Please Critique my work (be brutal)

Corweee · Apr 2020

I've been trying unsuccessfully to break into the game industry for years now and I've hit a wall. I've sent probably a hundred or more applications to game companies and usually either get a swift rejection letter or no response at all. The art tests I've managed to actually get (two) I've failed. I've been working in Maya and Zbrush for over 10 years now but obviously I'm doing things wrong so if anyone could pick apart my recent work and give me detailed feedback I'd really appreciate it. If not, all good too.

(I'm also aware that a lot of people are going through this as well, but I just figured asking for help couldn't hurt.)

Site: https://www.coreythomasart.com/
Artstation: https://www.artstation.com/corweee
Sketchfab: https://sketchfab.com/cwthomasart

These are all High Res screenshots from Unreal:

Render from Zbrush:

SketchFab:

model
model
model

Pirax · Apr 2020

I'd say your work is good, what kind of roles are you applying for, location also? Might help people here give more accurate advice :-)

Corweee · Apr 2020

I appreciate that Pirax! And I agree that's a good idea for context. So I've been mainly applying for character artist positions. I live in NYC and I'd say 80-90% of the jobs I applied for are on the west coast.

Joseph_Bramlett · Apr 2020

So a few things. Your website has to much on it to look at at a glance. I think your artstation compartmentalizes your work better and makes it easier to look at a t a glance. That being said I think a lot of your work on there should either be updated or removed as it doesnt seem of equal quality through out.

Without getting to lost in the weeds I think you have promising work. Your sculpts of the overwatch style characters are the best things in your portfolio. However that in itself kind of highlights a problem. Your textures arent as polished as they could be. A lot of generic grunge maps and kind of mismatched materials make characters like Red Fox come off as a lesser version of the Concept art where you should be aiming for your 3D model to be the definitive version.

With Red Fox specifically I think there are some short cuts being taken. The fur on the sleeves are pixilated when they could be hair cards, The clothing is a very simple sculpt when it could be simulated in Marvelous or have a more in depth sculpting pass. The face of this character is just slightly off with no expression in your renders. Overall everything feels low poly. Its something I struggle with but it makes the work come off as boxy and "gamy" and gets in the way of the look. This low res look carries over into the textures that also come off as rather low fidelity. For a portfolio piece I think 4k maps should be fine but your textures here arent quite taking advantage of that.

Your in a bad spot and its not because your work is so bad you should give up. Its because your work is so close to being good but theres just a bunch of little different things holding it back. Ide say go hard on one project. Find some really phenomenal concept or maybe even something to do a master study of and break it down into individual pieces. Make sure you plan out how to make each of those pieces the best they can be. Maybe it means learning a new program, maybe it means spending a month on look dev. But take a second and analyze what differences you see in your work in comparison to those who have the jobs you want. Compare your work to the overwatch models, hell see if you can find the overwatch models online and open them up in maya. See how they do it and then do it.

Corweee · Apr 2020

Thanks for the feedback Joseph! I agree that I should remove a lot of the older work off of my website. In regards to the Red Fox character the funny thing is I actually did download an Overwatch ripped model of Hanzo and studied it before working on this character, and on top of that I downloaded a character tutorial from Marc Brunet, who was a senior character artist for Blizzard specifically on Overwatch, to make absolutely sure I nailed the aesthetic because it was my first time attempting a stylized character like that. https://www.artstation.com/artwork/rvW4J . Maybe I just totally missed the mark on that one, haha. I've also been purposefully trying to keep the models and textures as optimized as I can while trying to keep as much detail as possible which is a hard line to walk. I definitely think going really hard on a character and not being too worried about the optimization so much as the final overall result couldn't hurt though.

pixelpatron · Apr 2020

You said be Brutal so here goes:

