CASIO sound synthesis

[horniger]

Hi

 

my name is Rainer and I love technology and history of sound generation.

 

Although I am very fascinated in the very early analogue tone generation (electro-mechanic, electro-magnetic, electro-static and tube- or transistor-based electronic), I am also very interested in the digital sound synthesis

(I did a lot of research, and wrote a presentation about the history of digital sound synthesis (in German: 'Als Kopieren noch schön war - die Anfänge der digitalen Klangsynthese' - viable here: https://media.ccc.de/v/vcfb2022_-_170_-_de_-_202210081030_-__als_kopieren_noch_schoen_war_-_rainer_siebert ) which I gave as part of the short conference 'Hard Bit Rock' on the VCFB 2022 (Vintage Computing Festival)).

 

I have a huge collection of CASIO keyboards and synths an am very interested in their sond synthesis and their great, unique sounds!
At the beginning, I was wondering about the very special sounf of the Casiotone CT-201 and CT-202. And later I understood, that these sounds come from a digital sund synthesis, like the FM synthesis...
While my research, I found this forum, but didn't managed it, to post my questions, although I had a lot of still un-answered questions...
Maybe I can post these now... There still are some secrets...

i.e. the CASIO MT-65 was discovered with the great NEC D931C LSI sound generator chip, which could be found in many other Casiotones. But there were also MT-65 models with the Hitachi HD44140 sound gen.
Are these identical? Did CASIO switched from NEC to Hitachi? Are the sounds identical?

Do you know a NEC D931G?

 

Does someone has the technical manuals / schematics of the early keyboards CT-201, CT-202 or M10...?

 

The CASIO HT series is actiually part of my focus. Does anybody has the techn manual of the HT-6000?
Which I found, that the HT-6000 (4 oscillators with 64 waveforms per voice) has four NEC D935G and the HT-3000 or HT-700 (1 oscialltor with 32 voices) just one NEC D935G. Is this true or are there different sound generator Chips in these keyboards.
Is it correct that the "new" SD synthesis of the HT series ist just a pimped Vovel Consonant synthesis of the early Casiotones...?

 

enough for now - thanks for reading

 

all the best

 

Rainer

http://horniger.de/

[Chas]

There is a thread in th forum where many discussed the wide number and different types of sound synthesis used in Casio/ Casiotone models. A lot of good information in there:

 

Is

[Chas]

The Casiotone 201/ 202 models use Vowel Consonant synthesis. This was also used in many early Casio/ Casiotone models. It uses some primitive waveforms and preset filters to generate the sounds. 

 

The Casios closest to FM synthesis are the Phase Distortion CZ models (and later VZ). These were very different beasts to the home keyboard Vowel Consonant Casios.

 

The model that had a Hitachi DSI fitted was the MT46 IIRC. The MT65/ 68 and CT405 all used the same NEC DSI and to my knowledge never came with a Hitachi DSI.

 

Yes, the Spectrum Dynamics HT series apparently uses an updated form of Vowel Consonant synthesis. The HT6000 is the only model with four DSIs fitted compared with all other HT/ HZ models that had just one. As you mentioned, the 6000 has four "lines" (Casio speak for "oscillators"), whereas all other models had a single "line"/ oscillator. The 6000 also had 8 filter chips making it fully polyphonic. The other models had just one main voice filter chip making them paraphonic. 

 

A while back I put together a video where I attempted to identify all the "Frog" sound models (as made famous in Michael Jackson's Thriller song). In that video I identified all the different DSIs used in the early Casio keyboards and I also showed the families of Casio keyboards they could be found in. That should help answer a lot of your questions 🙂

 

 

Edited November 1, 2024 by Chas

[horniger]

Thanks for your reply!

 

I really love your yt channel!

I previuosly watched some of your videos but didn't bring these togeteher with your account here

(I already read some of your posts here, and devoured the schematics you uploaded!)

 

My research did result in the same insights and many things I could add with information here.

 

I love those research about, which keyboard was used on this track, and which sound was played by which model...

It's really great, that the early Casio keybvoards were used on some famous tracks - but yeah: they were quite special and unique for its time! And affordable!

