5 Portfolio Review Tips Before GDC
quad damage
So. I review over 100 portfolios a year (in the past year the number is actually more like 500), either at GDC, online, or for job applications. Every GDC I take part in the "Killer Portfolio or Portfolio Killer?" panel and review session within the education summit, and run into some very common issues. Beyond the content inside the portfolio, there are many issues you can correct to optimize your review and make your work be as strong as possible.
I had posted this on Twitter, but wanted to share here as well. Here is the summary of those common issues I see in students portfolios on a regular basis:
I had posted this on Twitter, but wanted to share here as well. Here is the summary of those common issues I see in students portfolios on a regular basis:
Replies
Does polycount matter these days ? I mean characters with 25k tris vs 80-100k tris.
Of course, if you apply for Blizzard 100k tris characters it doesnt make sense, I mean overall
Thanks for the tips!
I think that is definitely part of what type of market you are targeting. Another problem I often see is that students often have work that are way below industry standard on a technical level. Without wanting to get into a rant, I really think that comes from archaic lessons being taught in schools and not so much a misunderstanding as much as it as misinformation. It's totally fine to have a portfolio of lighter meshes - but that is usually linked with handpainted textures and something like a MOBA or mobile type games. Then, it's fine to have something a little more dense, but it also needs to implement other common practices that justify that level of geo (pbr shaders, tiling materials that need geo support, normal maps that need geo for cleaner bakes, etc.) What generally looks like garbage is when the pairing of these things fall apart...like if you have a low rez model trying to support higher end techniques, or a dense model with baked in lighting and handpainted textures.
So, I guess my answer is, Yes it does matter. BUT. Context matters as well. If you're aiming to work in mobile or lower end games, lower polycounts are ideal but most likely need to be paired with assets of that caliber. If you're aiming at a higher end market, a high polycount doesn't matter too much (like 100k range) as long as everything else is to that level.
Edit> Beyond that, it does matter to show some level of restraint. Tying into the "its not a zbrush conference" comment, you need to show that you understand there are budgets even if they are a bit high these days. So, if you're aiming high, like Uncharted level, and its a 400k mesh...that's no good
First of thank you very much for the advice,
I had some questions because I think this is a very interestingly topic. I'm currently in my second year of my game development study(a very bad one haha) But I've been practicing a lot of character work for the past year. Learned a lot of software And now I think I'm ready to really start working on some characters for my portfolio.
When you talk about a lot of students being below industry standard on a technical level. I was hoping you could tell me a little bit more about this. Because I'm afraid I'm in the group of most. I know some of the more basic stuff like beveling an edge so you don't have a normal map seam when it's needed. keeping your texel density uniformly except for important parts like the face you could go a bit higher?
I was hoping that you could tell me some important technical things the students should know but mostly don't. Also I was wondering if you need to show Uvs when sending/showing a portfolio because I barely see anyone who shows them on artstation.
And just something I'm curious about is what do you hate the most what people do in their portfolio.
We had a talk about portfolios in class my teacher tried to convince us to put as much as you can on our portfolio. Even a fkng low poly vase and lots of your first work to show progression. Just thought it was a funny thing to share haha any other students with "good" advice from their teachers?
That advice you were given is probably the worst possible thing to tell someone and, honestly, is ridiculously common. If I had to guess, it's from wanting students to show that they learned something, that they have the relevant skills...the problem is that it just becomes noise. You end up with portfolios full of the same vase, the same swords, basic heads, etc. You mix all of that together, and rather than showing you can do everything, you only show that you can do many things just okay. In my years of doing this, I've never ever met another hiring manager or portfolio reviewer who said the opposite. It's always less is more and focus on one strong subject.
My biggest pet peeve, really, is not having game art in a game art portfolio. It's incredibly hard to even consider you for a role in a game company, usually an entry level job, if you've never textured anything. Plus, Game art is my world...i think being a game developer comes way before being a great sculptor because of the mind set you have. The technical interest and different skills it takes to make that pretty sculpt walk around in a game engine. Sculpting is fun, probably the most fun part of the job, but that's all you have it looks like it will be incredibly hard to fit you into the development process.
On the technical side, a lot of it comes from the way students are taught and how the system is set up in my opinion. I don't mean to be insulting, but, in order to be a full time teacher you basically need to drop out of current production to be effective..so...you end up teaching the last thing you knew, which is probably a step behind what everyone else is doing, meaning the students are catching up to old technology and best practices. Most of the time, the basics are there...you know how to bake maps, the general idea behind modeling, making textures, etc. But it's always behind. I rarely see a person's portfolio that is trying to break into the industry that are fully using modern techniques...it's usually like last gen work with a nod towards the future, but a general misunderstanding.
