Hoping someone can expand on the audio I/O. I did see the Audio Adapter page, I am guessing the small audio connection may be what I am after, it just isn’t clear what it is pushing out.
Ultimately, I am going to add a mic array to my project. A lot of the audio processing DSP chips (wake words, removing background noise, removing source audio, etc) seem to take in I2S.
Ideally, I would like 1) to capture the stereo out of what the device in desktop mode is generating. And 2) feed in the cleaned up audio from the audio IC from the mic array.
I think I read that not all of this is supported yet but I am curious what the future state plan is.
Ultimately I do need a microphone array for DOA calculation, beam forming, etc for proper wake word support. I do have a small baseboard that will connect via the HAT. I was hoping that the Pi Header would have similar I2S support.
However, it seems the datasheet may be a bit out of date? developer.particle.io shows the Tachyon does have what looks to be multi-channel I2S? Is this implemented currently?
Any guidance or thoughts here would be appreciated. I am looking at the ZL38063 to processes the raw audio, do echo cancelation, etc if that helps, but again open to any thoughts. I was thinking of using the Qualcomm DSP on the tachyon itself, they do have the Voice Assist stack, but I do not know how many hoops I would need to go through to get access to it, and I would also still need extra ICs for the mic array / connect to the MI2S and also the same to ultimately take the final sound and amplify it. ZL38063 + a simple amp seems to both be cheaper / easier as I get an out of the box tested solution.
That being said, this is not an area I know a lot about, so opinions welcome
Hey @RyanLM, I2S support on the 40 pin header is coming on Ubuntu 24.04, but I don’t have a solid timeline for you just yet. There are a few other hot items that are taking priority. I’ll keep you posted.