 

I still have some detail questions, if you won't mind:

 

The Hitachi HD44140 - you called it DSI - was found on some Casio keyboards (you're right - not on the MT-65 - my fault)

You call the sound synthesis "mixed multipart squarewaves"

But in my understanding, this isn't a real digital sound generation. With Synthesis and DAC afterwards?

On the Hitachi chip, the waves are being mixed and then amplified externally with some filtering?

Or am I wrong?

It's pretty much the same tone gernation like on most of the CASIO PTs und VLs...?

You wrote that the VL-1 usses a digital synthesis known as "Walsh Function".

Is there more information about this (maybe I should finish watching your great videos before).

I thought that the Consonant Vowel synthesis also uses some of the "Walsh Function" stuff...

 

Thanks about your information on the CASIO HT Series.

I was wondering why the NEC D935G just offers 32 different waveforms in the HT-700 / HT-3000 but 64 different waveforms in the HT-6000 - do you have an idea?

And I'm quite not sure, how the different sections (upper, lower...) were realized.

Unfortunately, I have no access to the service manual for one of these...

 

But I found the service manual for the Casiotone CT-6000, which also interested me always.

Three NEC D932G "Music LSI" with 17-Bit discrete DACs...!
Without accompaniment all three re used for melody, with just two of them and one for the accompaniment.

Was this still Consonant Vowel synthesis?

 

And what about the CT-6500?

 

Thank you a lot

 

[Chas]

@horniger - Hey Rainer, sorry for not responding sooner. It's been a busy past few days here. Also, thanks for the compliment about my YouTube channel! I'm glad you enjoy it 🙂

I'll answer your questions as best I can:

 

1. "The Hitachi HD44140 - you called it DSI - was found on some Casio keyboards (you're right - not on the MT-65 - my fault)

You call the sound synthesis "mixed multipart squarewaves"

But in my understanding, this isn't a real digital sound generation. With Synthesis and DAC afterwards?

On the Hitachi chip, the waves are being mixed and then amplified externally with some filtering?

Or am I wrong?"

First thing I need to point out, was that for some reason I was writing "DSI" when I actually meant "LSI". I must have been thinking of "DSP" and getting it mixed up, whereas I meant LSI as in "Large Scale Integration". LSI is the type of CPU/ processing chip(s) that are used in many Casios.

 

I also know very little about the Hitachi HD44140, and to my knowledge, the MT-46 might be the only model that uses this LSI.

 

For the early Casios/ Casiotones using "Vowel Consonant"/ "Consonant Vowel" synthesis, I would have been quoting from the excellent @CYBERYOGI =CO=Windler website (link below), where Cyberyogi has carried out extensive research into many of the classic and vintage Casio models. Check the link below for details on the very first Casio keyboard, the 1980 201 model, where you'll see in the text: "All sounds are based on 2 mixed multipulse squarewaves with independent digital envelopes those are sent through different capacitor filters. The resulting timbres remind to C64 or historical videogame musics. Casio called this system Consonant Vowel synthesis.":   http://weltenschule.de/TableHooters/Casiotone_201.html

 

Further details of this sound generation can be found here: http://weltenschule.de/TableHooters/Casio_CT-410V.html#ConsonantVowel

 

 

2. "It's pretty much the same tone gernation like on most of the CASIO PTs und VLs...?"

 

I'm not entirely sure, but I *think* the VLs and the PTs use a much more simplified version of Vowel Consonant. The description of the synthesis that I stated in my video was taken and quoted from within this link:  http://weltenschule.de/TableHooters/Casio_VL-1_PT-1.html

 

And if you watch my Casio VL-1 video, "The Little Synth That Could", you'll very clearly see on the oscilloscope displays the crude/ steppy digital envelopes acting on the sounds in real time.

 

  

 

 

3. "You wrote that the VL-1 usses a digital synthesis known as "Walsh Function".

Is there more information about this (maybe I should finish watching your great videos before).

I thought that the Consonant Vowel synthesis also uses some of the "Walsh Function" stuff..."

