Reached a milestone with my work Feedback appreciated!

Stoozy89

Hi there, my name is Stuart Wilson and I graduated from The University of Abertay Dundee, Scotland in 2012. Since then I have been working on my skills as a 3d artist. I have been working from the ground up and making sure that I fundamentally have a strong understanding of all the aspects of 3d artwork and continue to develop on my work.

Below I have a link to my portfolio, I'm needing feedback to help with my progress. Please don't hesitate to point anything negative as well as positive, this is all a working progress and any feedback is good feedback.

Portfolio: http://stuartboydwilson.co.uk/

Much appreciated,

Stuart.

Shiv

Compression is completely killing all of these images, cant see any detail in them till you click on them, I reaaalllly would go over and tidy that bit up first. It is your first impression atm.

Stoozy89

Do you mean the images or the video? If you have any tips that would be really appreciated, thanks!

DWalker

The lighting is really killing the images. The ambient is far too high, giving everything a flat, washed-out appearance.

The pistol has far too many polys that aren't contributing to the shape; given what you have, I'd put 6k as reasonable. Also, the bullets are in backwards.

The car's topology is odd. The front quarter panel is especially troubled, and the density of polygons in front of the door is too high. The front bumper could use some smoothing; it's the only angular part of the car.

Stoozy89

Yeah I agree with your point's, the trouble in terms of topology was when I was importing objects into UDK they weren't smoothing properly; may have been the setting's I was using at the time

Stoozy89

I basically had to smooth the object in Maya first which doubled the poly count pretty much.

Stoozy89

I'm grateful for the feedback, thanks! Do you have a recommendation for polycount limit for a vehicle game asset?

DWalker

Polycount for anything really depends upon platform, genre, and size. Racing games like Forza on Xbox One are pushing 1 million polygons/vehicle, while most Android games are still in the 1-2k range. As soon a possible, render the actual vehicle in an actual scene using the engine on the platform; if you're a hobbyist/student, pick an engine and use that for your tests.

In truth, though, polycount is no longer the limiting factor it once was; the visible difference between a 1 million polygon object and one with only 100k is far less apparent than the difference between a 1k object and a 10k one.

(Note that the above image is from an auto-decimating program; most artists would produce much better low-polygon models.)

In many ways, the problem facing modern artists is knowing when to stop; do you really need to model the tumblers in the driver's side door lock? Again, use the engine; look at the vehicle from the expected camera distance and the highest resolution and critically evaluate which details are important, or are even visible. This applies to both geometry and textures. Just because you CAN use 1 million polygons doesn't mean that you must or should.

Nosslak

The garage scene looks extremely similar to the one in Tor Fricks UDK Modular Masterclass. When I watch tutorials I try to learn the techniques and apply them to my own project. I would never put something on my portfolio that is almost the exact same as what has been shown in a tutorial. Also employers will probably have seen this scene many times (potentially better made) so they won't be impressed by it.

I like the models for the car and the plane but they are very wasteful topologically. You really need to learn to optimize your models and not waste polygons where they won't change the silhouette.

You also need to learn how to do specular and gloss maps pronto as without those you can't get any material definition which means you'll never get hired for texturing (at least for realistic games). You could maybe get by without them for very stylized art styles like the ones at the bottom, but I'd still say you should learn them. The cars material was pretty nice but for all I know the rest could've just as well have been made out of concrete as there was no shinies anywhere else.

I think you really need to work on your color theory as well as several of your models had pretty bad color schemes (mostly the plane and gun). The problem with the plane is that the colors are way too saturated and don't fit together too well. The gun has the opposite problem where almost everything is grey, you could've at least made the grip plastic brown or something. This is a great resource for designing color schemes.

I would strip out the gun completely as the design is kind of unappealing IMO, it has zero material definition and the dragon motif is IMO just plain ugly.

The stylized assets at the bottom and the car are IMO the strongest pieces at the moment but they need to be more optimized. The garage scene is pretty good but the only thing that shows an employer is that you can follow a video-tutorial that holds your hand for every step of the way. Deploy the knowledge you gained on something new instead.

Lastly I would also throw out the image system you've got for you're site as it's a disgrace. I had trouble figuring out how to get back from clicking an image which is never good.

GragGunslinger

The spec maps on all your props are sort of weak, and wind up making everything look very cartoonish, but the proportions don't match the material style, For example, the gun has no reflections or any wear and tear in the spec map and looks like something out of a cartoon.

Also your wireframe is really betraying you there, I would imagine it would be significantly lower poly. IMO it's your weakest piece.

Similarly your plane as a lot of the same issues as the gun, material definition etc. I agree your stylized assets and car are your strongest pieces.

