Hey guys, here is my latest piece. HP Should be about done. Just have to do some touching up and even out some edges I think. I'm still trying to wrap my head around how i can do a snub nose version of the LP without having different textures or a seam. But yeah, Snub Nose planned as well
Dem soft edges
Good as always man, love what you have done with that diamond pattern on the grip!(Mind sharing how your workflow on that )
Anyways, can't wait for the bake and textures!
Sure. I'd be happy to do a little write up on it. I'll have to share most the credit with the source i learned it from though Was largely derived from an image i found in How you model dem shapes thread a while ago.
Yeah, I forgot to mirror the bolt over to the other side of the grip. The grips will be mirrored and share the same UV space so there will be bolts on both sides. But that pic did remind me that i didn't put the bolt for the sight on the side of the barrel.
Just a quick ponder here while i was working on the LP. I always like the idea of mirroring UV's on any given part of a gun that couldn't be seen at the same time. But is that a better practice than just mirroring the entire piece outright? What i'm saying is, could i make the area in blue mirrored, Because most of those faces could never been seen at the same time, And do the top portion (in pink) as another UV island un mirrored. I'd imagine as long as its on its own Island, it wont cause issues with baking. But it is just a silly idea? should i just mirror right across the X axis and call it good?
LP Topology not final here. I know everyone likes to throw around the term "next gen" so i might as well. I plan on giving this gun a higher budget on tris as well as 4K maps.
Not thinking about it too hard, I'd say the image you have above makes sense, so the blue and pink areas will be separate islands rather than just having symmetry on one axis.
I say the above example works better mainly because the uv seams (and therefore hard edges) lie in better places than it would if you mirrored down the center. Don't be afraid to throw down quick UV islands and then test bake either.
Nice HP too, haven't seen any references of that gun that I can remember but is the trigger width right? Seems wider than in other guns I've seen. Also 4k maps is preeeeety huge, especially for a small gun. You can always scale it down after I suppose, but it's usually better to work at the highest intended level so you waste time on small detail that gets lost anyway.
@Bek: yeah, I can always test and change it I just like to gather thoughts while i'm working towards that point. And the gun is overall quite a bit wider than reality. The right would probably be closer to real (just doing a quick scale, Some things would be even thinner).
To my surpise as well, This guns trigger is just as wide as the trigger guard. So realitively speaking, The trigger on my model is too small! But it looks rather massive any bigger than it currently is.
You can see here that the side of the trigger width is about flush to the guard. But that was just too big on mine, Like i said. Thats also why i round the corners on the top, Instead of a boxier edge like reality.
It may also be worth noting, This gun is replicated A LOT. So some of my reference are inaccurate to the original, and different parts of the gun are based on both replicas and the original. I tried to find as much consistency as possible though.
@4K maps in general: Yeah, i know its massive That's kinda the point. I understand it wouldn't fair well in a game today, But its a real possibility that this gun will find its way into a game a couple years down the road. Even if tech hasn't jumped to that point by then, i can always scale it. But That is a good point, Being able to texture details that wont matter. But i feel like the process will go nearly the same, Just have more density and clarity to what details would have been put just the same in a 2K map.
Thanks guys. I hope to get baking done today and then let it be for a bit. Been putting of some homework a bit long than i like to admits while working on this
Yeah, I forgot to mirror the bolt over to the other side of the grip. The grips will be mirrored and share the same UV space so there will be bolts on both sides. But that pic did remind me that i didn't put the bolt for the sight on the side of the barrel.
Thx.
How are your PJ's ?
:poly121:
Ah i see, and since the right side wont be seen all to much, i guess a symmetry would work fine.
btw any reason you left out the ring at the bottom of the grip?
Yeah. I might add that metal ring on the bottom on the grip(frame) and put a lanyard through/around it if i get something that looks cool will have to consult with zbrush and see what i can work up. and figure out if i can have it sway and move in real time with its own physics.
And that sucks. You should buy a new pair! Skip your daily bottle of wine and you'll be well on your way !!
If its going to be visible from a FPV perspective, it should not be mirrored. If its going to face directly at the camera, scale it up 25% or so, if you're going to zoom in on it for an iron-sights view, scale it up another 25% or so. If it will rarely/never be seen, scale it down a bit. Everything else, maintain consistent uv density.
Mirror what you will only see from one side, no sense painting it twice. If you have asymmetric details that overlap a mirrored section be careful though, as your AO will be asymmetric as well, and with textures sizes as large as you're talking, there's is no reason to optimize the uv layout excessively. So make sure any unique AO detail has unique uv space as well.
