packages and how we approach projects. We’re starting to use Isotropix Clarisse a lot more, and more recently Marvelous Designer.” And though the department specialises in bringing the magic to animation, animators aren’t the only type of artist that Nelwan looks for, and generalists can actually bring in plenty of knowledge to the department. “We need both profiles. Our generalists flag potential problems at an early stage since their knowledge is widespread across multiple (overlapping) processes. We need specialists to bring out the magic and p ush for those extra miles. A really good project is often defined by those final details.” What’s important for any aspiring artist hoping to enter the industry is to “love and show passion in what you do, since that mindset will bring you far in everything!” So what should we expect from MediaMonks in the next year? “I think 2017 will see us continue to combine craft and technology to make great work. We’ve already released some amazing projects such as the Data Dress, where we partnered with Ivyrevel and Google to develop an application that tracks your whereabouts and activities to generate a unique custom dress based on the data. Another favourite is Audi: Enter Sandbox, where we partnered with POL and Audi to turn a physical sandbox into a virtual playground, by using a depth-sensing camera that renders your imagination into a 3D environment that you can explore with the Q5 in a VR driving simulator. You can expect more exciting projects coming from our monasteries, so stay tuned.” Beyond 2017, MediaMonks has plenty of aspirations. “We’re always chasing the bigger and better version of ourselves,” says Nelwan. “Winning awards and recognitions like the Advertising Age A-List [MediaMonks is part of the 2017 Production Company A-List, with honourable mentions including Framestore and mssngpeces] doesn’t change the way we work but it does affect morale in a positive way. We love recognition for hard work, and I think it’s very important you celebrate your victories.” It’s pretty apt then, that the MediaMonks ethos as he sees it is to “work hard, party harder!”.
39
YOUR BLENDER ISSUES SOLVED
22 PAGE
BLENDER SPECIAL
40
YOUR
BLENDER ISSUES
SOLVED PRO ARTISTS ANSWER YOUR QUESTIONS FROM SOCIAL MEDIA AND REVEAL THEIR EXPERT TECHNIQUES FOR TACKLING COMMON PROBLEMS
How do I groom hair in Blender? Nazar Noschenko
Freelancer artstation.com/artist/nazarnoschenko
In Particle mode, I add hairs with the aim to highlight the shape of the hairdo and give it a more natural look
The first step is duplicating the scalp area from the finished mesh and scaling it down under the mesh a bit. In the Particle tab I create a new hair particle system. I start from the nape area, adjust the needed length of the hair, and then set the hair number to zero. Next, I go into Particle mode. Using the Add brush, I draw the first level of hair, then with the Comb brush I pull all the hairs down and adjust them into the desired shape. Then I keep making new levels of the hairs up to around the crown area. To get a better idea of how it will look at render time I go to the Children settings from the Particle panel – here I set Children to Simple and increase the number of children. After that, I define the shape of the clump curve, radius and the roundness. The next step is the crown area. In the Particle tab, I create a new particle system with the same settings as the previous one. In my case, the crown hair divided into two halves, so I started from one side. In Particle mode, I start adding new hairs along the parting hairline. Using the Comb brush, I pull the hairs down and style them, as the concept
requires. Using the same method, I make another half. Now for the fun part – hair strands. I create a new particle system and, in Particle mode, I start adding hairs with the aim to highlight the shape of the hairdo and give it a more natural look. In the Children settings, I make the radius much smaller compared to the main hair and increase the clump factor. Next, I go to the Cycles Hair Settings panel. Here I set up the Shape parameter, which controls the transition in thickness between the root and tip and the root multiplier, which multiplies the particle size to give the strands thickness. In the hair render settings, I increase the steps up to eight so the hair curves are smooth. I do the same with each particle system. In the Node Editor, I mix the Diffuse Shader with two Hair BSDF nodes, of which one is Reflection and the second one is Transmission, and adjust their factor. Now I add two ColorRamp nodes and make two gradients in green and blue shades, then I connect Noise Texture to their factor. Using the MixRGB node I blend these nodes with ColorRamp. Finally, I connect the complete node to Diffuse, Transmission and Reflection colours, and I’m done! 41
YOUR BLENDER ISSUES SOLVED
Yanal Sosak
Character artist artstation.com/artist/yanalsosak
You should also look for inspiration on a wider scale. Inspiration is not program restricted
42
Where do you get inspiration? What workflow do you use? Part of what I do at the moment is scout Blender artwork as an art editor at BlenderNation so that I can post articles on them. To do so I need to have knowledge on where to find exceptional Blender work to show off to the community. Here are some of the sites I go to when looking for new and exciting renders: • blenderartists.org – A forum where you can post your artwork for critique and exposure. There are many different sections including WIP, focused critique and final renders. Some artworks get featured and can be seen on the top row of the website. • cgcookie.com/gallery – A website where you can subscribe and learn from many different tutorials. There is a section for images where you can check people’s renders.
• artstation.com – You can type Blender in the search bar and play around with the settings to look at the recently posted Blender work. • blendernation.com – The website for Blender news and artwork. • reddit.com/r/blender – A great place to find good art, just check the up-votes. • facebook.com - Blender groups on Facebook. The two main groups I am aware of are Blender and BlenderArtists (not the same as the website). A lot of members regularly post their artwork there – just check the photo section for faster viewing. That said, you should also look for inspiration on a wider scale – don’t limit it to Blender artworks. Inspiration is not program restricted.
My workflow for post-production: Step one: I use the Blender compositor. I have a setup that does the following: adding AO>Bloom>Color Correction>Chromatic Aberration and Lens Distortion (sometimes I skip the CA and LD part in Blender). Step two: Afterwards, I go into Photoshop and use it to do any quick fixes that I haven’t accounted for in the render. I then add a background there, sharpen the image, add chromatic aberration and some noise. It is important to be subtle with most effects. This workflow works very well for still renders, if you want to render animations then you can do the same for step one. As for step two, you can use After Effects instead of Photoshop.
What are some free add-ons for a faster modelling workflow? Zacharias Reinhardt Freelance 3D artist and Blender trainer zachariasreinhardt.com
A quick way to use Boolean operations in a destructive or nondestructive way
The number of add-ons that are available for Blender is quite enormous, really. It is nearly impossible to name all of the excellent add-ons and plugins related to modelling, so here is just a small selection of modelling add-ons that I use a lot in my own personal workflow. Some of them are shipped with Blender (and can be enabled under User Preferences>Addons). For more add-ons and in-depth information, take a look at bit.ly/2qdxIng or blenderaddonlist. blogspot.de. Internal means shipped with Blender and External means you have to download it from an external source. All of the following add-ons need to be enabled first: UI Pie Menu Official (Internal) A quick way to switch between all modes, like Edit Mode, Sculpt Mode, Object Mode and so on, with a circle-shaped menu. Press Tab to enter it.
LoopTools (Internal – Edit Mode>Tool Shelf>Tools) Offers a bunch of new editing tools like Flatten, Bridge and Relax. The option to make a circle out of the selected vertices is especially great. Bool Tools (Internal – Object Mode>Tool Shelf>Bool Tools) A quick and efficient way to use Boolean operations in a destructive or non-destructive way. Great for combining a huge amount of different objects with Boolean operations. This is a super useful add-on for quickly creating hard-surface objects. Carve MT (Internal – Object Mode) Select an object and press Ctrl+Shift+X to enable this add-on. Now you can easily cut rectangular, circular or polygonal shapes into your objects. This is a great tool for hardsurface modelling. F2 (Internal – Edit Mode) Adds a new face after selecting only one vertex in the v-shaped corner and pressing F. This is especially useful for retopology.
43
YOUR BLENDER ISSUES SOLVED
How do I create viscous PBR materials in Blender? 03 off the basic honey setup, add a Volume
Add volume to the material To finish
Reynante M. Martinez
Absorption node and set the colour to a darker and saturated version of the Glass BSDF node’s colour. Control how thick the absorption effect is by adjusting the Density value. In most cases, a value of 1 will work. Connect this Volume Absorption node to the Material Output’s Volume input. This will now add the effect of light interacting with a thick viscous matter.
Freelance 3D artist and storyteller reynantemartinez.com
To further enhance believability, add surface imperfections and variety, as you’d see in real honey
01 Glass BSDF shader with an IOR set to 1.484 Create the base shader To start, add a
(this may vary depending on the type of honey you want to create). Set the colour to something close to yellow and desaturated. We’ll enhance this in the following steps. For optimised rendering with Cycles, mix the Glass BSDF with a Transparent BSDF with a pure white colour, and use the Light Path node’s Is Shadow Ray output as the factor. This will tell the engine to compute the shadow rays as transparency, enabling faster render times.
04 step and to further enhance the
Create surface variety As an optional
believability of the material, duplicate the existing setup in step 4 but make the colours darker. Combine these via a Mix Shader and use procedural textures connected to the Factor input of the Mix Shader to control the blending of the two honey materials. This will enable you to add surface imperfections and variety, as you’d see in real honey.
02 Glossy BSDF node and tweak the
Add PBR Fresnel Reflectance Add a
Roughness setting to simulate the micro particle inconsistencies of the honey, but don’t make it too rough. Mix this glossy shader with the existing node setup in step 1, using a Fresnel Reflectance controller that consists of a Layer Weight node (to control the Fresnel offset), a Math node set to Power (to intensify the Fresnel offset) and a ColorRamp node to control the overall effect, attenuation and falloff. Unless you want your edges to be too reflective, you can lower the brightness of the right-most slider and make it greyish in value, as seen in the image.
44
45 5
YOUR BLENDER ISSUES SOLVED
Aidy Burrows Lead instructor at cgmasters.net bit.ly/2oL5ShQ
Create several entire building textures and place them all in a single texture ‘atlas’ to reduce draw calls
46
What’s the best way to make large buildings with high texture quality? I hate to say it, but it depends! Let’s unpack it a bit and take a few different examples to help build (awful pun intended) up some possible approaches. A tall glass city building can have a tiling texture at a fairly small size since much of the detail will be in the reflections. As for brick/stone buildings (assuming we’re limiting ourselves to image textures rather than procedural textures to keep things running fast in a game engine, for example) we can separate the layers and mix them together. Let’s take a look at a simple example from the game swarm I’m currently working on. We have a high-resolution 2K (2,048 x 2,048) brick texture that holds up even when the camera gets in close. In the first image if we look carefully we can see parts of the texture repeating about four times horizontally across the wall and about six times vertically. In the second image we can see a 1K grunge texture. The mapping of the UVs are scaled so that the texture only repeats once across all of the wall area that we can see. In the third image we overlay the grunge over the top of the tiling brick texture just a little. Hey
presto! Now the tiling that was much more noticeable in the first image is almost completely gone in this image. Of course that’s even before lighting and shading, which can add another level of variety to help large tiling surfaces go further. There is a normal and roughness map introduced here too, and although I’ve not done it in this case it is possible to use the same idea that we’ve done with the albedo/base colour texture with those normal and roughness maps. In the final image I’ve created some floating window meshes that can be placed on top. This is an extreme example since it covers much of the wall but hey, why not? Other objects could be basically anything you find in your building reference images. Finally, if it’s a background building that’s never even seen up close, then sometimes we might just use a low-resolution texture that shows the whole building from top to toe. More likely we’d create several entire building textures and place them all in a single texture ‘atlas’ to reduce those draw calls. As you can probably tell, these ideas can be applied whenever you’re trying to keep things optimised and make a little bit go a long way.
How do you handle night-time lighting? Gleb Alexandrov
Artist, founder of Creative Shrimp website creativeshrimp.com
Use the Filmic add-on to remove the ugly clipping of highlights in night scenes
Getting a good-looking night scene is a tantalising idea indeed. So many things can go wrong. Lamps can be too bright, to the point of clipping. Noise produced by Cycles’ render engine can go over the top. And the resulting render time can be much longer than you’d want. Fortunately, that doesn’t stop us from creating eye-popping night scenes, does it? Let’s go ahead and tackle the most prominent night-time lighting problems one by one to see if we can’t get yours up to speed. Use the Filmic add-on. This add-on enables us to display a wider luminosity range, compared to the default Blender colour management. What this add-on does is gracefully convert the scene-referred light values to the display-referred values, using the film emulation. In other words, it provides the view transform that looks just as we expect from typical photographic mediums. So use the Filmic add-on to remove the ugly clipping of the highlights in night scenes, or alternatively just use the RGB curves to reduce the intensity of the highlights.
