Art test question - is this normal? (Long post)
So I recently have been going through applying to a few jobs for a junior position. Some of which i've had good reception and made it pretty far but ended up not landing the job (no problem, still a compliment). I have this one that i've made up to the art test (I don't want to name the company) but I got approached for it because of my portfolio and time in the military due to me having a security clearance prior.
Beside those points though i'm currently in the process of taking this art test.. I need to make 2 assets which are pretty detailed in the sense they are both hard surface and need to include as quoted "All pipes, valve handles, gage faces, etc." same applies for both assets.. okay I get you want it detailed not a problem. I was only given one reference for each model, with one perspective and its low res enough I can't see the gauge faces they want shown or hardly in depth the switches, etc. I tried looking it up to really end up with nothing close to what i'm shown.
The second issue is the time given.. I'm not sure how I feel about this. It's already unpaid (so I get no money for my time doing this) and THEN I get an interview, that seems so backwards. Anyways, the test guidelines read "You have 4 days to complete both assets, not to exceed 12 hours total and fit the specifics for a game engine"...... what? so you want me to highpoly, lowpoly, texture, and render both assets in 4 days not exceeding 12 hours total on the test? how is that even possible?
I don't know their workflow, I go off my own which from my understanding is a standard workflow (highpoly, lowpoly, bake and texture, render, etc.) but I don't understand the logic here. If you want an applicant to show you the best work they can provide for you then why set a ridiculous standard? I voiced my concern to their lead artist (who I can't find any portfolio on, i've tried) saying in short "I think the the standard for this is a bit unrealistic for what you're wanting, because otherwise I'm just providing poor looking models that doesn't reflect the quality I can bring."
Well I receive an email regarding my concerns essentially saying "do your best and turn in what you got on due date or do nothing at all"....... okay. I have no clue how to feel on this really, and really my gut is just telling to back out because this company just 'feels' toxic. Not to mention the reply to my concerns (not by the lead artist mind you) "I appreciate you identifying your concerns and what you “feel” are unrealistic expectations in reference to our 3D Art Test however, if you feel that the expectations and amount of time provided to complete the art test are too unrealistic for you, then you have a couple of choices." just to point back at the response I was given...
For the record this is my first art test, but even then this don't feel right. As I've been offered art tests previously for really cool companies that the standards were laid out and I was very confident to deliver what they wanted.
thoughts?
Replies
First, my opinion is that you shouldnt do unpaid test. But probably you won't get any job if you dont, so who cares what I think about that.
Probably the intent of this test is not to determine what is the absolute quality you can output, but rather how quickly you can work, how well you can understand directions, how you manage time, and maybe even if you are professional enough to ask for clarification rather than guessing.
I think what they've told you is straightforward enough. It's not a knock against you to let them know, "im having trouble understanding some of the reference images and havent been able to find additional, therefore in interest of time I'll use my best judgement to fill in those details that are vague."
You can write out a full analysis when you deliver the test to cover what you did, what you werent able to do, what decisions you made and why...
I mean if that is something they hold against you, yes it's probably a toxic place.
I wouldnt worry too much. Just do your best, communicate the why of which choices you made. Don't think about, "how will I look," only think, "whats the best use of time I can make."
That was my thought as well for the test, I really ain't a fan of unpaid test but yeah I guess it just depends..
Perhaps it is, I guess it just sounded really weird in comparison. I wasn't really hoping to come off as complaining but trying to address my concerns and trying to make sense of what they want. Knowing I may not exactly be able to deliver two solid assets but more like 1 and half ya know?
I'll continue with it and see what happens though, cause perhaps it is what you listed and 'maybe' they know it's unrealistic but see what's given anyways? in which case, fine, okay. I'll do that though with writing out an analysis afterwards closer to completion.
Thank you @Alex_J
When I've worked with artist in the past, usually teh more experienced ones ask a lot more questions than others. And usually we try to get the art out in tiers, I guess you might say. Like, the first thing is that I want to get something in the game to start getting a sense of how it's gonna work. Maybe this a low poly blockout, maybe it's the hi poly, doesn't really matter just whatever can be made quickly and gives a sense of the final product.
Then to get the final result is a process of iteration but so long as the stand-in item is there there isn't such a big rush.
Anyway, I just mention that because it may help you prioritize your work to make sure that by the deadline you at least have the bare minimum ready - stuff that oculd be used - even if it hasn't gone through all the levels of iteration you need to really make it shine.
That's just how i prioritize things but I imagine two complete models that look rough but ready would be better than just straight up not delivering the complete order.
I suppose that would make sense and I would agree I guess it depends on the workflow expected. It just gets confusing in the details asking for high amounts of detail including filling in areas not shown (which I am fine with, I don't really mind filling in blanks even if it turns out its wrong). It's a very strange order to me though, short time frame but high detail but not really providing high res images to get those details so most of it feels like I'm making it up as I go but I have no choice.
