View Single Post
Old 26th July 2020, 10:06   #10  |  Link
vigan1
Registered User
 
Join Date: Jul 2020
Posts: 28
Yes, it's faster than UTvideo T1 but not like the graph on the website shows. But it's not faster than UTvideo T2.

Now UTvideo T1 is in version 21 and it's the same speed (almost) as magicyuv and it's opensource and FREE. UTvideo T2 v21 is faster than magicyuv.

There is a free trial version of magicYUV if you want to test, I was thinking it was the fastest codec, but it seems the graphs are outdated, as it is compared to UTvideo v17.

I respect ignus from magicYUV but my test showed me that UTvideo is as good, and free. Sorry Ignus.
EDIT : There is a free version of magicYUV the v1.2.

Decoding is supported, but encoding is not in FFmpeg. Because it is not opensource, and it was reverse engineered I think, but they failed to implement "multithreaded per-frame decoding" which make ffmpeg decode magicyuv slower than the vfw codec.
(This info might be outdated... maybe ffmpeg is right now).

EDIT : NEW TEST with 1080p Jellyfish 30 seconds video. I wanted to try on 10BIT DSLR videos, not screen shots.
Results are completely different than 8bit RGB !!!

Cineform 10bit422 Filmscan1 = 551MB
FFV1 8bit422 = 634MB
FFV1 10bit422 = 922MB
MagicYUV2 10bit422 =912MB
UTvideoPRO 10bit422 = 1180MB
x264 8bit422 SLOW --keyint1 =895MB
x264 8bit422 Ultrafast --keyint1 =932MB
x264 10bit422 Ultrafast with P-frames =1565MB
x264 10bit422 Ultrafast --keyint1 =1589MB

Last edited by vigan1; 27th July 2020 at 15:21.
vigan1 is offline   Reply With Quote