Sometimes too bright is okay. It may sound like we’re contradicting ourselves here, but sometimes the blown-out highlights are fine. Neither film, nor our eyes can handle some lighting scenarios. A night highway can be one of them. No surprise that the most extreme areas of the image can fall beyond the visible range. I would even recommend to emphasise these areas with the additional glow, optical flares and chromatic aberration (a lens distortion that gives objects colourful fringing). Regarding the render time, there’s not much we can do here. We can try the Branched Path Tracing (one of the two rendering options available in Cycles), but that requires that we first detect the cause of the noise. One way to do this is to enable the render passes (Diffuse, Glossy, direct as well as indirect). After that, you render out the image and check the passes one by one. And if you see that it’s the diffuse or the volume rays that are noisy, you then set the number of those samples up in the Branched Path Tracing tab. Often, that doesn’t help and we end up spending a night rendering 20,000 samples. Life is hard…
47
CREATE TE A CHARACTER C R ER WITH WITH BL BLENDER N 2.78
Expert advice from industry professionals, taking you from concept to completion
All tutorial files can be downloaded from: filesilo.co.uk/3dartist
Create a character with Blender 2.78 Learn how to render a still of a fully rigged 3D character prepared for animation or posing for illustrative rendering
I
n this tutorial, we’ll go over the entire workflow of creating a high-quality character for animation or illustrative rendering using Blender. You’ll be creating the model for your character, texturing it, posing it and then creating the final render. The aim of this tutorial is to give you an understanding of the full process of creating a 3D character, while learning tips and tricks along the way. Starting a character can be intimidating, but this tutorial will give you the knowledge you need to confidently create characters of your own from start to finish.
48
Research and planning As a 3D artist you need to
01 understand the character’s design and the original
artist’s mindset if you are basing the work on pre-existing art. It’s a good idea to check the artist’s other work and see if there are any sketches that may be useful, even if they aren’t directly related to the concept you chose. Take some time to break down the different aspects of the character and their accessories; try and find real-life equivalents where possible to fill any gaps in reasoning you may come across while building the character. In this case, we have a polished concept, an early T-pose sketch and references for the character’s body type. 01
22 PAGE
BLENDER SPECIAL
Create base mesh and refine With your
02 character planned out, begin blocking in the
character’s base mesh. You can do this in a number of ways; for example, you can use poly modelling, Blender’s Skin Modifier or Dynamic Topology. For this tutorial, we’ll use the Skin Modifier. While creating the base mesh, be sure to spend adequate time on the proportions, as this is the foundation on which you’ll build the whole character. If your proportions are incorrect, then no matter how many details you add, your finished character will not read the same as the concept. The image below, from left to right, shows the initial blockout using the Skin Modifier, applying the modifier and cleaning up the resulting mesh, and further refining using poly modelling. 02
Take some time to break down the different aspects of the character and their accessories; try and find real-life equivalents where possible to fill any gaps in reasoning 49 9
CREATE A CHARACTER WITH BLENDER 2.78
03
04
DANIEL PEDERSEN Tiger Samurai, 2017 Software Blender
Learn how to • Model and sculpt a character • Apply textures • Pose your model • Light a scene • Create shaders • Render a final image • Composite a render
Concept This amazing concept by Johannes Helgeson will be used as the base for our 3D character. Be sure to check out his other work at either www. artstation.com/artist/ helgesonart or www. helgesonart.tumblr.com.
Refine the head and retopologise Once you
03 have applied your Skin Modifier and have added basic features to the body, it’s time to refine the head. You can either poly model the face, or you can select the head region and split it from the body to be refined using Dynamic Topology sculpting. Splitting the head off will enable you to reduce the amount of retopology needed later on, since the body topology is already pretty good. Keep in mind that the main point of the head at this stage is to add all the basic features of the face. They can be further refined later using Multiresolution sculpting after you’ve retopologised. Once you have finished retopology and have re-attached the head to the body, you’ll have the base mesh.
05
Refine anatomy Now that you have a base mesh
04 for your character, the next step is to refine the
anatomy. When sculpting the anatomy, always use references and pay attention to the planes of the body. In this case you also have references showing how the original artist depicts the female form, which is extremely useful. Once you have sculpted in the muscles, be sure to smooth the character out using a layer of fat.
Create the mask There are a few ways to go about
05 creating the mask. Again, you can either use poly
06
modelling to create the mask directly or use Dynamic Topology. In this case the mask was initially sculpted using Dynamic Topology because it enables more creative freedom. However, dynamic sculpts tend to look lumpy and unpolished, so you’ll need to retopologise, then polish the sculpt using Multiresolution sculpting. Multiresolution sculpting enables you to reach a polished final result as well as enabling a greater control of sculpt resolution when navigating your scene or rendering. To make sculpting easier, the teeth are separate objects created using the Skin Modifier to be sculpted later on, also using the Multiresolution Modifier.
Character blockout Now we’ll create the first
06 blockout to check proportions and visualise the
YOUR
FREE
DOWNLOADS from filesilo.co.uk/3dartist • Textures • Tutorial screenshots
50
character. Use a simple armature to pose the body, then extract the clothes from the mesh by selecting the desired range of faces and duplicating them (Alt+P>Separate by Selection). The sword can be simple at this point, as it’s only for visualisation. Once you have your character in pose, you can scale the character’s armature directly to quickly correct any proportional issues. Make sure to scale symmetrically by selecting the desired bones on both sides of your armature and scaling from individual origins. If you like the changes you’ve made, reset the armature’s pose (Alt+G, Alt+R) and apply scale (Ctrl+A>Visual Transform). This will apply any scale changes you’ve made to the armature.
Proportions You now have a base mesh for the character. It’s a good idea to re-check your proportions before you begin detailing. This can be difficult since the character is currently in T-pose, but common measuring techniques such as measuring by heads can help you get close to the correct proportions.
Multiresolution sculpting enables you to reach a polished final result as well as enabling a greater control of sculpt resolution
51
CREATE A CHARACTER WITH BLENDER 2.78
07
08
Character tweaks Using the knowledge you have gained from your blockout, continue to refine your character. Start by blocking in all the accessories, then create another blockout and check your character against the reference. Continue refining, detailing the different elements one by one until you’re happy with the level of polish. Seek criticism and see what other people have to say. No matter how good of an artist you are, you may have missed something
07
09
10
that someone else will notice.
Set up deform bones The next step is to rig the
08 character. You should organise your rig by breaking it into layers. This makes it easier to work with and will help later when posing and animating. The image for this step (see FileSilo for the full image) shows the layers used to organise the rig. The first figure is a composite of the primary deform bones. The other figures show the individual bone groups. There are two additional layers not shown, which hide deform bones that you don’t need to see.
Rigging constraints Deform bones are the base of
09 every rig, constraints are the functionality. Constraints are used to automate motion across the rig, add complicated behaviours and make life easier for the animator. For this rig, the Inverse Kinematics (IK) constraint is used on the arms, legs and coat. The eyes use the Damped Track constraint to allow each eye to track to its own target bone. These are parented to a master Look Target bone. The fingers, spine and shoulder pads use the Copy Rotation constraint to allow for multiplied rotation across a series of bones. This speeds up animating actions like bending over or curling the fingers.
Developing character While working on the character’s poses, develop the character’s personality. With each pose, try to tell a story; try to think about what the character would do and how they would react in different scenarios. This will not only add depth to the character’s poses, but will enable you to add depth to the materials of the character. The history of the character will define how they look and act. Identifying features such as scars, weathering and damage to accessories and clothing add individuality. These are all very important to creating an appealing character.
Pose your rig Once you have the base functionality of
10 the rig in place, it’s important to work with your rig. This will help you find any mistakes. Put the rig through its paces, trying out different poses. If possible, animate it. As you work with your rig, you may find a need for additional functionality to streamline your workflow. Experiment with Blender’s constraints to create a personalised rig.
Lay out UVs Now we need to mark our UV seams and
11 organise the layout of the UV islands. In this case we’re
creating a character for still images or animation, so we can use more textures than we would on a game character. When laying out UVs, keep the highest pixel density on areas the viewer will see. For multiple objects using the same texture, use the Draw Other Objects feature in Blender’s UV Editor to view the UVs of all selected objects. For this character, we’ll use five textures, colour coded in the image on the right. The objects in red will be textured procedurally or by using box mapping.
52
11
06 12
06 13
Develop the character’s personality. With each pose, try to tell a story; try to think about how the character would react Character texturing Once your UVs are laid out, it’s
12 time to texture the character. Blender’s windows are
highly customisable, so go ahead and lay out your workspace. Keep the original concept open in a separate window in Blender as you’re texturing; this will enable you to constantly compare the concept and your textures. Within Blender you can colour pick directly from the concept for accuracy. When colour picking, try to select an area without highlights or shadows. In most instances, when a texture is too clean it’s not believable, so grunge up your textures appropriately.
Lighting setup Next, set up some lighting so that you
13 can see your character while you work. A basic lighting 06 14
setup will serve your purpose here. In this case a cold rim light, a warm fill light and another warm light on the mask was used. Make sure that you add an environment texture so that your reflective materials have something to reflect. Try experimenting with your lighting setup to find the look you like.
Base materials At this point you need to add
14 materials. The materials you create now are not final, but will serve as the base for your final render. When creating your materials, study real-life objects to achieve believable results. Also, analyse the concept closely. Your job as you go through the process of look development is to make your 3D model read the same as the concept. This can be challenging, as sometimes the materials in 2D concepts aren’t accurate to real-life materials. This is where logical reasoning and compromise come in. Try to find a middle ground between what is accurate and what looks right.
Final posing The first step towards developing your
15 final render is to put your character into the same pose 06 15
as the concept. Initially, just eyeball it. Once you have that base pose, go a step further using Blender’s background images feature. Assign the concept to the camera view using the Background Images tab in your options menu (brought up using the N key). Next, line up the concept with your 3D character and use the image opacity slider to compare your pose to the concept. Here you can see the original pass on the pose (left) and the pose after correction (right).
Box texture mapping Box projection texture mapping is a feature in Blender that enables you to map an image to an object using either world or object-generated texture coordinates. Using this with the Box projection option on your image texture enables you to apply textures without worrying about creating UVs. This is extremely useful for creating complicated materials that can be applied across any object. For more control over the rotation and scale of the texture, use the Mapping node. Your node tree should be set up in this order: Texture Coordinate, Mapping, Image Texture.
53
CREATE A CHARACTER WITH BLENDER 2.78
16
Final lighting and tweaking It’s time to craft your final render. You have all the
elements to create your image, so now it’s time to bring them all together. This step arguably takes the longest; this image went through 25 render versions. Compare your model side by side with the concept and try to re-create it as accurately as possible. This requires a lot of small tweaks across the character’s model, shaders and sometimes textures. Keep in mind that you’re creating a 3D model, so your details need to work in 3D space and not just from the camera view. 16
Daniel Pedersen I have over four years’ experience creating CG content. I enjoy almost all aspects of 3D content creation, and I’m always pushing to learn new techniques. When I’m not working on the computer, I enjoy being outdoors and love to explore.
Candy Knight, 2016 Blender
Simple lighting When trying to match your lighting to the original concept, you can overcomplicate the lighting on your model to the point that it will no longer look appealing and is needlessly complex. The initial lighting setup on this model had over eight lights and was far too ambient. By using so many lights, the shadows had been almost entirely eliminated, which resulted in a flat and boring-looking model. A simpler setup is almost always more effective. The final lighting setup contained five lights in total: three rim lights, one light on the body and one light on the mask.
