Using Marmoset vs UE4 for portfolio work?
polycounter lvl 10
So I'm a student studying at a school that focuses mostly on cinema quality CGI. I've been using Marmoset for my stuff, but one of my professors seems to be really against me using Marmoset. His arguments are as follows:
Now admittedly, I can understand a lot of the points he's making, except for 2. My understanding is that PBR is a universal concept, and that there shouldn't really be a difference between one PBR system vs the next (aside from the whole roughness/gloss thing). Another thing that confuses me is all of the people who use Marmoset in their portfolio.
Ultimately, my questions are these:
Is there a difference to how the shaders behave in UE4 and Marmoset?
Should I be using UE4 instead of Marmoset?
And if I shouldn't be using Marmoset, then what's the point of even having Marmoset?
1: Which person is more valuable to a gaming company, someone who understands the shader and lighting construction inside a gaming engine, say like UDK. Or someone who knows how to use Marmoset.
2: Marmoset shaders will not translate to a game engine, so ultimately the shaders and maps will need to be adjusted since they will not look right inside of the game engine.
3: Nobody in the gaming company cares anything about your ability to get "clean presentation-quality renders". They care about your ability to produce models and textures that look good inside the gaming engine.
4: Chances are you never end up where you think you will. So lets say you get an offer that is not in the gaming field. Would you NOT take it? How does Marmoset help you in a non-gaming field where you will use renders such as Mental Ray, VRay or possibly Arnold. Wouldn't having a better understanding of those production renders give you a wider usable knowledge base?
Now admittedly, I can understand a lot of the points he's making, except for 2. My understanding is that PBR is a universal concept, and that there shouldn't really be a difference between one PBR system vs the next (aside from the whole roughness/gloss thing). Another thing that confuses me is all of the people who use Marmoset in their portfolio.
Ultimately, my questions are these:
Is there a difference to how the shaders behave in UE4 and Marmoset?
Should I be using UE4 instead of Marmoset?
And if I shouldn't be using Marmoset, then what's the point of even having Marmoset?
Replies
http://www.polycount.com/forum/showthread.php?t=124949
If two people apply to me, one with all work done in Unity or Unreal, and one with all work done in Marmoset. I will hire the Unity/unreal one for sure if they are about the same level (and provided there are no other aspects/arguments). In fact even if the guy with the Marmoset work is a bit better than the one with in-game assets I'd still hire the one with in-game assets.
Because that technical understanding counts for much more than pretty looks for me. A lot of the work you end up doing usually involves a lot of technical steps and I want to be sure whoever I hire can handle that, and has the correct perspective on the pipeline/performance/large scale asset organization/other parts of the team and development process/etc.
A game artist has to not just be in it to make pretty pictures. If a game artist is either too afraid of a game engine, or thinks a game engine doesn't give good enough quality then I completely fail to see how such person would possibly be a good fit to work on a game... It gives the wrong impression.
So, use whichever one makes your stuff look the best. If marmoset gets the results you want, use it. With that said, it wouldn't be a bad idea to learn some basic ue4 because that's another skill to help with the job search. And hey, you might find you like it better for some things like large scenes.
production does not exclusively require cranking out assets assembly-line-style to stick into a game. especially in characters there's plenty of need to have good-looking presentation.
it seems your prof is arguing mostly from a technical POV. question is if you are aiming to go that route or rather want to focus on creating pretty things. not everyone in a production does lighting, shaders or rendering.
pluses for running marmoset:
- it's quick to get something good-looking
- less hoops to jump through - you don't need to mess with a full-blown engine
- visual quality stopgaps like texture compression or blocky shadows don't exist
- since it looks nice to begin with it's always a good reference to point to when we're talking prototyping/developing looks
Definitely being able to manage a scene (environment) involves more technical knowledge and factors to juggle, but what of the humble prop artist? Their main concerns are making efficient and pretty props. They can demonstrate technical ability through their highpolies, their UV's, flats, wireframes. This is smaller scale stuff, but involves technical steps nonetheless.
I doubt fear has much to do with it, more so effort. You could set up your model in an engine, or you could use toolbag which gives good results (and convenient ways to export those results) quickly. Depending on how many props you've got to show this could be a big factor. What if in your example the marmoset person had one prop shown in a game engine that, to your satisfaction, demonstrated their technical ability? Would that be enough to hire them (their work being slightly better) than the person with all in game-engine stuff?
