nice!!! looks like a lot of work! but cool. I have worked out a way of directly wiring audio into the exchange, what im going to do is record some through the exchange a direct and send to see if there isn’t any surprises??? as you say in the video with clipping etc, its not the neatest of signals through the exchange but simulating a direct recording might be a good idea. ill try to this today even
If you can grab a sample after it’s been through the exchange (as a .wav file), then I can run that through my Raspberry PI and see if it can decode it.
I will also,at some point, upload a sample of audio that the Raspberry PI has created, so you can play it back to the minicom for me
So I’ve done the GPIO stuff that @lookmumnocomputer requested. I think I’ve gone a bit overkill, but oh-well…
As usual, one of my crappy videos to demo it → http://www.marcnetsystem.co.uk/baudot_part3.mkv
I’ve basically used 15 output pins to show the different status values:
- Listening (capturing audio)
- Finding data in non-silence audio
- Data RX,
- Rx 0
- Rx 1
- End of audio block detected (pulses)
- Audio is too quiet, or is silent
- Clipping detected
- Transmitting audio
- Transmitting data (not padding, silence, etc)
- Tx 0
- Tx 1
- MegaHAL thinking
- MegaHAL saving
- MegaHAL running out of memory, so removing some stuff it’s learnt
There’s also a couple of input pins:
- Go to sleep (when set Low) / Wake up (when set high)
- Enable (high) / Disable (low) megaHAL learning.
Next step is to sort of the Raspberry PI booting and running the program automatically - then it’s done!
Just a short update. I’ve uploaded the videos here:
Part 1: Part 1: Decoding the audio on Vimeo
Part 2: Part 2: MegaHAL & Playing back audio on Vimeo
Part 3: Part 3 : Input / Output (flashing lights!) on Vimeo
I’ve also set up the Raspberry PI to load the program on boot. (rc.local for the win)