In your mirroring example image there, the bit closest to the FPV you have mirrored, don't mirror that. Map that unique like the top bit.
@EQ: Thanks! That really helps me with the UV part of the process. Was kind of an arbitrary process for me before. But that all makes great sense. I'll do my best to implement all those ideas. And thanks on the HP comment. Really means a lot to me. You pro's inspire me a lot, and its nice to here its lookin ok
@Polygoblin: Thanks man. And yeah. its polycount! That's why I'm here. Always good advice to be heard, and too much stuff to learn for what little time we have
As far as progress, I got nearly done with the LP, Before school work demanded my attention. (like 2 cuts and edge removals away, literally) but i should have the piece baked by or during the week end. I hope. Have an intuos5 medium due in tomorrow, so I'll be eager to get that worked into my texturing routine.
Hey, Quick question guys, I wasn't sure where too look for the answer sense its not normal/UV/baking specific.
Should i mirror the low poly geometry before or after baking? It seems like you would get normal errors if you didn't do it before, But i can see the normals baking over eachother and kinda wonking out.
any advice you guys have on that would be great. And i understand i could do it both ways, and figure it out, But my point in asking is to learn the right way. I could get a "Fine" result but if there was a better way to do it, I'd like to know it. So I just like to learn how to do it right
Just finished up the LP (finaly) and will be doing the UV's today/tomorrow and bake over the weekend. Should be good. Everyone brace for the tris count though. i know i'm going to get chewed a new one.
Ah. I see. Glad i asked! i had no idea. Thanks man. Will definatley do this. i assume no matter how many times you do that (1,2,3,4 times) That UVs on the outsides of the 1X1 will still use that space?
Err. No. That doesnt make sense...
I assume this is going towards :Mirror, offset, bake, Then offset back?
Yes, you could offset it by a 100 units, and it would still be fine as long as you don't start using numbers like "1,85".
And yes, mirror -> offset uv's -> do smoothing groups/doublecheck them -> bake.
Offsetting them back is optiona, the textures will show up fine (atleast in max/udk/most common engines/previewers)
(if this doesn't make sense why this is, look up tileable textures)
No, that makes perfect sense. Thank you. I understand tiling textures, But I guess I never thought of it in relation to an asset like this. As far as UV's.
Sorry about the massive delay here. System crashed a few weeks ago during finals week, So had to restore and get all the programs reactivated an all that jaz. So got back at this week, and finaly got to stickin it in the oven.
Disregard all the (obvious) bake errors, I'm pretty happy where its headed .
Looks good to me, just some very small nitpicking, except for point 1 it really doesn't matter all that much.
-You have some of that diamond pattern poking trough the hole
-The indent in the screw seems a bit large, and the edge on the indent seem a -bit sharp compared to the rest
That little circular indent seems a bit wonky, maybe remove it, and do it with ndo or whatever you prefer/use for that.
and maybe take out 1 or 2 segments from the trigger and put it in that thing that you pull backwards at the back (is it called a firing pin? dnno)
For that grip hole, yeah. I've yet to drop the bolt there. its going to be a copy of one of the bolts that's up near the pivot area, So that will get covered up. Same with that other wonky indent. It will have a bolt. I'm just being lazy as hell on my presentation atm :poly136:
I do agree on the lever (basically releases the barrel to pivot forward and eject the spent rounds). I already have a lot of geo in that piece though, In the bevels and the top portion due to it being right in the FPS view. What do you think of the last image? Do you think it could use segments from that point of view as well?
Ah. yeah. Didn't really notice the hammer. It is kinda low atm. I'm just sacred of adding geo, cause it know everyone here is going to beat me with a stick once I show wires/reveal the tris count
The indentation in the middle of the case may go all the way around the bullet, only partially, or it might be missing altogether, depending on the manufacture.
Wow. I must have been really lazy when looking for refs, Cause i couldn't find much on the round for it. Thanks for that. It really felt off. I think I'll redo the HP and rebake the bullet a bit more accurately. I did make it long though, to better fill the chamber.
So I had to revisit some UV stuff after finding some odd seams. Rebaked. And finally got onto texturing. Trying to really nail the material definition before i even get to deep into detailing and texturing. Which is why I'm posting. Any feedback you guys have on how the materials read would be great. Obviously, This is early wip, So lots of break up in the diffuse coming, Edge wear, Finger prints, all that good stuff ^^
Very nice bake man. Tough to tell how the materials read without some color and details. Looks good so far.