If you're familiar with skyshop I'd be keen to hear your thoughts on using that in this thread's context (portfolio renders) too.
Before marmoset 2 came out, easily 70% of the 3d guys did subpar presentation shots, and now with marmoset its very hard to distinguish in that matter, as it was a obvious indicator about user skill and artistic vision which is pretty much gone, which is a bad thing for the experienced and a good thing for the newcomers I guess.
But I think its about how you present yourself. Showing competence in shader setups etc can be very valuable in your portfolio, depending on the rest that is there, but I think nobody will give you a definitive answer here
as every person judging you is another human and valuables other things.
Your teacher is right however (apart from the texture part I guess) and the question is "at a school that focuses mostly on cinema quality CGI" - why do you even use realtime renderers ? If you come from a cinema school, you have to know your offline renderers no question. This sounds a bit like you have to know this stuff anyways and hope to take a shortcut here
Other than that, he does bring up some solid points, I would certainly recommend learning UE4/Cry/Unity to any students looking to get into the game industry, knowing how to work with a game engine is certainly very important. However, I don't see it as an either or thing, many people (including working professionals) use Toolbag as a quick previewer before sending assets off into game, mostly because its easier and faster than loading up the whole engine/level just to work on your textures. TB2 is something you can learn how to use in a couple of days, while most game engines take much, much longer to master, so again, I don't see it as an either or situation, on the contrary its easy to learn and use both together.
In its current state, UE4 is certainly more capable of presenting certain types of work, for instance complex environments, or anything animated, as we do not support animation or radiosity yet. But vehicles/weapons/props, small scenes, posed characters, etc can easily be shown off in TB2.
We've also talked to a lot of people in the CG/offline rendering industry that say they use TB for quick previs, again it can be faster to set up a shot in TB2 for a quick preview, where you can make adjustments to the lights etc in real time than it is to set it up in a traditional offline renderer and deal with long render times when you make a changes to your scene.
TL;DR: I think you would be doing yourself a disservice by using Toolbag and completely neglecting common game engines or offline renderers, but I think its misguided to avoid it entirely.
Yeah, in terms of hiring, having experience with a game engine will look better on your resume than not, but having your work shown in Marmoset VS UE4 generally will make very little difference to your chances of getting the job. I say this as someone who has done a lot of portfolio review prior to working at Marmoset.
The one who communicates the best art.
The same could be said about many different pieces of software and shader types. If you're producing art in one engine/renderer and your maps look good, then that should be enough to show that you could do it in another.
Err what? That depends on your role. In some cases it's harder and requires more effort to produce a "clean presentation-quality render".
Use whatever tool you feel comfortable with in getting the best result you can. If you're after CGI quality art then it makes sense to use what other CG artists are using.
If 2 people interviewed for a position with equal artistic ability with different presentation engines then the job should really go to the one who can communicate his artistic ability/personality better. Just because someone hasn't used X-Engine before doesn't mean they should be disregarded.
At the end of the day, if you produce good art in one package, then you can probably do it in any other. I've used UDK a quite a bit but now present a lot of my stuff in Marmoset purely for speed. My job requires neither, it's all in house, which you've got to learn either way. There is a wealth of knowledge available on the web to quickly get you into any package of your choosing.
What Earthquake says though is spot on, try not to limit yourself to a specific package, game art is far more software dynamic.
I'd personally invest time in making good art that can translate into any engine.
I'd argue that creating the best art that can run well ingame is the most important thing for a video game artist. Seems like everybody here is forgetting that. No, it;s not a small detail that is tacked on afterwards, this realization needs to be there permanently while you are making your art. When working with an engine like Unreal, you are confronted with that just a bit more than with an 'ideal' viewer like Marmoset.
So, this really depends on the situation.
Obviously, you can't make a full game in Toolbag, so if that is your goal, its a bad choice. Again, I don't think anything you can do in Max, Keyshot, Toolbag, etc can replace the experience of actually working in production, working with a team and making game ready, optimized work.
However, if we're talking about presenting your work for your portfolio, either work that you created in a production environment, or work you're creating specifically for your folio at game spec, there is very little, in terms of knowledge of optimization, that you gain by setting up individual assets in UE4 vs Marmoset for the sake of portfolio presentation.
