Pet Peeve: Sloppy work. Or: How Anal Am I?
polycounter lvl 16
Let me preface this for the people at work who will read this. I don't mean this against anyone's work in particular, this is just a general question, and is not meant to reflect on anyone.
Right now at work, we're doing some polish/performance passes, so I'm going back over other people's work, and making LODs, reducing material counts etc. As I'm going, I run across sloppy stuff periodically. Now, what I mean by sloppy isn't something that is drastically wrong with the model, but more things that are simply irksome. This is stuff like extra cuts/edge loops that don't add anything to the topology, aren't UV seams, etc., but are just there. Or places where the pixel densities are just slightly off from one part of the model to another. These are the kinds of things that will really not be visible by a player, and don't really affect performance or anything else.
I feel compelled to fix all of these minor nitpicks as I'm going through these things. My question is, does anyone else run into this? Am I just being too anal and need to relax, and trust that whoever's work it was was done that way for a reason? Or are you driven to fix these things too? Is it poor form? should I instead raise the point with whoever did the work in the first place? (sometimes that's not even possible)
Just curious about people's thoughts.
Right now at work, we're doing some polish/performance passes, so I'm going back over other people's work, and making LODs, reducing material counts etc. As I'm going, I run across sloppy stuff periodically. Now, what I mean by sloppy isn't something that is drastically wrong with the model, but more things that are simply irksome. This is stuff like extra cuts/edge loops that don't add anything to the topology, aren't UV seams, etc., but are just there. Or places where the pixel densities are just slightly off from one part of the model to another. These are the kinds of things that will really not be visible by a player, and don't really affect performance or anything else.
I feel compelled to fix all of these minor nitpicks as I'm going through these things. My question is, does anyone else run into this? Am I just being too anal and need to relax, and trust that whoever's work it was was done that way for a reason? Or are you driven to fix these things too? Is it poor form? should I instead raise the point with whoever did the work in the first place? (sometimes that's not even possible)
Just curious about people's thoughts.
Replies
If it takes alot of time,my time to fix then i get irritated.
It could be worst trust me.
Talk to Kevin about taking that extra time to be extra annal.
I absolutely hate redoing or optimizing other people's work, and really don't want to be that guy who needs other people to fix up his sloppy work.
Obviously this is why we all go through the polish/performance phase. You're going to find mistakes, quick/sloppy modeling, pixel density issues, etc.
Personally, when we're first given an area to work on, we're given a very rough prototype. The first phase for us artists is to block in the general shapes, get our gameplay features on the assets, get some base textures going, and pop in some quick/dirty LODs. That then goes back to get gameplay tested while we do the same to a new area...or move on to a finish phase on a previously tested area.
Ideally, if there are any more optimizations to be done *after* your finish phase is complete, the assets should be solid. I can totally understand your frustration if that's not the case here. Otherwise....meh, it's just the nature of the beast!
quoted for truth
if you build all your stuff properly the first time around you don't NEED a performance optimization pass ---> game ships 2 month earlier and you can spend 2 month at the beach !!!
*dramatizedoptimalcscenario*
and yes i agree this should be moved to technical talk ^^
Let me reiterate that while these things irritate me a bit, I'm certainly not getting all worked up about them or anything. I just see something and say "WTF? Why does it look like someone just took the finished model, and a cut tool, and went to work?"
The pass I'm doing now IS the optimization pass, so obviously it makes some sense for me to be fixing stuff like this, but my actual TASK on any of these assets is something completely different, and while it doesn't take TOO long to take care of these little things, it does take some of my time away from my actual task.
I'm definitely with Peris on this. I think a big part of why it bugs me, is that it would REALLY bug me if my work looked like this initially. I mean, sure, it might while I'm working on it, but when I'm ready to check it in, I definitely make a quick clean up pass to make sure junk like that is taken care of. While I tend to agree that early OPTIMIZATION is probably bad, I certainly try to make sure my models are at least decent before they go out.
But, point taken, I should chill and just take care of it without letting it bug me.
