[Coladam] [colecoprogramming] More about tone channels attenuation (volume) -and- digital sounds (PART 2)
newcoleco at yahoo.fr
Fri Dec 11 20:29:01 CET 2009
I did read about adpcm. As far as I understand it, it's a lossy compression, but not all the time the same quality, and it's very similar to my lossy compression technic which is a bunch of + or - values (I call them DELTAs), and these +/- values are based on a multiplicator given by the data (contrary to fixed values in my own lossy compression). For example, ADPCM will love using +3, -6, -3 and 0 instead of +2, -7, -4, -1, and it does that by saying something like : it's 3 times (+1,-2,-1,0) and these +/- values can be somehow encoded as one single byte, 2-bit each.
Because there is a multiplicator involved into decoding the adpcm-like data, this is not something we like when dealing with a Z80 cpu. A posibility could be precalculated tables in the decoder, like what I'm currently doing with a single one instead of many... or given table(s) within the sound data to play.
Have a nice day!
More information about the Coladam