Audio support - Githubissues

thejpster / monotron

A simple 8-bit home computer style application for the TI Tiva-C Launchpad

Apache License 2.0

191 stars 9 forks source link

Audio support #8

Closed thejpster closed 5 years ago

thejpster commented 6 years ago

We have some timers spare, and I'm sure we can generate some audio.

Once a scan line (so 37.680 kHz) set the PWM to output a level between 0 and 255. Add some filtering and buffering on the output pin and voila - 37.680 kHz 8-bit PCM mono audio. But as we have no room for samples, we probably need an API that lets you play beeps (square waves? maybe even triangle waves using some very basic maths?) at a given frequency for a given number of frames.

We probably want to set the audio output at the end of the scan-line to avoid delaying the video. Maybe calculate at the end of the scan-line and load at the start of the next scan-line.

thejpster commented 6 years ago

https://sensorium.github.io/Mozzi/ is a synthesizer for Arduino which may work as a reference. Has some output schematics too.

thejpster commented 6 years ago

See https://GitHub.com/thejpster/monotron-synth. Needs connecting to a PWM and running from the video interrupt. Not sure how to make it safe to generate samples under interrupt and change parameters in the main thread.

thejpster commented 6 years ago

Maybe the synth should be optional, incase they want to provide their own synth and just supply pcm samples?

thejpster commented 6 years ago

So, the LM4F120 has Timers, but they can't do up/down symmetrical PWM. The TM4C123 has Timers, plus two PWM peripherals which can do up/down.

So... I think we're going to have to switch to the TM4C123 as the official platform.

thejpster commented 6 years ago

Also, www.rakits.co.uk gave me a neat little amplifier, so I need to get that built up.

thejpster commented 5 years ago

Audio support is merged, and I tried it with the Rakits amplifier. Well, sound comes out, but it's very noisy. Interestingly, if I unplug the VGA cable, the noise drops right down.

Perhaps the impedance mismatch on the VGA lines is causing this?
Perhaps they just radiate more noise if they're conducting significant chunks of current? I think we're at about 8.1mA per RGB channel.

If I get the chance, I'll unpack my oscilloscope and see what frequency the noise is, which should give some clue to its source. Otherwise I'll just ignore it for now, and think about the ABI for driving the synth.

thejpster commented 5 years ago

ABI for driving the synth is in. Snake can go 'bip bip bip boooop'.