Composers Desktop Project Forum

At the terminal, enter "export CDP_SOUND_EXT=wav", then try your command again.  If it works, just add this export line to your .bash_profile (sans quotes) and this shouldn't crop up again.

Tricky, indeed!  I hadn't looked at Lua before, and from what I gather, luaposix (namely, setenv in the luaposive.stdlib module) would appear to be one way to avoid sending actual commands to the terminal.  But this still leaves out Windows . . .

I assume instructions for setting the environment variable are already included with the CDP tool.  Maybe the most "portable" way of dealing with the problem would be just to check that environment variable has been set with os.getenv("CDP_SOUND_EXT") and simply ask the user to set it before proceeding?

This is wonderful news!  Thank you so much!!

I appreciate your exacting reply.  It is very clear!  So while it is not possible to have the WORKSPACE files occupy a different folder than the location of the soundloom.exe executable, either workaround is definitely simple enough.  Thank you!

Thank you for the clarification/correction!

Thank you very much, Richard!

Both Sound Loom and Soundshaper have features that automatically insert the duration of the "input" sound file into the breakpoint file you are editing, with SS having additional features for scaling breakpoint files for input files longer or shorter in duration than the file the breakpoint file was written for.  But the actual process only understands literal time values.  To my knowledge, no such "shortcut" exists.

I recently acquired a fancy new desktop with a high-speed SSD "scratch" disk, and it'd be nice for files in Sound Loom's WORKSPACE to live on this drive instead of where the Sound Loom executable happens to be located (e.g., C:\Users\<username>\CDP7\_cdp\).

Is there a way to change SL's WORKSPACE directory?  (In Soundshaper, at least, this is easily accomplished by changing the "Temp" directory in the Environment settings.)

Excellent!  Thank you for posting your solution.  Others will undoubtedly find this useful.

It strikes me odd that the Renoise CDP interface does not set this environment variable automatically, seeing that most users of the Renoise interface are probably otherwise unacquainted with using CDP, and therefore should not be expected to know to do this.

I've never used the CDP Renoise tool, but taking a shot in the dark, it sounds almost like you might be having issues running the spectral processes, which expect and produce .ana extension (Analysis) files rather than normal sound files.

Does the Renoise tool use PVOC to automatically convert your sound file (.wav or .aif) into an analysis (.ana) file?

You have the basic idea.  Here's a NOTE DATA example pulled from the CDP manual, which can be used in a process like PROCESS > TEXTURE > of motifs > over harmonic field.

59 60
#9
0.0 1 60 0 0
0.0 1 61 0 0
0.0 1 63 0 0
0.0 1 64 0 0
0.0 1 66 0 0
0.0 1 67 0 0
0.0 1 68 0 0
0.0 1 70 0 0
0.0 1 72 0 0
#10
0.0 1 60 70 0.3
0.1 1 61 50 0.3
0.2 1 66 50 0.3
0.3 1 67 50 0.3
0.4 1 66 50 0.3
0.5 1 67 70 0.3
0.6 1 70 50 0.3
0.7 1 66 50 0.3
0.8 1 72 80 0.3
0.9 1 66 60 0.3
#3
0.0  1 60 40 1.5
0.34 1 67 45 1.5
0.67 1 72 50 1.5


#N specifies the number of event lines in the motif.  So in this example, there are three motifs, with 9, 10, and 3 events, respectively.  Note that time values may be all zero, in this case; although as a rule, time values must be in ascending order if specified or required by the particular mode.

I don't know how SC's sound algorithms compare to other programs.  (Maybe someone with firsthand knowledge of this can chime in?)  But a very large part of it, however, is the user.  Curtis Roads not only knows the nuts and bolts of DSP, but has cultivated a personal aesthetic in which to apply that knowledge.  In his hands, arguably any tool could be made to sound good.  In tinkering with different languages and systems over the years, however, I'm left with the impression that Csound and SC were (are?) easily top of their lot in quality and expressiveness.

As to whether CDP can rival SC in sounding "good", I feel strongly that the answer is Yes!  In transforming sounds, however, the nature of the input has a lot to do with what you get out--CDP won't magically sound great on its own.  Above all, it needs the user to apply it sensibly to the particular input sound, or at least recognize a sensible accident when it occurs!  ;-)  Even processes that, by design, "butcher" the input sound may be exactly what you need to achieve a desired effect.  CDP is as good as they come, imho.  For proof, listen to some of the transformations Trevor Wishart uses in his recent work.

Some of the core processes could be "ported" to SC either as a C++ plugin or by reworking the sound algorithms in sclang.  SC is basically a "realtime" synthesis language, however, so some of the CDP processes would not adapt well into SC (or any realtime system, inc. VST plugins) without taking this into account.

SC can be impressively good sounding.  But it's quite a different animal than CDP, which deals primarily with performing transformations on existing sounds rather than rendering new ones "on the fly."  This distinction means they are drastically different in their approach, even if they both deal with rendering sound data.  They are really complementary tools, if you are so inclined.

As an aside, Richard Dobson apparently ported some CDP sound algorithms to Csound, although I'm not informed enough about Csound to be able to tell you which ones (spectral processing?).

A minor update to Sound Loom, 17.0.0C, was posted to Trevor's website for both PC and Mac.

I'm assuming your two mono files are representative of your left and right channels, respectively?  If so, there are two CDP processes that'll do the deed: SUBMIX INTERLEAVE and INTERLX.  Within Sound Loom, I believe it is the latter that is invoked using PROCESS > MULTICHAN > Interleave Channels (WAVEX Formats).