ChucK 1.4.0.0 mixes up --in and --out args

[Acoustic Interloper]

I posted this in the ChucK mailing list, thought I should put it here in case anyone tries to use ChucK 1.4.0.0 to steer spatial sound:

Just upgraded my osx upgraded to 10.12.6 from 10.9.x, just installed chuck-1.4.0.0, am writing a new chuck patch. Lots of variables. Two days of debugging later:

chuck-1.4.0.0 is not finding dac.chan(N)s correctly with chuck-1.4.0.0, so I pulled a chuck-1.3.2.0 off of an old mac, and my patch works fine.

This is in a Live -> Soundflower -> ChucK -> MOTU driver signal path, which I took apart a piece at a time. When it looked like ChucK, I turned on verbose mode. When I run this command line with MOTU as my output:

/usr/local/bin/chuck.1.4.0.0 --verbose:10 --out:8 # Note: internal mic is the input, MOTU is output, str8.ck is this:

adc.left => dac.chan(2) ;
adc.right => dac.chan(3) ;
90000::ms => now ;

Nothing goes to the MOTU, and the diagnostics show this:

[chuck]2:SYSTEM): initializing audio I/O...
[chuck]2:SYSTEM): | probing 'real-time' audio subsystem...
[chuck]7:FINE!!): | initializing RtAudio...
[chuck]7:FINE!!): | | trying 8 input 2 output...
[chuck]3:SEVERE): | allocating buffers for 256 x 8 samples...
[chuck]2:SYSTEM): | real-time audio: YES
[chuck]2:SYSTEM): | mode: CALLBACK
[chuck]2:SYSTEM): | sample rate: 44100
[chuck]2:SYSTEM): | buffer size: 256
[chuck]2:SYSTEM): | num buffers: 8
[chuck]2:SYSTEM): | adc: 0 dac: 0
[chuck]2:SYSTEM): | adaptive block processing: 0
[chuck]2:SYSTEM): | channels in: 2 out: 8

Notice the "trying 8 input 2 output..." is wrong. There are symptoms that agree with this. If I increase --in above 2 and just do straight stereo in => out, it starts outputting to chan(0), chan(1), up through chan(N), and N is the --in command line argument.

This older version works fine, sends the mic input pair to the correct dac.chans:

$ ./chuck --version

chuck version: 1.3.2.0 (chimera)
mac os x : intel : 64-bit
http://chuck.cs.princeton.edu/
http://chuck.stanford.edu/

./chuck --verbose:10 --out:8 str8.ck
[chuck]2:SYSTEM): initializing virtual machine...
[chuck]2:SYSTEM): | probing 'real-time' audio subsystem...
[chuck]7:FINE!!): | initializing RtAudio...
[chuck]7:FINE!!): | | trying 2 input 8 output...
[chuck]3:SEVERE): | allocating buffers for 256 x 8 samples...
[chuck]2:SYSTEM): | locking down special objects...
[chuck]2:SYSTEM): | allocating shreduler...
[chuck]10:ALL!!): | adding 'Chuck_Object' (0x100712ff0)...
[chuck]10:ALL!!): | adding 'Chuck_Object' (0x1007130b8)...
[chuck]2:SYSTEM): | allocating messaging buffers...
[chuck]2:SYSTEM): | behavior: HALT
[chuck]2:SYSTEM): | real-time audio: YES
[chuck]2:SYSTEM): | mode: CALLBACK
[chuck]2:SYSTEM): | sample rate: 44100
[chuck]2:SYSTEM): | buffer size: 256
[chuck]2:SYSTEM): | num buffers: 8
[chuck]2:SYSTEM): | adc: 0 dac: 6
[chuck]2:SYSTEM): | adaptive block processing: 0
[chuck]2:SYSTEM): | channels in: 2 out: 8


"trying 2 input 8 output..." is correct for 1.3.2.0, but not for 1.4.0.0. Chans are buggy, at least on my mac.

[Armand]

Hello, any update on that? I too have lost a few days on it.
Issue is still there on both OSX 10.14.6 and Windows 10. I get an output on some channels but the wrong ones, in some kind of output modulo input way. There is basically no way of using command-line ChucK with a number of in channels diffferent from out, which is really a shame.
MiniAudicle doesn't have this problem, so I would suppose it's not that hard to fix.

Btw, if anyone knows a good way to declare a bug, I'm listening. I posted it on GitHub before finding this post. It's a really crippling issue. I'm a practicing artist in NYC and was hoping to project some code controlling my modular, which I clearly won't do with miniAudicle. It could introduce chuck to a lot of people. But most interfaces in modular have a different number of ins and outs, and I'm not voiding half my out channels...

[Acoustic Interloper]

I never heard back from them and, yes, it's a real PITA. I have been working more with visual digital art when I get time for that, and using Live for music. Used ChucK to generate some audio data for a data analysis course in spring 2018; it didn't require the command line args. I hope to actually get good at using Supercollider when I someday get time, probably when I retire. SC has higher quality ugens (IMHO) and an active developer community. I like ChucK's programming model better, but that goes only so far.

[Armand]

Hey, little update. I corrected the bug, the fix is in the master branch now.

True, SC is probably best for sound design. I like ChucK for its, like you say, paradigm and precision. If it works to begin with, I know it will do what I tell it, when I tell it when playing live. I might be wrong, but I figured I'd rather invest my time in that and let the CPUs get a few doublings so that sample-synchronous becomes really important.

[Acoustic Interloper]