 

It turns out that isn't correct. Again, if you read the email from Robin Whittle (quoted in the link above for Consonant Vowel), there was some suspicion that some of the early Casios used Walsh function. This led many people to believe, myself included, that those early VL and PT Casios (and also including their musical calculators and watches). Apparently they did not! And yes, I should go and amend the description in my video...

 

 

Regarding Vowel Consonant synthesis, this is a good description from a Reddit discussion group:

"Casio actually called it consonant-vowel synthesis, not vowel-consonant. It does use switchable lowpass filters but they're just fixed first order filters, not even as sophisticated as the tone filters in analog organs or whatever, and it generally sounds better to just bypass them entirely.

 

Basically it just stores two low resolution waveforms (16 steps each) and plays them together with independent amplitude envelopes. Typically one of the waveforms is used as the attack portion and the other is used as the sustain, so you can get a transient effect that's more interesting than just a single waveform. Each of the waveforms can optionally be modified to make a longer waveform up to 64 steps, with correspondingly lower pitch. This can be done independently for each waveform, so it's possible to have a sound that starts at a high octave (16 steps) and drops to a low octave (32 steps). Overall they have a very interesting sound that can't be accurately reproduced by anything else.

 

I think the consonant-vowel name simply refers to the fact that the sounds can have an attack transient with a different timbre, which is sort of roughly analogous to different phonetic sounds. Casio used the technology from about 1980 to 1987, but I think only the Casiotone 201 was marketed this way."

 

(Link to the discussion: https://www.reddit.com/r/synthesizers/comments/6p13pk/vowelconsonant_synthesis/ )

 

 

Casio also filed a Patent in June 1981 that looks to be describing Vowel Consonant synthesis. You can get a lot of information from there:

 

https://patents.justia.com/patent/4419919?fbclid=IwY2xjawGaWTlleHRuA2FlbQIxMAABHdodTc9g5GiYLmaVnSM6UzZP-jmOb-9YBOsJO5joh2Op1Cat5l0lMdo8Gw_aem_-QpSoHOkqQYp6kxEnjWfcQ

 

 

4. "Thanks about your information on the CASIO HT Series.

I was wondering why the NEC D935G just offers 32 different waveforms in the HT-700 / HT-3000 but 64 different waveforms in the HT-6000 - do you have an idea?"

The HT6000 was the flagship model (and the most expensive), therefore it got more features. Much the same practice that Casio did with other model series i.e. the CZ Phase Distortion synths. The HT6000 was also the only Spectrum Dynamics Casio with Ring Modulation. In the second bank of 32 waveforms unique to the HT6000, 0 - 15 contain noise (I'm assuming a white noise type of effect), and waveforms 16 to 31 have ring modulation. 
 

"And I'm quite not sure, how the different sections (upper, lower...) were realized."

 

You could split the keyboard (upper/ lower), and assign totally different tones to each split.

 

Unfortunately, I have no access to the service manual for one of these..."

 

I have a copy. I will upload it to the Forum files section once I've finished this response.

 

 

5. "But I found the service manual for the Casiotone CT-6000, which also interested me always.

Three NEC D932G "Music LSI" with 17-Bit discrete DACs...!
Without accompaniment all three re used for melody, with just two of them and one for the accompaniment.

Was this still Consonant Vowel synthesis?"

 

The CT6000 is basically a supercharged Vowel Consonant Casiotone. It layers the LSI's to act like a multi-oscillator synth and allows you to mix different tones together. That's why it sounds so thick and lush compared with all other Vowel Consonant Casios.

 

"And what about the CT-6500?"

 

The CT6500 is a pre-set version home keyboard version of the Phase Distortion Casios. It is possible, however, to program new tones into some of its locations by using a computer editor.

 

It also highlights Casio's sometimes confusing model numbering. The CT6000 was a triple LSI Consonant Vowel Casiotone, the CT6500 used the totally different Phase Distortion synthesis from the CZ synths, and the CT7000 used a single LSI Vowel Consonant technology as found in most of the common home Casios of the time (though with a digital sequencer and also programmable digital panning and effects). Even though the model number would make you think that the lower number would be inferior and the higher number more powerful, they are in effect three totally different models with one (the 6500) using a completely different form of synthesis. Meanwhile with Yamaha's DX series, it was very clear that the DX-1 was the biggest/ most powerful/ flagship, and then all the models numbered after it were sequentially less powerful i.e. DX-1>DX-5>DX-7>DX-9 etc.


