A new Voxel Engine for scultping please ?
Hi , voxel sculpting in blender is very nice BUT it is memory consuming i don't know what they have done to zbrush or 3d coat (mostly 3d coat) that it won't eat that much of the memory sir ! a 5mil mesh sculpting in blender is taking up 8 gigs of memory and it seems like the more voxels it became it takes quadrupling the amount of ram used by software as for a 64 gigs of ram we are able to only sculpt a 15 mil voxels model , i seriously begging blender staff to take a look at 3d coat because sculpting a 15 mil voxels in 3d coat only takes 2 to 3 giggs of ram and i don't seriously want to learn lots of programs as well as many other users , its maybe just because blender staff think thats true to be like that while others consider it as non industry standard software .
i have been told here to use more memory like 128 gigs or more but it just not the solution there will be always a dead end for the software to stop working ( also it isn't ideal to burn money on memory constantly)
in the picture below you can certainly see a 20 mill mesh is barely taking up any memory and 3d coat is mostly uses all of my CPU cores all together
7 Comments
"make it faster" is not a proposal. "make it 3dcoat" is not a proposal.
so what is it now is this suggestion getting remove ? also i didn't just simply said "make it faster" , i said maybe blender programming staff maybe think that this is just good enough no it isn't ,again i didn't said "make it 3d coat" , i referenced this program so we can see it is exists , almost every blender user is complaining about blender polycount issue sir
RCS if for making feature proposals. What is the feature proposed in your post?
Correct me if I am wrong, but your post is essentially asking Blender devs to look at another software, thinking they do not know Blender can't handle high poly modeling and that they could just copy things from elsewhere.
Hence, "make it faster" or "make it 3dcoat" could indeed sum up your post, which is hard to consider as feature proposals any dev can act upon.
Personally, I like to think the developers are the ones who know Blender's limitations way more than any regular user, they know why it is this way, and also what they can possibly do about it with their available manpower. They might not know the precise user experience (because they don't collect data and surveys like other devs would do), but performance is a constant issue on any software.
Looking at other softwares doesn't really help much. Patents and legal issues aside, you have to understand that you can't just copy whatever you want from anywhere. Much like asking a Swiss army knife to look to how a fretsaw works, it doesn't lead anywhere.
But hey, who knows, maybe a devs passing by could tell us more about it.
I'm not a dev myself, but for what little I know, I suspect it is hard to consider switching to or including voxel modeling like zBrush or 3DCoat do. Blender doesn't do only modeling, it also does animation. How does voxel perform in that area?
If you consider "just adding" a voxel engine to what we already have, you still have to remake the tools for that engine, and then maintain both the new voxel engine and tools in addition to the rest.
That raises the question: is that even feasible with the resources of the BF ? Doesn't sound so. Is that even worth it in the long run ? Maybe not. Are there other thing that can be done to gain performance that profit everyone ? I think so: #68873 - Highpoly mesh sculpting performance - blender - Blender Projects
Distilled down, your post says "Make it faster like 3DCoat". Show me in your 'proposal' where you propose an idea for how to do that. Show me in your proposal where you suggest some new aspect of sculpting that is unique and adds value to Blender.
You think they wouldn't just make it faster if they could just snap their fingers to make it happen? Performance improvements are an ongoing development task- speedups are ALWAYS happening. Either propose a solution for HOW to make it faster, or sit back and enjoy the ride like everybody else.
Ok it seems like they have done something in memory management in 3d coat cause everytime i do something all cpu cores start to firing up and the work continues but in blender in the other hand the more polygons even the idle ones they take up more memory and hek its not linear sir , its by power of 2 , a 5 mill mesh will take up 8 gb a 10 mill mesh takes up to 64 gb of ram it's just not logical , im not that programmer enough to address the issue correctly but at least im trying to bold it rather than pretending that it's not there , its very suspicious when i see the people trying to hide stuff , look at the market for industry level design large companies will merge to blender and they will support blender for good , but what blender dev's are doing ? making post processing in real time , they do have the time to improve it 3d coat came out long after blender . and what is going on with proposal thing sir ? im just telling my idea and wish for blender to became better , i believe blender dev's have still time to make it better
great, i'm super happy for you. However, RCS is not your personal soapbox. feel free to open a window in your house and yell your complaints outside to the people walking past, it will be just as relevant and maybe more cathartic.
@testure it's clearly visible that if blender community takes polygon count and industrialization of blender seriously , it goes against your benefits , i don't seeking for reasons here but I've already reported your behavior and the nonsense about propose thing , cause my propose was simple a new engine that handles voxels , thats it , i guess were not supposed to write the code here are we ? , and you may be better to do a little behave next time .