6 x faster than real time simulations
I propose the devs check out this video, investigate the paper it refers to and then see if it feasable to implement into blender.
I propose the devs check out this video, investigate the paper it refers to and then see if it feasable to implement into blender.
14 Comments
Please don't linkbomb with a YouTube video. Write a little text that explains the idea and how you'd envision it being integrated with Blender. As it stands, this isn't a proposal.
* @Jason van Gumster * I think it's unfeasable for me to re create the entire video, and a written explanation would be far less easy to understand. Let's not over police the forum.
* @3d illusions * Or you could write an explanation that people could understand and offer the video as supplementary support. This isn't a forum. This is a site for making well thought-out, actionable proposals.
By making a proposal here, you're essentially asking a developer to do work for you, sometimes for free. The least you could do is put the same amount of effort into making a proposal.
* @3d illusions * A good way to think of it as you are making a proposal to your boss. Cover what the current situation is, why this new situation would be better (think about how it will effect the user and what it will allow them to accomplish), and even if you only have limited knowledge on a subject, try to provide a rough idea of how or where this could be implemented into the existing software.
Remember, developers are busy and already have other stuff on their todo list than your proposal. If you want it to be actually considered, you have to convince them it's a good idea.
* @Zach Hixson * Ah ok, so this site isn't a good place to bring new innovation to the attention of developers. I'll be careful not to share such articles in future. Cheers.
* @3d illusions * It's the perfect site for that. Just put a little bit of effort in what you're sharing.
* @Jason van Gumster * Is there a site where I can share links to new innovations with just a brief outline like I did here in the title. It's not really feasible for me to delve into the paper to be able to define how the faster results are being obtained. I know for example Pablo was encouraging videos a couple of days ago, so I expected it would be ok to give a link to a video here. Perhaps it's better to just use twitter instead and hashtag with #b3d
With all respect, I do feel you're being more than a touch pedantic if you don't mind my saying. As long as the post provides a way of understanding the idea, that should be the main thing, not whether it requires clicking on a link, or whether the poster has duplicated the information in text form that the reader will more easily ascertain from the provided link.
Substance, not style.
Short Text?
Yeah, this is cool and all, but it's pretty new tech. Stuff like this isn't really plug and play atm. Just like FLIP fluid simulations have been around for ages and just recently been integrated into Blender in the form of Mantaflow, programming and integrating such a technology is no small feat.
* @Zach Hixson * I think half of these papers are written using Blender, I'm not sure if the Blender devs would be able to just use them though, whether they would need to implement from fresh using the information from the papers, or whether these papers have some kind of protection against developers using the findings without permission etc.
* @3d illusions * I agree, a lot of these are implemented using Blender, but just "getting it working," isn't the same as adding it to Blender. You need to develop a user friendly UI, make sure the code is integrated well with the rest of the codebase, and develop ways for it to interact well with the other systems. I'm not sure about the permissions thing though, that's a pretty good point as well.
How it works ?
* @Kdaf * witchcraft I think :)
* @Kdaf * You can find a link to the paper in the video description