[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Search]
Re: IBM ViaVoice and Alsa
- To: jasonw@xxxxxxxxxxx
- Subject: Re: IBM ViaVoice and Alsa
- From: "T. V. Raman" <raman@xxxxxxxxxxx>
- Date: Tue, 9 May 2006 20:23:43 -0700
- Delivered-To: priestdo@xxxxxxxxxxx
- Delivered-To: emacspeak@xxxxxxxxxxx
- In-Reply-To: <Pine.GSO.4.58.0605101242320.24415@xxxxxxxxxxx>
- List-Help: <mailto:emacspeak-request@xxxxxxxxxxx?subject=help>
- List-Post: <mailto:emacspeak@xxxxxxxxxxx>
- List-Subscribe: <mailto:emacspeak-request@xxxxxxxxxxx?subject=subscribe>
- List-Unsubscribe: <mailto:emacspeak-request@xxxxxxxxxxx?subject=unsubscribe>
- Old-Return-Path: <tvraman@xxxxxxxxxxx>
- Reply-To: raman@xxxxxxxxxxx
- Resent-Date: Tue, 9 May 2006 23:23:44 -0400 (EDT)
- Resent-From: emacspeak@xxxxxxxxxxx
- Resent-Message-ID: <W6giI.A.BgB.AzVYEB@xxxxxxxxxxx>
- Resent-Sender: emacspeak-request@xxxxxxxxxxx
Correct. Underruns will cause clicks. When you hear repeating
artifacts in sound processing programs, it is usually a result of
a "wrap around" bug ---
which are typically a consequence of someone reading a buffer
through to its end and then wrapping around to the beginning
(that's what produces the repeat)---
I suspect there is some bizarre interaction between the buffering
size I picked in the ViaVoice TTS server and what Alsa is trying
to do --- though at the time I didn't see anything in the ALSA
docs that recommended specific buffer sizes.
If you start the Viavoice server at the shell you'll see the
buffer size it's using. It looks suspicious because it's not a
power of 2.
I cloned the relevant code carefully from aplay.c --- in the
belief that it knew what it was doing, and in general that has
proven to be correct.
To understand how the buffer size to be used is determined, look
at the code in
alsa_configure (void)
One thing that might be worth experimenting with are the
start_delay and stop_delay numbers.
>From memory, I believe I originally had start_delay to be 0 and
this caused problems on the IBM X series laptops which is how I
went to start_delay=5.
It might well be worth trying a different value for stop_delay to
see if the wrap around bug goes away.
In general there is a fair amount of black magic involved in the
various parameters and knobs that alsa provides, and I dont claim
to understand all of them. So in the absence of complete
understanding, brute-force engineering rules, try it, if it
breaks, glue the pieces and try something else to see if it
breaks differently.
>>>>> "Jason" == Jason White <jasonw@xxxxxxxxxxx> writes:
Jason> On Tue, 9 May 2006, Lukas Loehrer wrote:
Jason>
>>
> The anonymous CVS is still not updated, but I did some tests
>> myself and have come to the conclusion that the xrun
>> events are not the cause of my repetition problems.
Jason> The usual symptom of xrun events is drop-outsand
Jason> clicks in the audio. My desktop system now has an RME
Jason> Digi9652 cards, which consistently produces xruns due
Jason> to IRQ conflicts in the machine.
Jason>
Jason> -----------------------------------------------------------------------------
Jason> To unsubscribe from the emacspeak list or change your
Jason> address on the emacspeak list send mail to
Jason> "emacspeak-request@xxxxxxxxxxx" with a subject of
Jason> "unsubscribe" or "help"
--
Best Regards,
--raman
Email: raman@xxxxxxxxxxx
WWW: http://emacspeak.sf.net/raman/
AIM: emacspeak GTalk: tv.raman.tv@xxxxxxxxxxx
PGP: http://emacspeak.sf.net/raman/raman-almaden.asc
Google: tv+raman
IRC: irc://irc.freenode.net/#emacs
-----------------------------------------------------------------------------
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"
Emacspeak Files |
Subscribe |
Unsubscribe | Search