The exception here would be if you're an environment artist/level designer and the work you're putting in your portfolio is a full on level, at that point performance and setting it up in a proper game engine is very important. Edit: or as Lee points out below animators, or FX artists, technical artists(whose main role is generally pipeline/optimization) or anyone who really can only do a specific type of work in the actual engine.
For your typical production artist (eg, character, vehicle, prop, weapon, etc), certainty its important to understand how to create optimized, game ready content, but in production you're going to have technical artists, TDs and engineers that you will work with to set up performance specs for your game. This isn't something that individual artists generally have control over, and if they do its either a really unorganized project, or you're talking about someone in a senior position or a technical artist, where the basic technical side of optimization is probably not a new concept.
Now, I'm not saying your average artist shouldn't be concerned with optimization, they absolutely should be, and I agree that making great art that runs well in a game scenario is a very important skill that every artist should have. However, optimization is rarely a simple task that one person handles. Optimization depends on many, many complex factors like specific engine, shaders, game type, hardware spec, resources devoted to other departments like sound, ai, and physics, and many other things. There's too many variables for an individual working on their portfolio content to possibly worry about or be able to figure out on their own. If you've ever worked on a project where the game and engine are in concurrent development, you know that optimization is not a straight forward, or linear task.
There is no inherent technical/optimization knowledge gained by simply loading a stand alone asset into UE4 vs Toolbag, UE4 is equally capable of loading poorly optimized art content. :poly124:
On the other hand, I agree that as a teaching tool to show artists the importance of draw calls, vertex use, etc, Toolbag is a poor replacement for the actual game and game engine your team is working on.
Environment artists would do well to show their work in-engine.
Animators would do well to show their work in-engine.
Character artists can pretty much do as they please.
why? well, environment artists need to show that not only can they create solid art content, but that it's usable at runtime and displays their ability to not only create the assets, but place them and light them appropriately in the engine.
Character artists, in my experience, have little to do with the pipeline outside of making the character and possibly skin them to a rig created by the animators. they're rarely responsible for final import into the engine, with the exception (maybe) of material creaion, which from my experience is handled by art leads and tech artists who then hand down preset materials for the artists to just plug maps into.
Animators will have to deal with import issues, any issues revolving around their rig and ensuring it works properly in-engine, making sure that they have their animations optimised for the engine etc.
yeah, character artists got it easy :P
TL: DR - you can show that your assets are optimised and are game ready without putting them in an actual game engine. you can't show that you understand engine specific features such as lighting and post processes without being in the actual engine. create your portfolio to best display your skillset at all times.
Yes and even then, while learning how to wrangle the specific light settings in an engine or how the lightmaps work is very important, lighting, composition and color grading are in no way engine specific skills. These are foundational art skills that can be demonstrated in any engine or rendering app.
I see nothing wrong with doing a scene in UE4/CryEngine/Unity, and then displaying individual props and breakdowns in Marmoset 2 or a viewport shader.
That said, getting stuff working ingame and showing them in context is also important, specially if you're an environment artist. Vehicle and character artists can possibly get away with Marmoset only.
With that said I would never build an environment in Marmoset because I can't do a lot of things that I can do in Unreal. You will get a better looking environment in Unreal then you will in Marmoset no doubt.
1. present the highpoly mesh +textures + fibermesh-hair with marmoset
2. present the optimized mesh + textures + haircards in Unreal
3. is it even nececary to show the plain meshes (so the polys can be seen
?
I haven't come across an applicant who is totally incompetent in several years - education in the UK has improved greatly - so the basics (polycount etc) are rarely a concern. its all down to how well considered the art is - namely, do the assets have purpose, place and history?
The real decider though - if they can't explain to me what they have done, why they made those choices and what they think they could do to improve on their work then an applicant is no good to me.
And yes. I want to see wireframes and UV layouts (so I can weed out the autounwrap monkeys)
Having said that, demonstrating your UE4 knowledge is very valuable as UE4 is super popular and ultra capable game engine, however this should not hold you back from presenting your work now!
Some people argue that Marmoset toolbag is not used in production...well of curse it's used in production but for presentation purposes, not for in game viewing and interaction!
I think poopipe is making a very strong point about the assets history, purpose etc. These aspects show how mature an artist is in the realm of applied arts such as games.