Tests with emacspeak and 4front soft mixer

["Greg E. Priest-Dorman" <priestdo@xxxxxxxxxxx>]

Hi Folks,

Longish post, so here is a quick summary:

Commercial soft mixer from 4Front on P166 and P266 debian linux system
is not robust enough to do ViaVoice speech.  The OSS (withought soft
mixer) works fine.  I have incuded details below, at the end I have
included a selection of error messages from /var/log/syslog.  


I have been playing with the soft mixer from 4front.  I have tried it
now on two systems.  Both "low end" by todays standards.  One a P166
with 64 megs of memory running Debian 2.2, emacspeak 12 and ViaVoice
TTS 5.  The other is the same except a P266.  

These systems both work well with the free sound drivers.  My
motivation in tring the commercial driver was to get the ability to
mix several audio streams. 

How I tested:

After loading the sound driver I started emacspeak, brought in a text
file (Mark Twain's "Is Shakespear Dead") and hit 'C-e b'.  In a few
seconds it started reading.

First off I was struck by the noticable drop in sound quality which
happened when I added the soft mixer.  The comercial sound driver
itself was very clear, however adding the soft mixer on my systems
caused a noticeable degradation in the quality of speech.

Once the system began speaking I started getting error messages on the
screen (not spoken, just printed) and in /var/log/syslog.
These were produced by the sound driver and referd me to one of the
README files that came with the driver, README.performance.  This
document states:

   First of all you should understand that the above messages are not 
   caused by a bug in OSS or a problem in OSS configuration. OSS just
   tells that the application you are trying to use does not work
   correctly. In some cases this happens because the application is
   buggy and sometimes your system/environment just doesn't meet the
   requirements of this particular application.

   IN ANY CASE THE PROBLEM IS SOMETHING THAT IS BEYOND CONTROL OF OSS
   DRIVER AND 4FRONT TECHNOLOGIES.

   For the above reason trying to fix the problem by changing OSS
   settings is completely useless. Such changes will not make any
   difference.

As I was using ViaVoice I will give IBM the benifit of the doubt and
take the second explanation (not enough horsepower) rather than the
first (buggy program).  Still, it seems odd that I can run ViaVoice
just fine on a P166 yet I cannot run it on a P266 with a software
mixer that is mixing nothing at the time.

In any event, within a few minutes on the P166 the sound driver
informed me in printed error messages that it was unloading due to
excessive processor load.  It did.

On the 266 I got similar errors.  Here however in stead of unloading
it occasionally got caught in a loop repeating the same sound over and
over (two words from the text) untill I removed the sound driver from
memory. 

I can only assume that to use the software mixer with ViaVoice you
need much more horsepower than a lowly P166 or P266 can provide.

I would be interested to know the outcome if someone could try this
with a P500 or the like.  However, if it does unload you loose sound
untill you reissue the "soundon" command.  If you do try it, set up a
root virtual console (like on Alt F2) so you have a place you can go
and restart sound.

If you are still interested, here are the some of the error messages
from /var/log/syslog:

on the P266

Nov 19 23:35:13 herbert kernel: SoftOSS: CPU overload. Limiting # of
voices to 9

Nov 19 23:35:13 herbert kernel: SoftOSS: Dropping sampling rate and
stopping the device.

Nov 19 23:37:39 herbert kernel: OSS warning: Application caused too
many output underrun errors on audio device #0.


on the P166

Nov 17 00:44:39 herbert kernel: SoftOSS: CPU overload. Limiting # of
voices to 9

Nov 17 00:44:39 herbert kernel: SoftOSS: Dropping sampling rate and
stopping the device.

Nov 17 00:44:58 herbert kernel: OSS warning: Application caused too
many output underrun errors on audio device #0.

Nov 17 00:44:58 herbert kernel: Look at oss/docs/README.performance
for more info.


Hope this helps,

Greg


--
 Greg Priest-Dorman
 priestdo@xxxxxxxxxxx


-----------------------------------------------------------------------------
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"