Nothing's changed, I just decided to do a quick vanity render of the highpoly.
I love it! I love the topology on what I assume is the gearbox at the bottom. It's tough to do!
I should probably give you something to improve on so the only thing I can see is the cooling fins on the engine at the front. They don't stick out enough for my liking and are quite thick. On a real engine they are thinner and longer to better disperse the heat, like this: I suppose depending on your optimal viewing distance it could get a little noisy though. I am nitpicking; it looks awesome!
@fdfxd2 That high-poly looks amazing! This is a sort-of-newbie question but are a lot of the shapes booleans (like that smaller block part behind the engine block) or intersection geometry?
For an update - still working on textures, any help would be...helpful!
@Orbit Nice start! Few thing that could be improved to me
I think the drawer's handle look too shiny, if you look at the reference, it's pretty dark and there's mainly a single round highlight in the middle which makes it a bit more part of the whole and not something that jumps directly to the eye.
Maybe also the gradients on those Bookmark are a bit too pronounced, I believe scene lightning will produce similar gradients as those observed in the reference without having to paint them. Also the colors might be a bit more dark.
The magnifying glass is pretty nice, I really love what you did with the shape. To improve it, you could maybe add some highlights on the body, particularly on the big bolt to make the different parts stand out a bit more.
Also, maybe you could add some dirt to most parts, like on the strips on the books, use a normal brush with low opacity values, take the color of the book with the pipette and fade those lines a little
Appart from that this is cool, maybe you could also change the shading of some elements to make them less specular (the letters for instance)
Hi, it's the first time that I participate in a challenge. I tried to get as close as possible to the concept even if with some difficulty. I hope I am doing a decent job .
Made a few changes on the colors and worked a bit on the lighting and toon outline (ended up using blender Freestyle render option and did nothing in UE4 in the end). I put it on artstation with more details : https://www.artstation.com/artwork/JlyGzR Tell me what you think
@simonBreumier Thanks so much Simon! I'll definitely try all of that out. I thought the same about the handles. Its my first try at PBR materials (of any sort) so I wasn't sure of the rough/metal values.
Yours is looking slick! Really reminds me of old Disney film style. Is that toon shader exclusive to Blender only or is there a UE4 equivalent that you would recommend?
About toon shading, I've also tried doing it in UE4. There's this really cool live explaining how to do toon outline/cell shading using post process materials: https://www.youtube.com/watch?v=cQw1CL0xYBE
I tried it on mine juste right now and looks good !
Hello everyone! I'm joining the party a little late but hopefully with plenty of time still to work and get some feedback
I love watching everyone's work progress.
I ended up choosing the motorcycle, which may come back to haunt me given how complicated it is. @fdfxd2 do you have any tips for interpreting the concept art? I see that you picked a bike to use for reference, but I'm having a hard time figuring out exactly what's what - even using references. The area around the engine is the worst. All of your forms look so polished and purposeful, how do you do it?
>I should probably give you something to improve on so the only thing I
can see is the cooling fins on the engine at the front. They don't stick
out enough for my liking and are quite thick. On a real engine they are
thinner and longer to better disperse the heat, like this:
- @Bostvic
Yeah I think my thought process at the time was "oh it'll have a smaller polycount and will be more justifiable to bake if it's a few thick strips instead of many thin strips." I still kinda sorta stand by that but, I'm probably going to move the current fins to a texture atlas anyway, so I'll consider making them thinner again.
>do you have any tips for interpreting the concept art? I see that you
picked a bike to use for reference, but I'm having a hard time figuring
out exactly what's what - even using references. The area around the
engine is the worst. All of your forms look so polished and purposeful,
how do you do it? - @kelly_caroline
I started by modelling a Honda CM 400 from a profile reference, since that's what the engine is based on and the engine's the one I'd probably have trouble modelling the most, then borrowed elements from the Yamaha Virago, the one that the other bits are based on. The hoverpads are just one big riff, taking elements from anything from that weird 1920s washing machine my cousins still use for some reason to, of course, jet engines.
