Video about 1-bit sound in Apple Lisa

Great. I'm really hoping you can exercise the techniques you used on PET, but on the Lisa... very interesting is the 3-bit volume control available on low level which is not available on other platforms of that era (from what I was able to research). In my experiments I was able to switch very fast between lowest and highest volume which created interesting echo effects.

Also funny how I found your article on the PET very shortly before publishing my video, glad I did because I was able to reference it!

Apple Lisa has a strong cult following but (almost) no demos, so there's a fertile ground and large audience if you ever find time to work in it.

Thank you for watching utz and for the kind words. And I'm sorry for totally missing the fact that you worked on the PET implementation with Shiru, I didn't realize that, I would've mentioned!

In any case yes, the volume gain definitely doesn't sound linear, and I'll do the measurements shortly and post here, good idea.

Also, Ray Arachelian is actively working on fixing the LisaEm based on my feedback, so hopefully soon it will be easier to compose stuff; see here for details https://lisalist2.com/index.php/topic,255.0.html

I guess with M68K and 6522 we can actually set the shift register to the highest shift frequency possible, hook up an IRQ fired when it gets empty, and generate bit stream for a few channels at once not in a timed loop, like we usually do, but for a few iterations ahead. So we'd have the traditional sounding beeper music that will still leave some time for other code to run.

Another idea is to have the same single voice music as I did on the PET, but modulate the volume with a timer driven rapid changes, like, triangle/saw/pulse, much among the lines of the AY-3-8910 envelope abuse. This approach won't bring the polyphony, but very complex timbres instead.

Either way, a good quality sound emulation is needed to be able to tackle the thing. I'm afraid I haven't seen an actual Lisa at any computer museum here, let alone someone's personal possession. It is even biggest challenge than the PET in this regard.

The increase in voltage looks linear (see attachment), but linear voltage increase does not translate to linear power increase, so probably that's why we hear relatively strong difference between quiet levels, and relatively negligible at the loudest levels.

That's for frequency 440Hz, 50-percent duty cycle, notes at individual volume levels played in quick succession.

Also, I am traveling for the next 10 days so won't be able to test anything on the Lisa. But I'll test when I'm back if needed.

I don't think there are 9 pulses; the vertical spacing looks arbitrary to me and does not really correspond to volume changes, something else must be causing the spacing (and quite possibly my cheap oscilloscope). There are clearly 8 levels heard from the speaker, and the code responsible for the above is attached.

Yeah, I don't expect people to drop everything and start working on the Lisa... it's been waiting 40 years, can wait a few more! I'll try to experiment myself a bit within limited capabilities. I'll be refreshing the thread with progress.

Ray Arachelian is working on fixing the emulator; he acknowledged my yesterday's feedback.

Lastly, just a minute ago I tried a crazy idea to simply keep shoving new and new series of 8 bits to the shift register without waiting for the sound to finish like below, and it's working — you can basically keep flooding the shift register without changing its state first in any way. Like this:

        MOVE.W #$20,D6
loader  MOVE.W #$FF,D5          ; load only 1s first, then change pattern with every iteration
pusher  MOVE.B  D5,SHR1(A0)
        MOVE.W  #$00D0,D4       ;set some arbitrary delay constant
ticker  DBF     D4,ticker
        DBF     D5,pusher
        DBF     D6,loader

It didn't choke or anything. Produced weird whizzing sound because we iterated through all possible 8-bit pulse train shapes.