[sdiy] Wavetable Design Update

[Dave Manley]

On 2/5/2011 10:18 PM, Matthew Smith wrote:
> Quoth Dave Manley at 06/02/11 16:25...
> ...
>> So what about running the wavetables with reduced depth? How
>> bad would it sound and what's the easy way to generate some
>> audio files to try it?
>
> I only went to 256 bit because it made the CPLD design easier! (Didn't need to figure out the logic to reset a counter before it overflows and goes back to zero.)
>
> Not sure what 128-bit would sound like at 'serious' bass frequencies, but then (I think I've said this before) I'm not trying to make a lab-grade signal generator so am not too fussed about waveform
> purity - so long as it sounds OK. (And if it doesn't? That's what filters are for!)
>
> Speaking of bass, I've tweaked my CPLD design both to fit it onto a smaller/cheaper device and also to add selectable sub-octaves. Just got to see if I've got sufficient resources left to make that a
> pair of sub-octaves (I think I have.) So CPLD is driving three EEPROM/DACs per note. The detunable second voice for the note will use a separate microcontroller, CPLD and EEPROM/DACs. Total of six
> signals per note.
>

You would only reduce the table depth at higher midi notes, and use
the full 256 entries at low frequencies.  The counter change is simple -
instead of adding 1 to the counter you add 2 for 128 deep, or 4 for
64 deep.

-Dave