> свой добавленный профиль имеет приоритет над "родным", автоматическим?

при прочих равных - да
значит надо сделать прочие не равными wink например, добавленному профилю отрицательных очков добавить

1,152

(3 replies, posted in Using SVP)

> So using an rtx 3080 and an 3950x, it does like like 5k videos aren't doable in real time.

I bet if you open it in a regular video player it will play fluently.
Even my 4800H+2600 laptop can do 2700p @60 -> 120 fps

- switch to NVIDIA CUVID in the LAV Video decoder properties
- switch SVP from "3D" mode to 2D
- enable NVIDIA Optical Flow in SVP

> в MPC-HC оно у меня не включается

Vapoursynth Filter надо поставить и включить, в описании написано
только на свой страх и риск, он пока весьма кривоват...

так что лучше в mpv смотреть "сколько fps выдаст 3080"

вот, например, ноутбучная 2060 прекрасно тянет ... 240p big_smile 360p уже нет.
только инициализируется эта вся шляпа атас как долго

h264 or h265 encoder?

> What should I do?

I wish I knew what encoder option is causing this on some files...

1,155

(6 replies, posted in Using SVP)

the same method as with Plex Media Player, I suppose

except you don't have to replace/remove mpv.dll, the one bundled with JMP is ok

> Create mpv's configuration file in "C:\Users\<name>\AppData\Local\JellyfinMediaPlayer\mpv.conf" with the following contents:

input-ipc-server=mpvpipe
hwdec-codecs=all
hr-seek-framedrop=no

these artifacts are from the encoder, not from SVP's interpolation

уже почти скоро будет... там больше организационных вопросов осталось, чем программных.
только батарейку жрет как тяжелая игра.

GPU; VRAM; CPU utilization; SVP index

424*240 - 23% (65% compute_1); 0.9 GB; zero; 1.0
640*360 - 9% (100% compute_1); 0.9 GB; zero; 0.76

1,159

(18 replies, posted in Using SVP)

> MPV should "only" have to passthrough HDR metadata

yep

> did you somehow manage to get the performance of the CUDA?

nope

> Is it possible to get smooth x2 playback with RIFE on your GPU on some resolution? For example 480p? 720p?

not a chance
---
240p works great, 360p is too much big_smile

1,161

(18 replies, posted in Using SVP)

BTW the new version have a tone mapping "shader" (in fact, an OpenCL routine) implemented in the Vapoursynth plugin, so both transcoder and SVPcast can now give satisfactory colors w/o using ffmpeg's "tonemap" filter (which is probably better, but _super_ slow).

1,162

(4 replies, posted in Using SVP)

this should be fixed now

https://www.svp-team.com/wiki/RIFE_AI_interpolation

1,164

(7 replies, posted in Using SVP)

should be fixed now

os? mpv version?

1,166

(7 replies, posted in Using SVP)

> in one HDR movie

what is the movie exactly?

check the performance stats - cpu/gpu load, droped frames, fluency test...
2-cores laptop haswell is probably too slow for 1080p

>  Ask the developer of the RIFE filter for VapourSynth to implement the code

actually there no changes needed in the "RIFE filter for VapourSynth" itslef
the problem is "ncnn with CUDA" is not completed yet, some functions necessary for RIFE are not implemented... and the last commit was on Jan 18.

> Implement the "original" RIFE in SVP - disadvantage: 2.5 GB of dependencies

... plus very long (re-)start time. a video player will hang for 5-10 secs on each seek, for example.

> how much time could be saved if this second operation was omitted

zero big_smile encoding via NVENC costs nothing...
even if you want good old software encoder then it will be like 10 times faster than "rife over cuda"

more like "RIFE over ncnn does not need these dependencies"
and the "official" ncnn only supports Vulkan as of now

> Great!!!

not great, because we definitely won't use 2.5 GB of dependencies required to run the "original" RIFE

there's a ncnn/CUDA (https://github.com/atanmarko/ncnn-with- … ocator.cpp) but it was last updated in January... so I dunno...

yeah, CUDA version is like 3 times faster than ncnn (in flowframes) and loads my GPU up to 50%

laptop rtx 2060 --> 704*528 @30->60 --> 24 fps transcoding speed
1080p @24->48 --> 4 fps
(fp32 = false, tta = false)

is it supposed to use all the GPU? hmm because I only see up to 10% GPU load regardless of gpu_thread value

---
flowframes shows the same speed and gpu usage so lets assume this is "normal"

bottom line - I think we could integrate this for the SVPcode usage

1,174

(18 replies, posted in Using SVP)

> any chance that it may be adapted to the chain MPV/vapoursynth ?

afaik, there's no maintainer for the mpv's encoding mode...

1,175

(18 replies, posted in Using SVP)

> as far as I understand it is now possible to encode HDR videos

only after HDR->SDR conversion

> Would it be possible to update SVP transcoder to get to use NVEncC instead instead of ffmpeg

SVP doesn't use ffmpeg directly but via the mpv's interface (cause we want to pass a video through the Vapoursynth filter)