guns, lots of guns
greentooth
fps weapons, looking for some feedback/crit. I feel like I'm still learning the ins and outs of baking normalmaps, but I've made good progress since starting on this stuff. Be brutal please 
this ones an oldie, diffuse lighting
I feel its a strong piece though, so would like to know where it could be improved aswell:
this ones an oldie, diffuse lighting
I feel its a strong piece though, so would like to know where it could be improved aswell:
Replies
I think maybe you're trying to rely on diffuse a lot while only really half trying to make a specular map. With current gen assets, espescially shiny metal stuff, the spec should have almost as much attention as the diffuse.... I'm actually really not sure what to say, like, there's no "mistakes" to point out on the specmap because all it really is at the moment is basetexture with some scribbles on it. You just have to like, actually make a spec map.
And while realism isn't always the goal, modeling things carelessly will degrade your observational skills as an artist which are extremely important for adding believability to a piece.
I wont do a full rant on this as you've heard it all before from me, but i think there is a place for sticking to refs, and there is a place for just doing the asset. In this case quickly looking over his stuff as far as accuracy is concerned, i dont really see glaring flaws. So while i agree its very important to know how to match a ref/concept, you reach a point of diminishing returns. Especially when you're doing real weapons, people get extra obsessive about things.... Which is good and has a place, i dont really see any of these assets as not matching refs to the point where it is a concern however. We could both spend a good deal of time going over it with a fine-tooth comb, and find some stuff, but its not like the magazine on his m4 is backwards or something.
What i'm saying is that from his work, i would say he put a good deal of effort and time into understanding his refs.
As far as presentation is concerned I'd steer clear of any light setup that shades to pure black, why not try experimenting with our new shader release: http://boards.polycount.net/showthread.php?t=72861 You can have alot of fun with tinkering ambient lighting in it :poly124:
The colt feels very flat, it needs some beefing up. Some simple vertex scaling might be enough. I also think the materials look too much like rock.
And some small issues like the magazine catch needing bumps.
The ??? I have no idea of model accuracy, but again, the materials feel off. Also, you could probably ditch about 3 loops at the rear of the main cylinder, and reuse a bit to smooth out the grip.
The M4 looks nice, but it's a bit too simple in places. It would really benefit from holes/dents in the stock and flash suppressor, more detail around the front sight, and some general detailing. The materials feel a bit off too, can we see some flats?
In general, I feel the materials could use some more work, and it seems as though you're not using any baked normal maps, only xNormaled heightmaps for ridges etc. You would probably benefit a lot from practicing on making metal textures for a bit.
Edit: wow I pause writing a post for half an hour and there's like 10 posts.
Harry, like i said, we could both pick stuff out, but i feel any inaccuracies we would find are generally minor, so the insinuation that it was done carelessly is off base. I'm not going to argue how accurate/inaccurate specific things are with you.
At a point in production you have to choose quality over accuracy, and while you surely will tell me that quality = accuracy(and i will agree to an extent), there comes a time when you're on a deadline where you need to make a conscience choice to either delay, dig back through your refs, rework things, or go on ahead and make what you have as good as it can be. Delaying is generally not an option in a professional environment, so knowing the point where you need to polish and stop worrying about accuracy issues is very important. Not to say that being accurate in the first place isn't important, on the contrary modeling things accurately, and managing your time well are just about the two most important skills you can have.
Again if his stuff was glaringly inaccurate, i would agree with you 100%, as it is i need to spend time researching your points to even have an opinion on the matter, and as someone who knows a little about guns, i feel when you're at that point, you've gone into excess.
I'm no expert so i'll leave that too the big boys but the 1911 seems too colorfull to me. typically you'd have the metal one shade and then the grip would be either wood like you've got or a dark plastic? the two toned metal just doesn't look right.
That's pretty much what i'm saying.
I mean it's obvious already that his intentions aren't to go back and fix this stuff up retroactively, but if you can nip stuff like what i mentioned in the bud before it's too late to fix then you have a "free" increase in quality. It doesn't actually cost any labour, just training your eye to see certain things and interpret refs properly - Which imo is the most important thing in general "art."