Hopefully this answers your questions, and hopefully one day there will be a comprehensive database of all the Casio LSI's used and the models they can be found in. Along with a comprehensive list of all the different sound synthesis methods they used and how those work. I'm slowly compiling such information, as are others, though there are many, many gaps as Casio have been rather secretive and vague with regards to specific details of their many different sound sources. 

 

 

 

Edited November 8, 2024 by Chas

[Chas]

Here's the download link for the HT-6000 Service Manual & Parts List:
 

 

Edited November 8, 2024 by Chas

[CYBERYOGI =CO=Windler]

Again, this sound source is named "consonant vowel" synthesis, NOT Vowel Consonant. 😡

 

The smaller Casios (PT-series, VL-1) were not walsh. They have only one multipulse squarewave with linear volume envelope and LFO. It may be that the idea of synthesizing timbres this way originated from Walsh (there was that Allen Organ patent lawsuit), but multipulse squarewave is not identical with Walsh. Layering multiple Walsh drawbars would still sum them with different volume levels, while within a multipulse (16 bit steps those each can be hi or lo) the steps always keep the same height (i.e. 1-bit signal of repeating 16 steps). I remember I read in a magazine(?) long ago an interview that Casio had invented multipulse squarewave coincidentally when they experimented with LCD control voltages for calculators (consisting of blocky 4 step (2-bit) waveforms) and noticed that they sounded like organ tones.

 

I also doubt that consonant-vowel or SD internally layers actual Walsh drawbars (which are a series of certain mathematically well defined special multipulses). It simply crossfades/morphs between 2 blocky waveforms (and than layers 2 or 4 of them having different preset analogue filters) to synthesize timbres. The internal basic waveforms were likely tweaked by ear and not planned to implement Walsh. Walsh makes only sense when you want to be capable to systematically approximate all repeating waveforms by summing a high number of multipulses in the same manner like conventional drawbars sum sines (implementing the Fourier series).

The HD44140 (in PT-7 etc.) is a specialized and badly crippled single-chip variant of the sound IC HD43720 (found in MT-36, MT-200, MT-90, CT-102). Its sound engine is quite unique but nothing great.

 

sound generator of HD43720

 

This sound hardware was apparently Casio's first cheap polyphonic sound ic with integrated DAC, chord and percussion generator. The main voice generator is obviously a polyphonic successor of the great multipulse squarewave sound engine of Casio VL-1, but unlike there, the blocky waveforms are bipolar, i.e. made from 3 levels {-1, 0, +1} (which IMO did not improve timbre quality). So the sound IC has 2 DAC outputs for a main and a sub waveform, those are mixed through resistors in a ratio of 1:64. With held monophonic notes the main waveform looks like a normal multipulse, while the sub waveform has 3 levels {-1, 0, +1} and looks like the tone. During falling volume envelope, the equidistant "0" sections of the main waveform rapidly move many times linearly up and down while the sub waveform amplitude slowly shrinks. I am not sure if these observations (seen on analogue oscilloscope) are only artifacts of sending upper and lower waveform bits to a separate DAC, or if main and sub waveform indeed get computed this way by 2 independent generators to combine into the envelope. Likely the main waveform only scales the envelope; in most preset sounds it is plain squarewave, but in 'elec. piano' and 'clarinet' it is a multipulse, so it may be that this is a very crude kind of phase distortion (FM) mechanism involving analogue multiplication of both components (main waveform has 2 diodes in series to GND, which may intend nonlinear behaviour), although it does not vary modulation depth to change timbre. (It would be interesting to build a more programmable version of this, to examine what else it can do.)

