Search options (Page 6 of 9)

Please share where you found this information.

It's easy, to upgrade to the TensorRT version 9.2 from here:
https://github.com/AmusementClub/vs-mlr … /v14.test3
you just have to download this big archive:
https://github.com/AmusementClub/vs-mlr … 4.test3.7z
and then copy pasta these files in "C:\Program Files (x86)\SVP 4\rife\"
https://gyazo.com/53e859fd5306eb8ce5f896bd87b22bd8

You can find some helpful settings here:
https://www.svp-team.com/forum/viewtopi … 352#p83352

You can add stuff called "shaders" to mpv & mpc-hc to enhance/upscale remove blur/remove noise/debanding like:
https://www.svp-team.com/forum/viewtopi … 264#p84264
that enhance your viewing experience while you watch.

Give a link where other guys can download such an example to compare if they see anything unusual.

Yep, that's a rife presentation video. Latest rife neural engine models have less interpolation issues. But never zero. It's still one of our best options for conversion on-the-fly while you watch the video.
You can try to follow this small guide to download latest models, and then resize the video down until you can test 120 or 240 fps smoothness.
https://www.svp-team.com/forum/viewtopi … 352#p83352
(if you tell the AI model to generate more frames, then it will improve the accuracy of the interpolation and your eyes won't notice any interpolation issues unless you spam pause every 0.5 seconds lol)
Have a look what it generated and compare:
https://drive.google.com/drive/folders/ … CKnKGieUtb

I think it's easier to test if u go to line 222 from generate.js and delete the following:
    if(profile.rife)
    {
    if(profile.rife_sc==6) rife_sc_algo = 1;
    else if(profile.rife_sc==8) rife_sc_algo = 2;
    else rife_sc_algo = 0;
    }
    else rife_sc_algo = -1;
then replace with following line for "scene change based on SVP's motion vectors"
    rife_sc_algo = 1;
or replace with following line for "scene change based on Nvidia Optical Flow's motion vectors"
    rife_sc_algo = 2;
or replace with following line for "scene change based on dumb frame comparison"
    rife_sc_algo = 0;

(manually editing the file this way allows u to use your own rife_sc value therefore removing that slight difference between rife_sc 6 and 8, thus making it easier for u to see the changes xD )

Yes, if that didn't fix it, it means you forgot to add the minus in front of the value.
Go to Application Settings, search resize and add:
-19201080
or
-12800720
or
-9600540

In some scenarios using fixed fps will make it go slower, so make sure you uncheck "Force the exact value"
For example setting a 240 fixed fps on a 23.976 video will make it generate 239.76 + 23.976 frames frames instead of rounding itself down to only 239.76 frames

Maybe try to select madVR in MPC-HC options and see if issue persists?
https://gyazo.com/509b01a01eb6dd413667b4dd3f6d819a

Don't think it's caused by a badly encoded HEVC[x265] file. Try to download other versions of the same video encoded in [x264] and compare.

For users that just started using SVP, it will tell them that the interpolation is working so they will focus their attention at the video smoothness.
For users that have HDR screens, it will help the user identify that he needs to enable his HDR to see better color accuracy and stuff because that specific video supports it.
You can disable it by going here and unchecking "Show OSD messages":
https://gyazo.com/3929e2194ac665b1fc67e8a258360b92

https://gyazo.com/3b2771c7a2dee8994b431a283f215f12
Go to Frame Size and change "Black bars detection" to "One Time Only" or disable it.

After further investigation it seems the v2 issue is caused by old .dll files.

Out of the box svp with rife model 4.9 v2 with tree wiggle present:
https://gyazo.com/67445b08ae0aa2a214e8cb054e051b1b
Out of the box svp with rife model 4.9 v2 with tree wiggle fixed:
https://gyazo.com/61d5906c4bbbe4887171fd5724c6d4ed

If u override vsmlrt-cuda and vstrt.dll from rife folder with:
https://github.com/AmusementClub/vs-mlr … 4.test2.7z
Issue gets fixed.
If u downgrade back to:
https://github.com/AmusementClub/vs-mlr … 14.test.7z
wiggle issue will reappear.

You can check the example here:
https://drive.google.com/drive/folders/ … sSi1Kl8tSn

Please compare the differences and address this Rife v2 issue when u have time.
Thanks a lot @Chainik!

Guide for updating RIFE to latest TensorRT from 3 days ago: https://github.com/AmusementClub/vs-mlrt/releases

Step 1: Download https://github.com/AmusementClub/vs-mlr … uda.v14.7z
Step 2: Go to C:\Program Files (x86)\SVP 4\rife\ and move the following to a backup folder: vsmlrt-cuda,vstrt.dll,vsmlrt.py
Step 3: Open vsmlrt-windows-x64-cuda.v14.7z and copy vsmlrt-cuda,vstrt.dll,vsmlrt.py in their place

Step 4: Edit vsmlrt.py replace line 1821
old:
tempfile.gettempdir(),
new:
os.path.expandvars("%APPDATA%\\SVP4\\cache\\"), #!SVP

Step 5: Edit helpers.py replace line 52
old:
return RIFE(clip,multi,1.0,None,None,None,model_num,backend,ensemble,implementation)
new:
return RIFE(clip,multi,1.0,None,None,None,model_num,backend,ensemble,True,implementation)

Step 6: Enjoy 10-15% performance drop caused by upgrading to latest code!

Noteable issues:
Your PC might not have the power to 2x a 4K video, and mpv might display desync errors that cancel SVP interpolation https://gyazo.com/a4714c85227b5b596a5f71084b5c2fae
so u should test it with a frc.frame.resize -19201080 or lower

Should SVP devs upgrade the out of the box vsmirt script so they will fix it for all users?
(knowing that non-lazy users might play around with v2 versions for that extra performance and run into that bug)

Never noticed, nevermind then. xD
If the upgrade to it would be safe and not create other issues, it would be nice for current SVP to be forward compatible with the future v2 models that they create.

I'm experiencing the same thing on my 4080. Our GPU is more bottlenecked by PCIe traffic flow than RAM at this point.
End goal should always be a stable product that users would brag about how good it is. "Random picture shake" effect when watching movies should not be one of it's downsides.