Aside from that I think it's valid for me to point something like that out, looking at your own work and figuring out things you misinterpreted about the nature of the object helps to actually train you to notice that kind of stuff first time around. The most common mistake is people making perfectly good meshes, modelling all the parts right, but not actually considering mentally the scaling of the object and how it would effect the materials.
It's the exact same in character art, for example, when you model clothing, the size of the folds is dictated by the thickness of the material. Same deal here, the kind of polymers they use cant really maintain that hard of an edge at that scale.
It's not really a finnicky searching-through-refs thing so much as a putting a little extra thought to add some believability to the piece thing.
It's pretty much the same thing as material definition, but part of the modeling process.
I'm sure most people wouldn't notice any problems as is, but i guarantee if the model was made with those considerations well in mind, the final result would be more convincing than if not (which conversely isn't always the case with technical, part-for-part accuracy)
sorry for TL;DR and don't get the idea i'm trying to be argumentative about anything other than the validity of the practise of observation in general, Like, that particular gun is just the example that is on hand.
In fairness, that is somewhat more pedantic than your example of clothing folds. I mean, I would never have known that the particular kind of polymer used to make those grips can't maintain a hard edge at that scale, for whatever reason, without actually researching the production materials for that asset (which is something I typically would not do). Whereas, I would imagine that the knowledge of heavier/denser fabrics having larger/deeper folds would be fairly fundamental, and more importantly, able to be understood through observation alone.
Not disagreeing with you, just trying to point out that in cases like this, that technical knowledge (which obv informs your work) is a lot more abstracted from what you can gain through pure observation?
http://api.ning.com/files/aE9pF2mAEmZdzFhe248mUqKm8tNYVn1xZiWtIX80sENP9vdYHMIKmLI1phSQFSfZYxyA-TxXpQIK-TY6kTcfMUOPJruCo6SL/tec9.jpg
if you can't see it instantly from that one pic, then idk, it's even MORE relevant that i mention it
...but these look as good as anything else. On a long day when I'm in a bad mood, I might gripe about a sharper spec on the metal of everything except the TEC9 and the M4 (sharper spec could really help on the frame and slide of the 1911), but honestly, you know, if I bought a game and these were the default weapon skins, I probably wouldn't waste time googling for retextures or model replacements.
So there ya go. It's 95% fully awesome - ship it.
...just sayin'
The whole Flashhider is messed up, and nothing to do with the real thing.
Most of all it´s missing the vent holes. Google Bird-Cage Flashhider for Ref.
Frontsight. People tend to fuck it up usally.. badly.
1. The "boxy" part right above the barrel (where the gastube leads in), should be thinner and beveld. It´s also missing the locking pin that holds the gastube in place (minor detail).
2. Box thingy right above that other (where the fold tapers). It holds the aiming rod of the front sight, and that box should also be a bit rounder/beveled.
3. The base of the Frontsight is all screwed. Usually the Frontsight is held in place by 2 pins running under/through a part of the barrel and go through 2 blocks on each of the rings of the frontsight:
it´s an airsoft part in the pic, but the RS one is basically the same.
Main locking pins in the fron and rear part of the lower receiver are missing.
Fireselector is somewhere between single and fullauto?? how´s that supposed to wort.
it´s either or
Nooks on the stocktube retentionring are missing:
Texture doenst look too cool either.
Basically a black base with random scratches all over.
M4´s arent black actually, airsoft ones are...
RS ones usually vary in shades of grey ranging from a yellowish gull gray too darkgrey and even a violett sheen.
Each parts/groups do have their individual hues and tones.
The "plastic" is actually fibreglass reinforced nylon, which should be dull and a darkgrey (but darker then the gunmetals color)
tone down the noise on the plastic parts aswell. the surface isnt THAT rough on RS ones.
same with the mp5k
I wasnt trying to say you were wrong, because you arent. Your example was quite clearly an observational mistake on my part, I just thought that your particular examples warranted further discussion. I wasn't dismissing the crit.
Also, I think that at least some of the 'flat spec' crits are down to the lighting setup in the renders, I'll try to get around to posting some viewport-shader screengrabs soon.