Red fox character has almost zero appeal factor, doesn't present well, and the void less expression shows little to no care about presentation.
The facets shown on the airship model really hurt it, also doesn't feel like a blimp or correct scale, lighting of this asset needs work.
Alien guy, again, appeal factor. head looks like a morphed sphere without much detail/work put into it. Not a fan of the design. Check out creaturebox designs. That's what a kickass cartoon style alien is supposed to look like.
WWE shotgun boy. Blackthorne's lame twin? Not very flattering of a character or pose. His arms look like an action figure, and this thing is startling realism and stylization in a weird way.
Wild West cyborg your strongest piece, lead with that.
Creepy monster is rad, more pieces like that. Telling a story, lighting is interesting, atmosphere and character.
Dragon looks like it's trying to do a crane kick, just looks dumb.
Cartoon style character, those feet look weird and the pose is like here! Loooky my feet. Looks like Sagat's lil cousin mixed with anime; not a fan.
Robot bust, could just be front and back
Weapons should be all in one area or grouped together
Drop all your 2D work, none of it is very strong, and actually brings down your entire portfolio (Only as good as your weakest image; and here is a full page of them). If you must keep, martini glass and pen and ink mask, and pencil zombie face portrait are decent but kinda look artschool studies work.
Apparel designs - drop this section. None of these will benefit you as a 3D artist too much, but there is a lot of unflattering work here. If you must keep, only conditions pain skeleton, CND on blue, conditions orange lightning bolt circle front only, and sphinx dog. Also don't crop these designs. Give them breathing room and negative space.

A few standouts, but a lot of average non-head turning work. Less is more and your portfolio is only as strong as your weakest piece. If it doesn't scream hire me I do badass shit! Then don't show it.

Corweee · Apr 2020

Thanks for the feedback pixelpatron. Most of the work you've referenced is old stuff that I'm going to be getting rid of, I was mainly focused on the newer stuff i.e. the images I provided in the post. I think for Red Fox I need to do new renders because these were before I did a full facial rig for him, he was more of a study on Overwatch style character design as well. The shirt designs were from when I was art directing for a merch store so I'm not really looking for feedback on that, but I agree I should probably get rid of that from the page. Also most of the 2D work has sold for money so I'm not sure about strength being an issue there, and again this was more about the 3D work. Most places I've applied to have "traditional art skills" as some level of requirement so I'll probably keep most of that in.

BagelHero · Apr 2020

I'm in the middle of writing up something more substantial, but I gotta disagree: your 2D items do nothing to help you. They are 1) mostly old and 2) amateurish. They bring down the entire portfolio and likely have potential employers questioning your judgment. Noone really cares if they were professional works or commissions if they don't look good and are 2+ years old at that. Nix them.

Corweee · Apr 2020

Thanks for the feedback BagelHero. All 2D work has been taken down and mostly all of my older 3D work. I also re-posed and re-rendered the Red Fox guy to hopefully give him more life/attitude. I personally think it looks better but ¯\_(ツ)_/¯.

BagelHero · Apr 2020

Hey, no problem. It was one of the more pressing aspects of your folio imho, now its easier to focus on the nitty gritty.

Your stuff feels like it isn't quite there and it's like... distressingly a little hard to pinpoint exactly why that is. On the plus side, that means you're really close! On the minus, that means its harder to figure out where you're going wrong. fwiw your sculpts are good! I did, though, notice that a lot of your prop models have pretty dire errors:

The gun... is definitely one of your better props! I would love to see the high poly, honestly! Unfortunately, the modelling and texturing both let it down.
The texturing isn't bad at all, but it's also not good. It looks like a first pass I guess. the materials are mostly looking good, but it's all very generic and maybe too weathered. It feels like its lacking in personality-- which you may not need or have time for when you're actually making a game, but you probably want for portfolio pieces. Toning down the weathering in areas all but where it's most handled or where shots have been fired, some hand-carved bespoke scratches and dents, and maybe some grease and fingerprints from use go a long way.

It also has... some concerning shading errors. I don't know if people have told you this recently, but it is probably better to have a little more geo than to use lots of long thin tris. If it didn't show up I wouldn't point it out, but this issue is showing up in nearly every one of your props and it shows in the bakes. I'm not even entirely sure what happened in that second example (the topo is mostly fine), but the shading errors definitely needed to be fixed before baking because it is unfortunately very visible.

