I’ve decided to rebuild the project next year (2020) since Max evolves and I can’t find appropriate ARTNET objects for this environment.
After studying what choice I have, OF seems a possible good option, however I should not go further without opinions from the OF community.
Circularium uses near than 6000 leds and 32 artnet universes. The sound is 8 audio-tracks spatialized. And it uses a hand proximity detection built over Arduino ton control the light.
My question is : it is consistent to rebuilt such an artwork in OF, keeping the idea that it would run on RPI ? I think about cpu calculation and optimization, light & sound realtime control. Sevral RPI? Other?
Also I’m not a pro computer scientist etc. but more a maker/hacker/artist using different creative coding tech such as Max, Processing, Arduino and else to achieve my projects. I already had some look at OF a few years ago and it seems approchable to me, probably with some help, of course.
2nd question : If I go in OF I’ll have to investigate in sound processing, artnet data and so on. Is it consistent or does it require years of low level coding practice?
Any advice is REALLY welcome to avoid going into a dead end.
Thanks for reading.
If you have very specific sound requirements, consider using ofxPD and running Pure Data/libPD from within oF. I’ve used it before and it’s excellent. If you have Max experience, pure data shouldn’t be much of a problem.
To add up to @ayruos, I believe you will be fine if you are doing it with 2 Pi, one for the sound and one for the visuals, although it seems like it would be possible with just one probably.
Then make them talk using the GPIO ports, using a serial port, it’s pretty straight forward.
So 1 Pi 4 for all the CPU processing of the leds, then the other one could be for the sound, using [this] kind of sound card. I haven’t tested this one, but i have seen it pass on the forum before.
@ icq4ever Thanks for your comment and pointing me to Art-Net 3
I though it was possible to compile OF and Contributing libs (ofx addon if I understand?) on any OS.
So this is not the case ?
Okay How do you know what addon can run on what system ?
@ pierre_tardif00
Yes I had this idea : running two RPI. So there should be some resource about how to sync them via GPIO ports, this is the idea? > using [this] kind of sound card.
this ? link missing
Hi,
for dealing with audio ofxPD is a very good option if you feel comfortable with pd.
If you want to make it fully contained I would suggest you to use ofxSoundObjects and/or ofxPDSP. the latter is way much more complex, which carries a complete suite of digital sound processing and synthesis stuff, it is really good, the former, which I wrote, is intended to be an easily extendable audio processing platform. Both will run on the PI.
As for the leds, I’ve done a lot of work on this area. While lately I’ve been using some dedicated controllers which receive artnet, I’d suggest you to take a look at the Teensy boards and this addon board for it. These are really awesome. I did a project where a single teensy3.6 was driving like 6000 rgb LEDs while doing sound processing. No computer, just the teensy. Quite awesome boards.
As for the ofxArtnet addon, I think there are a few different versions all called the same, so you’d better double check, but it should run on the PI,
@ roymacdonald
Thanks for these detailed information.
I’m interested by your lib since the idea is only to play DAW track as samples. I’ll have a look, thx for the link.
I also have to find a way to spatialize those samples/tracks throw out of 8 speakers.
About leds there is now an alternative to PJRC octo stuff, working on ESP32 : Yves Bazin from the FastLed community made a code allowing to run 80 lines of ledstrips on a simple ESP32, Yes it’s incredible but it works nice https://github.com/hpwit, there are more information on the FL reddit forum.
For Circularium I’m using 2 Artnet boxes called BC-216 that spread data to 32 WS28xx strips, now I looking for a way to connect those boxes to the RPI using ethernet cables and a good artnet library. Hiroyuki’s ofxArnet seems to do the job : https://github.com/hiroyuki/ofxArtnet/issues/26
As of spatializing tracks, if this means simply to play the track on different speakers then you can make that using my addon and its multiplexer for the output. there is an example that does this. if it means to process using some kind of algorithm that will spatialize the sound, then, I have not implemented such on my addon.
to be clear about spat : yes I only need to direct the sound level moving through the 8 speakers, no ambisonics effects and things … Can ofxSoundObjects do that?