Helps to spend a few hours periodically searching for each part of the motorcycles individually,
Useful keywords
"virago switch", "virago front forks", "triple trees", "cm 400 carburetor"
by the way when it comes time to model the carburetor, don't be like me and model it via traditional sub-d just do it through the boolean-dynamesh method, you wont want to kill yourself in the end, probably.
> That high-poly looks amazing! This is a sort-of-newbie question but are a
lot of the shapes booleans (like that smaller block part behind the
engine block) or intersection geometry? - @Orbit
Most of the parts are physically joined together, others are joined together using normal trickery, using the data transfer modifier and vertex groups. wasn't sure if they'd transfer well to actual normal maps, but they did.
Big "oof" so far on the texturing side. Focused way too much on getting the optimal texel density ratio between the parts, instead of just taking out even more parts out of the bake for the trim sheet/texture atlas. Probably should have added more rivets and welds on the highpoly too.
First of all you all are amazing artists, I'm so impressed by how amazing everyone is doing! It's because of seeing everyone being so positive and hard working that motivates me to finish my attempt to. Incredible stuff here!
Here's my attempt at the challenge. I chose the table since I really enjoyed the color palette and the interesting props. I had a lot of headache matching the scale of the props as well as the angle/rotation of them. I can't stress enough how helpful it was in the end to do measurements of props multiple times for the right sense of scale.
I decided to hand paint everything in Blender, but if I needed to copy or mirror a texture piece I used Photoshop.
Overall im very happy with the result, one concern is the AO in sketchfab wasn't as impactful as the one in Blender Eevee, but it's not a huge issue.Sketchfab screendump Tris: 6,442 Blender Eevee Material preview screenshot
@Donato I think it's looking really good! I'm trying to create the same scene myself, but I'm having issues with scaling. For me it's hard to tell how wide and tall the room is, but I like your interpretation. I am curious on your process for the bookshelves. Having the books as separate models seems resource intensive, both engine and time wise. However using a texture to create an illusion of books probably won't look as good. It's a problem I've been trying to solve. Are you planning on making a procedural texture for your books?
b_beauchamp Thanks a lot for your feedback. For books, I actually created three basic models. Then I created various collections using the three basic books and placed them trying to give as much variety as possible. I hope you find it useful.
As far as performance is concerned, the UE4 gave me no problems. In addition, it has an excellent automatic LOD system that reduces triangles without losing detail (especially for distant models because you can't see the details). As far as textures are concerned, here I have the biggest problems because, being completely amateur, I have tried to use only free software (Blender, UE4, Quixel Megascan) but I believe that there is not yet a valid alternative to Substance Painter that I don't have (there is Quixel Mixer but I'm finding it difficult to use). In conclusion, I am not trying to create a procedural texture, indeed I should still learn the basics I think. Up to now I have simply taken one or more materials from Quixel Megascans and applied them to the model. So would you have any suggestions for improvement?
@Donato I am an amateur as well, so take my advice and opinions with a lump of salt. Regarding your current textures, my tip would be to layer on some grunge maps and add imperfections to your textures in Mixer. Small things like dust, smears, and worn varnish can add a lot in terms of realism. Quixel has an "Imperfections" collection of roughness maps which you may find useful. I believe you can also tweak things in Mixer like color and height to help achieve a more unified theme. A simple thing to remember is that everything has a story. Consider the floor of the library. How would it wear over time? For my floor, I started with a basic wood texture. Since libraries don't tend to be the cleanest or best maintained places, I added varnish/oil stains to show age and grime. People tend to take the shortest path, so I clustered footsteps in the center area of the room and near the checkout desk. Dust accumulates where people don't travel, which is why the edges of the floor have it. This is just an example of how I work, but you may find it helpful. Below is the current version of my texture for the library floor in Unreal.