A character I put together based on a concept by Tim Von Rueden. I try to make my characters ready for production, so this character is rigged and ready for animation.
17
Clay Diver, 2016 Blender This was an exercise in creating a character as quickly as possible. It took three hours: an hour for modelling and the remaining time on lighting, shading and compositing.
17
Final render and compositing Once you achieve a look you’re happy with, it’s time
to render with a higher sample count and final dimensions, then put the final shine onto the image using compositing. This can be done using the powerful compositing tools directly inside Blender. Blender’s compositing tools work from a node-based system. In this case, the Glare node was used to add a slight bloom to the image, a vignette was added to help focus the viewer’s eyes on the mask, the RGB Curves node was utilised to adjust the contrast of the image and the background was added using the Mix node and the image’s Alpha Channel.
54 All tutorial files can be downloaded from: filesilo.co.uk/3dartist
Headed Out, 2016 Blender This entire scene started as a simple shading test where I was trying to create a unique balloon shader inspired by a vintage photo.
ACHIEVE A FANTASY FANTASY ILLUSTRATION STYLE IN IN BLENDER
22 PAGE
BLENDER SPECIAL
56
DAVIDE PELLINO Fantasy Castle, 2017 Software Blender, Photoshop
Learn how to • Create low-poly models • Make a grass field in Blender Cycles • Create materials in Cycles
Achieve a fantasy illustration style in Blender This fantasy castle is a study on how to achieve an illustrated, non-CGI looking render with Blender and Cycles
• Set up render layers • Work with Freestyle settings • Create sketch effects • Composite images
Concept As a big Miyazaki fan, I wanted to try something surreal that could remind people of Howl’s Moving Castle.
n this tutorial we will learn how to set up an efficient workflow in order to achieve a stunning non-photorealistic look for your Blender (Cycles) works, starting from the default empty scene up to the final rendered image. ‘Keep it as simple as you can’ will be the theme for this project; despite the amazing result, behind the scenes there’s just a low-poly mesh with standard Diffuse materials. Almost no textures are used and no unwrapping is needed, as all the elements involved are generated procedurally. We will proceed step by step through some key phases, such as: • How to model and compose a low-poly scene • How to create materials • How to set up the scene/render layers • How to post-process the final image In this way, by learning the basics of this process, you will be able to make it your own and apply the same passages to all your personal works. Please refer to the provided Fantasy_Castle_Example files whenever you need to, as they include all the features explained in this tutorial.
I
Set the scene We will start by spending a few
01 minutes on tweaking the rendering setup so that it fits
Model base objects Once the render settings
02 have been perfected, we can start to model the basic meshes we need for our scene. The main task of this project is to avoid using unnecessary geometry and objects. Aim to keep it as simple as you can, and get straight to the point. While modelling, if you can cut out some vertices/faces and your objects still work fine and smoothly, then always go for it. Start from basic primitives (cubes, sphere, cones), give them a defined and appealing shape, then add details as a finishing touch.
Model props and composite When the basic
03 shapes are done, we can focus on creating smaller parts/details as separate objects to add some extra variety to our subject. Try not to overextend your modelling at this point, as we just need a set of five to ten simple items to scale, rotate and displace all over the base. Now just spread your creativity: build up your subject by combining the basic shapes you’ve made, duplicate them if needed and create a more defined scene that has feelings and impact. Then place all your props until you are happy with the result. 02
our needs for this non-photorealistic project. We are going to use Cycles (with GPU in this case, although using CPU would not affect the final result in any way). By adjusting the Sampling values, the Light Paths settings and global World features, we will save a lot of time in the final rendering phase. In this way, on a NVIDIA GTX 970 (4GB) the render compute will take no more than 12 to 18 minutes to complete. 01
03
YOUR
FREE
DOWNLOADS
from filesilo.co.uk/3dartist • Tutorial screenshots • Additional sketches • Textures
57
ACHIEVE A FANTASY ILLUSTRATION STYLE IN BLENDER
Grass particle system We are now creating a
04 simple but effective grass particle system to cover the
05
Make the grass field
whole terrain. It doesn’t need to look photorealistic, so the setup is pretty straightforward. The most useful settings you can adjust according to your tastes are: • Emission tab>Hair Length • Physics tab>Brownian (to give more or less erratic movement to the particles) • Display tab>Display (aim for around 20 to 50 per cent) • Children tab>Display (try to keep it at 1/5 of render value) • Render tab>check Strand Render (since we are going to use Freestyle later on).
One of the coolest parts of this scene is the grass field. Here we learn how to make a ‘realistic’ terrain with grass, flowers and stones in a very simple format. Learning some fundamental skills could be incredibly useful not only in this context, but also for every scene in which you need to place some natural elements, as you’ll just have to repeat all these steps to generate your own customised grass field.
Flower and stone particles As far as rocks
05 and flowers are concerned, we can model some really simple natural elements and select them as the particles that will be emitted by our terrain. The setup is basically the same that we used for the grass field, with some fundamental adjustments. Head to the Number setting in the Emission tab; this time it can be a really small number, like 100/500 particles emitted. We just need the flowers and rocks to pop up in some random spots and not cover all the ground. Under the Render tab>Object, select the object you want to as hair particles.
06
Build the scene We have the base, we’ve got the
06 props, we’ve got the grass field, we’ve got rocks and flowers – now it’s time to build up our scene! With all the elements at our disposal, here it all comes down to a matter of taste. The only new things added at this point are a curved plane used as a background and a sphere used to simulate the moon. Once we are satisfied with it, we can work on the materials before moving on to set the lights and the camera for the render.
Base materials One of the big pros of this project is
07 that we use Diffuse materials for the most part. The
node consists of two Diffuse shaders going through a Mix node with a procedural Musgrave texture as Fac. Once you’ve got this, just duplicate the material and change the colour in the RGB node, creating as many tones as you need then proceeding to assign them to all the parts of your mesh. No UV unwrapping, no texture – just ‘one material to rule them all’. The only different materials used are a Glass Shader (clock’s glass) and a Glossy Shader (pipes and metallic parts). 04
58
07
08
Sketch material When we work on non-
08 realistic projects such as this, there are some cool workarounds that we can use in order to give some extra punchiness to our illustrated style. In this case we will use a special sketched shadow material (that was made by my amazing collaborator Luca De Felice and I) on a duplicate of the castle. Key features of this node setup are the super-light material, which only takes 10 to 30 samples to render; the B/W map that generates a two-tone scratch map, easy to use in compositing later on; and customisable sketches/ dirt textures.
Freestyle line set Now we want to achieve a cool
09 sketched style for our castle. Since it is possible to add
09
10
modifiers to the line set, we can easily create a fancy irregular contour. Under the Geometry tab, select Sampling>Backbone>2D Transfer>Backbone. Under the Thickness tab, select Distance from Camera. With just a couple of clicks it is possible to get a distanceresponsive line set that will generate random displaced strokes all around our subject, perfectly fitting the style in this project. (NOTE: line colour is set to white here just to be more visible in the screenshot).
Camera and lights Camera size is set to a 2:3
10 vertical ratio (2,000px to 3,000px), with a focal length set on 24mm. The lighting setup is pretty simple and is achieved by using just two planes with Emission shaders. There is a main light on the right, facing the castle with a soft, red emission that gives the highlighted areas the feeling of a warm sunset. Then there is a fill light on the left, with a smaller emit value, that creates a blue-ish rim light on the opposite edges.
Separate objects per layer Now we want the
11 objects to be arranged per layer in order, so that we can render them as separate images. In this case, it could be useful to apply this order:
11
• Layer 01, Foreground (subjects with Diffuse_color material) • Layer 11, Middleground (the moon) • Layer 12, Background (the sky plane) • Layer 05, Sketches (a duplicate of the castle mesh, with Sketch_material) • Layer 06, Freestyle (another duplicate of the castle and the moon, with a Holdout_material) Layer numbering is not mandatory of course, and you can place the various objects as you prefer as long as this helps you to keep them in order.
Cycles render layers setup Since Blender offers a fully customisable per-layer setup, we can spend some time creating a pipeline that will provide us with all the render passes we need in order to work on them separately in the compositor. In this way, by just hitting the render button one time, Cycles will generate a Foreground layer, a Middleground layer, a Background layer, a Sketches B/W map layer and a Freestyle layer.
59
ACHIEVE A FANTASY ILLUSTRATION STYLE IN BLENDER
12
13
14
Compositing This is the last step of our work, and it will transform a flat rendered image into a brilliant illustrated piece of art! You can use Blender’s compositor or any other image editor you like, such as Photoshop, Krita or Gimp. There’s no heavy post-processing in this work, since we will use some texture layers in Multiply/Overlay/Soft Light mode and then will proceed to a fast colour correction before calling it complete.
Set the render layer Now we can navigate through
12 the Scene tab of Blender and create as many render
layers as we need. There aren’t any specific rules to follow; what is relevant is to decide the objects we want to be considered on a specific layer, those to be masked and those to be excluded. Once they are finally set, we are ready for the render phase.
The render setup Before rendering, let’s have a
13 look at Cycles’ settings panel. Cycles really benefits from GPU rendering; if your graphics card can handle it (2GB or more will work perfectly), enable GPU. Next, have a look at the Output section; PNG with lossless compression is a 100 per cent safe choice when you need to keep quality at its maximum. Remember to check the Freestyle box if you are planning to use Freestyle, otherwise this render layer will only produce an empty image. And finally, under Film, enable transparency so that Blender can save PNGs with alpha channel information.
Blender compositing With all our render layers
14 ready, we can now use Blender’s compositor to overlay the images we have. Key steps are: • Foreground>Alpha Over>Middleground>Alpha Over>Background • Composited image>Multiply>Sketches layers • Alpha Over to add the transparent Freestyle layer Now all the pieces are glued together and you can work with filters, colour correction, blur and everything you want on each of them separately. If you are not going to use the Blender internal compositor, just save them as PNG files in the UV editor tab to edit them externally.
External compositing: Photoshop editing For
15 external editing, these are the steps you need to take. To begin, import all your PNG-saved renders into a new file (they will match perfectly). Then compose the base image as follows: Background>Foreground and Middleground> Sketches (Multiply)>Freestyle (Multiply). Now we can use some textures (clouds.jpg, old_paper.jpg, provided with this tutorial) to add some extra magic to our image. Just keep playing with the different blend modes until you are satisfied with it!
60
15
Find the workflow that fits your creativity… as long as you pursue your ideas, there are no right or wrong steps
61
PRODUCE A FANTASY ILLUSTRATION STYLE IN BLENDER CYCLES
16
Place additional details We can now work on some little details to make our
composition look even smoother. In this case there have been only two real adjustments. First, a slice of grass was cut and duplicated on a separate layer, then blurred and placed between the Foreground and the Background, to give an extra depth-of-field feeling. Second, a starry night sky was made on another layer by just painting random white/yellow dots all over the picture. As mentioned previously, both of these two layers are mixed by testing out different blend mode solutions. 16
Davide Pellino Davide is a self-taught Italian 3D artist, working as a freelance character and environment designer. He works mainly with Blender, with a mission to find new ways to achieve stunning non-photorealistic render results for his low-poly creations.
83030 - The Island, 2016 Blender, Photoshop This creation is an image of a fantasy low-poly seascape at nighttime.
Final notes As a self-taught 3D artist, I’d like to give one big piece of advice to everyone reading this tutorial: try your own way, try and try again until you find the workflow that fits your creativity. It doesn’t matter what software you are using – as long as you understand what’s beneath every step of this process, you can make it a part of your routine. Just remember: we are artists, we communicate something to people and as long as you pursue your ideas, there are no right or wrong steps you can make. Just focus on making your ideas come true.
17
Sir Squiggley, 2017 Blender, Photoshop Sir Squiggley, a low-poly mouse, sheltering under a flower during a rainstorm.
