Composers Desktop Project, still relevant ??

[cron]

spdif files from SPEAR.

They are not called spdif.
They are called (.sdif) .

Oh yeah! Good catch.

Xenakios: The only real CPU smasher I've found in CDP is the selfsim spectral process. Everything else (IIRC) will run real time or (usually) much, much faster using one core on my 4.5Ghz i5, with a few on the edge. I think the spectral waver program is an edge case, while in the time domain the pitch sync grains programs (on the rare occasions they don't crash) can also take a while. Selfsim takes another level of forever though...

[Xenakios]

The only real CPU smasher I've found in CDP is the selfsim spectral process.

YES, not shit! I think it's probably that one that is causing the weird benchmark results.

Anyway, making the test cases long enough, I was finally able to get a more sensible looking graph of the real time factor versus output generated results :



So, at about over 9 hours generated audio from a 4 minute test file, the real time factor seems to settle to about 90-100x. (If I omit the selfsim program from the test, I would guess these results will look much different...)

[Xenakios]

Sigh, yet another thing I really have to take care of before a release. For the benefit of both the code in the application and for the end user, I started working on a more fine grained system to deal with the different data types that can flow between the processing nodes. (So that's it's clear wav audio outputs go into wav inputs of the nodes etc.)



At the moment this is just something shows up in the node editing view, there isn't yet any logic to actually constrain the signal connections based on the formats and so on. Hopefully implementing this doesn't turn into a nightmare...

[CinningBao]

I'm not sure any one here is going to fault it as an idea or say it isn't worth doing - it probably would have popped up in a FR list anyway!

Looks amazing by the way, the ease of use will likely introduce CDP to a whole new crowd..

[Xenakios]

Still working on it. Finally decided I have to change how the audio rendering works. It used a working but relatively awfully designed mechanism since the beginning of development. However, since now there are also real time capable processing nodes involved, that old system became an annoyance to deal with...Implementing the new rendering system doesn't look too bad, but it does involve rewriting a lot of stuff.

[Michael L]

Thank you for your attention to that important detail. It will make a 'real' difference in usability!

[klovn]

Xenakios, any update on this?
Hope all is going well

[Xenakios]

Xenakios, any update on this?
Hope all is going well

Development is very very slow at the moment. (Read : not doing it at the moment.) I kind of slipped out of the "zone" working on this. But there's so much done already, I am going to continue at some point.

[ZebraTamer]

Back at it?

[klovn]

Back at it?

+1

[fmr]

Back at it?

+1

+2

[mcbpete]

Back at it?

+1

+2

+λ

[ThomasHelzle]

+ 42

[cron]

Sad to see that development on this one stopped. Even though λ didn't see the light of day, it's been wonderful seeing people get excited about the possibilities of CDP in this thread and elsewhere. Thank you for your efforts.

[Xenakios]

Sorry for no news for such a long time. I am still feeling a bit "off" developing λ...But I have spent so much time and effort on it already, I expect it will be revived eventually.