At 01:02 PM 11/5/2006, Tony Espy wrote:
>hi -
>
>i'm having problems getting splay to work with ASLA on
the Pepper Pad 3 (
>AMD Geode / FC4-based system ). i had this working last
week, but did a
>clean build on the cayenne_1_5_0 branch this weekend and
now i can't get
>MP3 playback to work via ALSA.
Hi Tony,
If this is a Helix ALSA device-related problem, I'll work
with you to
resolve it.
What is the version of your ALSA?
rpm -q --info alsa-lib
There were two check-ins 11 days ago to Helix ALSA device
code:
http://lists.helixcommunity.org/pipe
rmail/audio-dev/2006-October/000776.html
http://lists.helixcommunity.org/pipe
rmail/audio-dev/2006-October/000775.html
Could you do the following steps to trouble-shoot the
problem?
1. Check if your system's ALSA configuration is working:
speaker-test -c1 --device default
speaker-test -c2 --device default
speaker-test -c2 --device default --rate 48000
speaker-test -c2 --device default --rate 44100
2. A quick check if ALSA on your system is not set to
48000Hz Sampling Rate.
Change/Add the last line in this block to the relevant file
used by your
player:
~/.helix/HelixSDK_1_0:
~/.helix/HelixSDK_10_0:
~/.realplayerrc:
~/.hxplayerrc:
AlsaVaryingSampleRate=1
Now, play a clip to check if this fixed the playback using
ALSA.
3. Reverse the two check-ins to verify that it indeed is the
checkins that
is causing the problem:
cd audio/device
cvs -q update -D 2006-10-25 platform/unix/audlinux_alsa.cpp
platform/unix/audlinux_oss.cpp platform/unix/audUnix.cpp
pub/platform/unix/audlinux_alsa.h
pub/platform/unix/audlinux_oss.h
pub/platform/unix/audUnix.h
Alternatively:
cvs -q update -r 1.4.6.2 platform/unix/audlinux_alsa.cpp
cvs -q update -r 1.10 platform/unix/audlinux_oss.cpp
cvs -q update -r 1.7 platform/unix/audUnix.cpp
cvs -q update -r 1.4 pub/platform/unix/audlinux_alsa.h
cvs -q update -r 1.5 pub/platform/unix/audlinux_oss.h
cvs -q update -r 1.4 pub/platform/unix/audUnix.h
Then, rebuild and test.
4. Check if one of the check in is causing the problem:
There is more than an hour gap between the two checkins. I'm
trying to find
a "cvs update" command that will take both date
and time.
5. The most detail: It is possible that one of these diffs
is causing the
problem, but we'll have to find out why:
 -499,7
+617,7
unsigned int sample_rate = 0;
unsigned int channels = 0;
unsigned int buffer_time = 500000; /* 0.5
seconds */
- unsigned int period_time = buffer_time / 4; /* 4
interrupts per buffer */
+ unsigned int period_time; /* To set
to 4
interrupts per buffer */
switch (pFormat->uBitsPerSample)
{
-624,20
+742,20
if (err == 0)
{
- unsigned int period_time_out;
- period_time_out = period_time;
+ unsigned int period_time_out; // Output parameter
(plus input).
+ period_time_out = period_time = buffer_time / 4 ;
err =
snd_pcm_hw_params_set_period_time_near(m_pAlsaPCMHandle,
hwparams, &period_time_out, 0);
if (err < 0)
{
-715,18
+833,18
if (err == 0)
{
- err =
snd_pcm_sw_params_set_avail_min(m_pAlsaPCMHandle, swparams,
period_size);
+ err =
snd_pcm_sw_params_set_avail_min(m_pAlsaPCMHandle, swparams,
(snd_pcm_uframes_t)1);
if (err < 0)
{
HXLOGL1 ( HXLOG_ADEV,
"snd_pcm_sw_params_set_avail_min: %s",
snd_strerror(err));
-1025,21
+1256,50
int err = 0;
snd_pcm_sframes_t frame_delay = 0;
+ // Check that it is not SND_PCM_STATE_XRUN.
+ snd_pcm_state_t state;
+ state = snd_pcm_state(m_pAlsaPCMHandle);
+
+ if(state == SND_PCM_STATE_XRUN)
+ {
+ HXLOGL1 ( HXLOG_ADEV, "XRUN in
GetBytesActuallyPlayedUsingDelay()");
+ }
+
+ // snd_pcm_sframes_t frame_ready;
+ // frame_ready =
snd_pcm_avail_update(m_pAlsaPCMHandle);
+ // HXLOGL4 ( HXLOG_ADEV, "snd_pcm_avail_update
frame_ready : %d",
frame_ready);
+
err = snd_pcm_delay (m_pAlsaPCMHandle,
&frame_delay);
+ // HXLOGL4 ( HXLOG_ADEV, "frame_delay : %d",
frame_delay);
if (err < 0)
{
- HXLOGL1 ( HXLOG_ADEV, "snd_pcm_status:
%s",
snd_strerror(err));
+ HXLOGL1 ( HXLOG_ADEV, "snd_pcm_delay:
%s", snd_strerror(err));
+ }
+ else if (frame_delay < 0)
+ {
+ HXLOGL1 ( HXLOG_ADEV, "XRUN! frame_delay:
%d", frame_delay);
+ nBytesPlayed = m_ulTotalWritten;
+ retVal = HXR_OK; // Is this really OK?
}
else
{
- int bytes_delay;
+ UINT64 bytes_delay;
bytes_delay = snd_pcm_frames_to_bytes
(m_pAlsaPCMHandle,
frame_delay);
-
+ if (bytes_delay > m_ulTotalWritten)
+ {
+ // We have only written m_ulTotalWritten bytes,
but ALSA is
telling us that we have a delay of bytes_delay!
+ HXLOGL4 ( HXLOG_ADEV, "WARNING: Bytes
Delay > Total Written!
bytes_delay = %d, m_ulTotalWritten = %d", bytes_delay,
m_ulTotalWritten);
+ bytes_delay = m_ulTotalWritten; // Avoid a
negative result or
a large unsigned number below.
+ }
+ else
+ {
+ HXLOGL3 ( HXLOG_ADEV,
"snd_pcm_frames_to_bytes bytes_delay =
%d", bytes_delay);
+ }
nBytesPlayed = m_ulTotalWritten - bytes_delay;
retVal = HXR_OK;
}
-// HXLOGL4 ( HXLOG_ADEV, "nBytesPlayed: %llu,
m_ulTotalWritten:
%llun", nBytesPlayed, m_ulTotalWritten);
+ HXLOGL4 ( HXLOG_ADEV,
"GetBytesActuallyPlayedUsingDelay() nBytesPlayed:
%llu, m_ulTotalWritten: %llu n", nBytesPlayed,
m_ulTotalWritten);
return retVal;
}
Thanks.
--
Daniel Yek
>i have HELIX_FEATURE_ALSA defined in my .pf file.
>
>when I run splay, no sound is output.
>
>i see the following ALSA related output when I run
splay:
>
>Opening ALSA PCM device default
>
>the "===Stream 0===" output shows NumChannels
and Channels both are set to 2.
>
>i've also confirmed via 'lsof' that splay has the
correct PCM device open
>( /dev/snd/PCMC0D0p ).
>
>OSS output works just fine.
>
>any help would be much appreciated.
>
>thanks,
>/tony
_______________________________________________
Audio-dev mailing list
Audio-devhelixcommunity.org
http://lists.helixcommunity.org/mailman/listinfo/audio
-dev
|