17
Colour correction With all the pieces finally set in the right place, we can work on a
definitive colour correction in Photoshop. We can start off with using Curves to enhance image contrast and lift up some flat-looking areas. The Levels adjustment is useful for correcting the tonal range and colour balance by adjusting intensity levels of image shadows, midtones and highlights. Using Saturation, you can enrich the colours if you feel like they are too flat, and under the Photo Filter option we have selected a orange/red setup to make the image warmer.
62
All tutorial files can be downloaded from:
filesilo.co.uk/3dartist
The Sun Goddess, 2017 Blender A giant goddess statue suspended in the empty space under the heavy rain.
RENDER A PIRATE SHIP
64
Render a pirate ship Discover how to create a ship from blueprints, use UVs to curve wood textures and build foldable sails for the adventure of a lifetime
T
his tutorial will cover how to create your own 3D pirate ship from scratch. We will be using Maya for most of the modelling work, rendering in KeyShot and finalising in Photoshop. We will progress step by step, finding ways to add our own twist, tell the story we want to tell and make our own unique pirate ship with final renders. We will learn how to save time by creating a library of reusable parts and realistic details. We’ll use UVs to upgrade our wood textures and prepare materials for KeyShot. Finally, we’ll head to Photoshop to add some mood and atmosphere to the final image. Let’s get started!
65
RENDER A PIRATE SHIP
Start with the hull Everyone knows what a pirate
01 ship looks like, so there isn’t going to be too much to
GURMUKH BHASIN Discovery, 2016 Software Maya, Marvelous Designer, KeyShot, Photoshop
Learn how to • Use photographs and blueprints to create an initial block-in of a pirate ship • Take inspiration from many different existing pirate ships to help create your own unique vision • Use your UVs to your advantage to get your wood texture to curve the way you want it • Build a kit of parts that can be reused to save time on adding detail • Create and make sense of the insane amount of ropes needed to create a believable ship • Build sails that are both folded and deployed in Maya and Marvelous Designer • Prepare your materials and files to send over to KeyShot to render • Set up your lighting and materials in KeyShot • Finalise your images in Photoshop to add life, mood and tell your story
Concept Imagine saying goodbye to your loved ones, loading your belongings onto a ship made of wood and sailing off into the sunset. Adventures of a Pirate Ship is a series I created to evoke the sense of adventure that sailors must have gone through centuries ago.
design on your own. But it is also important to note that you don’t have to model something exactly – you can create your own version. Collecting reference images and drawings of existing pirate ships will help you to know where things go and roughly understand how things work. From there you can create your own versions of parts that you like from your references and fill in the blanks where you have to. To get started you can use a base image (I chose a profile view because it has the most information for a flat view of the ship) to start building off of, and once you have a solid 3D base to work from you can continue on your own, using your artistic intuition to decide proportions and how you want things to look.
02 06
03 06
Build to the human scale When creating a pirate
02 ship, it is important to show the human scale through details. These boats are huge, and it can be hard to get a good feeling of how big they actually are without the correct amount of details. Things like ladders, handrails, stairs, doors and windows reveal how big the ship is by relating it to a human scale, without having to show a scale figure.
Use UVs to your advantage Hand modelling the
03 wood planks that make up the hull would be very time consuming. Using a texture from textures.com for this step saves a lot of time. By flattening out your UV maps and then manually bending them when needed, you can easily control how the wood texture flows in the direction you need as if it were being built in real life.
Model simple and then adjust One of the
04 benefits of working in 3D is that you only have to build a part once and then you can duplicate it wherever you need. Model the parts flat and simply and then use tools like the lattice to move them into place and change their orientation after duplication.
04 06
06 01
Make it your own!
YOUR
FREE
DOWNLOADS
from filesilo.co.uk/3dartist • Photoshop file • Tutorial screenshots
66
When creating your pirate ship, it is important to remember that you don’t have to exactly re-create something that already exists. Remember that you are the artist and you get to choose how you want things to look. Adding your own changes and twists make the creation more fun and unique; just make sure you have a good story to back up your design decisions.
05 06
When creating a pirate ship, it is important to show the human scale through details. Things like ladders, handrails, stairs, doors and windows reveal how big the ship is Model your ropes square Start by modelling your
05 ropes square, then hit 3 on your keyboard for a
smooth mesh preview to make your ropes round. This makes it a lot easier to UV and use a rope texture from textures.com, and it keeps the poly count manageable since there are going to be a lot of ropes in your model.
06
08 06
06 07
06
Build the sails When starting the sails, it is
07
More ropes! When working on the ropes, it is really
08
Add the figurehead By adding a figurehead to
important to determine the height of the masts in regards to the overall balance of the ship. I decided to go a little taller with my masts and slightly larger with my sails for a more exaggerated look to my ship. I knew I would mostly be rendering this ship from a low view looking up with a fisheye lens, and the taller/wider sails would make for a more dramatic look. When modelling the sails, it is important to model all the parts that are needed in real life. The small metal rivets and pulleys will catch nice highlights when rendered from a distance. The ropes need to be modelled as if they are working for real, by wrapping around and tying to parts for a specific purpose. As you can see, I manually go in and move ropes around so that they bend around the parts they are holding down, instead of just letting things intersect. These adjustments are tiny and mostly won’t be seen from a distance. But by putting this much care into the model, the overall ship will feel more natural and believable in the final renders, instead of feeling too CG and fake.
easy to start to feel overwhelmed, especially when looking at your reference images and scratching your head while you’re trying to figure it all out. Start out simple and slowly build up the complexity. Find the main set of ropes needed for the sails, then move your way to the ladders and continue to the rest. Remember that you don’t have to model everything exactly as you see it, but you definitely want to do as much as you can to make it feel real and interesting overall. Sometimes there can be too much detail and the whole composition becomes disorganised and distracting, so use your artistic intuition and decide which ropes you want to leave out and which ropes make sense to show. You will have to hand place, rotate, scale, lattice and move individual parts so that they feel natural and believable – definitely take the time on these subtle elements as they will come together and really sell the believability of the ship.
the ship we can add another level of story to our creation. Do you want to go with dragons or lions for a more aggressive feeling, or do you want to go with an angel for a more peaceful story? When adding the figurehead, I searched online for a 3D model of an angel and found one made at Stanford University, free to use from graphics.stanford.edu/ data/3Dscanrep. I downloaded the model, removed the wings and the base, then latticed the model into place so it fit with the shape of the bow. 67
RENDER A PIRATE SHIP
Full ship with sails deployed After all of our hard
09 work, we are finally done with the modelling of our
10
ship. All of the sails, ropes, wood planks, cannons, life boats and more are there to make this ship exciting and believable. We are almost ready to start rendering, but there are a few more things to do before we send our model to KeyShot.
Fold the sails We didn’t do all this hard work just to
10 get one render from our ship. With a little extra work
we can create a version of our ship with folded sails. Using Marvelous Designer for this step, you will quickly be able to create realistic-looking folds and wrinkles. Start by creating a flat pattern of your sail that closely matches the shape when it is fully extended. Next, pin the sail at multiple points, so that when you start pulling the sail up into the folded position it will hold the rest of the sail down, and you will be able to control your folds with more precision. Keep pulling the rows of pins up to the top where they would be tied to the sail yard and let gravity naturally create the folds.
Full ship with sails retracted Once you’re done
11 folding the sails in Marvelous Designer, export your sails
11
as OBJs and bring them back into Maya. Now you can add the ropes that tie the sails up and adjust the pulley ropes to hang naturally. With just a little extra work, you now have a version of the ship that will render and feel new and interesting for an extra image or two.
Add some sailors The last thing we want to do
12 before sending our ship to KeyShot to render is to add a few sailors to the deck of the ship. You can use Daz Studio and pose some of the characters into positions to place around the ship. These are going to be used as base models to Photoshop over in your final renders, so they don’t have to be perfect. Get them good enough so that you can keep some of the clothing and anything else you can, without needing to Photoshop over them too much. 09
12
Fold the sails the way you want When creating the folded sails I wanted to keep the folds a little messy for a more dramatic feeling. In reality, the sails would be folded perfectly and not hanging around like I have them in my model, but I found it to be more fun to have them hanging unevenly and loose. This is another example of a place where you can add your own personal touch to the project to make it something more unique to what you want to create, instead of being dictated by what already exists in the real world.
68
12 13
16
14
Add your artistic touch to the sails In the real world the sails would not be transparent at all. But in our renders we can make them slightly seethrough to add more excitement and show the complex layering of parts in those areas. This is another way you can use your artistic intuition and break the rules of the real world to create the image you want to make. You are the artist and you can decide how you want things to look.
15
Jump into Keyshot 6 We are now ready to head over
13 to KeyShot and start finalising our materials and lighting.
It is important to clean up your model in Maya first, organise your outliner into groups that can easily be turned on and off in KeyShot and use Blinn in Maya to predetermine which parts of the model will get what material. This way you can easily drag and drop materials from the KeyShot library to test out different looks. KeyShot is an almost real-time renderer (depending on how heavy the scene is) and a really fun and fast way to try out different options for your final renders.
Look development in KeyShot Exploring different
14 render directions in KeyShot is so easy and lots of fun.
Once your basic materials are set up you should start playing with your lighting and camera position. Drag and drop different environment HDRIs into your scene to see which lighting setups work best with your shapes and materials. Once you stumble upon a mood you like and a story you want to tell, you can go back and adjust your materials to work better with the lighting environment you chose. Lighting plays a huge part in how your materials look, so it is always good to readjust your materials to look the way you want instead of worrying about them being physically accurate. Remember, the renders you get out of KeyShot are just a good base to start Photoshopping over to then create an original piece of art. You will be altering a lot of your final render to disguise and naturally adjust your image to tell your story and be more believable.
Final render parts for Photoshop We finally have
15 our render passes out of KeyShot and everything is
looking beautiful. The renders are an excellent base and we can now finalise our images by adding a bit of life, natural textures and controlling the focal point of the story in Photoshop.
Add photos and textures Start adding life to your
16 final image with the help of photos and textures. We
can use a dirt brush to lightly add dirt streaks that would appear from years of wear and tear to the boat and sails. This helps naturally break up surfaces that feel plain or have too many repeating textures. We can also start adding photos of ocean waves and mountain textures to make our CG-looking parts feel more real. Remember to colour correct the photos to match your base render so that everything feels like it is in the same environment and lighting conditions.
69
RENDER A PIRATE SHIP
17
Add depth and atmosphere After we have added all of our photos and textures, we
can now add in our atmosphere to push things in the distance further away and to subtly mute the details that are further from the camera. Use your Z-Depth Pass as a mask on the layer to gradually build up the atmosphere, and Paint Bucket the colour of the sky as your atmosphere colour. Make sure to lower the opacity of the layer so that you don’t make the haze too thick.
Gurmukh Bhasin Gurmukh is an architect turned 3D concept designer who currently lives in Los Angeles. His work draws from his life experiences, and provides a constant source of inspiration as he explores the untapped intersections between reality and imagination.
17
The White Rhino Concept Motorcycle, 2016 Maya, MoI 3D, KeyShot, Photoshop The White Rhino motorcycle was a quick, fun project. Creating my own motorcycle design was something that was at the back of my mind for quite some time.
Create more than one render After putting in all the hard work to build the pirate ship, you can get multiple pieces of art out of the project. Make a few different renders of the ship that tell different stories and show different adventures.
18
Weapons, 2016 Maya, MoI 3D, KeyShot, Photoshop Weapons are always fun little projects to make. They don’t take too much time to build and the images always look beautiful once created.
18
Finalise your image To finish the image, you can use a few layer adjustments to add a
bit of contrast and tie everything together. I usually use a photo filter to really set the mood, and adjust the brightness and contrast to make things more lively. Just be careful that you don’t blow out your light areas and lose detail in your final image.
70
All tutorial files can be downloaded from:
filesilo.co.uk/3dartist
SWAT Rhino R-2377, 2016 Maya, KeyShot, Photoshop The SWAT Rhino R-2377 is my vision of a future police vehicle coming to the LAPD in 2020. The truck is used to rush SWAT officers into dangerous situations and rescue those in need.
