Not sure how final everything is, but at the top of the wood grip, there's a little curve indent where the selector fits into. Just a minor observation, coming along great!
btw, thanks to everyone for the kind comments above. more work. fixed the bullets and barrel/laser. i don't have any real reference of the hinge or parts of the frame or the feed area, did my best. should get to the lowpoly soon
Ben tell me you're going to do PBR this time round? Your Atchisson was great work but it felt like it was held back as a pure show piece by staying with previous-gen DSGN maps and not going PBR (and by the 1024 textures as well).
Laurens, c'mon dude... Using diffuse/spec has nothing to do with pbr. The metalness workflow is just a different way to pack the content. Can we please stop perpetuating this myth?
Hm it looks like you should use a higher subdevision for your highpoly as there are all these pinchy artefacts due to the too low polycount of the highpoly. Your highpoly images look kinda like a baked lowpoly
If you turned a subdivision down to export to marmoset, then nevermind!
Don't be so scroogey on those polys man, you can make those guns really flawless, why not go for it : P Step away from those third person specs, it only hurts your portfolio without reason. You can always go down, nobody is questioning that
Also your edges are like over the top soft. If thats your style and it should appear a little stylized, I get that, but in case you aim for realism, I would use bit little less chunky edges definitely. I feel it hurts the model.
There is no reason to have perfect edge consistency either. Not all edges are the same width (in RL) Some are softer some are razor sharp, you can play with that. It feels like you tried making all exactly the same.
Proportions and all looks spot on man, but its your game, just want to help!
but if his high poly looks like a baked lowpoly.. when he bakes it should be all good yea? if this were a cinematic asset then sure but it works fine for a game model. i get what you're saying shrike but at the end of the day if this is intended to be something like a 3rd person weapon.. just adding more geometry for the sake of it looking smoother could be a waste of time imo. even in the fpv shot it looks pretty damn fantastic. the smoothing may help with a few pixels of shading but aside from that the gains would be pretty minimal
thoughts always appreciated and considered :thumbup: thanks for the replies everyone
didn't have a lot of time today again. usually i get up at 5:30 so I can work on shit but last couple days it's been more like 7:30 which screws me up. i wanted to get something done today though so i did the cylinder and hammer after work. altogether it's like 2k so far. been UVing progressively as well. normally i would do test bakes first but hey... nothings packed yet.
Laurens, c'mon dude... Using diffuse/spec has nothing to do with pbr. The metalness workflow is just a different way to pack the content. Can we please stop perpetuating this myth?
Oops! I really didn't mean this in a bad way, i genuinely thought he'd used old-school maps for the Atchisson since it didn't feel properly PBR to me. Looking at it again, I can see why I thought so; a lot of values are not conform PBR standards, it's done in still a bit of an old-school mentality that I saw many of our artists do when we just got started with it.
But whatever, if critique like that is not appreciated I'll shut up.
Oops! I really didn't mean this in a bad way, i genuinely thought he'd used old-school maps for the Atchisson since it didn't feel properly PBR to me. Looking at it again, I can see why I thought so; a lot of values are not conform PBR standards, it's done in still a bit of an old-school mentality that I saw many of our artists do when we just got started with it.
I actually went and checked his values and most of them are quite close to what I would expect for pbr, maybe a few % over in some cases, his wood material was like 25-29% sRGB which is 5-7% linear, so totally in line with the typical 2-8% insulators reflect in (plus its a lacquered material, so you get reflection from the wood itself and coating).
There was some minor stuff like subtle variation in the spec that doesn't make a whole lot of sense physically, but it wouldn't really make much of a difference to the end result.
But whatever, if critique like that is not appreciated I'll shut up.
Haha, no nothing like that, I thought it was just another "oh you're not using metalness, must not be pbr" reply that I see so often these days. Sorry if I came off like a dick. I would actually be curious to know which values you thought were wrong.
i think the latest rendering crazes can make us too uptight about values. i say as usual that we should compare results to references and go from there. sometimes i experiment and do things you shouldn't do, like put colors in dialectric specs. i suspect this is what laurens is getting at. i dunno though. maybe we'll find out later it's not so wrong. oil from hands. i do not know.
got the frame about half-done, meanwhile did a bake test for these parts
thanks guys. still crunching the lowpoly together, and still more to do.
i don't usually do HP reduction to get an LP, usually i do a blockout and then fork it into both the HP and the LP. on this particular project i went block > HP > LP, and well, it's just a bad way to work imo. huge waste of time deleting all this shit and stitching it together.
Looks neat but I'm really not liking the whole no subdiv hp thing. The idea is cool but it doesn't work too well for a piece this small imo. Bakes look super wonky in some places throwing the whole thing off. Since realistically this wont be used anywhere except portfolio (like that atchisson), I'd subdivide some areas for the sake of a clean looking bake.
sometimes i make geo mistakes, you can make bad geo with subdivision too. i don't see this as a battle of techniques, i still subd model all the time. but i didn't on this.
thanks harry. i've been pretty good about doing at least a little work on these every day so it keeps the clip up. reusing the texture (cobray > atchisson > this) helps too.
Is PBR like the new thing? or is using diffuse/norm/spec like going away?
PBR refers to a loose collection of rendering features (conservation of energy, linear space lighting, fresnel, etc) which won't be going anywhere because they're realistic elements of how light works. Diffuse/spec/gloss is just an alternative to color/metalness, nobody knows if one of those will come to dominate or if we'll just keep using both for a while, they're kind of interchangeable and it doesn't matter a lot.