I do agree that there is no viable alternative to Substance software currently. Quixel Mixer can do a lot, (Keep in mind I haven't used it much) but it does not give you the freedom that Substance Painter does. Paying for Substance is 100% worth it in my opinion. Substance software is industry standard for a reason, and the price is pretty affordable for what it gives. (If you are a student, they have a free personal license option as well.) Learning Substance Painter alone will definitely help your art improve. This is how I think about it: paying for the right tools can save you a lot in the long run in terms of time and frustration.
@b_beauchamp Thanks for your suggestions. You have a very advanced way of thinking. The reasoning you made regarding the example of the floor would never have occurred to me.
I will buy Substance Painter because I tried to add details to the textures using Quixel Mixer but I found many difficulties despite having followed their own tutorials to the letter. In the end I managed to add some details on this texture but I think that with Substance Painter I could have achieved more and also in less time (or maybe it's me who didn't understand something fundamental).
In the meantime that I watch all the Substance Painter tutorials, I applied the textures in the basic way to get an idea. I think the scene has a lack of detail. What do you think about it? This is the result.
Well, I think barring a big rework I'm almost there. I put in some of the feedback that I got and I hope it improved on it. I threw it into Unreal and put in a toon shader that @ThisisVictoriaZ and @simonBreumier mentioned and I think that took it up a notch although since I haven't really used Unreal the lighting is a bit dark or the materials are - not sure. Maybe its a simple fix that I'm missing. But anyways, here's my first modeling attempt after 8 years .
@Orbit that toon shader makes it looks so good!! I often have a hard time with things being too dark in unreal, something that helps alot with that is adjusting the enviornment color and the post processing volume settings.
@Donato you've got some really good dramatic lighting in the scene! The textures on your counter and floor are a bit too big but looks ok so far. Would love to see some horizontal separation of planks on the floor so it doesnt look like each plank is super long is that makes sense.
@Donato To echo ThisisVictoriaZ, I really dig the mood lighting you have going. Looks spooky! If you want you could render out the scene in different times of day. Your new table texture is a great improvement. I like how it looks like the varnish is stripped off and grime has settled in. I also agree that the textures are a bit big. Check out some reference photos and play around with the scaling.
I think Flipped Normals has a couple of good Substance Painter tutorial videos if you want to check those out. I also highly recommend getting a drawing tablet (sometimes called a "digitizer") if you don't have one already. It will allow you to work much more naturally when sculpting and painting. There are a few affordable options out there to pick from as well!
Regarding the detail level of your current model, I think some additional modeled details will go a long way. For example, compare your arches near the bookcases to the concept art arches. The concept art seems to indicate ornate details like filigree, carved capitals, etc. Textures will carry some of the weight in terms of detail, but the overall geometrical shape should be made more complex to sell the idea. The library concept art we are using gives off a steampunk vibe to me. I would look up some images of steampunk and Victorian interiors for inspiration on the types of details you could add.
@b_beauchamp and @ThisisVictoriaZ Thanks to both of you for your suggestions. Regarding the textures, I think I will take some steps forward now that I have Substance Painter. As a first test I redone the table again (@b_beauchampI tried to follow your advice to also think about the history of objects). In my eyes it seems like a huge step forward. This is the result (comparison with quixel mixer):
I will try to reapply the same workflow for all other textures as well. Regarding the complexity of the models I will try to increase the details by going towards a Victorian style. Thanks again for the advice. They are very useful for me even if they will probably seem stupid to you.
The table on the right definitely is an improvement over the left. Though something about the concept makes me feel like the tables and wood items would have more... "laquer", like a bit of a clear coat. Could just be me.
Though I wouldn't stress too much on the table texture as it's going to be covered by a table cloth if you're following the concept.
Slapped a bunch of random smart materials for now but at least the issue with the texel density has been resolved! (hopefully)
Decided to take you guys' (and others) advice on the heat sink, they're all nice and thin now.