Specialoffer for readers in North America
4 issues FREE W hen you subscribe
FREE
resource downloads every issue
Practical inspiration for 3D insp enthusiasts and ent professionals r
Order hotline o ne +44
(0) 1795 592 951 Online at www.imaginesubs.co.uk/3da *Terms and conditions
This is a US subscription offer. You will actually be charged £105 sterling for an annual subscription. This is equivalent to $135 at the time of writing, exchange rate may vary. 4 free issues refers to the USA newsstand price of $14.99 for 13 issues being $194.87, compared with $135 for a subscription. Your subscription starts from the next available issue and will run for 13 issues. This offer expires 30 June 2017.
Quote
USA3 for this exclusive offer!
MODEL & TEXTURE AN ASSET FOR FOR UNREAL ENGINE ENGINE 4
72
Model & texture an asset for Unreal Engine 4 Discover the current industry pipeline and tools for creating next-gen game assets using PBR
T
his tutorial demonstrates industry-standard methods of creating visually detailed assets for current next-generation game engines. In this tutorial, you will be learning how to create a highly detailed asset, along with using packages such as ZBrush to sculpt in extra detail. You will be learning how to create topology and effectively unwrap that topology according to the importance of the asset’s silhouette. We will also be covering baking maps for your topology and the importance of using a cage during the process of baking to achieve high-quality projections. Last but not least we will be using a texturing package called Quixel SUITE to texture the asset, followed by importing the asset and its materials into Unreal Engine 4.
73
MODEL & TEXTURE AN ASSET FOR UNREAL ENGINE 4
Gather your references When approaching
01 making a game asset to the industry standard today, it
02 06
is extremely important that you gather as many references as you can. This puts you in a great position to study that reference when you are initially blocking out the asset in real time. In today’s standard of quality, it’s important that you are spot on with the asset you are creating. It’s extremely important that you also seek reference from real-life images.
JAMES BRADY Rusty Farm Lantern, 2017
Software 3ds Max, ZBrush, Quixel SUITE, xNormal, Unreal Engine 4
Learn how to • Create high-poly models and add extra detail in ZBrush • Create topology and effectively UV unwrap your models
Block out your model When creating an asset,
02 it’s important that once you gather references, you begin to block out the asset while carefully studying the references you gathered. This enables you to get the proportions correct, along with saving time by avoiding any mistakes that would arise later down the line once you have moved onto detailing in ZBrush. Another point to remember is, if your blockout is exact and you are happy with it, you can always use it as the topology and save time by only having to unwrap it at a later stage.
03 06
• Bake and texture your model • Import your model into Unreal Engine 4 and set up the materials correctly
Concept I was inspired to make the lantern game asset as I started to imagine that it would be used by the player to light their way through the unlit barn that the lantern is hung on. I pictured a scene of the player going into the barn to get their horse.
Create the high-poly model Once you are happy
03 with the blockout, you can then begin to create the
high-poly model. This will be the model that you will bring into ZBrush afterwards for extra detailing. When creating this model, it’s important to remember that keeping your edges really smooth enables you to capture better detail on the normal map than if they were to be really sharp. This also allows the model’s materials to still appear crisp at a distance, along with avoiding any aliasing issues too. Remember to save a version of your high-poly model to use for setting up Color IDs later down the line.
Import your model into ZBrush When you
04 open ZBrush, click Import, which is located under Tool, to import your high-poly asset. This will bring up a window where you can select the high-poly asset you want to import. Drag it into the scene and click Edit to start working on the high-poly model. In this section you should split the model up into relevant subtools. This can be achieved by clicking Split To Similar Parts under the Split section of the SubTool tab. This means that any parts of the mesh that are duplicated can be split into a group rather than having to work on them individually.
04 06
06 01
YOUR
FREE
DOWNLOADS from filesilo.co.uk/3dartist • Tutorial screenshots • Unwrapped model • ZBrush ZTool • Lantern textures
74
Set up correct subdivision When creating the high poly in 3ds Max, it’s important to remember to keep the edge loops evenly spaced when subdividing your model. This means that once you start turbosmoothing your model, the polygon density will be evenly spread around the mesh, enabling you to cover more areas without having to climb through subdivision levels, thus saving time.
05 06
When creating this model, keeping your edges really smooth enables you to capture better detail and allows the model’s materials to still appear crisp at a distance Detail your model in ZBrush In this step, you will
05 start to add detail to your model in ZBrush. I would
06
06 07
recommend subdividing your subtools to get the best-quality sculpts. This can be done by hitting Divide under the Geometry tab. You should start by using the ClayBuildup brush to make welds around the sections of the model that are connected, as this makes them look more realistic. You can then use the TrimDynamic brush to break up the uniform pattern of the welds. You should repeat this process until all of the connected sections of the model have sculpted welds. This also blends the model’s high poly together when baking the normal map, along with catching the light nicely due to the exaggerated welding sculpts. In doing this, you also eliminate any visible seams between different parts of the mesh.
Import alphas Once you are happy with the first
06 pass sculpt and you feel that the high-poly model is fully welded together, the next step is to use alphas to start applying rust and other kinds of detail to the model. You can find a variety of free alphas on the ZBrush website. When you download an alpha, you can import it by clicking Import in the Alpha Quick Pick tab on the left-hand side of the UI.
Use alphas to apply extra detail Once you have
07 imported an alpha it’s time to start using it. It’s
08 06
important to remember to check your Z Intensity, which is located at the top of the UI. Adjusting this enables you to increase or decrease the intensity of the alpha when applying it to your model. You can also choose to Zadd or Zsub. These options enable you to choose if you want to sculpt inwards or outwards from the model. When applying an alpha, you should make sure that you are using DragRect, which is located under the Strokes tab. This enables you to click on a section of the model and by dragging the mouse outwards, you are adjusting the scale/size of the alpha when being applied onto the model.
Decimate your model When you are fully content
08 with the final quality of the detailing pass inside of
ZBrush, it’s important to remember to decimate the subtools of the model, which contain millions of polygons. This not only decreases the size of your high-poly model on export to OBJ, but it also decreases the polygon count of that model too, which means that it wont take as long when baking your maps. You won’t lose any of the detail since ZBrush’s Decimation Master crunches the polygons around the less dense area, leaving the more dense, focused areas with the main polygons. You can do this by first clicking Pre-process Current under Decimation Master in the Zplugin tab. This process allows Decimation Master to analyse the subtool you wish to decimate and come up with the best algorithm for doing so. Once the pre-process has completed, you can then move onto decimating that subtool. This is done by adjusting the percentage of decimation and clicking Decimate Current. 75
MODEL & TEXTURE AN ASSET FOR UNREAL ENGINE 4
Export your model from ZBrush In this step,
09 we will export your finished high-poly model from
09
ZBrush. To do this, click Export All SubTools, located under Decimation Master in the Zplugin tab. This lets you export the entire model as one mesh converted into an OBJ file. This saves time without you having to export each subtool individually. ZBrush exports models with triangles instead of quads, so it’s always best to finish your detailing before doing so.
Unwrap your low-poly model In 3ds Max, select
10 your low-poly model and choose Unwrap UVW in
the modifier drop-down list. The next thing you want to do is click on Open UV Editor and then use the Projection options on the right-hand side of the UI to unwrap your low-poly model.
Bake your maps with xNormal Once you open
11 xNormal, you should click on the High definition meshes tab and then right-click and select Add meshes. Select your high-poly model and then repeat this process in the Low definition meshes tab also. In the Baking options adjust the size to be 2,048 x 2,048 and switch Antialiasing to 4x. This will result in really crisp projections. Select Normal map and AO map and then select 3D Viewer. If you click Launch Viewer it will load up your high-poly and low-poly models and enable you to adjust the cage accordingly to make sure it projects everything in the final bake. Make sure you save your cage before exiting the viewer and then hit Generate Maps to bake.
10
Bake your Color ID Map in 3ds Max Open 3ds
12 Max and select the high-poly model that you kept
before exporting to ZBrush. Open the Material Editor and apply colours to the mesh according to what materials you want to create. Once you are happy with these colours, make sure that the Self-Illumination Color tab is at 100%. This eliminates any shading in your Color ID bake, which makes it easier to apply materials to the mesh later in Quixel SUITE. Make sure the low-poly model is sitting in the centre of the high poly to capture all of the colour information correctly. Select Render To Texture, enable Projection Mapping, select all of the high-poly model’s meshes and click OK. A distorted cage will appear around the mesh; this can be fixed by hitting Reset under the cage tab on the right-hand side of the UI. Make sure objects and sub-objects are using the same existing channel. Select Diffuse map and change it to 2,048 x 2,048. Click Bake. This will now bake your Color ID Map according to which colours you have applied to the high-poly model. Time to take it into Quixel SUITE!
Optimise your UV space When you approach unwrapping an asset, it’s best to take into consideration the silhouette of the asset and adjust the islands according to the most visible sections – the most visible parts or most detailed parts of the mesh get the most UV space compared to the smaller parts of the mesh that aren’t as visible. It’s also important to take into account your smoothing groups. They should always be adjusted according to how many UV islands you have on your mesh. This avoids any visible seams appearing on your mesh after you bake your maps and texture the final asset.
76
12
11
15
14
Quixel SUITE’s 3DO rendering tool Quixel SUITE offers a great rendering tool called 3DO that enables you to take really crisp renders of your textured asset for your portfolio. 3DO lets you add post-processing effects as well as apply different filters, increase the sharpness and even change the cubemap within the render tool. You even have a turntable, too.
Open Quixel SUITE Head back to Quixel SUITE and
13 select DDO. This will bring up a small window that
allows you to select the low-poly model, the Color ID and also the bakes that you made prior to this. Once you have all the maps selected, change the resolution to 2,048 along with making sure the Texel Density is also set to 2,048. This enables you to get some nice renders in Quixel SUITE’s built-in render system 3DO. Select Unreal 4 under the Export Target box and then click Create.
13
Texture with Quixel SUITE Once you have loaded
14 everything up in Quixel SUITE, it’s time to texture!
Select which materials you would like to use under the Material Preset option and then click Apply. This will apply that material to the entire mesh; instead, click the Color ID that you want to assign that material to under the Color ID palette. Repeat this process for your entire mesh until it’s fully textured. Once you are fully happy with the texturing it’s time to add some extra detail such as dirt and rust.
Add extra detail This part is really fun. Select the
15 Material Browser and choose any kind of dirt or rusty metal that you would like to apply to the mesh. Don’t be afraid to increase the intensity on the albedo or normal map to really make those materials pop on the mesh. Once you have selected the materials of your liking, click on the Edit DynaMask option, which will bring default options showing how you can apply that material to the mesh. If you want to use these, select once and then select Apply Mask. This will then apply that material to the highlighted sections on the mask. Applying dirt to the areas of the AO bake is also a great way to ensure that it hits the correct areas.
77
MODEL & TEXTURE AN ASSET FOR UNREAL ENGINE 4
16
Save and export your materials It is now time to save and export those materials
for Unreal Engine 4. Once you have saved them, close Quixel SUITE and then open the PSD files for each of the materials. Save them as a TGA file but remember to invert the Roughness map. Unreal Engine 4’s Roughness map is read from black being too shiny to white being really rough. Once all the maps are exported as TGA files it’s time to import everything into Unreal. 16
James Brady I was first introduced to videogames by my father, who would sneak me into his office after hours to play old-school games such as Marathon or Syndicate on the classic Macintosh computers. I knew from this moment I was destined to be a videogame developer.
Harry Potter - Deathly Hallows Spellbook (Fan Art), 2016 3ds Max, ZBrush, Quixel SUITE, xNormal
Use UE’s post-processing to add after-effects
I’ve always wanted to do a personal art piece dedicated to Harry Potter. Thankfully I managed to get some time to make the Deathly Hallows spellbook!
In Unreal Engine 4, there is an option to add post-processing to your scene. This can be achieved by selecting the PostProcessVolume and dragging it into the scene. It offers loads of options ranging from changing the colour temperature to adding a vignette or even adding bloom to your scene. This is a great way to achieve the mood that you want to portray in your scene.
