Closed GoogleCodeExporter closed 9 years ago
also Elapsed & Remaining Time
Original comment by ved_k...@ymail.com
on 9 Sep 2011 at 7:53
I use this custom 2pass commands with the 64 bits version of ffmpeg:
ffmpeg.exe -i "C:\movie001.avi" -vcodec libx264 -b 3072k -bt 3072k -threads 4
-pass 1 -acodec libmp3lame -ab 96k -ac 1 -ar 44100 "C:\movie001.mp4"
ffmpeg.exe -i "C:\movie001.avi" -vcodec libx264 -b 3072k -bt 3072k -threads 4
-pass 2 -acodec libmp3lame -ab 96k -ac 1 -ar 44100 "C:\movie001.mp4"
I wrote the number of my processor threads (4) and it make the first pass at 30
fps and the second at 24-26. (Using a 720p raw avi video).
But using bencos witht the same video i get very very slow encoding speeds (3
fps).
So i think it doesnt detect properly my cpy threads or the program needs a
option to disable the use of preset, profiles and tune, in order to get more
speed.
Original comment by nuclears...@gmail.com
on 9 Sep 2011 at 8:02
bencos use higher quality presets than ffmpeg's defaults..
in my tests I didn't see any (speed) difference using 32bit or 64bit.
about progress I'll see later, if I finish the threaded encoding.
Original comment by s1r...@gmail.com
on 10 Sep 2011 at 1:42
I made some tests with the same video enconding first with bencos in the
maximun settings (high10, film, very slow) and with ffmpeg simple input, and
the quality result was aprox. the same. The idea is to add the option of not
using the custom presets for people with slower computers be able to code in
decent speeds. I have a i5 processor which is ridiculous and when im enconding
hd video i need to use the vp8 codec for not taking years. (Sorry for my shitty
english).
Original comment by nuclears...@gmail.com
on 10 Sep 2011 at 6:05
presets are for h264, why are you talking about vp8?
if it's too slow, why not using a faster preset?
also vp8 is mostly single threaded: it never used more than 1 core on my i3.
also, create new issues for new issues.
about progress bar, it's only cosmetic. you can see in the realtime logs where
the encode is.
Original comment by s1r...@gmail.com
on 14 Sep 2011 at 5:44
Original issue reported on code.google.com by
nuclears...@gmail.com
on 8 Sep 2011 at 10:44