@Donato I like the new table it looks nice but I agree with the above comment, the scene doesn't give me the vibe of having super grungy furniture. Id decrease the grunge to being mostly on areas where people touch it a lot so the legs and not the top. If you can in substance painter it might give you some good results if you layer some materials with smart masks on them, I can tell you have some kind of mask the follows all of the curvature. Something like the ground dirt mask or an inverted moss from top mask could be cool
@fdfxd2 You're right on the table. I had thought of adding a thin sheet of marble. However your model is fantastic .
@ThisisVictoriaZ I tried to improve the textures and add more detailed models. I'm trying to learn substance painter at best. On some models I applied the ground dirt mask or curvature generator as you suggested and tried to think about how the object is used. For example, on the library, where in theory the books are inserted and removed, I tried to simulate the damage that comes after years of moving books. I also included other decorations. I hope to be making some progress. Here are some pictures:
@Donato The wood textures (in general) seem to be a bit noisy - indicating that its weathered, aged wood for example - but the geometry doesn't play into that. There are no weathered/broken/worn edges, especially on corners (where you'd expect wear and tear).
The wood textures I think need to be brought onto the same level as the marble and the floor, both of which look great. Its cliche but if you look at some of Blizzard's texturing - even stylized - there's a contrast between noisy/busy areas and areas where the eye can rest. In the case of Substance (which I'm still learning as well), if you apply a black/white mask to your materials you can mask off areas which could be rougher/noisier or more worn down - like edges/corners and leave the bigger panel area relatively clean with a bit of vertical wood grain coming through.
@Orbit Thank you so much for the suggestion. I redone the wood of the wall trying to respect what you said. If I did not do well, forgive me but the inexperience and the translation of google translate into Italian are not always understandable, they do not simplify my understanding. However I also show you the material in substance painter
If I'm wrong let me know (and please be patient and don't hate me).
My suggestion is somewhat related, I feel like individually baking and texturing every pillar and wall is probably not the the most efficient way of doing things.
Modelling and texturing a single 2k plane, that contains all the details and patterns you'd use on most of these walls,pillars,even props like tables/chairs, then uv mapping those props to work with the trim sheet, is how i'd go about texturing them.
all the shiny post processing is to distract you from how awful the aluminium material is
@fdfxd2 Thanks for the advice. It seems to me a difficult technique to master. I will try to use this technique in the next challenge starting from scratch. However your model looks fantastic. Personally I would have added a slight layer of dust and rust.
I continued to work on the textures trying to follow your suggestions by remaking the walls once again. In addition, I redone the texture of the books and added a thin marble slab to the tables. Once again I have to ask you for feedback.
Well it's looking as it should material wise as far as I can tell.
I'm not really an environment artist but I think letting in a bit more light through the windows, making the floor shinier, increasing the indirect lighting intensity, would improve the scene
Another change to the aluminium texture, probably wont be the last as I feel it's a bit too grungy, decided to bake the stitches of the seat. Not that many noteworthy changes to the model itself but over the past few days i decided to mess around and greybox a background scene.
@fdfxd2 Personally I tend to go towards a darker atmosphere so, for me, the second render, with the darker background and light, represents a nice step forward. In addition, I would add some signs of wear on the seat for example.
Another change to the aluminium texture, probably wont be the last as I feel it's a bit too grungy, decided to bake the stitches of the seat. Not that many noteworthy changes to the model itself but over the past few days i decided to mess around and greybox a background scene.
You tell me if it was a good idea
first one by far, you can see more of of the texture work!
This is my final piece. I made some changes trying to give more variety. I also added it on Artstation, if you want to see something more (https://www.artstation.com/artwork/v2Ae9a). Thanks everyone for all the advice .
^ I think it could have been better if the floor was glossier and the windows had an even stronger influence over the scene but otherwise decent scene.
(not sure if the sketchfab link works but you can find it in the artstation page if it doesn't)
I'm also calling my project done. I'm really happy with the leather and brass textures, not as happy with the aluminium textures but i think the modelling is strong enough to make up for that ( I should probably learn how to texture metals properly one of these days... )
Hi everyone, this is my rendition of the wildcard hard surface interior. Everything is rendered in UE4.
