HI Mark, Ok, some news from the front: I've got what I want going on a pi with alsa's snd-aloop module (see https://www.alsa-project.org/main/index.php/Matrix:Module-aloop). This sets up an input device which loops back to an output device. So I used chuck --probe to get the corresponding dac number for the input, sent chuck into that by specifying --dac, and then set up darkice to listen to the output device.

From there the data gets sent to icecast on the same pi. The whole system seems to run stably and generate audio constantly (I've got a shell script restarting the chuck process after it finishes (c. half an hour for each go).

Since I now have an rpi with a bunch of cobbled together configuration I'd like to formalise it and make it repeatable, so I'll be looking at your and mheuer's docker work :) Thanks, Stuart

---------- Forwarded message ---------- From: Mark Cerqueira To: ChucK Users Mailing List Cc: Date: Mon, 16 Jan 2017 20:01:24 -0800 Subject: Re: [chuck-users] ChucK => internet audio stream? <snip> Stuart - I think you could accomplish your idea with some work over my Docker image. AFAIK, the --silent flag does not mean no sound is produced; it just means no audio is outputted during the running of the program. You could set something up like:

1. Let ChucK produce 30 seconds of audio (with --silent it should take less than 30 seconds to produce 30 seconds of audio) and save it to an audio file. Look at export.ck in the chuck-renderer repo for how to do this. 2. Streaming clients grab this 30 seconds of audio and start playing it. 3. While the above is occurring, you produce another 30 seconds of audio. 4. Send the next audio file to clients and they queue it up after the first segment ends. 5. Rinse, repeat.

This has the added benefit of letting ChucK shut down between producing segments to do garbage collection. :) You might need some annealing to get the separate audio to sound "together" but this should get you a "Hello, World ChucK stream."

Good luck!

mc