Re: [Traverso-devel] Traverso audio backend & custom audio app

[Remon]

Hello Niklas,


> I am currently looking for sequencer engines that are lightweight and
> simple to use since I am about to write an own application geared
> towards layer-based sample editing. 

I'm not familiar with sequencer engines, and neither with layer-based sample 
editing, would be great if you could elaborate a little on it :)

> so, I am asking if you
> guys can advise using your audio-backend for other
> projects as well or more importantly: is your audio-backend currently
> stable and useable enough to be used in external applications? 

It's as stable as I can get it currently!
If it would be 'the' solution for you application and your needs is a 
different question.

libtraversoaudiobackend is basically a wrapper for the alsa driver, and less 
evolved, the jackd engine.
It's a callback based framework, and thus seperates the logic between audio 
backend and application a 100%
It's design is lockless, (un)registering a 'client' to it (which holds the 
callback function) is garuanteed not to block the audio stream.
For this, I created a real time event dispatch system, which depends on Qt 4's 
meta object system. 
There's only support for audio, no midi support (yet).

That's largely that's about to say about the audio lib of Traverso. Great 
thing about is, that when another audio backend emerges in linux, it's very 
very simple to adopt it in Traverso, either being it a callback based one or 
not!

> or is 
> there anything else you can adcovate? I can explain the details and
> benefits of layer-based sample editing in sequencing applications
> lateron,  but I am currently short on time. Any help&suggestions are
> greatly appreciated!

Well, if you 'need' for example outside processing in your application, jackd 
would be a great choice.
I didn't like to depend on just jackd, but wanted support for alsa as well. 
The reason is that jackd is for professional use 'only', it's rather hard to 
get used to it.
Launching an application, and having all the audio hardware (from simple to 
pro audio cards) setup directly, and hitting the space bar makes you 
recording multichannel seemed very important to me.

If you don't need low level low latency operation, and target for a KDE 4 
application, you might have a look at phonon for example.

Hmm, there has been some talk on a 'linux audio api', but it might take a 
while before it's there, if it's ever gonna happen lol

Another audio framework is portaudio, it seems it's cvs version is becoming 
usefull, but I never see it mentioned or used, so I'm not sure how useable it 
is.

So it depends rather a lot what you need. Traverso has more stuff that might 
be interesting to use. (what about integrating layer based sample editing 
into traverso lol)

Greetings,

Remon

P.S.
Although a bit rough and less stable, there is a cvs based ready to run binary 
posted on the download page.
It could give you an idea of the power of it's interface...