Hi. I found another things interesting. I compile mplayer2 with "use mpg123 as backend". But mplayer2 plays correctly!!!! It seems this is a problem of mpg123, not the library?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Compared to mplayer? Perhaps even using -ac mpg123? Or at least: Do you mean truncated in output from the loudspeakers (that's what the other truncation bugs are about) or truncated when storing to a uncompressed file?
I.e., is for you
mpg123 -w out.wav azure.mp3
aplay out.wav
OK? Instead of aplay, actually looking at the waveform with something like audacity would help, too. If it's "just" about what you hear, not about what is decoded, it's important to know which audio output method is used. Pulseaudio? Alsa? The former via the latter? SDL? Since this is gentoo, I suppose this depends on your preference ...
What we know is that some audio output methods have trouble with playing the last bits, but usually on Windows or Mac.
Btw, "azure" sounds fine for me on my box, so I need more input from you to see what's broken.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
So it is just about what I hear. mpg123 -w/aplay out.wav works.
But mpg123 azure.mp3 has the truncation problem.
My gentoo has pulseaudio running. mpg123 is also built with pulseaudio support. Kmix do not report it use alsa so I think it is using pulseaudio. (but not 100% sure) The buffer size of INTEL driver is 2048. Any additional information I need to provide?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Am Wed, 20 Mar 2013 02:50:09 +0000
schrieb "Broken.zhou" zhou13@users.sf.net:
My gentoo has pulseaudio running. mpg123 is also built with pulseaudio support.
So, to be sure: You get the same with
mpg123 -o pulse azure.mp3
? And it's really obvious that the word is cut before the end, I presume (so I can't miss it when it happens)? I'd have to check on an ubuntu install with pulse ... or install that thing here.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I finally got around fixing that. Turns out we simply need to call pa_simple_drain() when closing the output. Somehow I thought it would be default behaviour to play back what has been provided before closing. Well, that were the old times of simple audio APIs like OSS (not the emulation), that simply worked for simple things and got hairy if non-simple things. Now you gotta fight to get the audio played back at all.
Hi. I found another things interesting. I compile mplayer2 with "use mpg123 as backend". But mplayer2 plays correctly!!!! It seems this is a problem of mpg123, not the library?
Compared to mplayer? Perhaps even using -ac mpg123? Or at least: Do you mean truncated in output from the loudspeakers (that's what the other truncation bugs are about) or truncated when storing to a uncompressed file?
I.e., is for you
mpg123 -w out.wav azure.mp3
aplay out.wav
OK? Instead of aplay, actually looking at the waveform with something like audacity would help, too. If it's "just" about what you hear, not about what is decoded, it's important to know which audio output method is used. Pulseaudio? Alsa? The former via the latter? SDL? Since this is gentoo, I suppose this depends on your preference ...
What we know is that some audio output methods have trouble with playing the last bits, but usually on Windows or Mac.
Btw, "azure" sounds fine for me on my box, so I need more input from you to see what's broken.
So it is just about what I hear. mpg123 -w/aplay out.wav works.
But mpg123 azure.mp3 has the truncation problem.
My gentoo has pulseaudio running. mpg123 is also built with pulseaudio support. Kmix do not report it use alsa so I think it is using pulseaudio. (but not 100% sure) The buffer size of INTEL driver is 2048. Any additional information I need to provide?
Am Wed, 20 Mar 2013 02:50:09 +0000
schrieb "Broken.zhou" zhou13@users.sf.net:
So, to be sure: You get the same with
mpg123 -o pulse azure.mp3
? And it's really obvious that the word is cut before the end, I presume (so I can't miss it when it happens)? I'd have to check on an ubuntu install with pulse ... or install that thing here.
mpg123 -o pulse azure.mp3
has the truncation problem but
mpg123 -o alsa azure.mp3
works fine. So it's a problem of pulseaudio I think.
I finally got around fixing that. Turns out we simply need to call pa_simple_drain() when closing the output. Somehow I thought it would be default behaviour to play back what has been provided before closing. Well, that were the old times of simple audio APIs like OSS (not the emulation), that simply worked for simple things and got hairy if non-simple things. Now you gotta fight to get the audio played back at all.
Anyhow, please test current http://mpg123.org/snapshot and report.