Suzanne, Blender’s time-honored test-object, was added in 2002 – she got a default UV map when all the primitives did, in/around 2017. Recently the UV map’s been improved, including centering the face/head island, and that version’s been available in the 4.0.0 - Alpha main Daily Builds for a while now. But it’s not quite symmetrical, the centerline verts aren’t 0.500000 on the X axis.
I asked in its developer projects topic, and the responses were spotty – one dev dismissive, one offering to put in a fix if I provided it. Which I did, over a month ago now. Not only has nothing happened, but the dev who offered to help’s had no public activity there since his response to me (hope he’s okay).
It seems to me that if your UV map has an island for a symmetrical part of the mesh centered on a given axis, then it’s fundamental to good UV design that the island also be symmetrical, and its centerline verts be exactly centered on that axis.
What is the margin of error ? how much asymmetrical the UVs are ?
Can you demonstrate the issue it’s causing and how your fix solves it ?
It’s hard to judge if you’re being over picky or if it’s worth insisting … Sometime we have to let it go, at least it’s a bit better and TBH I never been bothered by the old one either But I agree UVs were bad
If you’re really asking…without malice nor ill intent, I’ll say “absolutely, without question.”
That monkey has hardly any practical use to anyone, except as a placeholder object to use when testing or demonstrating something that requires an object more complex than a cube or a cylinder.
Literally zero minutes should be spent by blender staff on this issue. I would rather they spend that time correcting tool tips.
MysteryPancake wanted to improve Suzanne’s UV map, two other devs worked with him to make it the best it could be. Unless we’re all on the autism spectrum, the whole project can’t be autistic over-precision. I suggested (and provide) an easily-implemented improvement that would bring the map in line with a basic principle (is it though?)
@sozap most of your questions are answered in the project topic I linked to above (my fix blend file can also be downloaded from there). My point is that this is a basic UV map design principle which, if there’s no reason not to, should be adhered to in Suzanne’s new UV map. With the exception of a single, simple change to get it in, all the work’s already been done. But nobody seems to be taking that last tiny step.
From what I know personally, for characters you do not have to be so perfectly accurate in your coordinates. This is because organic modeling (such as characters) favors inaccuracies and either hides them or accepts them as normal. You need to have a balance between getting the job done while being somewhat well organized and well presented.
I am not saying that the data should be a horrible mess (that indeed looks bad), so there can be lots of steps to be done to make things appear better, such as smoothing the UVs or making them more aligned, or get the islands grouped into in a logical sequence.
But in general terms your motivation was legit, but it would count as hyper-detailed. Don’t get discouraged if the proposal is shot down, usually there is a percentage of fail-successful attempts that happen with such improvement proposals.
Having it symmetrical at least would at least give the impression that the BF cares a lot about addressing the little things. I know it might seem pedantic, but it is not like you are unwrapping a full human model here and the actual unwrapping part will just take a minute or two.
If anyone has Maya or Max, can they confirm if Autodesk created a proper map for the teapot?
IIUC Blender added default UV maps to the primitives because something then upcoming required UV mapping, which made it problematic that the primitives didn’t have it.
The UV mapping didn’t have to be good, and sometimes wasn’t (I remember the UV sphere’s map was painful to look at). Later there were improvements made, but not to Suzanne’s.
Unlike Suzanne, the teapot predates Autodesk – also, do Maya or Max have a need for default UV mapping on their primitives?
You’re missing something: the May 30th 2023 commit you’ve linked to is in the 4.0 alpha, not the 3.6.1 master, and that commit links to the same pull request 106652 that I linked to in my topic-opening post above. My contributions to that pull request are from June 6th to June 22nd, and (unless I’m missing something) don’t appear to be in anything yet.
The two UV maps are visually distinct – I’ve just downloaded the latest 3.6.1 LTS build (c284cfcd0830), it’s got the older map.
I get that Blender’s new projects format is necessary, and a significant improvement in a lot of ways, but with the dates obfuscated unless you hover over 'em things that should be up front and obvious (when something happened) are difficult to see. I’ve checked my account settings, can’t find anything to make the dates always-visible.
It’s been so long since I used Max, that I honestly cannot remember the situation with primitives’ default UV mapping. I know I never needed a teapot with pristine UVs in Max. Like the monkey, the teapot was just used as quick placeholder when you needed a non-symmetrical object to test something like light positioning, a particle setup, a shader highlight, whatever.
It could also give the impression that BF cares more about the rather pointless little things, than the little things that actually have some use in production.
Sort of reminds me of the “let’s update the Mac OS Blender icon” discussion, which went on for TWO YEARS.
Yeah… I don’t mean to diminish your feelings in any way, @KickAir_8P, but this really feels like a non-issue. I don’t think it matters if Suzanne has a “production ready” UV map because Suzanne isn’t a production model, it’s a placeholder in every sense of the word. I agree with Thorn that asking developers to spend time on this is a poor use of their time
As would I, if they weren’t already doing it. Since they’re improving Suzanne’s UV map, I believe my suggestion should be included in the improvements. They specifically stated a symmetrical face/head island as one of the goals of the pull request – until what I’ve proposed (or something substantially similar) is included, they’ve failed their goal.
It’s a test-model – that includes being a placeholder, but isn’t limited to that. One of the dev comments on the pull request is about consideration of its use as a UV test model.
Well, I did ask if I was getting persnickity about this, because I’m aware of the possibility. But it really feels like my question isn’t getting considered appropriately, because the arguments against my point repeatedly contradict the evidence cited, linked to, generally available, etc. Suzanne isn’t just a place holder, it’s a test model. The devs aren’t being asked to devote time and energy to this out of the blue, it’s part of an existing pull request that has already been acknowledged by them as having sufficient value to commit (on May 30th as stargazer linked above).
All I’ve done is propose, (and provide) a solution that meets one of the pull request’s stated goals, which the implementation currently in 4.0 alpha doesn’t (they wanted that island symmetrical, it isn’t). If there’s something wrong with my solution, fine, but if not then I don’t see why this can’t be implemented.
That having been said, the dev who agreed to commit this hasn’t been visible anywhere I’ve seen since he posted that agreement. He was the instigator, without his input the others involved . . . well, aren’t getting involved. It looks to me like it only needs the last tiny step to commit, which any of them can do in a couple minutes, but I don’t know enough about the process to be sure of that. And I don’t wanna be more of a noodge on the pull request than I already have been.
hahaha , at least if you provided an updated mesh then it’s up to them to decide what to do with it…
and if current UV brings further complains they can probably retrieve your model from the commit logs.
Suzanne in many areas is less than perfect anyway, sometimes these flaws are even good for testing purpose ! Like the disconnected eyes
Finally an inspirational gif to help you cope with devs silence …
And the trigons in the mesh – the “flawed” topology is currently considered part of its utility as a test model, etc (as people who suggest replacing with an all-quads version are oft reminded). That’s why I started my top post with a reminder of Suzanne’s history: the mesh hasn’t changed since 2002, the UV mapping is a far more recent addition.
As for exhortations to “let it go”? Giving up too soon is a guarantee of failure. As mentioned above, the deadline for changes to 4.0.0 is August 23rd. It’s literally more than three weeks too early to give up on this.
MysteryPancake has reappeared on projects.blender, taken my fix, and improved on it before commiting. @Xeofrios for MP’s cite.
4.2.0-alpha+main.3be6dbb02a06-windows on the Daily Builds (went up about an hour ago) doesn’t have this in. Don’t know much about this part of the process, or when it’s likely to hit the alpha+main.
UV design that the island also be symmetrical, and its centerline verts be exactly centered on that axis.
But I agree UVs were bad 