It's hard to be specific, but using my own barrier of entry for what I look for on my own team for an entry level position:
-Understand of pbr materials using metalness or spec/gloss scale.
-Proper mesh density to support deformations, have interesting silhouette, and create cleaner bakes
-Creation of more advanced objects, like hair and eyeballs.
-Proper fold work that is game friendly (distorts properly, works in different poses.
-Overall appealing character design (iconic silhouette at a distance, use of color, clearly defined purpose)
At the end of the day, I don't care about software. I can personally teach you, or get you trained up on ZBrush, Substance, Marvelous, etc. but if you lack the basic understanding it's a way bigger mountain to climb.
UVs, I dunno, I don't care much about them...especially now when most people paint in 3d anyway. But, it is important to know proper seam placement, and things like that...and stretched textures will be obvious even without seeing your unwrap.
Thanks for the reply,
I will definitely listen to the podcast today. When you talk about proper fold work that is game friendly could you tell me a bit more about that I've heard that folds should be placed at the sleeves going down with a slight curve(no idea how to explain it better haha. And that you shouldn't add big folds on pants except for the: crotch/knees/end and butt area? ofc with the exception of baggy clothes. But How do properly distort clothes I've never heard anything about that yet but it sounds really need.
But here is something I'm really worried about and that is Proper mesh density to support deformations. I have barely any idea what do when it comes to geometry for supporting deformations. Which I feel like is a really big problem for a character artist. I know some of the really basic stuff like, Add loops around the arms and knees and other joints. Try to follow human anatomy for edge flow. But It's still a really gray area for me I was hoping you had some tips about it.
Lately I've been considering to learning how to properly rig and skin characters. But I think that's going take some serious time, and I only have around 9 months left to get my portfolio in order to start finding an internship. But I also think this will be a really useful thing to learn so I can see what work and what don't. Or is this a really bad idea to spend time on and should read through certain guides or watch videos about it?
Also so for all the questions I'm trying to leech as much need information out of you now that I have the change haha.
Yeah if any change there is even one dutch guy that reads this don't go to deltion college. Second year and we haven't even talked about baking yet or a more advanced modeling way than extruding.
No worries.
For proper fold placement, it's really about what the character will be doing in game, but generally you want to properly convey tension and gravity through your garments as well as create a believable silhouette and general clothing construction (i.e. your shoulder seam is properly placed.) Generally speaking, most "bunching" are going to happen at points of articulation...like elbows, knees, waist....but there are lesser folds that gives character (like zig zag / compression) to the clothing and is the result of the more major compression happening at the joints.
Density depends on the level of detail the character needs to be, there really isn't a magic number, but at the very least you want 3 loops per joint (one per each side, and one in the middle of the joint) I usually go a bit higher and even add cuts int he geo to retain the fold once a leg or arm bends.
If you want, take a look at my Gumroad. There's some free models you can download to get some ideas. They're old, but the general idea is the same, just needs more supporting geometry in most cases: https://gumroad.com/gavingoulden
Rigging, I think it's useful to know enough to teach yourself about proper modeling, but it's a complete waste of time (right now) If you're not planning to do it as a job. Character TD's are rare...being a rigger isn't what most people "want" to do...so if you want to do that, focus on rigging and become really good at it. If you're juggling character modeling and rigging, expecting both skills to be at an employable level...one or both of them are going to suffer. It's best to pick one and become good at it. That being said, I (and probably lots of game character artists) have rigged their own stuff...but thats usually like a simple FK rig and not a complex system that takes a lot of time. My point is, though, 9 months isn't a lot of time...if you're just getting comfortable with character work, and it's what you want to do, focus your time on that because right now any distraction is going to reduce how much you improve that one major skill.
Thanks a lot for the time spent on answering questions!
I'd like to ask, does it make sense and / or help to include something done with pre- pre- pre last gen technology(I mean pencil & paper) into portfolio as character artist?
Some pieces like this:
( two iterations, so that none suspects that it's been overlayed on photos )
For your question, you might not like the answer I have. Personally, I scroll past these types of pieces when reviewing character art candidates. While the skills of understanding anatomy and translating that in your art certainly isn't lost, it just doesn't have much bearing on the job itself. For example, I'm about 15 years in and have never worked with a character artist who used traditional art on the job or had been tasked to. Painting textures, sketching out ideas, drawing logos..sure..but never like actual portrait or life drawing work that is more involved. I think some sort of artistic ability is just assumed when seeing a good portfolio of game art. At best, it's a nice to have, but to me it's usually noise and would rather see more in game art.