I'll try, may as well but right now its looking like I may only get one fully complete and the other one will look half baked or barely textured with care. I don't know, roll with the punches I guess and see what happens. If it flops it flops, but I hope to at least have something at least good to show for it.
Appreciate the advice dude, I'm pretty new to the whole art test area so I was going in blind not really knowing what to expect. So sorry if I sound a bit scatter brained at the moment.
Honestly, to me it sounds like they just want you to do as much work for them as possible, for free. I understand my opinion is super biased, but it's because of my personal experiences, where I've done art tests and I saw my assets in the company's game or done art tests and the company "could not open the UE4 project" while I checked the version and tested in multiple different PCs and laptops, and opened it properly. It's just my 2 cents though.
Again, I'm probably getting a bit bitter about this, but I don't think companies value juniors anymore (at least in Europe). For interns it might be better, as they might be getting you for little to no money. But no matter if it's a start up, indie or a well established gaming studio, they only care about getting a senior that suits them, establish their art style and block-outs, and then they outsource everything.
And then 'the industry lacks talent'
Sucks to take art tests for free, but if you are starting out it's pretty much a necessity. I've honestly never seen an art test require two unique assets though. Deadlines can be tight, but they are usually reasonable if you are not working on something else fulltime. Makes me wonder if you are approaching it the wrong way. I haven't seen the test so just spit-balling, but perhaps they don't expect you to bake out every component individually, but recycle gauges and pipes on the same texture sheet. Might be good to ask for additional clarification.
I agree, it does suck taking art tests but I understand I'm new so I get it and I'll play the hiring game for a newbie, doesn't bother me too much. I have no clue really as to the work culture as I've tried looking it up through Glassdoor getting some moderate reviews on all sectors but nothing really too in-depth. I also agree it feels so strange with two assets? and I would say I agree with recycling components (which I am) but the two models are different...
1.) is like a water-pump valve, with all sorts of gauges, wires and connectors
2.) and the other is like some sort of generator with entirely different pieces I can't really switch between the two.
For all I know I could be approaching the test wrong and if I am I'll take that hit, failure to ask deep questions. But my assumption is it's for virtual training (that's the game produced by the company) in VR and you give me two assets... I'm going to make them detailed to their real life counterparts. As an artist i'll do my best to translate that work into the vision laid out but with the guidelines.. but I have no clue what 'their' method is though as I haven't been able to speak with any artists on the team so it's all jumbled. That's one of the few things I'd ask if I had that chance.
I'll gladly share the asset I do get done (hopefully) as I've been crunching over 12 hours for about 3 days now, tomorrow being due date. Hopefully I can follow up with some details as well with the company too while i'm at it.
A huge thing new people do, is assuming things. Do not assume, ask.
@Larry Will do. Thank you for the advice by the way.
@neilberard same to you!
Update:
Failed the test, no feedback from the artist or anything on my work (completed one asset out of two) and the hiring manager had to call and tell me instead of the actual people... nice. I'm more relieved than anything I suppose feeling like I dodged a bullet. As the further I worked on the test the more I felt this wasn't really suited towards what I actually do. oh well, back to personal practice.
Learned a few things from this experience so I won't chalk it up as a total loss, thanks again for all the advice and feedback everyone!
This is one of the models I was able to finish. 2048x2048 textures, tri count I can't remember but I rushed to try optimize a bit more of it. Understandably it looks fairly 'meh' (hell I think it looks terrible, but got to throw in something. Not like I didn't mention with the standards it was gonna look bad). Time constraint was too much it felt like, lack of references (I was only given 1 blurry image) tried my best I suppose with what I had.
What was your process for making that?
DId you bake normal maps? Did you breakdown how much time you spent on each part of the process?
What is the intent of this model? Is it a background prop, or a hero asset that the camera would be really close to and it needs to look hyper realistic?
That's a bummer. Hopefully you'll do better next time. Some things that stand out to me.
-The normals look soft / off on the pipes, are you using hard edges where they intersect?
-One that note, I don't think the lighting presents it very well, it looks like you are shining a flash light straight at it.
-I feel like there could be a lot more subtle variation of the paint color, Rust, peeling, dust accumulation etc. A lot of it feels too shiny, some AO would really help.
This is sci-fi, but I think it does well to illustrate how you can add a lot of interesting details in a material.
https://www.artstation.com/artwork/0nqPZ8
@Alex_J I modeled the highpoly, made the lowpoly, brought everything into substance and baked the textures, textured and then setting everything back up in Blender to duplicate and fill the empty spaces with recycled pieces. Reduced anywhere I felt could be reduced more for polycount. Finally I brought it all into Marmoset Toolbag and quickly render out to stills for submission. Because I was running out of time I had to quickly just setup lighting and get all of it sent.