Also, presuming this would be a first-person weapon; of ALL the places to skimp out and just float/clip some faceted geo, this was not it:

The apex crate study probably needs to go. Few reasons: if fanart isn't technically as good as or better than the in-game model, its probably a mistake to show it on your folio because it's much easier to directly compare and find faults with it. There's also a lot of pinching and other shading errors, one of which messed up some of the procedural maps and your normal map too (the bottom lip of the crate). this is a problem with a LOT of your props as mentioned, but it's super apparent here.The last reason is this:
To your credit, this is similar to how it's bashed together in-game. But it is likely that was out of necessity and you never usually see it this closely (and also the texture is cleaner), so you copying it and pointing it out is possibly not a great result of the study.
It's harder to cap, but the shading error issue is also visible on Red Fox's blade, hilt, and sheath. Makes it look blocky, but I also worry that's a bit inherent to the way you've interpreted the design. You need to figure out what is going wrong and fix your workflow to remove this error. It is alarming how often it shows up.

Moving on a bit, I'm going to focus on your characters individually for a bit. Starting with Red Fox, as it's a bit easier to directly compare to Marc's tutorial results & overwatch and see what doesn't look quite right

The skin on marcs looks really nice, yours looks a bit blown out. I think the albedo should really be a bit darker overall and more saturated, with more focus on the underlying color temperature. The baked lighting pass should be more subtle too, probably.
You really did need to exaggerate your usual forms a bit more to match the concept I think. A lot of the appeal is lost by toning it down. The proportions on the collar of the jacket, how he appears to have an incredibly puffed out chest at rest, the turned up sleeves... they all feel a bit 70% of the way there in your sculpt and model. Overwatch legs are also absurdly long, so altogether making some things more realistic it looks like... some normal dude cosplaying a knock off Overwatch character which is an odd look!
Studying some of overwatches materials might have helped a lot with bringing this together more; they rarely have extremely pure-looking metals, they're often frosty metals, and more often than not they're painted. Since this guy is very genji-esque, it's likely they were going for something like the exposed metal of genji's pecs-- not so much glossy or brushed. Paying a little closer attention to the hard surface areas and what makes them appealing I think also would have helped. A couple of things are just not quite right in shape or proportion and it does the model a disservice.
It would have been a nice touch to make the padded collar and sleeves like... actually fluffy. A combination of fuzziness on the base, then cards or layers to make it plush would have really been something worth having on your folio I think. Making it just hard plastic like this feels like a cop-out when you have the time to make it stand out. At least making it look really soft a-la original tracer would have at least been a style matching thing, but as it stands the sculpt is good in that area, then the final model is dull.
in conclusion: what other people said.

Assassin:

Kind of a bad look that it's one of your oldest characters and "your first post on Artstation" and its the first thing you show. Ideally, this should cycle out of your portfolio soon and be replaced with something even better
Weak points are the back hard surface and the fabric areas; strong points are the arms and upper chest.
Missing a lot of details from the concept if I'm thinking of the right one. It really did have some (excuse my french) really sexy proportions and shape language that you've lost, as well as the "feet" actually being digitigrade deer hooves (so, he's supposed to be standing on his toes). You took a lot of the concept too literally, the camera is positioned so we're looking down at his lower half. His legs would be a lot longer, and I think the front of his loincloth thing is symmetrical. Also, the pants area was probably intended to be something like tattsuke bakama, though granted seemingly without the pleats. Still, the type of fabric and the way it folds and sits is quite different to what you've gone for. The little things add up. I'm sure you know a lot of this by now which is exactly why it shouldn't be in your folio much longer if you haven't been hired.

Wild Westbot

Definitely your better character.
Leather jacket is gorgeous, the best piece of material definition on your folio without a doubt. the sculpt is maybe a little rough but you clearly show you can make it work during texturing, which is perfectly fine imho. Pants are also nice, even with the stitching looking at bit off on the crotch.
Design of the feet/"shoes" kind of lets the rest down a bit but compared to the other errors common up to this point this is like incredibly easy to look past. My brain keeps wanting to see something like this, and I feel like it would be easier to come up with a design that rigs up and moves more realistically too without having to either be inflexible or bend. the key part is the way the upper foot is a straight line, I guess it needs either a curve or to have two straights meeting where the arrow is pointing. Anyway, too late now probably; something to think about in the future is all.
If all your heads are this elongated in profile, you probably need to bring up some more proportion refs while sculpting. He's a robot, so it's not a problem here really, but if this is something that gets repeated in all of your characters it would be a concern.
not sure about the weathering or color blocking on the hand or torso. Also feels like the tubes on the torso could have been baked down to simpler geo, too. Nitpicking stuff mostly.