People should do a good job from the beginning... and care all the little details. We are making art.
Optimization is always good, if you are customed to do optimizations from the start, it's work you save.
sighs...
A list of my pet peeves when working with other people's files (some of these are very 3ds max - centric so bear with me):
- People who save our their max files with a single viewport maximized, usually in "User" mode. Upon toggling to a four pane view, you discover that ALL four of their viewports are in some form of user mode. Protip: If you toggle back to four pane and then save out your 3ds Max file in the perspective viewport before closing, you get a nice little thumbnail when browsing the files later and if someone else opens your file their first impression is a nice neat workspace. I mean, they are already fixing your shit, at least give them a nice starting point.
- People who do not name objects in their scene and do not use named layers.
- Same for Photoshop: Name your layers and use named groups!
- People who check ALL of their iterations into source control, in the same place making it nearly impossible to tell which one of your 10 files is the most recent that is the asset being used in game. If you feel the need to back up your iterations in source control, at least make a "working" or "back" folder for your older iterations. Also, if you are using 3ds max, why in the hell do you feel the need to save in iterations, anyhow? Set your autoback to something reasonable for you and your workflow and never worry about saving iterations again!
- File, asset and package names that make no sense. At least name it something that gives me an idea of wtf it is.
- People who save their source on their local drive, and never check it into source control, forcing you to track down who actually did the work in the first place as well as the source files. Bonus points if you had your PC upgraded or your HD go bad, making your source files unrecoverable. Kudos if you quit the company or, better yet, were fired; at least you got the last laugh.
there are still people around who work with 4 viewports open??
exactly! Plan your stuff out well from the beginning, don't just make something saying you'll optimise later. You'll save alot of time if you plan stuff out well.
Why exactly are you guys saying that optimising from the beginning is a bad thing? Are you afraid optimising will make the art look worse? Because what i see in most cases with unoptimised props is that they could be made way cheaper and at thesame time better looking.
If you don't want to feel like a stooge for ratting out your coworkers I totally understand, you should probably fill them in on their gross incompetence, in a nice way. I prefer blunt force trauma, but not to the head. You want them to remember what you just told them. Also remember, the tire iron is for the 3rd offense not the 1st.
Working on pre-rendered backgrounds made me soft... I try to be not as harsh on people as I have in the past but a lot of what AZ mentioned bugs me still.
I do agree that any job worth doing is worth doing right and hold myself to that standard. But I also forgive my coworkers minor mistakes. Also what looks like madness to me, might be another persons sound method.
BUT you can turn a 2hr project into an 8hr by needlessly worrying about building a technically solid master piece no one will ever appreciate after "show edges" is turned off.
With all that said, it sounds like some bad habits might be forming and it would be worth mentioning so it doesn't grow. I'd suggest changing the name of this phase to "final pass" or "oh shit the game is borked pass" because "optimization pass" might lead people to think they can be sloppy at first and that they have time later. Letting people share in the glorious LOD making process lets everyone see the good along with the bad, and might improve work all around.
For me its better to assume there isn't time for clean up later and that I need to build it solid "enough" the first time without killing myself making it "perfect". It's easy for me to slip into personal project mode and spend way too long perfecting something, I've learned to cut myself off.
I don't think most people realize their little sloppy mistakes can hurt the project/affect other people.
Edit: As Vig said, besides fixing other people's work, you'll want to prevent people from making needless mistakes later on down the road. Otherwise you'll be back in their scene files later on deleting edge loops.
Shit that's what I'd do if I were in your position. Just think how they'd feel if you didn't tell them and you fix it or someone else fixes it. It's a matter of professional pride for them, and yours if you are a senior position or manager, you lead by example and positive, constructive feedback, unless there isn't time for it... then you just let them know you had to tweak it, but still.. let them know.