I didn't get to talk to anyone, I did explain I had to push about 8 hours or more for about 3 days to get this one asset in particular at its current state with all the things I said I did (if that's slow then i'm surprised because I felt I was rushing a lot).
As for the intent, no clue. The guidelines read "Model as much as possible – even areas not visible in provided reference. Fill in any gaps in data or reference to the best of your abilities. Model should be consistent with current generation real-time/game engine standards while maintaining as much fidelity and detail as possible." so.. if its being used for some sort of training for the player then I suppose it would be considered a hero object and would need lots of detail. As the test even listed "including labels and serial numbers, ect." that of which I couldn't even see... because blurry reference.
Again, didn't really get to talk to the lead artist at all besides one message.
All good. Did the most I could with what I was given. Believe me, I know its bad and know I could have fixed a lot but no time sadly, too little in all honesty. Especially for 2 assets. I do agree the thing you posted looks great but realistically I imagine it took this person much longer than 3-4 days to make, something I didn't really have. The lighting was thrown together in less than 10 mins, I had AO but again... just crappy overall. Really wasn't happy about the result at all but:
¯\_(ツ)_/¯
Plenty of other jobs out there, this one wasn't for me and i'm totally okay with that.
I find it great that you posted it here for others to see. It's nice seeing art tests and their requests/restrictions.
I don't have much industry experience but here's my 2 cents on this, just as food for thought. There might be lots of info packed together. Keep in mind that I do this with good intentions ✌️
I'll just assume that this is a prop for games, but the same principle could be adjusted for other areas as well. I think you fell in the trap of trying to model the real thing. If they gave you a blurry image, it means they want you to capture the essence of the prop. That translates to it being as complicated as it needs to be, to make you understand its purpose. Also keep in mind that everything you create will be seen on a screen. So if it's difficult to read, it can probably be done better. Doesn't matter if it's matching the picture 100%. The only case would be for manufacturing, but they would not give you vague stuff.
I think the real test here was to make this look as believable as possible, and as modular as possible, while still adhering to the core values; shape and composition. Modularity would save you hours here. Baking all the pieces cost you as well. You could achieve a similar result with smoothing groups and some bevels. Have you ever experimented with weighted normals? You could also post some wireframes, so that we can tell you if you could reduce the geometry further more. Cylinders take up a lot of tris pretty fast, and by the lack of jagged edges in them I dare say you could have reduced those 😅
As far as shapes and composition. I think you have quite a few thin objects there which do not read well, even on my huge TV for a screen. Don't be afraid to make them bolder, thicker. Spreading things out would help as well (busy vs eye resting areas). Stuff seem clumped together towards the middle. The upper area definitely needed some love too. If you thought that the shadow would cover everything in that area, you're right. So you need to work on lighting. I'm not sure how it's done in marmoset. but you need to increase the indirect lighting (or sky light, or environment light, or light bounces?) so that your darkest spot is still properly visible.
For texturing, if you wanted to use a 2k texture, you could explore making an atlas with 3 materials, and the 4th tile to add dirt/damage/leakage or graffity decals to give it some life. As far as colors, I'd say 70% red, 20% silver and 10% blue. Rule of thirds must be applied in texturing as well as in all the aforementioned areas.
From this prop I get the feeling that you ended up looking too closely on what you were making, but I can also see your dedication and commitment to it even though it was probably something not many people would find interesting. I hope this info was not too overwhelming and made some sense 😅 Good luck on your endeavors and future tests!
I'd try to get in touch with some experienced enviro artist and take a query what workflows they'd use for asset like this and why.
Personally, I'd question a lot because I although I am not an experienced environment artist, I'd expect to be able to make like 3 items like this in a few hours, with more complete texturing job.
Like mentioned, it's not necessarily a matter of skill or motivation, just knowing where to focus.
I wouldn't bother with a high poly and baking for something like this - but it's not my area of expertise so I don't think its worth me saying exactly how i'd do this or that. My way is probably really different from enviro artist working at ubisoft or somewhere.
It is too bad people dont reciprocate 1% of the time you sacrificed and at least offer some feedback. "Not a realistic use of time" is the usual excuse but if thats the case they shouldnt be handing art test out like candy. Just a matter of corporate psychopathy if you ask me.
Anyway, Im not entirely convinced the test was very unrealistic in terms of time so I wouldnt dismiss it.
@Larry That was the goal and yet hopefully someone may find this helpful not only from my mistakes and choices but the discussions can give some solid insight. I'm fairly new and trying to get into the industry but I'm very hopeful. Always appreciate the advice as this stuff is a constant learning experience and I think that goes regardless of how long you've been doing it.
