Re: A89: >1-bit sound on the 89

[Prev][Next][Index][Thread]

Re: A89: >1-bit sound on the 89


Henri Moilanen wrote:
> 
> On Sat, Apr 15, 2000 at 11:35:05AM -0700, Seth Peelle wrote:
> >
> > Forgive me for my lack of calc hardware knowledge... I understand that
> > link port 1-bit sound is created by rapidly switching the link on and
> > off (correct me if I'm wrong).  Are there any possible methods for
> > making sound with greater bit-depth (i.e. 2-bit, 4-bit, 8-bit, etc.)?
> > That would create some interesting potential for sound/music synthesis.
> 
> Well... One could put some microcontroller as a gateway between TI89 and DAC
> (DAC is a hardware chip which converts digital signals to voltage levels
> which could be then amplified and feed to a headphone or ... speaker (I
> remembered the word!!! :)). That would need much hardware to do the job, but
> then... if the microcontroller was programmed to do some synthesis, it would
> be quite a good way to produce music, as the TI89 wouldn't need to do more
> than to output the correct notes, lengths, and waveforms to the controller.
> (Samples need too much space to be stored even on TI89, yes it is possible,
> but the length is minimal, decompression eats calculating power, and quality
> is terrible).

Hmm, interesting idea, even if it's not exactly what I had in mind...
but it would probably be complicated hardware-wise, like you said.  I
was thinking more in terms of software-based, so that the user could
simply d/l the stuff to the calc & all he/she would need to assemble is
the speaker.  That is a neat idea, though.

> Other way, which could also be used, is to output variable length pulses to
> the link port. Then if the overall pulse frequency is high enough, and if it
> was filtered away, then the variable length pulses would end up being (on
> average) different voltage levels. They could then be pulsed to generate the
> sound.
> 
> For example, if there was four different lengths of pulse: 1/4 2/4 3/4 4/4
> Then 1/4 would be "one fourth signal level" etc.. so this would be 2bit
> sound. There is major drawbacks in this (PWM Pulse Width Modulation) scheme.
> There is the pulsing frequency for the variable length pulses which could be
> heard on the background noise for the inteded sound. And it takes quite a
> lot of CPU cycles to generate the variable length pulses.
> 
> Well, on the above explanation the variable length pulse means that the
> overall width of the "pulse" is fixed, but the amount of time the pulse is
> high, i.e. "1", i.e. 5V, is variable.
> 
> I'm really sorry for the poor explanation, but as English is not my native
> language... I really can't explain things very well. Maybe someone will
> clarify/correct my explanation.

I thought you did a great job, especially considering that English is
one of the most difficult languages to master!  I can tell all those
English classes they give from a young age over there really pay off :) 
Too bad they don't do something like that here in the US...

This sounds to me like a similar method to the one used to create
grayscale -- am I right?  Wow, that _would_ require an extremely high
frequency.  What is the highest frequency that could be generated on the
calc, approximately?

> > On a side note, and out of sheer curiosity, is there any reason in
> > particular why a disproportionate amount of TI calc programmers are
> > French?  Not trying to discriminate or anything, just simply wondering
> > :)
> 
> That is mystery for me too... I've been wondering why there aren't any other
> Finns on the TI89 scene... :-) Am I alone in the big world? 8-]

I'm surprised, too, considering their humongous contribution to the old
demoscene.  Of course, that was all based on computers that were both
widely-available and advanced at that time -- and graphing calcs aren't
exactly the most advanced computers around ;P

> Henri Moilanen

BeaT (Seth Peelle)
beat@drigon.com
http://www.drigon.com/~beat
References: