Measurements. There may be a difference since each of them was done once.
CPU: Ryzen 5800X
RAM: 64GB
GPU: RTX2070 Super 8GB
TestSource 720p
H.264 YUV420 8bit 23.976fps 30sec 1280x720
TestSource 1080p
H.264 YUV420 8bit 23.976fps 30sec 1920x1080
Output 4K(Proteus)
ProRes422(YUV422 10bit) 23.976fps 30sec 3840×2160
v2.6.4
Measured time from “Processing Started” to “Processing Ended” in log.
v3.0.0.40a
Directly execute ffmpeg.
-sws_flags spline+accurate_rnd+full_chroma_int -filter_complex veai_up=device=-2:model=prob-3:scale=?:compression=0.24:details=0.24:blur=0:noise=0.12:halo=0:preblur=0,scale=w=3840:h=2160,colorspace=bt709:iall=bt601-6-625:fast=1 -color_primaries 1 -color_trc 1 -colorspace 1
720p to 4K is scaled 4 x and then reduced to 4K.
1080p to 4K scaled 2x.
720p to 4K
v2.6.4: 7m34s
v3.0.0.40a: 8m58s
1080p to 4K
v2.6.4: 5m52s
v3.0.0.40a: 6m22s
The conversion speed seems to have dropped from v2.6.4 to v3.0, but if the processing has been upgraded from 8-bit to 16-bit, the drop is acceptable.
