Saint haven point is a very good one. Front running modellers, like Kent Trammel have moved from box modelling to sculpting because its faster and allows for more creativity. Blender’s current modelling tool set is … excellent. I love it. But what would set Blender more distinctly apart and above would be powerful sculpting. Already its a very complete and powerful all purpose tool set and I’m incredibly happy with it.
But having a compelling sculpting capability would be, to my mind, icing on the cake. Having sculpting at the level of Zbrush or Muddbox would encourage a ton of artists to jump ship. And to be honest, although I’ve not tested the sculpting or Dynotopo much, it actually doesn’t seem that far off now, given the sculpting work produced by Michaelis and Doris and others in the last two years. A deluge of great sculpted work has come down the pike. That’s a statement of the attractiveness of sculpting as a tool as well a tribute to its relative level of effectiveness.
and therefore it naturally beckons to the next level of ability - the ability to jump from moving a couple hundred K of polys to throwing around a hundred million polys in real time. It’s one project that really could use specific donations and exclusive funding.
I think thats not fair. Whether blender developers listen or not to the users you can see it easily if you visit #blendecoders via irc or the dev mailing list.
The problem is that that no coder is forced to work on an area of code. I hear all the time about what user need but people fail to consider what coders want too. Blender developers are free to work in whatever they like as long it benefits blender, this is why blender is so awesome because it keeps its developers excited.
Now if you want a specific set of features and improvements there is a process for that too, its called bug/ issue reporting. You can also visit #blendercoders and voice your concerns or register to the mailing lists. These are the places developers frequent and not BA forums because these are their main ways of communicating with each other, the same way BA is the main way for you to communicate with other users.
Most likely developers are already aware of these issues, the only thing that can help here is for you to collect money so that BF can hire a developer to work on those things. Or else you just wait.
HAHAHAH Michalis ? He was A TROLL before BLENDER SCULPT HAS Dynotopo he only love Zbrush , Zbrush this and that he was very boring making those comparation . He think the program would do the work for you hahaha .
I know becuase I discussed with him those days ago
Yes, the sculpt tools have some limitations which can make them quirky. Can you take a couple of levels of multires out using the shrink wrap? (I’m sure you know this - duplicate the mesh apply the multires at a lower level (level 3 maybe - a few thousand polys or so) subdivide upto your final level (same as original), shrinkwrap to the original (after multires modifer in thr stack) then apply the shrinkwrap. This should get you to where you were with maybe only 4 levels of multires). Performance might improve a bit so you can keep working, although not always in my experience.
I always thought that’s what the reshape button was for. If you duplicate the mesh apply multires at a lower level, subdivide up the the same level, then click the original shift click the new one and click reshape, I thought it would give you the same result as the original. Unfortunately the mesh explodes cause (I assume) the new vertices generated from the subdivide have a different order. It would be cool if that could get fixed, as reordering the vertices doesn’t seem to work (I assume because the reorder happens before the mulit res)
Anyway…Regarding the performace of sculpting, there were a lot of threads about it a while back. I always got the impression there were some technical issues deeper in blender that made it hard to fix (although I could be wrong about that). Hopefully projects like the depsgraph refactor could help this in the future, but I have no knowledge of such things so that’s entirely supposition on my part.
Eitherway I’m not sure why you are having so many problems rendering, I’ve rendered sculpted mesh without trouble. Maybe try a different build or something like that.
So further I just did a quick experiment with an old mesh. Normally I don’t render multires meshes, I retopo etc, etc. But this one is from a while back and it was originally just rendered with multires on at about 2 million faces (iirc it even fit on my gpu). I simply divided upto 8 million faces (nothing more) switched to cpu and hit render.
It took 2:30 to synchronise the mesh and about 5 min before I saw anything start rendering. It’s using less than 6 gig RAM and less than 4 gig swap as reported from my system monitor. heres a screen shot. This is why I’m wondering if maybe it’s a build problem, or something, cause the numbers you reported were huge.
thank you for the advices. i have done all except disabling outline. my point of this thread was not to aquire a workaround, i have rendered easily the mesh when the multires is applied, but my point was to restart the discussion of the quirks of multires so that we can get a better workflow when the developpers know where to look when they decide to work on it…
this one quirk of multires i encountered is a big issue: when multires is applied, cycles renders this mesh in less than a minute. if the multires is not applied, blender builds a huge swap disk of 84 gigs, this is the issue i was pointing out. there is a problem i wanted make aware. and, i wanted the discussion, so we can see what should be done… yes workarounds, i learned a lot, many are vital, but my point was to make aware of what quirks are there… using 84 gig to render a not that big mesh with multires is a big issue… i tried today again, blender did not even stop at 84 gig it still kept building swap disk until i broke off at 150 gig… see my point?
