Search options (Page 58 of 355)

включается 13-й если хуже чем порог m1

The thing is SVP doesn't even know the total number of frames when transcoding cause mpv doesn't provide this value to the Vapoursynth filter.
So even if mpv is actually losing the last frame - there's nothing we can do about it.

BUT there could be another reason - look through the transcoding log, do you see any warnings like "Invalid video timestamp" or "non-strictly-monotonic PTS" ?
If the source reports zero frame duration for some frame then the output frames durations will also be zero, leading to these warnings, and I'm not sure what will happen next with those frames in the encoder.

> But, turning on dithering seems to have fix the problem anyway!

... which means you aren't using 10-bit output actually

> Perhaps there is some way to only enable that setting if it is a 10-bit video?

as a quick fix:

    AVS.push("ConvertBits(" + (media.p10 === true ? 16:8) + (media.src_fmt.endsWith("10") ? ", dither=0":"") + ")");

dunno, I can't see any banding in 10-bit mode

---
note that by default SVP only allows 10-bit output up to 72 fps (because of performance issues)

---

btw I just found that 10-to-8-bits conversion (if required) is made w/o dithering, it's probably a good idea to enable it
SVP 4\scrips\generate.js, line 356, replace

    AVS.push("ConvertBits(" + (media.p10 === true ? 16:8) + ")");

with

    AVS.push("ConvertBits(" + (media.p10 === true ? 16:8) + ", dither=0)");

or "dither=1"

it will install and register madVR and enable it in the MPC-HC player

> I didn't make that connection till I saw

he is using ffdshow, not AVSF, and ffdshow doesn't support 10-bit at all

could you share this sample? (torrent/magnet url is good too)

indeed... I remember I fixed this months ago, but it appears my changes to AVSF were lost for some reason

download, build, install...

you probably shouldn't install not-yet-released Ubuntu if you don't know how to build things

"обрезка", очевидно, кадр обрезает, уменьшает его площадь в пикселях

потому что тут важен общий поток "выходных данных" ~ ширина*высота*кадровая_частота

and your question is... ?

trudhd isn't allowed in mp4