[Fwd: Re: [Gwc-general] Cursor stops too soon on 0.20.10]

[John C. <ci...@pu...>]

Jeff,

I just installed GWC 0.21.03 beta.  It looks like the cursor
stalling has been resolved to a great degree.
With 20 seconds or more of audio on the screen, the cursor
appears to make it all the way most of the time.  I notice
as I zoom in more and more the end point gradually creeps
away from the end, but still not bad.  I think one thing
that made it less bothersome in the .19 series was that the
cursor was erased after the last sample was "played". This
didn't bother me and I wouldn't mind if the cursor was again
designed to disappear after the sound card claims that all
of the bytes have been processed (as long as it comes
reasonably close to the end as it does now.)
But it's definitely much better than 0.20.10.  It's very
close to 0.19.7 which was the last 0.19 series that I had
been using.

I wondered if this situation is soundcard-related. I tried
installing gwc on my backup server which has a cheap onboard
VIA sound chip.  Using GWC 0.20.10, the degree of cursor
stopping seems about the same as my es1371 so I don't
understand this. Speaking purely from ignorance I wonder if
it is somehow the driver reporting the early finish?
I am using the stock ess soundsystem on both machines by the
way.

Anyway, it's pretty good now. I can use this. And with the
channel-selection issue probably fixed, it should be much
more pleasant to work with.
Thanks for the quick service!

John


Jeff Welty wrote:
> Ahh, the cursor -- my nemisis.  This has been the constant thorn in 
> GWC's (and my) foot for a long time.
> 
> GWC handles playback with 2 timers -- one for feeding audio to the audio 
> device, and one for redrawing
> the cursor.  The audio playing timer just delivers bytes to the audio 
> device at regular intervals to make
> sure the audio buffers never empty.  But the cursor timer must know how 
> many bytes have actually been
> played in order to get the cursor positioned in the window.  So it 
> queries the audio device to determine
> how many audio bytes have been processed.  I think what is happening is 
> the audio device is saying all
> the bytes have been processed, but what it really means is all the bytes 
> are in my buffer, not necessarily
> processed (played).
> 
> I am more aware of this than before, because I recently (last week) 
> upgraded to Fedora Core 3, and have
> the native alsa audio, and the cursor is stopping for me to soon, or not 
> even showing up at all if the sound
> segment is too short (under 1 second).
> 
> In the short term, I'm going to investigate this a little more, 
> hopefully get to the bottom of it and fix it. In the longer term, I'm 
> going to think about using Jack for the audio playback.
> 
> Thanks again for your feedback.
> jw
> 
> 
> John Cirillo wrote:
> 
>> As a result of a recent system upgrade on Debian I now have GWC 
>> 0.20.10 instead of 0.20.3 (I think)
>> One thing that happens with some severity now is that the cursor stops 
>> about 2/3 of the way across the wave during play.  It just doesn't 
>> make it to the end of the waveform.
>> The side-effect is that I'm not convinced that the cursor is really 
>> keeping up with the played sound at any time.
>> Has anyone else noticed this?  Has the redraw of the cursor been 
>> changed between these two versions?
>> I am at a loss as to what to do now.   I like some of the other 
>> features in this release, so I plan to keep using as long as I can 
>> stand it.   I hope it's not a hardware demand issue. I'm using a 3DFX 
>> Voodoo 3 which isn't so fast by today's standards.
>>
>> John
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by:
>> Power Architecture Resource Center: Free content, downloads, discussions,
>> and more. http://solutions.newsforge.com/ibmarch.tmpl
>> _______________________________________________
>> Gwc-general mailing list
>> Gwc...@li...
>> https://lists.sourceforge.net/lists/listinfo/gwc-general
>>
>>
>>
>> !DSPAM:4348c03684927296568094!
> 
> 
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Gwc-general mailing list
> Gwc...@li...
> https://lists.sourceforge.net/lists/listinfo/gwc-general
>