From: BlueChip <cs_bluechip_at_webtribe.net>
Date: Wed, 03 Sep 2003 01:18:14 +0100

>>WAV ...how simple does it get?
>Yes, it's simple, but the serial communication with the MAS is too slow
>for 44kHz 16-bit playback.

That is a shame ...but only a limitation, not a show-stopper.
Do you know what the datarate limit is?

It will be nice to use the digital I/O as a serial port to allow the
control of external devices - LCD -> PAL/NTSC display converter anyone?

>>MOD ...What is so difficult? - I used to play MODs on my old 286!
>The MOD format has the same problems as the WAV playback. But sure, it
>might be possible to mix the MOD channel data with the SH1 and then send
>the WAV data to the MAS.

Pretty much what I was thinking - I like the idea of using my recorder as a
composer - something to do whilst sitting on the train - for composition I
could use low bitrate samples (although strict MOD actually specifies 8KHz,
8bit samples) and then use higher quality samples for the final mix-down
(on my PC).

>>MIDI ...just like MOD with preset sample bank
>See above.

I do fully appreciate that there are and will be virtually insurmountable
problems - well, without SOME level of hardware mod - which is way beyond
_me_. But I can also appreciate how much more CAN be done - I've just
never been a settle-for-second-best kind of guy. Single parent family,
anal retentive, you know the social stereo-type ;)

I suppose what I am trying to say is, Thanks for your support.

>>The reason "we don't have enough OGG support provided by hardware" is
>>because we do not have the MAS SDK and therefore cannot use 50% of the
>>hardware - LOL
>In the MAS case, the problem will not necessarily be CPU power, but rather
>(lack of) memory. The MAS has 4 or 8k of memory for both program and data.
>For example, the Tremor OGG decoder has hundreds of kilobytes of lookup tables.

First off, I am very happy to use this subject for little more than
acedemic value.
I never have or had or plan to have any plans to even glimpse at the OGG
code - simply no personal interest.
But academically, it really depends on how well we can get the two
processors to talk to each other and precisely what has been implemented in
the MAS instruction set. There is no reason at all to exclude complex
functions such as SIN or COS, as they will already be featured on the
silicon elsewhere - but without more info, it's ALL guesswork at this level.

BC
Received on 2003-09-03