The unused 11-bit DAC output port makes me conclude that for 8 note polyphony (3 bits) each polyphony channel including envelope has 8 bit. The sub waveform of one note is up to 5 steps high (seen on analogue scope), hence the sub waveform is 3-bit while the main waveform has likely 63 steps and thus 5-bit. In preset sounds 'piano' and 'vibraphone' the sound ic pulls pin 11 lo to enable a discrete external filter that makes them duller and adds a little resonance.
 
preset sound     sub waveform     main waveform     envelope
piano     11111100     11111100     decay 1s
elec. piano     1--01-00     011X01XX     decay 0.5s
organ     1---1-00     11111100     hold + slow attack
oboe     1-----10     11111110     hold + slow attack
clarinet     11-100-0     11110010     hold
vibraphone     1---0---     X111X111     decay 0.5s + always sustain
strings     100-----     10011111     hold + slow attack
elec. organ     1-101--0     11101110     hold
bass & chord     11110010           (capacitor sustain)

 

The complete waveforms on analogue oscilloscope look very distorted by capacitors, thus the bipolar multipulse patterns were hard to identify, but they seem to be indeed identical with their own "sub" waveform, which shape is better visible. The "-" indicates the intermediate level (center line) between 0 and 1.

 

The "main" waveforms here are only an approximation. Their "X" sections move rapidly. The "0" section amplitude moves with envelope. But these don't strongly change the complete waveform anyway but mainly affect envelope.

 

The volume envelopes are linear with very few internal parameters. Unfortunately they decay too fast. Sustain decays 0.5s after key release.

 

pinout HD43720

 

The Music LSI "Hitachi HD43720" (54 pin SMD, pins count anticlockwise from below the left stub pin) is the squarewave sound IC of various early Casio beginners keyboards. It contains 8 polyphony channels for melody voice (4 of these can be re-routed as chord & bass voice) with linear envelope. Additionally there is a percussion generator with waveform outs for {base, snare, cymbal+hihat} and corresponding trigger pins for external analogue envelope circuits. Also the 4 separate chord & bass outs have each an envelope trigger pin. As a real sound IC it needs to be controlled by an external CPU through a data bus. The integrated DAC the integrated DAC outputs its lower bits (at increased level) only on a separate pin; also the highest bit is separately output in normal and inverted. All these have to be combined through an external voltage divider (small resistor network with ratio 1:64), which was likely done to reduce noise. The only 11 DAC bits also have separate output pins (not used in any instruments I know).

 

Each waveform is made from bipolar multipulse squarewave (8 steps long) and a simple volume envelope that does not change with note pitch. Each step can have only the height {-1, 0, +1}, i.e. protrude fully up, fully down or be zero. The linear envelope and low bit resolution make it sound rather artificial, but the unusual sonorous waveforms are unique and the high internal clock rate (1.65 MHz) prevents cold aliasing noise. Unlike the more advanced D77xG there is no spread scale, i.e. holding the same notes in different octave (except the highest note) causes no phasing (but phase changes among trilled decaying notes, thus there are no octave dividers involved).

 

It is unknown if synthesis parameters are send by the host CPU to the sound IC (like in D931C), or if it only can select 8 preset sounds from an internal rom (like done in early Yamaha keyboards). The fact that all known instruments with HD43720 contain the same preset sounds suggests the latter.

 

See attachment for tech details.

Casio_MT-36.html

Edited November 9, 2024 by CYBERYOGI =CO=Windler

[horniger]

Hello Chas

 

no problem because of a few days delay - we all (unfortunately) have a life apart from the Casiotones...

 

Thanks again for all the information and also for providing the CASIO HT-6000 circuit diagram!

A lot of things have become clearer to me.

It's really annoying that Casio made such a secret of the sound generation - either because they didn't want to offer too much scope for patent disputes or because they didn't think it was worth mentioning...

Let's see what else we discover and collate here!

 

 

[horniger]

Hello CYBERYOGI

 

Thank you for the detailed explanations.

 

I found the Hitachi HD44140 in some instruments, where was the HD43720?

 

Is a clan synthesis according to Walsh something like additive synthesis with a lot of (sine) waves?

 

I'll post my understanding and lots of new questions right afterwards...

[horniger]

In sound synthesis I make a distinction between real digital sound synthesis with somehow (FM, PD, LA, additive,...) generated waveforms,
which are then converted by D/A converter (DAC) into analog audible signals and directly generated - for example by the mix of several (analog) pulse waves.
On every (former) Casio keyboard there's always written 'Casio Computer Company' and I interpreted this to mean that Casio did not really produced instruments with analog sound generation (except for the sound generation of the accompaniment).
Is that true?
 
