It affects both, but ALSA API does the setup of stream parameters once,
but OSS API does this multiple times because there is no way to distict
the configuration phase and the working phase, thus we must reconfigure
after settings of all parameters.
If you strace aplay then you'll see that one A
It's related to the snd-ice1712 driver and hardware which has the cs8427
transciever. I've added cs8427_timeout module option, so you can
lower the reset timeout value for this chip. I left the safe value 50 (0.5
sec) as default. If you set this timeout to very small value, then
the S/PDIF outpu
0.000531 ioctl(4, SNDCTL_DSP_RESET, 0) = 0
0.000232 ioctl(4, SNDCTL_DSP_GETFMTS, 0xb668) = 0
0.510506 ioctl(4, SNDCTL_DSP_SETFMT, 0xb668) = 0
0.506515 ioctl(4, SNDCTL_DSP_STEREO, 0xb668) = 0
0.504940 ioctl(4, SOUND_PCM_READ_RATE, 0xb668) = 0
0.065604 io
I used the latest alsa-tarballs to upgrade alsa-lib, alsa-driver and
alsa-oss.
To make it short: The problem is even worse now !
This is the output from strace (relevant part):
.
.
0.000531 ioctl(4, SNDCTL_DSP_RESET, 0) = 0
0.000232 ioctl(4, SNDCTL_DSP_GETFMTS, 0xb668) =
I really want to resolve this bug, but there has been little help so
far. I haven't got any response from Perex, too. He was asigned to this
bug via alsa bug report...
Please, someone needs to tell me wich modules are connected with the
functions that took so long, so I can check revision of tt
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Lars Heineken wrote:
| I'll figure out how the script works and make the trace when I get home.
| (~9 hours).
All set, here we go:
The script calls this line:
sox 01.wav -t ossdsp /dev/dsp
After adding an "strace -r" to the begi
Clemens Ladisch wrote:
It seems play is a shell script that calls sox to play the file.
Please try to strace sox.
Oh, you are right ! I guess that was the reason why there was so little
alsa related stuff in the trace..
I'll figure out how the script works and make the trace when I get home.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Takashi Iwai wrote:
| try strace with -r option and check which call takes too long time.
I have tried to run the application with the longest delay (~5 seconds)
using 'strace -r' wich is an X-application. Unfortunately, it spills out
a lot of text (~
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Here is one more mail related to the problem. He uses a 2496 card, too:
- - He's experiencing the problem on a 2.4. kernel, but using a rather new
alsa-version (1.02). I guess it is related to updates made from alsa
0.96 to 1.02.
http://www.music.colum
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
- - I have tried to use a non preemptive kernel and as a suprise to me,
the problem is still there.
- - After googeling around, I found this post to [linux-audio-dev]:
http://eca.cx/lad/2003/11/0081.html.
What's strinking is, that he uses a 2496 card, t
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
.. some more infos on the issue:
- - The delay does not affect output via alsa directly, only the
OSS-emulation.
- - Using XMMS and jumping around in the song (via the main slider)
creates the same delay every time the player jumps to the new position.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi !
I switched from 2.4 to 2.6 a few days ago and noticed this strange
problem. Whatever program uses OSS-emulation takes 1-3 seconds to open
/dev/dsp and start playing sound. It seems that this has something to do
with the preemptive kernel option, b
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Did someone take notice of my mail, at least :-)
Lars Heineken wrote:
| I tried some capturing through the spdif input today, and it didn't
| worked, but after some googleing found this mail from jaroslav:
|
| | Known bug. The problem is
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi !
I tried some capturing through the spdif input today, and it didn't
worked, but after some googleing found this mail from jaroslav:
| Known bug. The problem is that playback and capture channels are
| assymetric for the envy24 chip and we cannot d
I think we need a similar extension to support Mandrake, after
installation from the sources all modules in modules.conf where gone,
except the ones from alsa.. :-)
Included is a patch that fixes alsaconf to work correctly under fedora
core 1. Otherwise alsaconf happily creates a modules.conf fi
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Takashi Iwai wrote:
| At Sun, 09 Nov 2003 21:41:58 +0100, Lars Heineken wrote:
|
|> Hi !
|>
|> I've been a member of htis list some time ago, and now I'm facing a
|> new problem: Playback through my Terratec EWX 24/96 (ice1712)
perience than me, to
find the reason for the poor ice1712 sound quality, I consider serious.
Regards,
Lars Heineken.
---
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments
Hi !
I finaly managed to capture sound from spdif-in using my Terratec EWX24/96 (ICE1712)
Doing it this way: arecord -D cards.pcm.iec958 -c 12 -r 44100 -f S32_LE test.wav
// all other ways failed, as you can only record from spdif when using S32_LE and 12
channels.
My problem now: This is a
Hi !
I finaly managed to capture sound from spdif-in using my Terratec EWX24/96 (ICE1712)
Doing it this way: arecord -D cards.pcm.iec958 -c 12 -r 44100 -f S32_LE test.wav
// all other ways failed, as you can only record from spdif when using S32_LE and 12
channels.
My problem now: This is a
19 matches
Mail list logo