@4k maps for a pistol. I've stated this before, even at 2k x 2k there are more pixels in the texture map of the pistol than the amount of screen pixels that can be displayed for it at 1080p. You could make a 16k x 16k map for that sucker and at game res, in first person view there will be no visual difference then 2k, maybe even 1k map. I know the next gen promises bigger numbers, but authoring textures at 4k can be a major resource sink compared to a 2k or 1k maps. With that said, its for your folio, so GO CRAZY! :]
ATM its 4K X 2k DNSG. I hadn't found a reason not to stick with that, until i actually read one of your other posts in another thread about the same thing you just stated, And that logic about pixels in relation to screen space is perfectly sound! Because of that I'm definitely resizing to 2k X 1K.
And thanks on the bake. I hear you as far as not being able to judge the textures yet. Just trying to develop the look of the material as best i can between just a few layers of texture and the correct balance between Diff spec and gloss values.
ATM its 4K X 2k DNSG. I hadn't found a reason not to stick with that, until i actually read one of your other posts in another thread about the same thing you just stated, And that logic about pixels in relation to screen space is perfectly sound! Because of that I'm definitely resizing to 2k X 1K.
And thanks on the bake. I hear you as far as not being able to judge the textures yet. Just trying to develop the look of the material as best i can between just a few layers of texture and the correct balance between Diff spec and gloss values.
More to come soon!
I think most people forget screen pixel resolution when coming up with texture size numbers. Just trying to spread the love.
nice material definition you have so far! I personally like the metal definition you have so far, i cant seem to even nail these simple types of materials
I would say that you could break up the definition at the screws and maybe the grip could use some color as well so that not everything will just look like metal and provide variety to the revolver.
Working hard on getting it just right. Not sure if I'm succeeding. Its hard to judge objectively when your so close to a piece. So. Obvious WIP but any feedback is coo.
Also. A question, Does the color in metals like the below exist mainly on the diff or the spec? i have a bit in the diff and and a bit less on the spec. Not sure if its strong enough or not. Its an extreme example, But I think you guys will get what I mean.
My monitors are both pretty shitty so I'm not sure how levels/layer opacities are reading on proper panels. Will have to balance them all out on cintiqs at school one of these days.
That was a great read. I appreciate the link. But it rendered more questions than clarity for me . Is it a hard fast rule that conductors Diff/Spec correspond more or less in Hue? Obviously there are exceptions to every rule, But for the most part, is this how it goes? right now i have a red/brownish diff with a primarily blue specular. Does it look funky cause I'm neutralizing the color?
I just took a look at Fletches AR15 and and he used a blue hue on diff and a red/brown on spec, and his piece turned out quite nice. Maybe this is because his weapon is coated with a material that reacts more like a dielectric? EQ's camera model follows this suite with bare metals having similar hues, But the coated metals having opposites. I guess my question is, is this Webley coated? and does it appear Conductive or not?
From all the images I've seen of the gun, it appears to have a redish diffuse and a blue spec. But maybe the variable lighting in images has been throwing me for a loop. maybe someone can tell me which looks more natural and true to the gun.
I believe the gun is painted, so only where the paint is wearing thin will there be a more "conductor" like appearance. The reflections/spec are cooler than the diffuse/reddish color of the gun metal, though it's hard to tell if it's actually blue or white, I'd try a red diffuse and whiter spec. In the red diff/blue spec example the red diffuse is getting over powered by the blue spec.
In Jfletcher's gun, the red spec is not very powerful, just enough to offset the cooler color in the diffuse. Also some engines add spec differently, so different engines may look very different. It can either be additive specular or gamma corrected.
Hey man, its awesome to see how well made and how all that time you spent on the gun is paying off, you are on the final stretch keep up the awesome work!
Hey guys. I was curious if you have any recommendations for a lighting set up when texturing within marmoset. It's hard enough to define materials even if you don't factor in that those definitions are a dependent on the light hitting them. So am i best off with a pure white? value of 1? Is there a certain sky that is good to texture with without "Distorting" what your working on?
Was going through them, again, And just thought I'd ask this question. Found a cool on that made me want to make this a BSI tribute piece The revolver in BS1 was a webley style design, so that excuse might hold air.
Anyway. Did some gloss/spec balancing. Gonna leave that for now and finish edge wear. Then at school maybe i can balance it all out, Hit the grip/bullet, and get close to wrapping it up.
Replies
Well done!
@Pedro: Thanks! I appreciate that. I hate to pry, But any tips or advice? I really like getting picked on
get baking!
