I've rolled back to the version from emacspeak 35, and nuked the
#ifdef blocks that were clearly not useful to anyone.
I've also checked in the last snapshot of the "new version" as
new-atcleci.cpp.
I may not have much more time to look at this, so if you or
someone else here has time to play with it -- go for it.
In the new-atcleci.cpp what was done was:
1. Take aplay.c from alsa 1.0.25
2. Copy out the xrun and suspend routines from that version into
atcleci.cpp
3. Update the corresponding pcm_write function in atcleci.cpp
>>>>> "Tim" == Tim Cross <tcross@xxxxxxxxxxx> writes:
Tim> Hi Raman, I will experiment with past revisions and try
Tim> to identify the 'best' version.
Tim>
Tim> Note that I've been using the 'previous' version
Tim> (previous to recent changes) for quite some time under
Tim> pulseaudio with no significant problems apart from
Tim> outloud segfaulting a few times a day. For me, the
Tim> responsiveness was only a little less snappy than it is
Tim> with the recent updates and did not suffer from the
Tim> repeated utterance artifact issue. The main potential
Tim> advantage of the recent changes was that it appeared to
Tim> make the server more stable.
Tim>
Tim> Bart, I believe that if you add a .asoundrc to your home
Tim> directory, outloud will use native alsa and not use
Tim> pulseaudio. However, you will run into device contention
Tim> if you have another device trying to access the sound
Tim> hardware, so make sure all other sound apps are stopped.
Tim>
Tim> I will post the output from the 'good' version in case
Tim> that can help with diagnostics etc. However, I may not
Tim> have access to do this until Monday.
Tim>
Tim> Tim
Tim>
Tim> T. V. Raman writes:
>>
>> On the plus side, I can now produce the problem on my home
>> machine if I tweak my .asoundrc -- so there is
>> card-specific asoundrc complexity biting again.
>>
>> An interesting test to try is to ask mplayer to play .flv
>> files using -vo null -- and see if the audio stutters ---
>> if it does, then you also see the buffer underrun and
>> consequently the artifacts Bart and Tim have seen.
>>
>> I'm almost tempted to roll back all the way back to last
>> week at this point.
>>
>> Bart/Tim, since yyyyou initially said things got better
>> for you, let me know the checkin that that corresponded to
>> -- for no wI'll go to 7548.
>> ---
>>
>>
>> >>>>> "tvr" == T V Raman <tv.raman.tv@xxxxxxxxxxx> writes:
tvr> Could you try with pulseaudio out of the picture? I'm
tvr> not using pulse on any o fmy machines, and am not seeing
tvr> breakage with char echo --- also, each commit Imade was
tvr> working fully on my machines when I committed them.
tvr>
>> >>>>>> "Bart" == Bart Bunting <bart@xxxxxxxxxxx> writes:
Bart> Hi, Rev 7548 is the last good commit. After this the
Bart> artifacts are still there but character echo is almost
Bart> toatlly broken.
Bart>
Bart> Here is the output when starting the outloud script:
Bart>
Bart> bart@xxxxxxxxxxx:~/emacs/lisp/emacspeak/servers$ ./outloud ALSA
Bart> <-> PulseAudio PCM I/O Plugin Its setup is: stream :
Bart> PLAYBACK access : RW_INTERLEAVED format : S16_LE
Bart> subformat : STD channels : 1 rate : 11025 exact rate :
Bart> 11025 (11025/1) msbits : 16 buffer_size : 5513
Bart> period_size : 1378 period_time : 125000 tstamp_mode :
Bart> NONE period_step : 1 avail_min : 1378 period_event : 0
Bart> start_threshold : 5513 stop_threshold : 5513
Bart> silence_threshold: 0 silence_size : 0 boundary :
Bart> 1445199872 allocating 2756 samples output buffered to
Bart> waveBuffer with size 2756 ALSA <-> PulseAudio PCM I/O
Bart> Plugin Its setup is: stream : PLAYBACK access :
Bart> RW_INTERLEAVED format : S16_LE subformat : STD channels
Bart> : 1 rate : 11025 exact rate : 11025 (11025/1) msbits :
Bart> 16 buffer_size : 5513 period_size : 1378 period_time :
Bart> 125000 tstamp_mode : NONE period_step : 1 avail_min :
Bart> 1378 period_event : 0 start_threshold : 5513
Bart> stop_threshold : 5513 silence_threshold: 0 silence_size
Bart> : 0 boundary : 1445199872 %
Bart>
Bart> On Wed, 7 Mar 2012 18:08:32 -0800, "T. V. Raman"
Bart> <tv.raman.tv@xxxxxxxxxxx> wrote:
>> >>> Tim and Bart,
>> >>>
>> >>> That is indeed strange --- am still not able to
>> reproduce >>> the issues.
>> >>>
>> >>> I did one more clean up pass -- do an svn update and
>> try >>> again.
>> >>>
>> >>> If the problems persist, run the tcl script at the
>> shell, >>> and send me the device configuration it
>> outputs.
>> >>>
>> >>> I'm typing this on my laptop with outloud doing the
>> >>> talking, and I dont have character echo going away --
>> nor >>> do I see the stutter Bart mentioned when using
>> voice-lock
>> >>> --
>> >>>
>> >>> --
>> >>>
>> >>> --
tvr>
tvr> -- Best Regards, --raman
>>
>> --
>> Best Regards, --raman
>>
>> -----------------------------------------------------------------------------
>> To unsubscribe from the emacspeak list or change your
>> address on the emacspeak list send mail to
>> "emacspeak-request@xxxxxxxxxxx" with a subject of
>> "unsubscribe" or "help".
>>
Tim>
Tim> -- Tim Cross Information Technology University of New
Tim> England Phone: +61 2 6773 3210 Mobile: 0428 212217 Fax:
Tim> +61 2 6773 3424 E-Mail: tcross@xxxxxxxxxxx Web:
Tim> http://www.une.edu.au/itd --- Please avoid sending me
Tim> Word or PowerPoint attachments. See
Tim> http://www.gnu.org/philosophy/no-word-attachments.html
Tim>
Tim> Any fool can make things bigger, more complex, and more
Tim> violent. It takes a touch of genius (and a lot of
Tim> courage) to move in the opposite direction. âAlbert
Tim> Einstein
Tim>
Tim>
Tim>
Tim>
Tim>
--
Best Regards,
--raman
-----------------------------------------------------------------------------
To unsubscribe from the emacspeak list or change your address on the
emacspeak list send mail to "emacspeak-request@xxxxxxxxxxx" with a
subject of "unsubscribe" or "help".
If you have questions about this archive or had problems using it, please send mail to:
priestdo@xxxxxxxxxxx No Soliciting!Emacspeak List Archive | 2010 | 2009 | 2008 | 2007 | 2006 | 2005 | 2004 | 2003 | 2002 | 2001 | 2000 | 1999 | 1998 | Pre 1998