[MPlayer-dev-eng] [MPlayer-cvslog] r34099 - in trunk/libass: ass.c ass.h ass_bitmap.c ass_bitmap.h ass_cache.c ass_cache.h ass_drawing.c ass_font.c ass_font.h ass_fontconfig.c ass_fontconfig.h ass_library.c ass_library.h ass_parse.c...

[Reimar Döffinger]

On Sun, Sep 18, 2011 at 02:41:00PM +0200, Erik Auerswald wrote:
> >My concern now is mostly about making sure that no significant
> >optimizations are lost when removing our heavily modified mp3lib.
> 
> ffmp3float seems to be faster on my system:
> 
> ==========================================================================
> Requested audio codec family [mpg123] (afm=mpg123) not available.
> Enable it at compilation.
> Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
> AUDIO: 44100 Hz, 2 ch, floatle, 192.0 kbit/6.80% (ratio: 24000->352800)
> Selected audio codec: [ffmp3float] afm: ffmpeg (FFmpeg MPEG layer-3 audio)
> ==========================================================================
> AO: [null] 44100Hz 2ch floatle (4 bytes per sample)
> Video: no video
> Starting playback...
> A: 271.7 (04:31.6) of 358.0 (05:58.0)  0.9% 0%
> Audio output truncated at end.
> A: 271.7 (04:31.6) of 358.0 (05:58.0)  0.9% 0%
> 
> 
> BENCHMARKs: VC:   0.000s VO:   0.000s A:   2.433s Sys: 276.795s =  279.228s
> BENCHMARK%: VC:  0.0000% VO:  0.0000% A:  0.8712% Sys: 99.1288% = 100.0000%
> 
> ==========================================================================
> Trying to force audio codec driver family mp3lib...
> Opening audio decoder: [mp3lib] MPEG layer-2, layer-3
> AUDIO: 44100 Hz, 2 ch, s16le, 128.0 kbit/9.07% (ratio: 16000->176400)
> Selected audio codec: [mp3] afm: mp3lib (mp3lib MPEG layer-2, layer-3)
> ==========================================================================
> AO: [null] 44100Hz 2ch s16le (2 bytes per sample)
> Video: no video
> Starting playback...
> A: 271.6 (04:31.6) of 538.0 (08:58.0)  1.1% 0%
> Audio output truncated at end.
> A: 271.6 (04:31.6) of 538.0 (08:58.0)  1.1% 0%

Yes, but they are outputting to different formats.
I guess where Carl sees the performance loss is when you need s16le in
the end, e.g. if you add -af format=s16le (or something like that, do
not know the exact syntax).