17
Silent Hill Homage - Unreal Engine 4, 2016 3ds Max, xNormal, Quixel SUITE, Unreal Engine I’ve been a huge fan of the Silent Hill franchise from the very first game, released in 1999. After the cancellation of the upcoming game, I felt compelled to make a homage in UE4.
17
Import your final asset into Unreal Engine 4 Once you have opened Unreal
Engine 4, import your low-poly FBX file along with the TGA material files that you saved above. Right-click in the Content Browser window and select New Material. Name it to your liking. Once you open the Material Editor window, drag all of your final materials into it and link them up to the correct nodes, for example Albedo map to Albedo, Normal map to Normal and so forth. Once this is complete, hit Save, then apply that material to your FBX file. Once you drag your FBX file into the scene, you will see the asset lit and fully textured in Engine.
78 All tutorial files can be downloaded from: filesilo.co.uk/3dartist
Rusty Garrett 50 Turbo, 2017 3ds Max, ZBrush, Substance Painter I used this personal art project to learn Substance Painter along with getting a better understanding of how metal decays under damp weather conditions. This piece was a lot of fun!
Discover another of our great bookazines From science and history to technology and crafts, there are dozens of Future bookazines to suit all tastes
Techniques
Our experts The best artists from around the world reveal specific CG techniques
Maya Mike Rutter www.thedigimonsters.com Director and cofounder of The Digi Monsters, a UK digital art studio for games, VFX and VR
YOUR
FREE
DOWNLOADS from filesilo.co.uk/3dartist • Tutorial screenshots • Maya file • SVG file
80
MAYA
Create a motion graphics ident C
hannel idents are often a cracking opportunity to stretch your creativity to the limits, showcase your skills and learn new techniques while producing striking and memorable visuals. An ident’s purpose is to portray the identity for a channel, helping to communicate the values and personality of the channel to the viewers. This helps the channel stand out and makes it recognisable and memorable through branding. In practice, this often means a wide, flexible brief for artists, so do make the most of it when these opportunities arise. Our client is Monster Four, a fictional streaming TV channel that is modern, punchy, cheeky and a little bit monstrous. The client is not afraid to try something abstract or visually edgy and it should appeal to a young-ish audience, aged 16 to 35. Our brand pack has a logo, light brand guidelines, suggested colour palette and a lot of flexibility. At the start of this process, consider what resources you may have at your disposal. It’s easy to get carried away with ideas worthy of a Pixar short, rather than something appropriate for the brief and achievable by the deadline. For this tutorial, we’re going to produce a short 3D animation, achievable by one artist with one machine. Our weapons of choice are going to be Maya 2017 using MASH networks and the Arnold Renderer, focusing on production
and rendering in Maya. We will convert the channel logo from the brand pack into a 3D asset that we can use within Maya. Maya 2017 ships with two very significant updates for us: the MASH Mograph plugins and the Arnold Renderer, both now as standard features. MASH uses procedural techniques more often found in Cinema 4D and Houdini, but now available to Maya-based studios. MASH networks can integrate with a lot of other Maya features, so it’s good to understand what they can achieve. Leveraging the immense power of Arnold for look dev, lighting and rendering, we will explore how these new features can help us develop our TV channel ident and deliver a professional look and feel. So we’re about to invite a bit of chaos into our production lives. By its nature, it can be difficult to completely re-create our results, but why not go further and change the models, logo and colours, or tweak the settings? It won’t take much to produce your own original piece.
01
Create a new Maya project Maya works best with
a standard project structure. Launch Maya 2017 and from the File menu select Project Window. From the window, click New next to Current Project. Set 3DA_Ident as the current project. Click on the folder nav icon next to Location. Decide where to locate your folder structure and hit Accept. Save
regularly and version up, but enable AutoSave as a last resort. Click on Preferences. Under Settings, select centimetres and 30 fps. In Files/Projects enable AutoSave, select 10-min intervals and choose Project. Select Save. Close Maya, then relaunch. If Maya crashes first, you will lose these settings.
02
Your first MASH network Let’s make sure
02 everything works. Check the MASH and Arnold plugins are loaded by going to Windows>Settings/ Preferences>Plugin Manager. Ensure mtoa.bundle and MASH. bundle are both set to Loaded and Auto Load. Maya crashes can unload these plugins, so you may need to repeat this step. Set your workspace to MASH. Create a Poly Platonic Solid from the Create menu>Polygon Primitives. Select Icosahedron from the Platonic Type dropdown and hit Create. Rename your shape Glows_Input_GEO. With the shape selected, from the MASH Editor or menu, select Create Network. Done. Well, not quite, but it will improve.
Assign Arnold material and preview In Outliner,
03 rename the new MASH_Repro to REPRO_Glows_
GEO. With this shape still selected, select Smooth Mesh Preview in the Shape tab in the Attribute Editor. Select Display Subdivisions, set Preview Levels to 2 and Render Levels to 4. Right-click in the viewport and select Assign New Materials. Assign aiStandard and on the Shader tab, click Presets>Ai_ Incandescent_Bulb>Replace. Select Arnold>Arnold RenderView. In the IPR window, hit the play icon and you should see your orbs. We now have an input shape, a MASH network and a resulting Repro Mesh. It’s building.
03
Plan your production Try to produce quick storyboards and descriptions, collect reference samples and think about what you’re trying to communicate every time. Even with the most flexible briefs, your ident should still develop over time; think Start, Middle and End. Try and block in your timings roughly in After Effects early, as this can give you a good outline for your key sequence frame ranges. To secure buy-in from the client and help co-ordinate teams, you may need to produce style frames or pitch materials to keep everyone aligned.
Use MASH networks Let’s look at a simple
04 MASH network. Create a poly sphere, set
subdivisions to 50, 50. Scale the sphere much bigger than the orbs. In the MASH Editor, select the Distribution node. In the Attribute Editor, up the Number of Points and set Distribution Type to Mesh. Scroll down and open the Mesh options. Middlemouse drag our sphere from Outliner to the Input Mesh field. Play with both the Method in the Mesh options and the Number of Points. On the MASH waiter, add a Random node. Dial up Position and Scale to see what happens. 01
04
81
TECHNIQUES
06
Create your actors We need to repeat steps 2 and
05 3, naming them ‘powers’ and ‘builders’. We want a
Scene lighting strategies
distinct visual read for each element. Instead of Incandescent, when creating a new material from the Presets menu select a shiny dark material for the ‘powers’, like Shiny Plastic, and a shiny polished metal for the ‘builders’, such as Wheel Rims. Keep organised and rename your materials, MASH networks and Repro Meshes.
Front lighting shows detail but will flatten your image, side lighting shows off a model’s structural form while back lighting will silhouette the model and be dramatic. Lighting from below can be sinister sinister,, whereas top-down lighting is more standard. When lighting your scenes, try and think about the mood you are trying to achieve. Colour variation in your lights is good, so try something warmer on the left and cooler on the right. White is very neutral and good for fill lighting. Try not to blow the highlights or have fully black shadows; you can tweak this in post.
Create the 3D logo In a new scene, go to Create>
06 SVG. Select Import from the Attribute menu tab and
navigate to MonsterBoss.svg. Set SVG size to 100. Under Geometry, set Curve Resolution to 20. Scroll down and set Extrusion Distance to 30. Enable Bevel, reduce Bevel Divisions to 2. Tune your bevel profile and size. We need a fairly dense mesh – we’re going to explode it! Set Extrude Divisions to 30 and turn Deformable Mesh on. Adjust the Deformable Mesh settings for a cleaner result. Save your mesh and import it into your main scene. MMB drag the logo into the MASH networks, replacing the sphere if needed.
07
Set the scene Create a simple stage with an infinity
07 curve and area lights for lighting. Create a Cube and
scale it up. Delete each face except the bottom and back. Select Mesh Display>Reverse to reverse normals, then hit 3 on your keyboard for a smooth mesh. Add edge loops to tighten the curve if desired. Select Arnold>Lights>Area Lights. Scale up the shape in the view, about the size of the logo. Head to Arnold> Arnold RenderView to start a new IPR session. You may need to really crank these values, I have intensity at 2000 and Exposure at 5. Light your scene as desired.
Tune your MASH networks Select the MASH
08 waiter for Glows. Under Mesh Distribution, change
the Method to Voxel. Under Voxel Size, set to 2 and Max Voxel Count to 200. On our Random node, set Position X and Z at 400, and Y at 100. Scale X, Y and Z at 20. In the Strength rollout, set Random Strength to 0.25. In Outliner, select glows_input_GEO and uniformly scale up and down until you’re happy. Select the MASH waiter for Powers. Under Distribution, set Number of Points to 200, the Mesh method to Scatter and Push Along Normal to 20. Adjust the Random node and tune as before. 05
82
08
09
Use trails to make connections Select the
09 ‘builder’ network, up the Amount to 400, then set the method to Scatter and Push Along Normal to 2. Add a Trails node to create the scaffolding, choose the mode ‘Connect to Nearest’, and set Max Trails 1000, Scale around 0.13 and Count to 4. Create a NURBS circle, MMB drag this from Outliner to Profile Curve, then set Samples to 9. Add a Trails node to the ‘powers’ network. Choose Connect to Nearest, set Max Trails to 100, Scale to 0.1 and Count to 1. Use the same Profile Curve.
Build and explode the logo In a new scene, import
10 the logo mesh. Duplicate it with Ctrl/Cmd+D. Create a
poly cube and from it, a MASH network. MMB drag the logo into the Mesh tab, set the Method to Face Centre and tick Flood Mesh. Hide both the logo and Repro, leaving the second logo mesh. Select the MASH waiter and add an Explode node. MMB the logo mesh into the Exploding Mesh box. Right-click and create a Falloff Object. Add a Random node, crank the position and strength. Scale the Falloff Object to encapsulate the logo, and animate its position. Alembic Cache the build.
10
Baking networks – Alembic Caches Adding more and more nodes and creating more and more geometry gradually slows down any scene and introduces instability. If you are happy with a network (or other animations/ simulations in general), a good step can be to create an Alembic Cache that can be used downstream. This ‘bakes’ the animations into a cache file that you can import or reference back into your scenes, without the complex networks needed to support the original animations. As they are much less complex for Maya to deal with, you generally regain scene performance and stability, and there is much less to go wrong.
11
Animate your sequence Adding Signal nodes to your
11 networks adds motion to your elements quickly. Try
different Signal Types and adding additional Signal nodes. If you have audio, explore the Audio node. Most values in the nodes are keyable, just right-click on a value to set keys. Animating the node Strength and Random Strength alters the influence of each node. Adjust these in the curve editor to perfect timings. Import your Alembic Cache to include your logo element. With the element selected, in Channel Box/Layer Editor select the Alembic input to expose the offset to get your timing.
Render and post After this whistle-stop introduction,
12 take the time to really explore your settings, as well as
the look and feel – keep rendering style frames. Don’t be afraid to experiment or step back again, embrace the chaos and layer up the complexity. Set up sweet camera angles and add motion to them. When you’re ready to render, select Render Settings and set your frame size. Change the Frame/Animation ext to something other than Single Frame. Edit your frame range. Adjust the Arnold samples if you’re getting a noisy image. Select Render>Render Sequence to render your sequence, repeating with extra cameras. 12
All tutorial files can be downloaded from:
filesilo.co.uk/3dartist 83
Subscription offer
Subscribe and save 20%
*
Every issue, delivered straight to your door Never miss an issue
Delivered to your home
Get the biggest savings
sure to get every single one
Free delivery of every issue, direct to your doorstep
Get your favourite magazine for
“@3DArtist We love using #blender &
“Just arrived! The print version is so
#cycles in our studio! Looking forward
to the new #3DArtistMagazine.”
get home to give it a good read.”