I committed a mindset where I needed to make exactly what I saw and odds are (considering I failed, hah) that wasn't the point, which ya know oh well. It is what it is, but looking at it now I think I would have approached more of a trimsheet approach? I recently started doing that when I started doing environments (followed some Polygon Academy tutorials, love that guy. Hope to learn more from him as I go along). Not sure about weighted normals, or atlases?.. gotta look into that.
As for my tri count...... I'd rather not talk about it heh.. just kidding, it wasn't great. Cylinder shapes absolutely murdered me here on this. Because a good majority of this thing was all round shapes I believe I was ranging from 40k to 50k which is terrible in this instance. Felt like having a stroke just seeing that going up the more I was trying to recycle those bolts around. Though lately I looked into how artists on Ghost of Tsushima got around making convincing ropes while making them not so heavy on tris.. so that's something.
My lighting was garbage, I was aware of that. Lighting has been my biggest weakness for a long time but I'm trying to be better at it. The rule of thirds you mentioned just dawned on me and I feel silly for never really taking that into account so I will take more notice of that. Thank you!
I won't lie I was dying inside making this thing... I try but I admit it wasn't really interesting to me but I still 'tried' to care. Regardless I dropped the ball a lot in the process but that's alright, learn from it. Failure can be fun sometimes. I say that because I've been trying to shake the defeatist attitude and move on. Appreciate the post dude! and thanks! :)
@Alex_J You may be right, and that what's already been said I just failed to see the point or the intent of what I was suppose to do. I'm slowly learning this environment artist path and as much as it is frustrating I do enjoy it. I like a lot of peoples work I see on here and I can pick apart some things to translate into my own works later on. It is certainly a journey, but at least I'm having fun so that's good. Thank you as well! living and learning.
@Mr_Nova So I don't know what specific company this was for but if it was military simulation or training, my guess is that their workflow doesn't use a high to low poly workflow. I used to work for a company that was very similar.
4 years ago I was given a similar test for a military simulation company. When I received the vague instructions and ridiculously short amount of time I was given. I had the same exact thoughts that you had. My thought was 'you want me to make a High poly and bake it down to a low poly, texture and render it all within X amount of time?'
I then had a thought. Is that really what they were expecting?
After being frustrated and trying to figure out how I was going to do this for 20 minutes, I decided to reach out to a contact of mine that was working at a different military simulation company doing similar stuff. I was vague and just asked him if at his job they used a high to low poly workflow. Much to my gut feeling, they didn't! I'm so glad I asked. It should be noted that I would have asked the actual company itself, but I was given the test and instructions on a Friday and was asked to have it completed before my interview on Monday.
Needless to say, I did the art test as a mid to low poly asset and passed the test and took the job.
After working there for 3 and a half years, I learned that almost all the other military simulation and virtual training companies in town (there's quite a few here) had a similar workflow, skipping the high poly to low poly all together and just going straight into the "mid" poly modeling, uvs, simple bake, and textures.
My guess is this sounds like what was to be expected and "assumed" to be done at the place that gave you the art test, especially if there was no language in the instructions indicating anything about a high poly. I really just think that they don't know any better and didn't know to clarify it. It would probably be like that with other things on the job too, so take that for what you will.
It wasn't all bad for the 3 and half years I worked there. I got used to it and just accepted it for what it was. It was also my first time working in a studio environment side by side with others. I learned a lot of soft skills. I learned a few tricks here and there, but I really wasn't stretched as far as hard skills go. I also worked small side contract gigs and personal pieces on the side to itch that scratch that I still had.
Funny enough, one of the personal pieces I did landed me my first gig in AAA. They reached out to me and everything!
All the best moving forward!
@Tobbo First of all, your work looks fantastic! This place I recently tried for wasn't the same as yours I believe. However it was a military VR/simulation training company (which is pretty cool.)
That makes sense, I kind of wish I would have known that, or at least would have been told something about their workflow.. I do feel relieved I'm not the only now that was confused by the guidelines at first. This may sound silly but I never really looked into mid-poly modelling but having done some browsing I may try to adapt this method for environment props perhaps? not quite sure yet. I wanna be fair to the job and say maybe they didn't know how to communicate it and me being new as well wasn't asking the right questions.
Appreciate you going out of your way to give some insight as well, considering you've done this particular work before. So I got a few tricks to learn and just expand skill sets. I'll eventually land something that I can perform well in the industry.
Thank you and all the best for you as well!
Thanks for sharing. Not sure how anyone could 'pass' this test without being able to ask questions and clarify workflows (like, y'know, a real job...) anyway. Sounds like you've got the right attitude; I didn't, when I was in a similar situation 🤣.
Definitely look into mid-poly modelling. As an enviro artist, you'll need to produce art in an efficient, iterative way that's as non-destructive as possible.