That being said, some art directors enjoy seeing sketches as it shows the thought process and - really - more accurately resembles the bulk of work you would do with them. If you were going for a concept art type role, it makes more sense, but if you're trying to land a job, it's best to stay focused and keep it more straightforward.
While I feel like I've progressed way beyond what I was taught at university, I'm still certain that there are a ton of things I should know.
Do you have any tips on how someone outside of the industry can keep oneself updated (to avoid this problem)?
Hey,
I took a look at your Gumroad and downloaded most of it I also watched your podcast I would recommend everyone to watch it very helpful and fun to listen to. I was looking at your Argonaut model (looks amazing) and I had two questions about it.
1: You had all the stitches in Zbrush in the hp was this just for the hp render or did you bake them down in the normal? And if you did wouldn't it be easier to do it in the texture so you don't have to paint every stitch in the albedo?
2: What software do you use
And a technical question I've got this idea in my head and have no idea if it's correct. When you are working with a character, for example, a ripped shirt. Should the ripped shirt be together with other parts that need an opacity map? Or is it okay to combine it with other parts that don't need an opacity map.
@Sunray -
1. There's a few different ways to make stitches...a) You can model them out and bake them down into your low poly model - you can use the stitch vertex color as a mask to avoid painting each one by hand and essentially just apply a material to them. b)You can just do it in the texture by hand..to me it's time consuming and can be more inaccurate since the height and detail of it is relying on 2d information instead of a 3d source. c) You can do a tiling stitch texture and apply it to a floating model over your base one, or as a 2nd UV set in your base model (but that takes special work if your viewer isn't set up to render 2nd uv sets in that way...)
2. Get outta here.
3. Generally, your renderer is going to be analyzing chunks of your model and the materials assigned to it. If a larger chunk has a material assigned to it with alpha enabled, the entire chunk is still processed as it's being checked "is this part invisible or opaque?" So, because of that, it's better to keep alpha objects split from ones that don't need that information - you may be causing a performance hit in other ways (like adding a separate material) but you gain more by reducing that process on screen. Sometimes, it's unavoidable...like if you have a ripped up shirt, it doesn't make much sense to have matches of an alpha material mixed in..or a material for the sleeves. But. If you had something like a jacket with a fur collar, you may want to put that fur collar on a separate material so only that part is being processed.
Thanks for all the info you've given.
I've been a big fan of watching your work, Jon and Layna's for quite a while. I attended the Gnomon event you had down in Hollywood last year. Since then, been huge fan of following, growing and just wanted to say keep it up!
Anyway, do you agree with this approach for jobs?
Not knowing artists and only letting your art speak for you? Posting online (ArtStation, Polycount and etc)? Being humble enough that companies just come to you?
Coming back from living in LA, I won't have a method of networking with artists like I did at Gnomon. I also can't go to conventions like GDC for quite a while. I am curious how you might suggest artists like me might make it not knowing anyone at any company in an over saturated industry? I live in a city where there is 0 industry. I've been given art tests from AAA, indie, interviewed and such all over. Basically where I knew no one.
My situation is I keep succeeding in industry, getting a job, shipping titles, then the contract ends and it's back to McDonald's and back and forth. Short contract or internship successes. I have some industry experience success but being a junior makes me maybe less trustworthy compared to a senior?
Anyway, how do you make it longer term if you know no one?
Definitely keep it up all on my own improving my portfolio and studying.
Really, that's a tough spot to be in, and I can relate to it. Years ago, I had moved from a town with zero industry to Vancouver which, at the time, was a pretty big hub in the games industry.It wasn't easy, though, like you mentioned I basically worked odd jobs for a year or two in between low paying contracts to make ends meet. The best thing I can say, since everyone's situation is different, is to treat yourself like a brand....get your name out there, post work everywhere you possibly can, and update regularly. The more active you are on social media, here, artstation, etc. the more you will get noticed and the more opportunities should come. So, while I don't agree that you should wait for the jobs to come to you, I do think you should promote yourself as if you are trying to gain clients...in this case, I would say "Don't be humble" and really make an effort to promote your work.
From personal experience, that's basically what I did. When moved, I knew zero people. Literally no one, and at the time there were no resources like there are now. I just got my name out there...constantly...did the online groups, posted here a lot, took part in competitions, wrote articles on process, etc. Basically anything to stay active and keep my work "out there"
The only thing I will suggest is even if you're going for a particular role, include game art outside of that role only if it meets the quality bar of the studio you're applying for.
When I applied for a weapons art position I had an environment and a concept character bust in my portfolio. It only seemed to be a positive in my case.