I am wondering if this piece (or any part of it) would be acceptable to have in my portfolio when applying for a junior environment artist position. If not, what should I improve for next time?
I have also thought about breaking some elements out as separate pieces for my portfolio, like the chair, lamps, and book trim sheet. I am curious as to your thoughts on this.
A personal critique:
I spent quite a bit of time trying to figure out the scale of the scene, but I never got it quite right. Adjusting the scale constantly meant I had to redo several pieces and do some patching in the end.
After doing a preliminary block-out, I switched to working on some of the more detailed pieces as a change of pace. This ate up a lot of time early on, when I should have been building the major parts of the scene.
I learned about trim sheets toward the end of the project. I used my first trim sheet to texture the roof and upper walls. If I had known about trim sheets at the start of my project, I would have made a new one at the start of the project to give the walls and roof a more cohesive look. For practice, I created a custom trim sheet to texture all of the books, and I think the result turned out quite well.
I would appreciate any feedback you may have. Thanks for reading!
Replies
I should probably give you something to improve on so the only thing I can see is the cooling fins on the engine at the front. They don't stick out enough for my liking and are quite thick. On a real engine they are thinner and longer to better disperse the heat, like this:
I suppose depending on your optimal viewing distance it could get a little noisy though. I am nitpicking; it looks awesome!
For an update - still working on textures, any help would be...helpful!
@Orbit are you planning on adding a shader or something to make the black outlines? I think that would do a lot for your piece
https://www.artstation.com/artwork/JlyGzR
Tell me what you think
Yours is looking slick! Really reminds me of old Disney film style. Is that toon shader exclusive to Blender only or is there a UE4 equivalent that you would recommend?
I love watching everyone's work progress.
I ended up choosing the motorcycle, which may come back to haunt me given how complicated it is. @fdfxd2 do you have any tips for interpreting the concept art? I see that you picked a bike to use for reference, but I'm having a hard time figuring out exactly what's what - even using references. The area around the engine is the worst. All of your forms look so polished and purposeful, how do you do it?
Yeah I think my thought process at the time was "oh it'll have a smaller polycount and will be more justifiable to bake if it's a few thick strips instead of many thin strips." I still kinda sorta stand by that but, I'm probably going to move the current fins to a texture atlas anyway, so I'll consider making them thinner again.
>do you have any tips for interpreting the concept art? I see that you picked a bike to use for reference, but I'm having a hard time figuring out exactly what's what - even using references. The area around the engine is the worst. All of your forms look so polished and purposeful, how do you do it? - @kelly_caroline
Helps to spend a few hours periodically searching for each part of the motorcycles individually,
Here's my attempt at the challenge. I chose the table since I really enjoyed the color palette and the interesting props. I had a lot of headache matching the scale of the props as well as the angle/rotation of them. I can't stress enough how helpful it was in the end to do measurements of props multiple times for the right sense of scale.
As far as performance is concerned, the UE4 gave me no problems. In addition, it has an excellent automatic LOD system that reduces triangles without losing detail (especially for distant models because you can't see the details). As far as textures are concerned, here I have the biggest problems because, being completely amateur, I have tried to use only free software (Blender, UE4, Quixel Megascan) but I believe that there is not yet a valid alternative to Substance Painter that I don't have (there is Quixel Mixer but I'm finding it difficult to use). In conclusion, I am not trying to create a procedural texture, indeed I should still learn the basics I think. Up to now I have simply taken one or more materials from Quixel Megascans and applied them to the model. So would you have any suggestions for improvement?
I do agree that there is no viable alternative to Substance software currently. Quixel Mixer can do a lot, (Keep in mind I haven't used it much) but it does not give you the freedom that Substance Painter does. Paying for Substance is 100% worth it in my opinion. Substance software is industry standard for a reason, and the price is pretty affordable for what it gives. (If you are a student, they have a free personal license option as well.) Learning Substance Painter alone will definitely help your art improve. This is how I think about it: paying for the right tools can save you a lot in the long run in terms of time and frustration.
