kern/107516: emu10k1 - skips, clicks and lag after a day of heavy usage

[Ricardo Nabinger Sanchez]

On Sat, 6 Jan 2007 13:52:00 +0300 (MSK)
Yuriy Tsibizov <Yuriy.Tsibizov at gfk.ru> wrote:

> I think I fixes most low-bitrate playback problems I had in emu10kx.

Tested, and there are still some issues:

* roger.wav: no distortions so far, but the play command takes longer to end:

% /usr/bin/time -l play roger.wav 

Input Filename : roger.wav
Sample Size    : 8-bits
Sample Encoding: unsigned
Channels       : 1
Sample Rate    : 22050

Time: 00:00.88 [00:00.00] of 00:00.88 ( 100.0%) Output Buffer:  19.51K

Done.
        3.01 real         0.01 user         0.02 sys
      1852  maximum resident set size
       143  average shared memory size
       124  average unshared data size
       128  average unshared stack size
       484  page reclaims
         0  page faults
         0  swaps
         0  block input operations
         0  block output operations
         0  messages sent
         0  messages received
         0  signals received
       290  voluntary context switches
        19  involuntary context switches

* chat1.wav: there's no initial glitch, but while the sound is playing back,
a small and progressive distortion is noticeable.  It sounds like amplitude
clipping, and reminds me of the noise of loosely connected analog audio
cables.

* _roger.wav (22050 Hz, 16 bits, signed, stereo): no distortions/glitches.

* offset_sine.wav: no initial glitch (expected), but the same type of
distortion with chat1.wav, only louder.

* sine.wav: no initial glitch (it happened before), and the same type of
distortion with offset_sine.wav.


Forcing sine.wav, offset_sine.wav, and chat1.wav to stereo output doesn't
help to avoid the distortion, the distortion only gets complex, with a
behavior similar to constructive/destructive waves.

With XMMS I noticed noises right after hitting play and stop, and when it
goes to the next music.  The sound reminds me of LP player (vinyl records),
because it's very quick and sounds similar.  There's no such noise in the
remaining of the musics, though.

Is there any way to test using cat directly to device, to rule out problems
caused by either sox's play or XMMS?

Regards.

-- 
Ricardo Nabinger Sanchez     <rnsanchez@{gmail.com,wait4.org}>
Powered by FreeBSD

  "Left to themselves, things tend to go from bad to worse."