Hypothetical 6502-based synthesizer calculator?

[Joseph_21sv]

Now that HP have been producing business calculators which are based on the MOS 6502 CPU or one of its variants, although neither is based on the exact 6510 variant which appeared in the Commodore 64 series beside the famous 658x/8580 SID, for over 10 years, that has made me think that there may be a new open space in the market for a handheld calculator with polyphonic audio synthesis. But of course, there will be little point to the feature unless the end-user actually has a documented way to program multiple voices. Imagine, if you will, HP producing a 17bii+, currently 8502 (double-speed-capable 6510) based with beeper speaker, which uses an upgraded SID chip to generate sound. It could be numbered, if you wouldn't mind the pun, 68B as 68 is the next multiple of 17 above 64, not to mention that it would only help such a programmable sound generator be useful if the user had at least 68K of memory to use for data storage.

[Garth Wilson]

I don't know what you'd want a musical synthesizer calculator for, but modern 65c02's are powerful enough to do the job without the SID, just by sampling the waveforms based on timer interrupts, and feeding the samples from tables, polyphonically. I've run about 125,000 interrupts per second on a 65c02, and the fastest 65c02's are about 40 times as fast as mine (200MHz, about 50MIPS, and an interrupt sequence that's only 35ns long). CD sampling rate for high-quality audio is only 44,100 samples per second.

I do emphasize the "c" in 65c02 though, for "CMOS." It has a ton of benefits over MOS Technology's original 6502, which I outline at http://wilsonminesco.com/NMOS-CMOSdif/ .

[timofonic]

(01-31-2016 04:47 AM)Garth Wilson Wrote:  I don't know what you'd want a musical synthesizer calculator for, but modern 65c02's are powerful enough to do the job without the SID, just by sampling the waveforms based on timer interrupts, and feeding the samples from tables, polyphonically. I've run about 125,000 interrupts per second on a 65c02, and the fastest 65c02's are about 40 times as fast as mine (200MHz, about 50MIPS, and an interrupt sequence that's only 35ns long). CD sampling rate for high-quality audio is only 44,100 samples per second.

I do emphasize the "c" in 65c02 though, for "CMOS." It has a ton of benefits over MOS Technology's original 6502, which I outline at http://wilsonminesco.com/NMOS-CMOSdif/ .

People do chiptune concerts with gameboys. Why not use maths in live to make music? I see it as a nerdy version of a pianist

[Joseph_21sv]

(01-31-2016 04:47 AM)Garth Wilson Wrote:  I don't know what you'd want a musical synthesizer calculator for, but modern 65c02's are powerful enough to do the job without the SID, just by sampling the waveforms based on timer interrupts, and feeding the samples from tables, polyphonically. I've run about 125,000 interrupts per second on a 65c02, and the fastest 65c02's are about 40 times as fast as mine (200MHz, about 50MIPS, and an interrupt sequence that's only 35ns long). CD sampling rate for high-quality audio is only 44,100 samples per second.

I do emphasize the "c" in 65c02 though, for "CMOS." It has a ton of benefits over MOS Technology's original 6502, which I outline at http://wilsonminesco.com/NMOS-CMOSdif/ .

HP have actually been making one or more models of graphing calculator with a fully-programmable beeper continuously since 1988. It would only be logical for them now to put a polyphonic synthesizer chip, such as the 65c02, in a calculator and have a tracker for it pre-programmed into firmware. However, a sound chip intended as non-sampling, such as the SID, makes more historical sense than a sampler chip when one is building on an 8-bit processor such as the 65c02. Moreover, it is sort of an excuse, if one must be needed, for a non-graphical calculator to have more than 32K of user memory, because most of the 8-bit home computer models with polyphonic synthesis had that much. As to why someone might want to buy a polyphonic synthesizer calculator, outside of nerdiness, they could simply be squeamish about sinking the money (and space) into a full-sized general MIDI keyboard or a keyboard purist who cringes at most of the samples that might come out of general MIDI keyboards or realize what it would take for a calculator to be seen in stock at a Guitar Center.

[benbradley]

I remember when the SID chip came out, a description and the whole data sheet was printed in Polyphony magazine (originally started by PAIA and largely aimed at DIY synthesizer makers, it later became a mainstream magazine for keyboard players, Electronic Musician). It used DDS/NCO for the oscillators and thus had high resolution (something like 0.01Hz) frequency control. It would have been interesting for use in synthesizers and music-generating expansion boards for other computers, but Commodore chose not to sell it to others, but to only use it in its own computers, most notably the C64.

What the OP describes would be "neat" in a retro way but I can't imagine anything like that happening today. I'm a bit surprised even the 8-bit processor architectures have survived with all the ARM chips being so cheap and ubiquitous. Everything the SID did and much more can be done in software on an ARM.

[Garth Wilson]

(02-01-2016 09:30 PM)benbradley Wrote:  I'm a bit surprised even the 8-bit processor architectures have survived with all the ARM chips being so cheap and ubiquitous. Everything the SID did and much more can be done in software on an ARM.

It can be done on a 65c02 also, now that it's plenty fast to do it without the SID. I suspect the reasons for the 8-bitters surviving is that they're simpler to program, take very little silicon real estate, and that a lot of things have no use for GHz ARM performance. I've brought a lot of products to market using 8-bit PIC microcontrollers, and 98% of their operations were 8-bit-only, and they never even did a multiply (although the last one did a divide). The 65c02 is still being produced today, to the tune of over a hundred million units a year, virtually all of the hidden at the heart of custom ICs in industrial, automotive, appliance, consumer electronics, game, toy, and even life-support products.

[Paul Dale]

(02-01-2016 09:30 PM)benbradley Wrote:  I'm a bit surprised even the 8-bit processor architectures have survived with all the ARM chips being so cheap and ubiquitous.

The tiny eight bitters are cheaper -- a lot cheaper. They are physically smaller and usually require very few additional components to get them working. They'll be around for a long time yet.


Pauli

[Joseph_21sv]

My bad, the 65c02 is actually used in conjunction with a PSG which allows 5 (up to 10 by combining audio channels)-bit samples in the NEC PC Engine/TurboGrafx-16. That processor is the Hudson Soft HuC6280. If Epson is still making it, it won't be too much more than just enough to advance over what is in the HP 50g. Imagine now any one of HP's 6502 calculators remade with this processor and a pre-programmed 6-channel tracker.