Because I didnt know it was out yet ben
yeah i'm not a gun nut or anything but i think they look cool, defo aesthetically pleasing to me anyway ^^
think everythings already been said but yeah, lighting or spec seems a little weak for the most part and some of the textures are looking a little to stone-like (the 1911 mostly). less rock detail on the diff and more oomf on spec is all i really got...
The best tip on this i can give you is this: Think of the spec of a surface seen in glaring light. If you see this this in nature you will see that alot more detail is usually revealed. a surface might seem smooth but in a bright contrasted light you might reveal alot of surface imperfections. Basically that means your assets will look alot more interesting if you bring this thinking into making your spec maps. So put some microscratches, discolorations, fingerprints, .. whatever in there that are not on your diffuse and it will look more interesting and more believable.
edit: Although I will say there was alot of good critique on the previous page, not much seems to have been transferred to this most recent asset
You could try adding some more lights with color to your scene to give the guns a little more volume, just for show
fixed. Also, its lovely saying there was good critique which wasnt heeded, but if I knew what I missed, I wouldn't need to be posting. The main gist of crit as I understood it was stoney metal textures, and work more on spec/put more unique detail into spec. I think I've done both of those things with this piece. What did I miss?
If you're talking about the soft white line part around the actual buttpad, I'm not entirely sure but its probably due to either the lighting setup or a normalmap error. Both are kinda pointless to fix as they wont matter ingame. The "ingame" angle where you can see the loller-normalmap power is precisely that, user error on my part.
(disclaimer: missing fire-select lugs are due to a "wrong layer" accident before rendering)
:P
just don't go botchin' it up with no texture shmexture
i can't really crit the gun itself much... i'll leave that to the rest of these jokers.
this version - 7400 tri, 1024x2048 diff/norm/spec/gloss
also, the whole texture just need some more lovin', not sure what it is but it seems real bland and kind of undercontrasted.
For the metals try darkening the diffuse alot more and using the specular, preferably a colored spec map and maybe a envmap to drive the texture as metals generally have a low diffuse and derive their surface texture from reflections.
a bit like this:
Is there any reason, though, why the barrel and front sight is much lighter than the main body?
it just doesn't look almost finished. Like, it barely looks like you've done any actual work on the texture. Regardless of however long you may have actually spent on it - it looks closer to "started" than "finished"
Because it happened to be the first usable image I found? :P
Anyway I think the thing harry was trying to get at was your materials are hard to read and the ware and tear doesn't look thoughtout.
I would add scratches to the metal parts as everyone know that as soon as you get something metal you scratch it within 10 seconds of owning it. Due to the level of dirt on the weapon it looks like it has gone through a lot and so would be covered in small tiny scratches.
My main beef with the texture is the dirt though either tone down the dirt on the magazine or up the dirt in other places to fit.
On this image when I said the dirt doesn't fit on the front grip I was thinking about how that area would get dirty and it would have to be transferred through the players hand. So maybe when it is beeing animated ask for the idle animation and bake an ao map with the hands covering the front grip, then you can use this ao bake to help define areas for the dirt. Use the blend if options under the layer options in photoshop.
Im not sure what you're asking me to do. You ask me to spoonfeed you advice and finish by acting like you're above it? If you're just going to make an excuse as to why you can't go back and look at it again then firstly, I don't understand how you could have blown your entire allotted time with the result you have here, and secondly, you should make it clear you can't change anything else on it.
Or, you know, just don't feel compelled to mention how professional and deadline-oriented you are when it's not yet relevant.
It just doesn't look finished. The materials look like they're really roughed out. I can't crit any details when the details aren't there. I'm really not sure what to tell you if you look at these materials and consider them finished. The texture exhibits none of the imperfections and detail real world objects do. I can see you've overlayed some grunge but the forms are hastilly slapped over the whole thing, without any inkling of an effort to massage the detail to conform to the actual geometry of the object. The gloss and specular levels seem totally constant, and outside of hue I can't recognise any efforts at material definition. Hopefully this is more helpful.
I suppose the best way to put it is, it looks like a procedural texture.
PS. im not sure exactly what the last guy is trying to say. In this context "plastic" and "polymer" generally refer to the same shit.