Lol. Fair enough. Gonna start strippin edges off the HP.
Good as always man, love what you have done with that diamond pattern on the grip!(Mind sharing how your workflow on that )
Anyways, can't wait for the bake and textures!
1 minor thing, on the grip's right side, the diamond in the center, does it really end like that? (the circle indent)
might just be refs, but looks different here; http://www.classic-firearms.co.uk/uploads/1/0/8/3/10836860/6843544_orig.jpg
Thx.
How are your PJ's ?
:poly121:
LP Topology not final here. I know everyone likes to throw around the term "next gen" so i might as well. I plan on giving this gun a higher budget on tris as well as 4K maps.
Not thinking about it too hard, I'd say the image you have above makes sense, so the blue and pink areas will be separate islands rather than just having symmetry on one axis.
I say the above example works better mainly because the uv seams (and therefore hard edges) lie in better places than it would if you mirrored down the center. Don't be afraid to throw down quick UV islands and then test bake either.
Nice HP too, haven't seen any references of that gun that I can remember but is the trigger width right? Seems wider than in other guns I've seen. Also 4k maps is preeeeety huge, especially for a small gun. You can always scale it down after I suppose, but it's usually better to work at the highest intended level so you waste time on small detail that gets lost anyway.
No point in texturing the same thing twice
To my surpise as well, This guns trigger is just as wide as the trigger guard. So realitively speaking, The trigger on my model is too small! But it looks rather massive any bigger than it currently is.
You can see here that the side of the trigger width is about flush to the guard. But that was just too big on mine, Like i said. Thats also why i round the corners on the top, Instead of a boxier edge like reality.
It may also be worth noting, This gun is replicated A LOT. So some of my reference are inaccurate to the original, and different parts of the gun are based on both replicas and the original. I tried to find as much consistency as possible though.
@4K maps in general: Yeah, i know its massive That's kinda the point. I understand it wouldn't fair well in a game today, But its a real possibility that this gun will find its way into a game a couple years down the road. Even if tech hasn't jumped to that point by then, i can always scale it. But That is a good point, Being able to texture details that wont matter. But i feel like the process will go nearly the same, Just have more density and clarity to what details would have been put just the same in a 2K map.
Thanks guys. I hope to get baking done today and then let it be for a bit. Been putting of some homework a bit long than i like to admits while working on this
Ah i see, and since the right side wont be seen all to much, i guess a symmetry would work fine.
btw any reason you left out the ring at the bottom of the grip?
PJ's are busted still
And that sucks. You should buy a new pair! Skip your daily bottle of wine and you'll be well on your way !!
Basic rules of thumb:
If its going to be visible from a FPV perspective, it should not be mirrored. If its going to face directly at the camera, scale it up 25% or so, if you're going to zoom in on it for an iron-sights view, scale it up another 25% or so. If it will rarely/never be seen, scale it down a bit. Everything else, maintain consistent uv density.
Mirror what you will only see from one side, no sense painting it twice. If you have asymmetric details that overlap a mirrored section be careful though, as your AO will be asymmetric as well, and with textures sizes as large as you're talking, there's is no reason to optimize the uv layout excessively. So make sure any unique AO detail has unique uv space as well.
In your mirroring example image there, the bit closest to the FPV you have mirrored, don't mirror that. Map that unique like the top bit.
Highpoly looks great, should bake really well.
@Polygoblin: Thanks man. And yeah. its polycount! That's why I'm here. Always good advice to be heard, and too much stuff to learn for what little time we have
As far as progress, I got nearly done with the LP, Before school work demanded my attention. (like 2 cuts and edge removals away, literally) but i should have the piece baked by or during the week end. I hope. Have an intuos5 medium due in tomorrow, so I'll be eager to get that worked into my texturing routine.
Should i mirror the low poly geometry before or after baking? It seems like you would get normal errors if you didn't do it before, But i can see the normals baking over eachother and kinda wonking out.
any advice you guys have on that would be great. And i understand i could do it both ways, and figure it out, But my point in asking is to learn the right way. I could get a "Fine" result but if there was a better way to do it, I'd like to know it. So I just like to learn how to do it right
Just finished up the LP (finaly) and will be doing the UV's today/tomorrow and bake over the weekend. Should be good. Everyone brace for the tris count though. i know i'm going to get chewed a new one.
Looking nice btw, hurry up with that LP and bake
Else if you don't do this, it will bake multiple times to the same UV's and that will cause artifacts.
picture to make it more clear:
http://img526.imageshack.us/img526/4296/uvthing.jpg
Err. No. That doesnt make sense...