BINGO. We need to find these four-pane fuckers and put them against the wall.
http://boards.polycount.net/showthread.php?t=50588
omg dont you dare muck with Layer 192!!!!!! nooooooo
I'm pretty anal about my PS docs, so much to the point of Photoshop not allowing me to make more sub-groups. stupid ps. Gotten back into the kick of merging shit lately though, 8gb files suck!
this topic in general is pretty touchy - i agree that one should always "work smart," but sometimes, that mesh was good at the time. As projects go on, ideas change, budgets may change, and you just have to adjust. Sucks to have to go back over work, but if you didn't, the quality may not show through as well as it would after going back over things.
may be frustrating, but its worth it to take a second look at stuff after a few months [if time permits]
wasted uv space gets my goat the most though.. im no genious at it, but if its uvmapping individually all the leaves on a plant when they all look the same.. grrrr
I hate hate hate people who have highpoly or lowpoly scenes full of "polySurface164" or "Object26" in "Layer08" (or worse, no layer at all) in Max or Maya. They are clearly not being considerate of anyone else in the workplace - while a lot of studios you might end up just focusing purely on your own stuff a lot of the time, there is always a time when someone else (or multiple people) are going to need to open and edit your files, and if you don't keep your stuff organised then you're just wasting their time, and probably your own if you open up a file 6 months down the line to find a hundred objects with no identifiable name.
It literally takes 2 seconds to name a mesh when you create one, and most apps will sensibly number duplicate meshes if you array/instance them, so there's really no excuse for unnamed objects or layers. That's my main pet peeve.
(also, moved the thread
I don't run into naming issues as much (well, spelling. . .grrrrr) but the one that bugs me is a single max file with 150 modular pieces that fit together on a 10' grid, but with each object placed at some arbitrary x y z coordinate. And then people wonder why our modular pieces have gaps or overlap or whatever!
EQ, I totally agree on loops like that. I'm not an optimization nazi, here for every last triangle I can get. What I'm looking at are (as far as I can tell) TOTALLY useless edges.
Here's an example I mocked up. This is similar to a planter bed I found recently:
Maybe I'm missing something, but is that cut at all necessary? Again, totally unnoticable by a player, and adding 14 tris isn't going to kill us by any means. . . but why is it there? I'm not questioning anyone's competence, I'm sure it was just a mistake. But it does bug me a little when I find stuff like that.
A: Its a messup /holdover from initial box definition settings
B: The artist thought it would be a vertex lit asset and took a random stab at where to put a cut
But as you said in your first post
The player won't notice it, and it doesn't present a performance concern, so really the stress and effort you put yourself through is only to assuage your own neurosis. (ie its not a waste per se, but you could likely put the energy spent to better use. )
And of course I would inform said person (if they were still here), and have many times in the past, that wasn't really a question in my mind.
Ya know.. I'm pretty sure when ya bitch about someone else's workflow... someone else is even bitching about yours. thats life for ya.
Well, if your mesh is being vertex lit, some cuts like that may be necessary, although the cut would be better placed more to the center of the object were that the case. I am thinking that is just a stupid, sloppy artist.
Another minor pet-peeve of mine is the cylinder that has the vertex in the center of the planar caps. You know if you remove that vert you would save a whopping two triangles!!!
the vertex normals are nicer with a vert in the middle
Like I said one mans chaos is another mans method.
Welcome to art lead-land.
Extra pet peeve to add: Have some friggin balls about collapsing psd layers. You don't need 140 layers to make a diffuse texture. Finding any of those elements that you need to change becomes almost impossible, so just collapse them. I collapse by matched surface type. All brick on one layer, all dirt on another, etc. I
f I'm nervous about a correction that I'm making, I may use one or two adjustment layers, but not more than this. I've seen files with 9 adjustment layers just to make something a little darker and more blue. DO NOT do this.
becomes a bit weird.
I tried to stay away from that. There were many times when I wanted to say to some people,
( even those senior to me)
'Why don't you try this way' and am sure they would like to say the same to me, but we shyed
away from it overall for the sake of team harmony.
I think you will find in you avergae games company, nothing is perfect and people are loathe to
even speak, let alone speak out:)