Blackpill: Why 8 Bit?

8 bit sounds good. Not just in an arty way. Your ears can easily pick out the audio, it just comes with faint background noise. Providing you filter the output a bunch and your input signal is loud enough, this background noise/distortion is only really noticeable as the audio fades out.

Counter-intuitively, 8 bit sounds analog. I like to slap a hefty low pass filter on an 8 bit signal to smooth over the janky corners, giving it a similar warmth to BBD delays (sample rate reduction only adds to the illusion). Largely it's the treatment of your signal either side of the code that defines its character.

You can store twice as many 8 bit samples as 16 bit ones, letting you record twice as much audio. Computers being binary creatures, memory is arranged in factors of two... 2, 4, 8, 16... Every programming language and compiler I know only allows variables in 8, 16 or 32 bit formats, so your choices are few.

That said, with a bit of finagling it's technically possible to store (for example) three 10 bit samples across one 32 bit variable for a memory/resolution compromise.

Why? Because then I can use the term "lofi", excusing all other sonic deficiencies in the process.

I sort of joke but the angular edges of 8 bit audio provide the perfect cover for maniacal cutting and shunting. For the many FX I make where sections of audio are cut up, sped up, reversed and switcherood, those choppy transitions would sound drastically worse at 16 bit, as your ears get accustomed to smooth hifi textures before confronting the violent segue.

Finally, I'm falling in love with the STM32 Blackpill board. It's fast, cheap and has enough internal RAM for long delays. For the life of me I can't get decent, low noise 12 bit readings from the ADC. So perhaps the real world limitations came first and all this justification is in retrospect, but that doesn't mean it's wrong.

Listen to the noisy 12 bit ADC in this vid, honestly (Bluepill is comparable to Blackpill) ...