Fairlight CMI 3D Waveform (Page D) on Atari

[UNIXcoffee928]

I would like to have a 3D waveform tool, like Page D on the Fairlight CMI for the Atari. (Youtube has some Fairlight videos, if you have never heard of it)...

 

I'd also like to be able to use it not just for sample data, but to perform a statistical analysis of any given data sent to the POKEY for sound output.

 

Is there anyone out there who has done anything similar?

 

Can someone post algorithms, or references to mathematical info that would be relevant for doing such a program?

 

Essentially, it is a monochrome 3D terrain type of display (sort of along the lines of Battlezone, but raster), made up of wireframes, which are not surface-mapped. It would be very nice to do this in close to real time, although a static display would be fine to begin with.

 

I remember that Antic had an issue a long time ago on a similar sort of display, done in BASIC (yes, I also remember that it was slow...). Does anyone remember what issue that was printed in?

 

Something like this would be really cool.

 

In fact, a pseudo-emulation of the Fairlight on the Atari would be an over-all great music sequencing tool. Most likely it would cause quite a stir, as the Fairlight is a legendary piece of kit.

 

Here is a link to some manuals & ROM dumps for the Fairlight, for anyone interested in learning more about it. As an aside, they do cover every command, so if it were to be reverse-engineered, these manuals would be the tools to use.

 

 

L8R.

[analmux]

I would like to have a 3D waveform tool, like Page D on the Fairlight CMI for the Atari. (Youtube has some Fairlight videos, if you have never heard of it)...

 

Why would you need 3D for displaying waveforms? What is is used for? Only thing I can think of is using it for showing dynamics of the sound spectrum (instead of the waveforms itself).

 

I'd also like to be able to use it not just for sample data, but to perform a statistical analysis of any given data sent to the POKEY for sound output.

 

I'm not sure what you mean by statistical analysis, like the average of the spectrum of a sample of f.e. one second length...or spectral analysis?...

 

Most sounds of the Pokey are simple squarewaves, even the noisy distortions. So, the standard sounds would be easy to analyze, as the squarewaves (& derived waveforms) are handled by timers. A wave-analyzer would be very simple. You wouldn't need to do any complex waveform analysis etc. The only thing needed would be an eavesdropping program/process when executing a music track, which reads the values sent to the pokey-registers by the main music executable, and generate waveform gfx straightforwardly/immediately from that.

 

On the other hand, there is some analogue behaviour in pokey supporting sawtooth- and triangle-waveforms, and a lot of variations/combinations of them. These would need a lot more analysis. You are lucky though that there are nearly zero pokey tunes that make use of these features

 

 

In fact, a pseudo-emulation of the Fairlight on the Atari would be an over-all great music sequencing tool. Most likely it would cause quite a stir, as the Fairlight is a legendary piece of kit.

 

Not sure what you are after, but Fairlight has 8 independent modules of 8 bit sampling, each with their own RAM and processors. This is far from possible on atari to emulate. So, are you after emulating this sampler or giving the atari/pokey the same controls of the CMI? Or just a baby-version of the CMI?

[UNIXcoffee928]

Well, it depends on a few things...

 

There are a few threads that relate to this little excursion:

 

- "How Do You Pass Info Amongst Two or More Atari800Win Emulation Instances?"

- "8-bit synth?"

- "600XL NOOB"

 

Depending on what we can do with inter-emulator message passing (and the next logical step of having a physical Atari 800, with it's four ports available, to be used as a hardware console connected to a PC hosting multiple instances of the emulator in a "virtual network of emulators" type of environment) ...this alone will determine the overall adaptability of the Fairlight to an Atari environment.

 

Of course, there are many other application for the usefulness of a virtual network of emulators, but specifically regarding the Fairlight, the following applies, as it would overcome the deficiencies of the single processing Atari 8-bit...

 

The way I see it is this, in no particular order:

- Get a fast, small, & standard method of message passing amongst emulator windows on the PC desktop

- Hook up the 800 to the PC's serial port via the 850

- Decide on a standard, fast, & easy to build MIDI interface for the Atari community

- Reverse engineer the Fairlight interface into standalone Atari executable modules, one per Fairlight page

- Integrate a short duration sampler. Many Amiga Parallel port samplers could probably work

- Add Atari-specific POKEY controls

- Find out if LPT POKEY works with an Atari instead of a PC, scale it up with more POKEYS to quad-quad (16)

- Make knobs/sliders/switches box to connect to all 4 joystick ports, using all available lines at once

- Upgrade the physical Atari 800 to dual POKEYs to overcome bandwidth limitations

- Upgrade standard 800 to utilize expansion RAM, for both handlers & sample space.

 

With such a setup, you would be well on your way to having a very interesting use for your Atari, and you would have a tool to add into your existing MIDI studio that would provide sounds that you can't get elsewhere. Imagine the richness, weirdness, and brilliance of 16 POKEYs layered together?!? Each with separately controllable parameters? It would be unlike anything else, especially due to the polynomial nature of the duty cycles.

 

As far as the 3D display, well it's cool.

 

= )

 

 

Nobody but the Fairlight users have it, and don't forget, with a real Atari in the mix you can use a lightpen, no prob.

 

Check the manuals on "Page D" in the above link for a full intro to 3D waveform editing, and check Youtube for videos of it in use.

 

Other than that, as usual with any hack, the following answer always applies... "Why not?".

 

= )

 

 

L8R.

Edited March 6, 2008 by UNIXcoffee928

[Rybags]

Probably the closest thing to Fairlight is the Covox upgrade, which can provide 4 D-A convertors with 8-bit resolution per channel - there are a few threads here about it.

 

The upgrade could easily be scaled anywhere from 1 to 8 or more voices, but space is a big consideration if it were to be built internally.

 

The other consideration is that it has to be 100% software controlled - realistically doing much more than 4 voices in software with a decent sampling frequency becomes a chore due to lack of speed and lack of RAM.

 

Such an upgrade could be built externally to use joystick ports, but would probably need 3 ports. You need 8 bits output, 2 bits for chip-select and 1 bit write strobe.