But anyways, here's my first modeling attempt after 8 years
@Donato you've got some really good dramatic lighting in the scene! The textures on your counter and floor are a bit too big but looks ok so far. Would love to see some horizontal separation of planks on the floor so it doesnt look like each plank is super long is that makes sense.
I think Flipped Normals has a couple of good Substance Painter tutorial videos if you want to check those out. I also highly recommend getting a drawing tablet (sometimes called a "digitizer") if you don't have one already. It will allow you to work much more naturally when sculpting and painting. There are a few affordable options out there to pick from as well!
Regarding the detail level of your current model, I think some additional modeled details will go a long way. For example, compare your arches near the bookcases to the concept art arches. The concept art seems to indicate ornate details like filigree, carved capitals, etc. Textures will carry some of the weight in terms of detail, but the overall geometrical shape should be made more complex to sell the idea. The library concept art we are using gives off a steampunk vibe to me. I would look up some images of steampunk and Victorian interiors for inspiration on the types of details you could add.
Though I wouldn't stress too much on the table texture as it's going to be covered by a table cloth if you're following the concept.
@ThisisVictoriaZ I tried to improve the textures and add more detailed models. I'm trying to learn substance painter at best. On some models I applied the ground dirt mask or curvature generator as you suggested and tried to think about how the object is used. For example, on the library, where in theory the books are inserted and removed, I tried to simulate the damage that comes after years of moving books. I also included other decorations. I hope to be making some progress. Here are some pictures:
The wood textures I think need to be brought onto the same level as the marble and the floor, both of which look great. Its cliche but if you look at some of Blizzard's texturing - even stylized - there's a contrast between noisy/busy areas and areas where the eye can rest. In the case of Substance (which I'm still learning as well), if you apply a black/white mask to your materials you can mask off areas which could be rougher/noisier or more worn down - like edges/corners and leave the bigger panel area relatively clean with a bit of vertical wood grain coming through.
My suggestion is somewhat related, I feel like individually baking and texturing every pillar and wall is probably not the the most efficient way of doing things.
all the shiny post processing is to distract you from how awful the aluminium material is
I continued to work on the textures trying to follow your suggestions by remaking the walls once again. In addition, I redone the texture of the books and added a thin marble slab to the tables. Once again I have to ask you for feedback.
I'm not really an environment artist but I think letting in a bit more light through the windows, making the floor shinier, increasing the indirect lighting intensity, would improve the scene
first one by far, you can see more of of the texture work!
(https://www.artstation.com/artwork/v2Ae9a). Thanks everyone for all the advice
The voting thread for Challenge #67 is now live! Check it out here, and vote for what concepts you'd like to work on during July and August: https://polycount.com/discussion/220827/the-bi-monthly-environment-art-challenge-concept-thread-july-august-67
Hi everyone, this is my rendition of the wildcard hard surface interior. Everything is rendered in UE4.
I am wondering if this piece (or any part of it) would be acceptable to have in my portfolio when applying for a junior environment artist position. If not, what should I improve for next time?
I have also thought about breaking some elements out as separate pieces for my portfolio, like the chair, lamps, and book trim sheet. I am curious as to your thoughts on this.
A personal critique:
I spent quite a bit of time trying to figure out the scale of the scene, but I never got it quite right. Adjusting the scale constantly meant I had to redo several pieces and do some patching in the end.
After doing a preliminary block-out, I switched to working on some of the more detailed pieces as a change of pace. This ate up a lot of time early on, when I should have been building the major parts of the scene.
I learned about trim sheets toward the end of the project. I used my first trim sheet to texture the roof and upper walls. If I had known about trim sheets at the start of my project, I would have made a new one at the start of the project to give the walls and roof a more cohesive look. For practice, I created a custom trim sheet to texture all of the books, and I think the result turned out quite well.
I would appreciate any feedback you may have. Thanks for reading!