Re: Re: looking for some info on Magic Voice

From: Grósz Attila <>
Date: Sun, 23 May 2010 15:55:34 +0200 (CEST)
Message-ID: <>
Oh, those dumps look familiar :) They were done by me and sent over to Stefan - who at that time appeared to be the sole person interested in the MV stuff - while implementing the Magic Voice emulation in YAPE.

There is no real magic going on, the data is sent uncompressed to the 8706 (the decompression is done in software by the ROM) but with a variable bit length. Stefan should have the bitlengths posted somewhere on his site. All in all, that was the biggest challenge to figure out, I remember asking myself the very same things back in the day like you do now :)

From there, the actual synthesis is relatively easy, since the S&S patent has all the right formulae to be used for the filter.

I can forward a brief summary I mailed to one of the MESS developers a few months ago, once I am back from holiday.


   <>, ezt írta: 
  On Sonntag 23 Mai 2010, you wrote:
> The pinout was posted on this very list many moons ago by Richard Atkinson:
> Be aware that the 8706 was only the glue logic for the V364, the C64 one is
>  more complex (though only with familiar items) with two TIA 6522's (among
>  others).

BTW, since you have already implemented it... you probably know this.... i 
have implemented the necessary i/o stuff so the software runs (talking about 
v364 here, i am using that for a start since it is a bit less complex, as you 
said) and i can "see" the commands and the data sent to the t6721 chip. the 
data written to the 8706 looks like this: - 
however this obviously isnt the actual parcor data. stefan uhlmann vaguely 
describes how to "unpack" the data, and shows the actual parcor data as a 
result: ... however, 
there seems to be some step missing between the data written to the 8706 
(which i assume would be the "unpacked" data) and the final parcor data. two 
questions arise at this point, which you might be able to give a hint on :)

- where exactly does the "magic" happen that will transform the data written 
to the 8706 to the parcor data shown on that webpage? is it the 8706 doing 
some bit shuffling magic? or is it happening inside the t6721 ?

- what exactly IS that magic? the mentioned webpage mentions things like 
swapped nibbles and reversed bit order, but so far i couldnt make any sense 
out of it, and was unable to produce the final parcor data shown there from 
the data written to the 8706.

can you shed some light on these? :)


My aim is to produce games that have graphics and audio that will rape your 
eyeballs and take your ears from behind with a strap-on 
<Jeff Minter>

       Message was sent through the cbm-hackers mailing list

30-50% kedvezménnyel kínáljuk az Osiris Kiadó 250 féle könyvét, amelyek között tudományos és felsőoktatási könyvek, szépirodalmi és ismeretterjesztő kiadványok találhatók. Az akció június 30-ig, illetve a készlet erejéig tart. Válasszon a Bookline ajánlatából!
Message was sent through the cbm-hackers mailing list
Received on 2010-05-23 13:55:34

Archive generated by hypermail 2.2.0.