How to watch AI Passthrough vids with HereSphere or locally?

boboweb : I've been testing out "Goodbye Greenscreen", an After Effects plug in for $80. Haven't compared it to others yet, but what piqued my interest that it combined a "difference matte" using a clean background plate and AI. Most of the scenes I'm interested in trying this on do have a locked down camera where you can sometimes extract a clean background plate. I'm on an M1 MacBook, so you're might get a faster render with the best PC graphics card, but I'm not sure how much the "AI" part uses the GPU. As I said above, on the lowest quality setting 1 hour took 10, so if the best GPU cut that in half it's still a bloody long time. I've been thinking about maybe making some VR PMVs and using it sparingly, but to try and replicate what SLR is doing and doing a whole library takes a lot of time and resource. Maybe worth it to convert some of your favourite scenes.

While doing research for it I kept seeing CapCut being mentioned, but I decided to avoid it because there were many reports of it being Chinese spyware (it's made by the same company that created TikTok).

sunkenlust basically all these AI models use the GPU and if they don't they are exceptionally slow. Also a 4090 is 611% faster then an M1 (quick google search) so that's still pretty fast for a background removal program.

However, more generally I think a lot of time is wasted in re-encoding the video with the green background. It would be much faster just to generate an alpha masked that is applied on top of a video (similar to what SLR is doing) but no player supports local masks like that yet.

fenderwq : Looking at the "Goodbye Greenscreen" product page again, and it does confirm what you're saying. It says it's hardware accelerated for on Nvidia GPU on Windows and M1/M2 and AMD on Mac.

boboweb We probably will, likely not perfect for porn as it's usually not in the dataset but soon enough

boboweb Hello. I have personally focused on creating passthrough videos, primarily using JAV content. I have tried utilizing numerous libraries available on GitHub, as well as paid background removal services like After Effects, DaVinci Resolve's Magic Mask, and RunwayML.

After Effects performed surprisingly poorly, and the GitHub libraries were not suitable for VR videos (similar performance to CapCut). The services that I was relatively satisfied with were DaVinci Resolve and RunwayML. With an RTX 4090, DaVinci Resolve completes the process quite quickly. However, DaVinci Resolve struggles to track the person if the female performer transitions from the cowgirl position to the doggy style position.

RunwayML is convenient and tracks people more naturally than DaVinci, but it has a 4K limitation and a 15-minute time constraint. I have heard that if you convert the entire video into frames and remove the background in a photo format, u2net_human_seg can be useful, but I have not utilized it myself.

I have converted JAV 8K videos ranging from 1 minute to about 15 minutes into passthrough format. However, after SLR started mass-producing AI passthrough videos, I discontinued my work.

If you want to remove the background of videos not available on SLR, as someone with some experience, I recommend using DaVinci Resolve's Magic Mask and RunwayML.

Thanks for that, Koreanman, that was on my list as something to try. Could that be done with the free version of DaVinci? Just by watching the SLR AI passthroughs, I'm pretty sure the matte streams are lower resolution than the source material, so a 1K or 2K matte video might be acceptable on an 8K video. I think DaVinci has a free-version 4K export limit. I don't think I'm going to render hours and hours of video, but I was thinking about creating PMVs and doing select clips.

sunkenlust I understand that DaVinci Resolve's Magic Mask requires the Studio version to be installed. The Studio version is quite expensive unless you're doing professional video production work. Therefore, I recommend searching the internet for ways to try it for free before making a purchase.

boboweb why would you want to put it in there? Doesn't seem like the right tool to me.

boboweb Just saying you'd probably be better off just to do it in the command line (or a bat file) and process everything in small sections. Creating masks takes a long time and it is good to be able to resume where you left of or stop processing if you need your pc for something else etc. You can also just process all videos in a folder using this way stuff like that. No real advantage of putting that in Automatic1111 since there is no real use interaction/ different use case.

koreanman ...maybe assuming you don't mean the still encrypted video files...?

fenderwq I think DeoVR does, that is how SLR can even do AI Passthrough, Alphapacking, Chromakey.

boboweb

koreanman ...or able to afford their cameras... then its "free" usually...