@Vibe_Studio via Twitter
@pablander via Twitter
“Really having a good time with issue 90 of @3DArtist magazine.” @cochesaurus via Twitter
Subscription offer
MOST
GREAT
Subscribe and save 20%
One year subscription
Automatic
renewal – never miss an issue
saving 20% on the retail price
Instruction to your Bank or Building Society to pay
direct to your doorstep
0 1 8
UK £72.70 (saving 20% on the retail price)
8
Europe
USA
Rest of the world
Visa
Mastercard
Amex
Postcode
Account Name
Expiry date Sort Code
Account no
Signature
I enclose a cheque for
Date
Imagine Publishing Ltd
£
Signature
Date
Your information Name
Address
Email address Postcode Please tick if you do not wish to receive any By post By telephone By email
Please tick if you do not wish to receive any promotional material from other companies By post By telephone
Please post this form to 3D Artist Subscriptions, 800 Guillat Avenue, Kent Science Park,
Order securely online www.imaginesubs.co.uk/3da Enter the promo code PS172 to get these great offers
Speak to one of our friendly customer service team Call 0844 249 0472
These offers will expire on Friday 30 June 2017
Please quote code PS172
WACOM MOBILESTUDIO PRO 16
Wacom MobileStudio Pro 16 How does this high-end tablet fare with a typical 3D artist’s workflow?
T
he core of Wacom’s focus is directed at the needs of digital content creators. It clearly understands the effective interaction between stylus and work surface that we need to work our magic. We will start with the perfection that is the new Pro Pen 2. It has the proven shape of Wacom pens you know and love, but packs more refined control. The Pro Pen 2 has 8,192 levels of pressure, 60 levels of tilt recognition within a 40-degree range and, being battery-free, never needs to be recharged. The work surface is a bright LED display with an aspect ratio of 16:9 matching that of 1080p movies. The total viewing angle is 178 degrees (89 degrees/89 degrees) so you can have it flat on your desk and see what you are working on. It can be tilted horizontally or vertically based on your composition demands; this also makes it perfect for right or left-handed use. You can work with the screen using the stylus, which should be second nature if you have ever used a Wacom tablet or Cintiq. Some of the parallax that was there in previous versions is virtually gone, because the tip of the pen looks like it is on the surface and the response is second to none. You also have the option of multi-touch controls which are super useful for zooming in and out, rotating images, typing with the on-screen keyboard and many other useful gestures. The MobileStudio Pro is a powerful portable workstation fuelled by sixth-generation Intel Core processors. Mine has the i7-6567U, which has plenty of power to run the core applications I use: ZBrush, Maya, Photoshop and KeyShot. I have also tried it with many other applications, and for the bulk of the work that I do, the MobileStudio Pro 16 handles better than any high-end laptop I have used to date. What also helps with 3D functionality is that the MobileStudio Pro 16 comes with a professional-grade graphics card from NVIDIA: the Quadro M1000M. The Pro 16 has 16GB of RAM, which is comfortable to work on for demanding projects. It also comes with two or three cameras: a frontfacing 5MP camera; a back-facing 8MP camera; and Intel RealSense R200 3D scanning camera. The latter is really handy for scanning objects in 3D for reference, or retopologising for use as 3D assets, or measurement. The battery life is comparable to high-end workstation-grade laptops – about two to three hours when using it for painting in Photoshop, or modelling in ZBrush or Maya; a typical tradeoff for
86
the ability to use powerful desktop applications and a full-fledged OS. The battery also charges quickly at about two hours for a full charge. As for portability, the MobileStudio snugly fits into most laptop bags that can handle 17-inch laptops, and the weight is 4.45 lbs (2,020g), which is easily manageable. With my experience of working on tablet computers since the debut of Windows for Pen Computing in 1993, I can finally say the dream of having the capability to do my work on the go has been realised with the Wacom MobileStudio Pro 16. This tablet makes no compromises and sets a milestone. The Cintiq Companion 2 had come close, but the MobileStudio Pro with the large screen hits every mark. As an instructor at two art and design colleges in LA, an author and a freelance concept designer, I use the MobileStudio Pro 16 daily to teach my classes, critique student work by zooming into areas of discussion, do paint-overs, modelling demos and write while on the road. I also use it when I am at a client location to show and discuss work or iterate on designs. At $3,000 it is priced on par with high-end laptops from Apple, Microsoft and others, however in addition to high-end laptop capability, you are getting the Wacom-perfected stylus interaction and a touch interface. Ara Kermanikian
The dream of having the capability to do my work on the go has been realised. The tablet sets a milestone
MAIN The MobileSudio Pro 16 is 0.75 inches thin and has a brilliant 15.6-inch UHD touch display for working and presenting on the go BOTTOM LEFT Perfect for use on the road; here I am on Naoshima Island in my hotel room at the Benesse House starting a project for one of my classes BOTTOM MIDDLE ZBrush base mesh of an exoplanet creature I am sculpting for a class BOTTOM RIGHT The newly enhanced Wacom Pro Pen 2 has 8,192 levels of pressure and comes with a futuristic case and three extra nibs
Essentialinfo £2,750/$3,000 wacom.com Windows 10 Pro (can be used as Cintiq attached to a Mac or Windows PC) 6th generation Intel Core i7-6567U CPU 512GB SSD, 16GB RAM Memory NVIDIA Quadro Graphics Card M1000M 4GB GDDR5, 16:9 UHD (3,840 x 2,160) Bluetooth 4.1, Wi-Fi 802.11ac Wireless B3 USB Type C Peripherals Cameras 8MP back facing, 5MP front facing, Intel RealSense R200 3D scanning camera Price Website OS
Summary Features Performance Design Value for money
Verdict The Wacom MobileStudio Pro 16 is the ideal tablet for mobile designers, artists and creative professionals
87
SCAN 3XS CLASSIC 3D
Scan 3XS Classic 3D A great all-round specification for a mid-range 3D design PC, but it’s the NVIDIA Quadro P4000 that steals the show
T
he entry-level workstation market is being turned on its head at present. Pricing for six and eight-core CPUs is being driven down, partly thanks to AMD’s aggressive pricing of its newly-launched Ryzen processors. The availability of a solid eight-core CPU at a considerably lower price point than anything else on the market has forced Intel to reduce its prices, too, making workstations more affordable across the board and bringing significantly more bang for the buck to entry-level systems. It’s also the season for new graphics cards, with NVIDIA dropping upgrades to its whole line with Pascal and AMD about to follow suit. It’s the first time we’ve had an NVIDIA Quadro P4000 on our test bench, using the same 16nm Pascal architecture that resulted in some impressive performance gains from the P5000. In keeping with the Quadro naming structure, this mid-range card is a single-slot 1,792-core graphics card with 8GB of GDDR5
88
memory (a slight downgrade from the GDDR5X on the P5000). Its 1,200MHz GPU clock rate is about 20 per cent lower than the P5000, but the overall gains from Pascal still mean some hefty rendering power, almost double the performance over the previous generation, from a quoted 2,600 to 5,300 GFLOPs, and almost doubling the texture fill rate to 134.6 GTexel/s. Scan has used an Intel Core i7-6800K processor in the 3XS Classic 3D, a six-core, 12-thread Broadwell-E processor with a standard base clock of 3.4GHz. A modest 4GHz overclock has been applied, perhaps less than you might expect from a typical gaming PC, in order to maintain reliable running. In a similar vein Scan has also opted for a high-end air cooler, in the form of a Noctua NH-U14S, rather than an all-in-one liquid cooler. With the Corsair Vengeance DDR4, 32GB memory comes as standard, with more available if needed as a buy-to-order option. A 240GB Intel 600p NVMe PCIe M.2 SSD came in our test rig, although Scan informs us these are likely to be swapped with better-performing Samsung drives in the near future. Nevertheless, the 600p delivers some superb results, with over 1,500MB/s read speeds, although it’s less capable at writing. A 2TB Seagate hard disk is included for additional storage space. As expected, the overall build quality meets all the standards we would demand from a professionally assembled workstation. Cables are tied neatly out of the way including all the data cables, but there’s still plenty of space for user upgrades inside the Corsair Carbide 330R Titanium chassis if needed. It’s generally very tidy with plenty of internal airflow. The Corsair case comes with noisedampening foam on both its side panels, which results in close-to-silent operation, at least under medium load during our test run. The combination of an efficient air cooler and reduced-noise chassis meant this system was perhaps slightly quieter than some water-cooled rigs we’ve seen, without the clunk of a pump kicking in when first turning it on, or the typical subtle sloshing of liquid. As we’ve seen before, an overclocked CPU can push a graphics card to new heights, and Scan’s 4GHz overclocked CPU really makes the most of the P4000, even performing better than the P5000 in HP’s Z640 rig in
The Corsair case comes with noise-dampening foam on both its side panels, which results in close-to-silent operation
some tests. ArionBench results, indicating CUDA performance, are also significantly higher than some of the M4000-based rigs we’ve seen, while LuxMark, which measures OpenCL, showed results roughly double those of the same reference system. Great stuff, mostly thanks to NVIDIA’s efforts with Pascal. As expected, CPU results in 3ds Max 2017 fall down compared with the ten-core 6950X systems that we’ve seen, particularly on the longer HDTV underwater render test that took an extra minute to complete. But having six cores still gives an advantage over a typical quad-core system, with the overclock driving 3D and CPU-bound tasks even further. It all adds up to a well-rounded system, performing well across the board and backed up by Scan’s typically great service. Orestis Bastounis
MAIN The sensible chassis choice quietens operation with noisedampening materials on the sides
BOTTOM LEFT A large air cooler can be as efficient as water or hydro cooling, and simpler as well
BELOW Six cores drive CPUbound rendering tasks in the Classic 3D
ABOVE A PCI Express SSD provides much faster transfer speeds than traditional SSDs
Essentialinfo Price
£2,499 inc VAT
Website
www.scan.co.uk
CPU
Intel Core i7-6800K
RAM
32GB DDR4
GPU
NVIDIA Quadro P4000 Intel 600p 2TB Seagate
SSD HDD
Summary Features Performance Design Value for money
Verdict Scan’s 3XS Classic 3D delivers where it counts. Great performance, fast storage and near-silent operation
89
CREATE THE IMPOSSIBLE www.photoshopcreative.co.uk
Available from all good newsagents and supermarkets
ON SALE NOW
Striking imagery Step-by-step guides Essential tutorials
PHOTO EDITING
DIGITAL PAINTING
PHOTO ART
TOOL GUIDES
BEGINNER TIPS
BUY YOUR ISSUE TODAY Print edition available at www.myfavouritemagazines.co.uk Digital edition available for iOS and Android Available on the following platforms
facebook.com/PhotoshopCreative
twitter.com/PshopCreative
The inside guide to industry news, VFX studios, expert opinions and the 3D community
092Community News
Festival film favourite VFX Find out how this award-winning short film used Blender and Cycles
Almost all computer-generated images in Circle were made with Blender, such as the bubbles, the campfire scene with all the grass and trees and a few set extensions like the wooden ceilings and walls
094 Industry News
Arnold 5 out now Solid Angle’s latest release is available, plus Ridley Scott’s RSA Films launches VR division
096Social
Alexander Heringer, Director, editor and producer of Circle
Readers’ Gallery 90
To advertise in The Hub please contact George Lucas on 01225 687311 or
[email protected]
The latest images created by the 3dartistonline.com community
91
COMMUNITY NEWS
The look dev of the memory bubbles took the longest time
How this film festival favourite blended live action with CG Director Alexander Heringer reveals how Blender, Maya and 3ds Max helped to add incredible emotional prowess to short film, Circle
A
hmad Saleh has suffered losses, personal losses. aim was to melt the reality and the bubbles together so that The conflict in the Middle East has been tragic for they didn’t appear out of place and take the viewer out of the Student Academy Award-winning screenwriter, the film. We wanted to make it possible for the audience to and he’s tried to cope using various creative methods. One be in the moment in order to reflect and connect their own such method resulted in a poem. Alexander Heringer met experiences with the story.” Saleh at Cannes in 2013, and it was from that point that Parts of the woman’s life can be seen in these floating Heringer and Saleh worked together, adapting the poem into spheres, and Heringer explains that it was Blender and a short film script called Circle. Cycles that helped create these incredible glowing bubbles: “Almost all computer-generated images in Circle were made Circle is the story of a woman who experiences a tragic loss and tries to escape into a seemingly perfect past. with Blender, such as the bubbles, the campfire scene with However, she ends up losing herself in vivid memories and all the grass and trees and a few set extensions like the starts to neglect her life, and her wooden ceilings and walls. Cycles perception of reality becomes gave us incredible render results. With distorted. “It is this painful feeling of its amazing real-time previews it was a constantly running in a circle, which pleasure to use, especially for the look inspired us to call the film ‘Circle’,” says development of the bubbles.” Heringer, who is director, producer and The VFX department consisted of editor on the short. just three 3D artists: Michael Circles are a predominant theme Wuensch, Mitja Mueller-Jend and throughout the film, both as a Johannes Wuensch. Heringer metaphor and a visual manifestation in explained which tools were used by ‘memory bubbles’. “We wanted a the artists where: “This might be certain dreaminess conveyed through difficult in a larger production, but for Alexander Heringer, the bubbles and, overall, a wellus it worked out quite well. Michael director, producer and balanced picture,” says Heringer. “The started creating the animatic with 3ds editor, Circle
Cycles gave us incredible render results. With its amazing realtime previews it was a pleasure to use, especially for the look development of the bubbles
92
60 computers were used over three months to render Circle
Max, simply because it’s the 3D animation software he is most familiar with. After finishing the previs, Michael handled all the 3D tracking in SynthEyes and compositing in Nuke. We tried to use Blender for the 3D tracking, but since it lacks features such as being able to detect zooming, we couldn’t really use it.” There were also other challenges in the film, Harbinger explains. “After quite a few experiments of simulating falling leaves for the opening of the short film, we realised that the desired particle simulation was hard to achieve in Blender. The capabilities of Blender’s particle system just weren’t good enough to get a convincing and natural look of falling leaves, and also didn’t have a decent control over their flow in the wind. So we handed the simulation over to Mitja who then used Maya. “Since another challenge was to make the long camera movements smooth to keep the viewer in the story, we stabilised all shots in post afterwards. Unfortunately, because of the long tracking shots, stabilising was challenging and sometimes we had to zoom deeper into the picture than we wanted to.” The film ran in 2016, collecting a string of festival awards, in cinematography and post-production. “They also applauded the performance of our main actress, the distinct style of atmospheric and emotional storytelling and the composition of the film,” says Heringer. “We received a review from an IMDB critic, which everyone on the team should be proud of: ‘the beauty of the technical delivery, and the delicateness of the music, performance and concept, all work together very well to produce a film that is achingly beautiful in so many regards, and really hard not to love.’” You can watch Circle at vimeo.com/210245233 .