Dragon Hero

Not sure it should be on your folio anymore, but darn if it isn't cute. I'm kinda 50/50 on this one because the concept and your re-interpretation are sweet, but it's old and technically speaking not... great. Maybe you could just keep revisiting this once a decade haha. I'd also be really happy to see this up like... on your personal site or a blog or twitter or something. It's definitely something you should have SOMEWHERE I'm just not sure here is the place.

Some other points that I thought about while scrolling:

Your proportions and anatomy might need a little bit of work overall. Your torsos feel a tad long, and it makes the characters seem very top-heavy in an unbalanced way. This is more of something to check yourself on the next time you're making a character. A pull-up-ref-just-to-be-sure, kind of thing.
It might be nice to see more human faces (different ages? Races? Genders?), right now you only really have the one and it's a little like... Do you have a problem with making more than one kind of face? I don't know just looking at your folio, so it might be a good idea to try and make room for more kinds of characters.
Your lighting could kind of use some oomph. It's fine but maybe do a bit more color grading and experimentation before settling on a final. Wild Westbot and the Combat Shotty are definitely the better of the bunch.
Really studying some more out-there styles and making notes on them might help you with sculpting/modeling what you see in concepts, instead of what your brain thinks you see (the same way a painter may have to study color to see that a tree they're painting is less bright green and brown, and more a varied palette of warm greys, oranges, and vibrant yellows; we may have to train our brains out of sculpting "human leg", or "generic finger" when we're seeing something much more highly stylised).
You just need more stuff, and, more importantly, to take more care while doing that stuff. It might be good to nail down what kind of studio you're going for (I get the vague feeling... sci-fi semi-realism is where you're headed :P but hey, I could be wrong! ) and really try and nail down what they do, then do a bunch of that.
Right now its all feeling a bit sparse, so when I see "2 years ago" part of me is like... why isn't there anything newer? Did he peak 2 years ago? How do I even know he can still make this? Realistically I'm hoping you're just busy with life and work since you're not having luck getting an industry job. But a batch of really new stuff that's better than the old stuff would go a long way just in proving that you're not stuck.

I'd like to see you around here more too. If you are able to put aside time to do a longer project again, maybe people here can help you really polish it while you're still working through it, rather than a year after the fact. It's also important to point out that even though this is a huge, intimidating wall of a post: it's like this because I had to dig a LOT harder to try and find why people could be passing over you. I ended up picking up on a lot of little things instead of one, easy, big thing. It's not hopeless at all, and I wanna see more from you. Hoping this current crisis with covid and stuff passes over soon so you have a fair shot again. Good luck, and hope something in this mess can help!

Corweee · Apr 2020

Thanks for the detailed and honest feedback Bagel, it's definitely appreciated.

Wow a lot to unpack here haha:

