hello,

David Baelde a écrit :
> Hi,
> 
> I'm glad that you could work around the glitches and that you're
> enjoying liquidsoap -- but, about your bug report, I confirm that you
> put your finger on dirty things and I'll fix that soon.
> 
> On Wed, Dec 3, 2008 at 1:40 AM, okay_awright <[EMAIL PROTECTED]> wrote:
>> -the input.harbor on_connect and on_disconnect callbacks are really
>> great. However, it could also be useful to have this kind of mechanism
>> when this function does not produce any stream anymore, without
>> disconnecting first, and then resumes streaming.
>> e.g. a track played through the Winamp DSP plugin or whatever is now
>> over: no sound is broadcast but the source is not disconnected: it could
>> be handy to notify any third-party software when it happens, in order to
>> take any appropriate measure.
>> Maybe there are other ways to achieve this but I wasn't able to find out
>> how.
> 
> If I understand correctly, you want to know when the harbor client is
> streaming silence. You can't get anything better: with an icy/shout
> stream, we can't distinguish silence inside and outside a track. A
> solution is to use the on_blank() operator that will call a given
> function when a certain amount of blank is detected on a source.
> Another solution is to use strip_blank() that will make a source
> unavailable when it is streaming too much silence; that would be
> combined with a switch above, transitions, etc.

yep, exactly. I knew about the on_blank feature but AFAIK there is no
way to get a callback when the stream is up again.
>From the logs, it appears liquidsoap was aware of the transient 'no
stream available' situation:
2008/11/30 15:22:55 [src_5502:2] Feeding stopped: Mad.End_of_stream
2008/11/30 15:22:55 [threads:3] thread "harbor source feeding" exited
but it didn't process this signal as a proper disconnection.

> 
> In case I haven't said enough, there is a summary of blank detection
> operators there:
>   http://savonet.sourceforge.net/doc-svn/blank.html
> 
> Concerning LADSPA I don't have time to check again right now, but the
> CPU consumption did not seem absurd to me. In any case, I'm afraid we
> can't do much better without changing the whole architecture --
> especially since you already asked gcc to optimize. You talk about
> similar apps doing 10-20% CPU: are they using LADSPA as well? That
> would be a concern. Do you know a better lib?

I do understand, I knew LADSPA plugins are usually not designed to be
used for real-time processing (too few, too little choice). The test
system was a P4 2.6 w\ 2Gb of RAM (the liquidsoap server is a P4 2.8 w\
2Gb of RAM) but running Windows XP and either DirectX or VST plugins
(there are tons of such plugins as you already know).
I've first benchmarked a chain of 3 generic VST components from
Steinberg (expander->compressor->limiter for three distinct bands) and
then the DirectX version of Izotope Ozone (really impressive BTW). And
none stressed the CPU above a mere load of 25% (15-20% was the norm).

Sorry, I'm rather new to the audio realm of Linux/Unix and LADSPA is the
only unified library that comes to my mind.

> 
> I'm not the expert about the internal sound processing operators. I'm
> sure they can be done better, what we have now is only the first
> stable implementation -- although it's already far from naive as far
> as I can see. You can put a ticket for our idle days, somebody may
> pick it up.
> 
> Cheers,

thanks for your answer

-- 
best regards,
sincèrement,

okay_awright
<okay_awrightATddcrDOTbiz>
Public PGP key on request

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Savonet-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-users

Reply via email to