I totally agree, I’d love to see many areas of blender become less quirky. The thing is I have a considerably lower powered machine compared to yours, but with the mesh above, multires on at level 6 with around 8 million faces, I don’t get figures anywhere near those levels.
Doris,
I don’t intend to go into the matter of multires oddities because I don’t know anything about it, but what you are reporting looks like a real bug.
So to speak, in the past I had a similar behavior from Apple Safari (!!), which restarting in some circumstances made swapfiles grow like crazy until I had to reboot…
So it would be a good thing if you could investigate further, maybe trying to import that object in a new scene, or make copies or links of it, I don’t know actually… but I hope you get the point.
yes kettlefish, that was my point, i am glad you see what i meant…
ok, i understand sourvinos, you ask me to try to circulate the problem closer,i will try tomorrow,apple safari wow, so maybe it is a programming error that can occur in every code… i will try to circle it closer
The question isn’t the difficulties of understanding the issue,but the fact that this problem,like spike in multires,are well known…these issues will not be fixed because someone opens a thread on blenderartist…without a coder who’ll work for some months on multires things will remain like now for years.
For people interested maybe is better to start a campaign to attract some coder(with a donation system)
I only partially agree, it’s important that we try to spot the issue or isolate the situation in which it occurs, then, as it has often happened, we can see one of our coder commiting a fix.
1+ on that…
Instead of complaining and pointing at devs…community should take iniciative, contact experienced developer (Nicholas Bischop?), come up with goals and start kickstarter campaign.
Main purpose for that campaign should be reliability/performance for multires and brushes… then via stretch goals we could get all fancy sculpting features
How does it sounds to you?
Not to be a pain, but if donations are coming in due to Blender’s use in developing game content in an arrangement with a game development company, I think that call for that extra money to be used for extra-high-detail sculpting (not really useful amongst current Blender game artists) is going to bend a few people out of shame. Myself included.
Whilst I want Blender to focus more on game asset creation tools and features, they are getting a pass on not doing so because they’re focus (and their donations) are coming from people / projects focused on movies & vfx. It’s not even a slap, but a kick in the face for monies derived from Valve to not be put toward features / bug fixes related to game content creation. DOTA 2 is, after all, not really a target for 1Gb sculpts.
This I would agree on and think would be perfectly reasonable. Personally, I’d suggest a horror short film taking inspiration from H.R.Giger’s work. You can combine the camera tracking improvements with sculpting, cycles, fluid, and composition fixes/upgrades. Hell, done well, it could serve as a basis for a game based on the film (media tie-ins are big these days).
With all due respect, I too have talked to the man and he is/was a big fan of Blender’s sculpting and, especially, dyntopo. Yes, he thinks ZBrush is the bee’s knees and has been quite vocal about how he feels about sculpting being practically dropped from maintenance/support with bugs being ignored in release after release. Thing is, frankly, the man has created more interesting and worthwhile busts/figures using the dyntopo/sculpting tools than the vast majority of those that are calling him a troll. And now doris, another talented sculpter actually using the tools, is referencing one of these long-term bugs.
Liking & comparing Blender to another application is not trolling. Doing so purely to generate a negative response is. Micahlis, whilst at times annoying, is complaining because he wants Blender improved. I would posit calling the man a troll because of this is more trollish than anything he’s done to date.
I don’t have much good news to share. It’s absolutely true that multires and all the other sculpt tools in Blender need more work. I can’t speak to any other developer’s plans, but speaking only for myself I don’t expect to do much to improve the situation in the near term.
The issue is that making significant contributions to Blender (including bug fixing) takes as much energy as a “real” job. I already have a job though, and at the end of the day I don’t have enough mental energy to get my head into an entirely different code base written in a different programming language.
The tricky thing with paid Blender work is that it tends to be a freelance model: money from the development fund or Google Summer of Code or Kickstarter lands you a month or three of work, and then that’s it. Some people like working freelance, but I’m not one of them. I like a salary and benefits and the general comfort of knowing that I can keep showing up to the same job for as long as I like.
Maybe someday there will be companies hiring people to work on Blender full time, but I don’t think that day is here yet (at least not in the US, no idea what it’s like elsewhere in the world.) Many companies like to use open source, but it’s a rather smaller group that has the confidence to pay developers to write code that then gets freely distributed to their competitors.
