Re: Crystal's new chip

> Huh? I'm not sure what this sentence means. (Really I'm not being facitous! :)
>
> I was originally refering to your need to change the "frequency" of the
> background noise, and like I said I don't have the schematic here, but I
> know they either used the all_in_one noise generator chip, or they did it
> themselves using shift_regs/xor_gates. They then piped the output into a
> low pass filter section, then controlled the attack/decay with one of those
> current control amplifiers. I wasn't sure where the adjustable clock, going
> into the noise generator, was.

        Yeah, but it's not done that way for the background sound/drone. That's a whole different circuit which is a DAC -> LPF -> Level Shifter/Limiter -> VCO I think we both know what the other is talking about now, but we got mixed up somewhere along the way....

        There is no adjustable clock for the noise generator -- it's generated internally, in fact...

<snip>

> Easily done with pitch bend...

        Ahhhhh, I'm obviously a bit out of my league here. I've never touched anything MIDI at all. I'm not even sure what pitch bend is....

> No need, you're missing the beauty of the synthesizer chip. It has pitch
> bend. I can't remember what the midi specs is, but the resolutions is real
> high. Something link 4096 or 2048 steps between half notes! Whether it's
> done with a PLL (most likely) or whether they just tweak an analog clock
> inside the IC, I don't care, the point is you have control of the frequency
> of the played waveform with very high resolution.
>
> If you find someone with a midi synthesizer, hooked up to a midi keyboard,
> and press a note and play with the pitch bend control you'll see what I
> mean. You can't here any stepping of frequency between one note and the
> next, even though it's being fully digitally controlled.
>
> As far as stepping through the wavetable ourself, I don't think that would
> be possible. The table needs to be accessed by the synth chip in a truely
> random way, in order to play back the 32 simultaneous sounds. And it's kept
> in a "proprietary compressed" format. You might have access to that info, I
> obviously don't, and don't need it. I would use the synth chip it's self to
> save the samples. I'm not concerned with how it saves them. I'll just take
> what ever data it saves in RAM, and ROM it.

        .....not stepping through it, just controlling how fast the chip steps through it (i.e. the clock frequency.) My idea of a wavetable was that it was just that -- a lookup table for values of the sound/wave vs time. You can represent a sine wave of one frequency with a certain wave table. You can represent a sine wave of twice the frequency with the SAME wave table, but stepping through it twice as fast....This sort of implies that the signal be periodic, because if you want the signal to last any significant length, you just keep repeatedly looping through the table.
 
> All that *and* wavetable synthesis. Looks like a very nice chip, they told
> my friend that they would be available near the end of the month (or maybe
> they already are -- I'm having dinner with him tonight, I'll find out), but
> we're definitely going to grab some.

        Hmmmmm, maybe I'm misunderstanding what wavetable synthesis is (i.e. it's not just a lookup table, which could be implemented on a DSP with RAM/ROM) looks like I need to do a bit of "homework."

> The external ROM looked very standard in it's interface. Just a basic
> parallel data and address busses.

        It looks like this is a new chip. The one that I'm looking at has a built-in ROM. Ignore the part number I gave you, then....

Joe
  
Received on Mon Jul 7 14:38:08 1997