I assume this is going towards :Mirror, offset, bake, Then offset back?
And yes, mirror -> offset uv's -> do smoothing groups/doublecheck them -> bake.
Offsetting them back is optiona, the textures will show up fine (atleast in max/udk/most common engines/previewers)
(if this doesn't make sense why this is, look up tileable textures)
Disregard all the (obvious) bake errors, I'm pretty happy where its headed .
Lots to fix. But got excited. So here you go.
LP+Normals
*Too many images, See below*
All comments/feedback welcome. Idk if its lookin so shitty no one wants to comment, so on par that it doesn't stand out, or what. Throw me a bone.
Edit: Shell casing is pretty ugly. Might go back and remodel that HP or add some geo to the LP or something.
*Too many images see below*
-You have some of that diamond pattern poking trough the hole
-The indent in the screw seems a bit large, and the edge on the indent seem a -bit sharp compared to the rest
That little circular indent seems a bit wonky, maybe remove it, and do it with ndo or whatever you prefer/use for that.
and maybe take out 1 or 2 segments from the trigger and put it in that thing that you pull backwards at the back (is it called a firing pin? dnno)
For that grip hole, yeah. I've yet to drop the bolt there. its going to be a copy of one of the bolts that's up near the pivot area, So that will get covered up. Same with that other wonky indent. It will have a bolt. I'm just being lazy as hell on my presentation atm :poly136:
I do agree on the lever (basically releases the barrel to pivot forward and eject the spent rounds). I already have a lot of geo in that piece though, In the bevels and the top portion due to it being right in the FPS view. What do you think of the last image? Do you think it could use segments from that point of view as well?
the hammer?
And thanks.
The indentation in the middle of the case may go all the way around the bullet, only partially, or it might be missing altogether, depending on the manufacture.
Thanks!
@4k maps for a pistol. I've stated this before, even at 2k x 2k there are more pixels in the texture map of the pistol than the amount of screen pixels that can be displayed for it at 1080p. You could make a 16k x 16k map for that sucker and at game res, in first person view there will be no visual difference then 2k, maybe even 1k map. I know the next gen promises bigger numbers, but authoring textures at 4k can be a major resource sink compared to a 2k or 1k maps. With that said, its for your folio, so GO CRAZY! :]
And thanks on the bake. I hear you as far as not being able to judge the textures yet. Just trying to develop the look of the material as best i can between just a few layers of texture and the correct balance between Diff spec and gloss values.
More to come soon!
I think most people forget screen pixel resolution when coming up with texture size numbers. Just trying to spread the love.
I would say that you could break up the definition at the screws and maybe the grip could use some color as well so that not everything will just look like metal and provide variety to the revolver.
Looking forward to see your texture work!
Also. A question, Does the color in metals like the below exist mainly on the diff or the spec? i have a bit in the diff and and a bit less on the spec. Not sure if its strong enough or not. Its an extreme example, But I think you guys will get what I mean.
My monitors are both pretty shitty so I'm not sure how levels/layer opacities are reading on proper panels. Will have to balance them all out on cintiqs at school one of these days.
I just took a look at Fletches AR15 and and he used a blue hue on diff and a red/brown on spec, and his piece turned out quite nice. Maybe this is because his weapon is coated with a material that reacts more like a dielectric? EQ's camera model follows this suite with bare metals having similar hues, But the coated metals having opposites. I guess my question is, is this Webley coated? and does it appear Conductive or not?
From all the images I've seen of the gun, it appears to have a redish diffuse and a blue spec. But maybe the variable lighting in images has been throwing me for a loop. maybe someone can tell me which looks more natural and true to the gun.
Sorry to go so image heavy on you guys.
Single white spotlight in marmoset used here
In Jfletcher's gun, the red spec is not very powerful, just enough to offset the cooler color in the diffuse. Also some engines add spec differently, so different engines may look very different. It can either be additive specular or gamma corrected.
Was going through them, again, And just thought I'd ask this question. Found a cool on that made me want to make this a BSI tribute piece The revolver in BS1 was a webley style design, so that excuse might hold air.
Anyway. Did some gloss/spec balancing. Gonna leave that for now and finish edge wear. Then at school maybe i can balance it all out, Hit the grip/bullet, and get close to wrapping it up.
@Stromfreek: Thanks man !
for lulz:
VERY INFORMAL EXPERIMENTAL BIOSHOCK INFINITE TYPE LIGHTING. Doesn't represent the texture well at all.