Voice synth broken on Linux?

Bert Freudenberg bert at isg.cs.uni-magdeburg.de
Mon Oct 16 09:07:01 UTC 2000


On Sun, 15 Oct 2000, Kevin Fisher wrote:

> Speaker default say: 'This is the default voice.'.
> 
> In the Windows VM, it works great.  However, the Linux VM doesn't fare so 
> well...the voice is broken up badly.  Actually, broken up is the wrong 
> description...you only hear fragments of each word, seperated by long delays.

There is a tiny bug in the Sound support which only shows on certain sound
cards (like SoundBlaster Live). If you hear similar dropouts in the
Midi synth (PlayWithMe 8) then this is the problem for sure. 

You need to change one line in src/unix/sqUnixSound.c, function
auPollForIO(). The semaphore needs to be signalled as soon as any fragment
is available. So change the comparison to

	if (info.fragments > 0)

That did the trick for me ...

-- Bert





More information about the Squeak-dev mailing list