I like the extra control you get with D/S/G but i also realize it lets me do unrealistic things sometimes compared to a metalness workflow. I've been meaning to try working with metalness, i actually just keep forgetting lol, but i understand it fine.
The big myth is that the d/s/g/n workflow is somehow "not pbr". That is simply not correct in any way, its really important to understand that the d/s/g/n workflow is simply a set of inputs, just like the metalness workflow, neither are mutually exclusive or inclusive to PBR.
When using logical values, the d/s/g/n workflow can be just as physically accurate as the metalness workflow, and actually moreso on a technical level, for instance non-metals in a metalness workflow typically get tagged as uncolored 4% reflective while in reality various insulators reflect in the 2-8% range. Also, while UE4 has to some extent popularized the metalness workflow with their PBR system, Cryengeine uses the specular workflow for their PBR, and we at Marmoset support both workflows.
Now, it can be argued that if your texture inputs for the d/s/g/n have illogical values, then you're breaking the system. This applies to the metalness workflow in exactly the same way though. I see a lot of inexperienced artists using completely random values in their metalness map, which essentially breaks the whole pbr system.
What really makes a pbr system is two things:
1. Shaders that simulate the physics of light and matter as well as possible
2. Art content that is calibrated and uses logical values as much as possible
Whether you're using full color specular maps or metalness maps has no bearing on whether the end result is physically accurate. Both systems can be broken to the point where physical accuracy is lost, and both input systems could be implimented in a NPR (non-photo realistic) rendering pipeline.
I hope this helps. I'm actually working on an article that explains a lot of this (and how to easily convert content from one workflow to another). Though much of what I've written here is explained in my previous tutorial as well: http://www.marmoset.co/toolbag/learn/pbr-practice
Replies
Looking forward to seeing progress.
Not sure how final everything is, but at the top of the wood grip, there's a little curve indent where the selector fits into. Just a minor observation, coming along great!
Keep it up, nice work as always!
Cheers!
Nice work! Interesting how the cylinder opens up.
Oh, Jack, yes it all makes sense now.
teeny update
realized like 5 minutes ago that the barrel is the bottom one, not the top, lol, so i'll have to close that up.
The bullets for this gun are rather odd; they're cylinders rather than the traditional shape, and they protrude a bit from the clip.
The barrel is not the top cylinder, but the lower one; the upper cylinder is for the laser sight.
didn't really have a morning today so i just did lowpoly/uvs for a couple parts after work
heres another shot of the HP
If you turned a subdivision down to export to marmoset, then nevermind!
Don't be so scroogey on those polys man, you can make those guns really flawless, why not go for it : P Step away from those third person specs, it only hurts your portfolio without reason. You can always go down, nobody is questioning that
Also your edges are like over the top soft. If thats your style and it should appear a little stylized, I get that, but in case you aim for realism, I would use bit little less chunky edges definitely. I feel it hurts the model.
There is no reason to have perfect edge consistency either. Not all edges are the same width (in RL) Some are softer some are razor sharp, you can play with that. It feels like you tried making all exactly the same.
Proportions and all looks spot on man, but its your game, just want to help!
didn't have a lot of time today again. usually i get up at 5:30 so I can work on shit but last couple days it's been more like 7:30 which screws me up. i wanted to get something done today though so i did the cylinder and hammer after work. altogether it's like 2k so far. been UVing progressively as well. normally i would do test bakes first but hey... nothings packed yet.
Oops! I really didn't mean this in a bad way, i genuinely thought he'd used old-school maps for the Atchisson since it didn't feel properly PBR to me. Looking at it again, I can see why I thought so; a lot of values are not conform PBR standards, it's done in still a bit of an old-school mentality that I saw many of our artists do when we just got started with it.
But whatever, if critique like that is not appreciated I'll shut up.
I actually went and checked his values and most of them are quite close to what I would expect for pbr, maybe a few % over in some cases, his wood material was like 25-29% sRGB which is 5-7% linear, so totally in line with the typical 2-8% insulators reflect in (plus its a lacquered material, so you get reflection from the wood itself and coating).
There was some minor stuff like subtle variation in the spec that doesn't make a whole lot of sense physically, but it wouldn't really make much of a difference to the end result.
Haha, no nothing like that, I thought it was just another "oh you're not using metalness, must not be pbr" reply that I see so often these days. Sorry if I came off like a dick. I would actually be curious to know which values you thought were wrong.
got the frame about half-done, meanwhile did a bake test for these parts
Nice modeling.
i don't usually do HP reduction to get an LP, usually i do a blockout and then fork it into both the HP and the LP. on this particular project i went block > HP > LP, and well, it's just a bad way to work imo. huge waste of time deleting all this shit and stitching it together.
Is PBR like the new thing? or is using diffuse/norm/spec like going away?
pic
Most of these wont be too noticeable with textures but those that will, will make the whole thing seem half assed.
e: annd you ninja'd me with basic textures pic
dialing in values here
Can we get another Rising Sun camo skin?
PBR refers to a loose collection of rendering features (conservation of energy, linear space lighting, fresnel, etc) which won't be going anywhere because they're realistic elements of how light works. Diffuse/spec/gloss is just an alternative to color/metalness, nobody knows if one of those will come to dominate or if we'll just keep using both for a while, they're kind of interchangeable and it doesn't matter a lot.
I like the extra control you get with D/S/G but i also realize it lets me do unrealistic things sometimes compared to a metalness workflow. I've been meaning to try working with metalness, i actually just keep forgetting lol, but i understand it fine.
Someone asked me something similar in a pm, so I'll repost it here as well.
Sorry for tl;dr Ben
Thanks for taking the time to post wire-frames and everything as you make this.