Posted By
Charlemagne
on 2018-05-26
14:42:58
|  Re: New feature of LUMYDCTT: (Music) ZackZack
V2.8.2:
- Killing some bugs.
- You can Play/Pause and Stop the playing!
gerliczer sent me a private message in Hungarian and I hope he won't be so angry with me, but I try to translate his message to English, because it's very important:
"Well the TED is not so accurate in controlling pitches. It can be seen perfectly in 'Ultimate Map' by SVS: 'Deviation between obtained and aimed frequency' column on TED sheet. And this deviation would be more inaccurate if we didn't use the whole 10 bits long register value. Telling the truth, it is only noticed by who has got an extremely good ear for music, but 200 bytes is not a huge quantity of sacrifice on the altar of exactitude. What's more coders usually don't save values of the 10 bit registers, but only the 8 bit increments for initializing the sound table. After all, the space of generating code and datas will be used as workplace or temporary data storing so this area can be 'recovered'."
Thanks Stinaris, as you can see what gerliczer said... ...OK I understand what you want to express, but please have a closer look at "eremgyujto" sheet of music in the picture above and tell me how you store the values IF somebody like using not only the normalized values of sound notes(?). "Eremgyujto" (Mac Coin) uses two sound channel side by side and the difference is only 4. OK I understand that it's enough to store only the first channel and adding 4 to it when playing.
But it seems a little bit AI... I am going to think about how to optimize (or compress) DATA and DB lines after converting music score.
Thank you both for the new ideas!
|