Get in touch…
www.3dartistonline.com
@3DArtist
Facebook.com/3DArtistMagazine 93
INDUSTRY NEWS
The Sequence Group boosts its team The Australian office adds executive producer Melissa Benavides to its roster
The preview session at NAB 2017 was livestreamed to fans
Foundry previews Nuke 11 at NAB
The next release of the industry-standard compositor is demoed alongside new post-production cloud service Foundry has previewed the next version of Nuke and shown off new post-production cloud service Elara for the first time at NAB 2017. The initial look at Nuke 11 included new feature Timeline Disk Caching as well as improvements to Live Groups and Lens Distortion, plus a huge update to the latest industry standards via the VFX Reference Platform 2017.
Foundry introduces subscription model for Modo Following in the footsteps of recent industry trends, Foundry has added subscription plans for the first time to coincide with the release of Modo 11. At $59 a month or $599 annually, a subscription will let users access the latest versions, effectively budget their software costs and scale with project demands.
Elara was also debuted at NAB. The cloudbased service for post-production centralises infrastructure, creative tools and pipeline. Studios will be able to access Foundry and third-party applications, integrated cloud rendering, scalable storage and compute from their browser. Foundry demonstrated how easily a virtual studio can be set up with central pipeline, storage and resources, without worrying about up-front investment or management complexity. One demo illustrated how a VFX short film can be delivered through Elara using a wide variety of creative tools including Nuke Studio, Katana, Houdini and V-Ray. The session at NAB concluded with a panel on how cloud technology could change the VFX industry, hosted by Foundry CTO Jon Wadelton. Foundry also released Modo 11, with improved auto retopo features along with MeshFusion workflow enhancements, Mesh Light and occlusion baking improvements and much more.
Creative studio The Sequence Group has appointed Melissa Benavides as executive producer in its Melbourne facility. Benavides has a decade’s worth of experience across highend animation, experiential events, motion graphics and live-action VFX across the Australian creative industry. Achievements include Australian television’s largest live graphics package for the Logie Awards, the opening titles for the 30th anniversary of Neighbours and work on TV shows including the Catching Milat mini series and INXS: Never Tear Us Apart. Ian Kirby, founder and creative director at The Sequence Group, says of the appointment: “Melissa is an incredibly important addition to the Melbourne team. At Sequence, we don’t just produce one kind of content. From ice rink projections to hit show opening titles, we deliver across the board. Melissa is perfectly capable, no matter the project.” Melissa Benavides comments: “I’m thrilled to join the Sequence team, strengthening the output both in Australia and across the North Pacific at Sequence Vancouver. I look forward to delivering new and unique projects from animation to live action and beyond.”
Benavides has also previously worked with VR, live projections and experiential projects
HAVE YOU HEARD? The late Carrie Fisher will not be appearing in Star Wars: Episode IX whether in CG or previous footage
94
Anima 2.5 Material system has been revamped and AXYZ Design’s ready-posed characters are now supported AXYZ Design has given Anima 2.5 support for its Metropoly ready-posed and ready-rigged characters for renders and animation, respectively. All actors have also been visually upgraded: high-quality normal maps are now supported and lighting quality has been improved too. The Custom Actor and Motion Clip FBX have received a revamp along with the material system, with new presets including skin, leather, fabric, glass, hair and plastic. In addition, multimaterial restrictions have been removed. You can now also change the texture quality used by the 3D viewports depending on your system specifications, and actors can store multiple-quality versions of textures inside of resource files, improving performance. Visit secure.axyz-design.com for more info.
Render by Pavel Huerta. The classroom is part of a project he is working on to show off furniture designs
Arnold 5
thinkingParticles 6.5 out now
Solid Angle announces new version to coincide with the Autodesk release
New feature list includes procedural deformation
With the release of 3ds Max 2018, Autodesk has replaced the inclusion of mental ray in 3ds Max with Arnold. Also announced was a new release, Arnold 5, which has several new built-in physically accurate shaders: Standard Surface, Standard Hair, Standard Volume and Utility shaders. Arnold 5 also includes several sampling optimisations, including a new two-dimensional dithered sampling that improves the visual distribution of noise at lower sample rates, and significantly improves sampling results with soft shadows, indirect illumination and depth of field. Arnold is available standalone or as plugins for Maya, 3ds Max, Cinema 4D and Katana.
Software shorts Onion Packer This handy tool keeps all of your textures organised and ready to use with a built-in 3D texture preview and integration with all of your favourite 3D applications. You can drag and drop textures into your folders, which are then tagged for you automatically. With a search bar, you can find the right texture in seconds.
Cebas’ thinkingParticles version 6.5 introduces VolumeBreaker, ShapeNoise, FlowEmitter, ParticleLight, HydroField Flow as well as several updates. VolumeBreaker Clustering allows fragment particles to be set into different breakage clusters and appear more natural or have specific patterns. ShapeNoise provides procedural deformation and the power to add a deformation to any particle shape’s area and edges. FlowEmitter will make fluid effect animations consistent and allows for more flexibility. ParticleLight and the PLight Operator can turn any particle into a real Omni light type along with proper shadow casting. HydroField means more Fluid and Rigid Body ‘collisions’ with much better buoyancy and interaction. For more information, check out cebas.com.
With ithFlowEmi FlowEmitter, er,cons constant an flowcannowbe flow can now be achieved. achieved.Ima Image e courtesy of cebas Visual Technology. All Rights Reserved
Bringing you the lowdown on product updates and launches SceneTerrain 1.0
Replay for Modo
The Blender add-on for landscape generation is now considered stable with the main features in place. Population layers is a new addition, removing the previous default random weights that came with vertex groups in version 0.9. Two new tree species with three variations each have also been included, now with tweakable random colourisation.
This convenient new application creates macros and scripts in Modo without you having to write a single line of code. Replay records commands, which can be easily reordered by dragging and dropping. Macros can be played step by step for debugging and the finished result can then be converted into a Python script.
DID YOU KNOW? Ridley Scott’s company RSA Films has launched a virtual reality division, headed up by Jen Dennis
95
Share your work, view others’ and chat to other artists, all on our website
Register with us today at www.3dartistonline.com
Images of the month These are the 3D projects that have been awarded ‘Image of the week’ on 3DArtistOnline.com in the last month 01 Rhinoceros Beetle by Konstantin Gubry 3DA username kostia gubry Konstantin Gubry says: “The task was to show the beauty, subtlety and complexity of nature, and draw attention to the amazing world around us. Protect nature.” We say: Great depth of field, excellent lighting and intricate texturing make this render from Konstantin a winner. There’s a lot of talent on show here and the oblique angle really adds to the piece.
02 Pantone Style by Marcello Ganzerli 3DA username Lello.guru Marcello Ganzerli says: “The image represents my passion for interior design and pop art. I love Andy Warhol’s art, and during a design session I mixed models to obtain an image in his style.” We say: This is such a cool idea for a render – it’s quite extraordinary. The colours are great and the composition emulates Warhol’s style nicely to result in a really ar tistic concept scene.
Image of the month 01
03 Marceline by Harold Hernan Tovar Delgado 3DA username haroldtovar Harold Hernan Tovar Delgado says: “This 3D artwork is based on a Marceline illustration done by Visark. I really liked the way she looks in Visark’s concept so I decided to make a 3D version of it.” We say: This is a lovely adaptation from Harold that really does justice to Visark’s initial concept, with a nice hair groom and strong lighting adding to its considerable appeal.
04 C O N T A G I O N by Tomasz Artur Bolek 3DA username TomaszArturBolek Tomasz Artur Bolek says: “The main concept was to create a simplified urban structure for experiments with lighting and reflections. It was modelled in Blender and then rendered in KeyShot.” We say: We absolutely love clean, futuristic, hard-surface modelling like this scene by Tomasz – it’s very easy on the eye. Good choice of colours here and the application of materials is strong throughout.
03
96
Pug by Rhenan Fidelis 3DA username Rhenan Rhenan Fidelis says: “Personal work that I really enjoyed, practising character design and fur. Original concept by the great artist, Gil Rimmer.” We say: This is a really well-made, lighthearted cartoon render that’s full to the brim with character. It’s not always easy adapting cartoon designs into 3D renders, but Rhenan has done a superb job here.
02
HOWL by Herman Serrano
04
3DA username RawSunlight Herman Serrano says: “This was a commission to create a large surreal piece – something that would feel iconic on first glance but have lots of details for the viewer to discover.” We say: There certainly are loads of details to discover when you study this awesome scene. This is a really interesting render boasting massive detail and excellent colour choices. 97
LIGHTING
Artem Glazov artem_glazov.artstation.com
Incredible 3D artists take us
behind their artwork
LIGHTING I decided to make a bright Scandinavian bedroom where
outdoor lighting is combined with warm lamps and LED backlight. The interior lighting is a simple Corona light with a rectangle or disc shape, at 3,800 Kelvin for a warm light. A Corona light is used for outdoor lighting, but for a more natural-looking light I rotated this at an angle to the floor.
98
Artem is a 3D artist working in an architectural studio as a visualiser and designer Software 3ds Max, Corona
Renderer, Photoshop, Color Efex Pro
Scandinavian Bedroom, 2017
