[linux-audio-user] Some music made with Linux

[Pete Bessman]

At Sun, 22 Feb 2004 17:15:11 +0900,
Patrick Shirkey wrote:
> 
> What is their style?

The kind which implements a solution with an order of magnitude more
complexity than mine.

> Seems to me that you have definitely implemented a nicely coded 
> interface that could relatively easily be integrated with linux sampler. 
> Don't sell yourself short.

That's not so simple.  It could be done, but the only code I would be
reusing is the grunt gtk stuff, which a trained monkey could churn
out.  My gui code is conceptually separate from the patch management
code, and it depends upon the patch infrastructure having a particular
interface in order to work.

The gui code is also the least complex part of the program,
logic-wise.  Given that, I think it would be much easier for somebody
with knowledge of LinuxSampler internals to retrofit my code than it
would be for me.  In fact, I might be a rather poor choice because
I've come to expect a certain patch infrastructure and I'd have to
unlearn my old habits.

> Currently there is no actual gui although and as Mark said Rui is 
> looking into the qt version.

Ugh.  I don't even like GTK, but I dislike it less than I dislike QT.

> Assuming you have some spare time it would not take you very long to add 
> the calls for each button and place that shell up for one of the main 
> developers to make some changes to the backend. Then you will be 
> transformed from "man working alone on small project which accomplishes 
> some new things" to "part of professional team creating new project that 
> amazes a whole sector of the professional audio world".

I've been working on Specimen for the past 12 hours (forgot to sleep),
cleaning it up monstrously (which excised about 500 lines of code, I
think), and adding about 1000 lines of code worth of improvements.  I'm
about halfway done with my Things I Must Complete list.  When I'm not
directly working on Specimen, I'm indirectly working on it by trying
to teach myself DSP (hard!) and all the associated math and physics.

I really simply do not have the time to help LinuxSampler unless I
divert time from Specimen, which I'm not keen on doing.  And I don't
have the foggiest idea what they're doing anyway.  Maybe I'm a retard,
but I don't see why direct-from-disk streaming takes months of
planning.  Intelligent use of rudimentary thread pools and
disk-caching should get the job done.  I'll probably regret saying
this later, but I think the whole deal could be implemented in 1000
lines of code.  It's on my TODO list, but I think the sound sculpting
and usability aspects are more important than that given todays RAM
quantities.  50 megs will get you 5 minutes of audio at 16/44.1k.
That's a _hell_ of a lot to work with.

Like I said, I'm a complete neophyte so I don't know what I am talking
about by any objective evaluation, but I can't for the life of me
figure out why they have twice as much code for a program that seems
to do half as much.  Couple that with the fact that it's written in a
language I don't like, planned for a widget set I don't like, and
moving too slowly for my infinitesimal attention span, I can't see the
incentive for contributing.

What it boils down to is, LinuxSampler is supposed to be awesome
within a year.  Well, I'm interested to know what Specimen will be
like within a year.  Hell, within 6 months.  And I'm gonna find out.

[pb]