The shading issues. Usually this happens with triangles at poles and usually I don't know how to fix it without adding in more geometry which I typically don't want to do for the sake of poly count. I've tried using smoothing groups with hard and soft edges and splitting the UVs along those hard edges which usually fixes these issues but then I get jagged edges in the bake from substance. This might also have to do with the way Maya smooths edges? I'm not sure. If anyone has a work around I'm all ears. Typically I just try to put them in places that aren't overly noticeable and ignore them if they are on faces that are flush because they always get resolved from the normal bake.
Usually when I'm retopologizing I try to minimize geo at all costs. Might not be doing me as much service as I thought.
Ok so the shotgun. That was more of a speed exercise for me, I did the whole thing in about 3 days. I definitely could have spent more time on the textures but I was just trying to nail the overall concept which was actually pretty plain for the most part. I think for the triangle issues here I also just missed a couple spots that needed a beveled edge, which is typically around curved or cylindrical surfaces. I might go back and adjust that on this one.
Loot crate. This was also more of an exercise. I went into Apex and took a bunch of screen-grabs of the asset and tried to re-create it as closely as I could. I zoomed into the smaller pieces to show the detail of the mesh but the texture definitely does look shitty up close, so I'll probably take that render out. I might keep this piece in just because in the YouTube vid it shows that I can also do all the of the particle and shader effects of the prop in Unreal as well.
Red Fox. For the skin that might just be my post process and lighting, because the texture itself isn't very blown out, and I used a lot of the techniques he was using
For the materials I didn't want to do a 100% knock off of Overwatch but my own take on it, that's why I pushed the PBR materials a little harder, in the new renders I just did I did tone them down a lot more and I'll admit it looks much better.
for the jacket fluff I tried to make it look as much like the bomber jacket fluff as I could, I even added a slight emissive and I think even a fresnel pass in my shader, maybe I need to intensify it more to get the result.
. For the proportions I actually tried to match soldier 76's proportions because the character designs were similar, but while also referencing human anatomy proportions which might seem why the legs look a little shorter. I've noticed that the Overwatch concepts are actually a little different from the final result in terms of how they interpret some of the proportions, plus in terms of actual human proportions the upper body is shorter while the legs are elongated, 3 heads from the top verses 3 1/3. (This is an older version of the project btw.)
Assassin. I actually show this first because it has almost 100 likes and was the most popular on my ArtStation, but I totally get what you're saying. I referenced Hanzo directly for the proportions so again I think there might be a disconnect from the way the concept translates into 3D as far as the proportions are concerned. The fabric is asymmetrical in the concept and there wasn't anything for the back so I had to make it up myself. (This is an older version of the project btw.)
WildWest character. This was for an Art Station challenge and I had more of a direction as far as polycount and texture resolution and map count was concerned. This is also the only one using my own original concept. The boots were hard to manage because I did try to make them look as close to cowboy boots as possible but also have them look mechanized and not like I put a metal shader on cowboy boots so that was definitely a challenge. I did have animation in mind and made an attempt to have a split near where the toes were but it still might not have been enough.
Dragon hero is my oldest piece. I've contemplated taking this out completely, the proportions are really bad but I feel like the hard surface is really strong so it's a hard one to judge. Maybe I'll take it out for now and then revisit it again in another 10 years haha.
For anatomy I have a HUGE folder of anatomy ref and many different model references. I would hope after 10 years my anatomy would be pretty solid but I guess there's always room for improvement.
For faces I actually have a couple faces I've been working on. I'm revisiting a portrait of David Bowie I started a while ago and I just started work on an African American male face so that will be up at some point. I agree they need to be more diverse.
As for lighting I have an Unreal scene I've built out specifically for showing characters and it's the same for Red Fox, Wild West Robot and the Combat Shotgun haha. I adjust the light positions depending on the model but that's about it. Might need to dig deeper into that as well.
As far as my general modeling I completely agree I should do more, and more out there stuff. Currently I don't have a "style" or full direction per se, I just see something I like and take it as a fun challenge and think "I can probably model this pretty well." I would like to do more Red Fox style stuff as well. I like that from the concept I chose he was part of a team so maybe I could try and add on to that with other characters of my own design to see what I could come up with.

Your assumptions are also totally correct as far as the sparseness of my work is concerned. I've been extremely busy the past 3 or 4 years. I work at a small agency in NY as a Technical Director doing mostly AR/VR and I'm basically the work horse. Most of this stuff was done in my limited free time. I love character design but I don't get to do much of it at work unfortunately. I'm definitely trying to do as much as I can in this short amount of free time I currently have.

To sum up, thank you. I'm actually humbled you took the time to write such a detailed and honest post about my work. I haven't had a good critique in a long time. Back when I first started I was almost exclusively posting in ZbrushCentral forums and I think this may actually be my first post on polycount in regards to my work. The community here seems way more active and engaging which is refreshing. I will definitely be sharing my work here more often and hope to see you pop up in my posts in the future!

Please Critique my work (be brutal)

Replies