I would recommend you try to get your assets into a game engine, mental ray renders and out of engine renders don't really tell me anything about your game art workflow.

Stoozy89

Thank you for all the feedback and advice for colour Schemes etc. The garage was influenced heavily on the tutorial you speak off, I was trying to take what I learned and creating my own scene. Since my main interest lies within Environment's and the Unreal Engine do you all think I should focus on that rather than trying to push to everything including vehicles, guns etc when I know I don't feel confident with it all?

As for the material definition, the plane and guns were done in Marmoset; I think I could have made it all look much better in the Unreal Engine.

The colour theory and texture work I do realize needs some work, which is what I'm studying further at the moment.

Essentially my main interest lie in the follow: Modelling, Texturing Environments/Props/Unreal Engine more specifically the Material Editor. Do you think these are all possible right now to study and get really good at, or am I piling far too much work on myself?

thanks, Stuart.

Stoozy89

Thanks GragsGunslinger for the advice on showcasing my work I never thought using Mental Ray or Marmoset were true representations of my work. From now on I'll focus on showcasing in the Unreal Engine.

Stoozy89

Nosslak would you be able to point me to a good source regarding gloss materials. I understand better about specular maps but never delved into anything to do with gloss maps (preferably relating to the Unreal Engine if possible)

thanks.

Stuart.

Stoozy89

If you know a good place to create good game asset tree's that would be highly appreciated, I've looked literally for years and never found anywhere that shows off a good process to doing it

regards.

Stuart.

Nosslak

Stoozy89 wrote: »

Thank you for all the feedback and advice for colour Schemes etc. The garage was influenced heavily on the tutorial you speak off, I was trying to take what I learned and creating my own scene. Since my main interest lies within Environment's and the Unreal Engine do you all think I should focus on that rather than trying to push to everything including vehicles, guns etc when I know I don't feel confident with it all?

I'd say that the best case scenario is that the person looking at your work isn't familiar with the scene and might be impressed by it and in the worst case scenario people will recognize that it's essentially just a copy (with some small additions and extra saturation) from a tutorial that you use to present as the very best art you've created. Recruiters will have seen a lot of portfolios so the scene will probably have popped up a few times so if they realize it's from a tutorial they might think everything else is as well and just close your portfolio altogether. So just take it that piece out.

If you look at a tutorial about how to make a shotgun apply the knowledge you've gained to creating pistol or rifle. If a tutorial teaches you how to model a car, make a motorcycle instead. Don't just follow along and create the exact same thing as you won't learn as much from that (IMO), instead adapt the knowledge to something new and show that you've learned something beyond just blindly following a tuturial.

I'd say focus on doing what you like the most and it'll show in the quality of the art.

Stoozy89 wrote: »

As for the material definition, the plane and guns were done in Marmoset; I think I could have made it all look much better in the Unreal Engine.

Why would you say that? Saying something like that just makes me think 'Then why didn't you?'. I can't find a good answer for that beyond potentially laziness as you obviously already know UDK. I am not a great UDK user myself but I'd say you'd have a hard time making your art look as pretty (or prettier) in UDK as compared to TB1 (especially compared to TB2). If anything the models will probably look worse in UDK, due to compression and other issues. To be honest though the quality of the rendering isn't the problem here, the quality of the models and textures are.

Stoozy89 wrote: »

The colour theory and texture work I do realize needs some work, which is what I'm studying further at the moment.

That's good to hear!

Stoozy89 wrote: »

Essentially my main interest lie in the follow: Modelling, Texturing Environments/Props/Unreal Engine more specifically the Material Editor. Do you think these are all possible right now to study and get really good at, or am I piling far too much work on myself?

If you love what you do it's not too much to pile on as lots of folks on this forum are doing that already. Go join the Monthly Noob Challenge and you'll probably learn a lot!

Stoozy89 wrote: »

Nosslak would you be able to point me to a good source regarding gloss materials. I understand better about specular maps but never delved into anything to do with gloss maps (preferably relating to the Unreal Engine if possible)

thanks.

Stuart.

Here's a tutorial for you, it's for TB1 but most should apply to UDK as well.

Qinzhu

The polycounts on many of the models are enormous. The gun especially not only is there a huge waste in the barrel but the grip looks like a few thousand on its own. You need to optimise the mesh if you are displaying it as an example of a low poly game mesh, also there is an implausibility. Round bullets (which have already been pointed out to be backward) being fired out of slim rectangular vents.

The design and overall look you are trying to achieve is nice and robust but as has been pointed out you need to go back and refresh your understanding of maps.