Here's the result of my research. Please confirm or deny:
 
1. consonant vowel Synthesis (rectangular / staricase)

• digital sound synthesis (with D/A converters - DACs)
• dual rectangular (?) staricase / stair-step waveforms (?) with 16 (?) steps
• each with own digital ASDR (?) enevlope with ? steps
• eight notes polyphony

Casiotones

• Casiotone 201:
  • NEC D771G + NEC D772G (29x2 sounds, 8-note polyphony, 4 octaves), keyboard and switchboard controlling
  • 14-Bit hybrid DAC (CASIO CT-201), 12-Bit discrete DAC (CASIO CT-201/b)
• CASIO M-10, Casiotone 301, Casiotone 401:
  • NEC D773G (23 sounds, 8-note polyphony, 4 octaves, 12-Bit discrete DAC ), keyboard and switchboard controlling
• CASIO MT-30, MT-31, MT-40, MT-41
  • EC D775G (22 sounds, 8-note polyphony, 4 octaves, 14-Bit discrete DAC), keyboard and switchboard controlling
• Casiotone 101, Casiotone 403 (early version)
  • NEC D776G (15 sounds ?, 8-note polyphony, 4 octaves), keyboard and switchboard controlling
• Casiotone 202:
  • NEC D788G + NEC D789G (49 sounds, 8-note polyphony, 4 octaves), keyboard and switchboard controlling ?
• Casiotone 403 (later version), CASIO MT-60
  • NEC D990G (15 sounds ?, 8-note polyphony, 4 octaves), keyboard and switchboard controlling
• CASIO CT-310, CT-320, CT-405, CT-430, CT-410V, CT-605, CT-610, CT-620, CT-810, CT-606 ?, CT-7000 ?, CK-500
• CASIO MT-52, MT-65, MT-68, MT-85, MT-86, MT-100, MT-210, MT-400V, MT-800, MT-500 ?
  • NEC D931C - 'melody source' (20 Sounds ?, 8-note polyphony, 5 octaves, 14-Bit)
  • 3x4049 (Inverter) + Resistor-Array - 14-Bit discrete DAC
  • NEC-D930G - CPU (keyboard and switchboard controlling, analog accompaniment: 6 voices: 1 bass, 1 arpeggio, 4 chords, drum trigger)

 
Do the two waveforms of these earlier consonant vowel instruments really consist of staircase curves and not simple pulse waves?
So more like those consonant vowel waves

http://horniger.de/musik/ConsonantVowel02k.jpg

 

than these consonant vowel

http://horniger.de/musik/ConsonantVowel01k.jpg

?

(unfortunately it is impossibile to add images here...)
What resolution do these waveforms have? In other words, how many steps (Y direction) and what length (X direction)?
How many and which waveforms were there for each of the two waves?
How is the envelope constructed (ADSR?) and what resolution (x,Y) does it have?
 
Were the instruments with two LSIs necessary to generate more different sounds?
Do these have more basic waveforms?
 
2. consonant vowel Synthesis (sinusoidal)

• digital sound synthesis (with D/A converters - DACs)
• dual sinusoidal (?) staricase / stair-step waveforms (?) with 16 (?) steps
• each with own digital ASDR (?) enevlope with ? steps
• eight notes polyphony

Casiotones

• Casiotone 501, Casiotone 601, Casiotone 701, CASIO CT-1000P, CASIO MT-70
  •   Hitachi HD43517 (20 sounds ?, 8-note polyphony, 5 octaves)

 
Were the (staircase) waveforms on these instruments simply more sinusoidal?
Or were there other significant differences in the basic sound production?
And / or was there only more filtering afterwards...?
How many and which waveforms were there for each of the two waves?
 
3. consonant vowel Synthesis (simplyfied)

• digital sound synthesis (with D/A converters - DACs)
• dual rectangular waveforms 0/1 (?) with 8 (?) steps
• each with own digital linear (?) enevlope {-1, 0, +1}
• eight notes polyphony

 Casiotones

• CASIO MT-11 ?, MT-20, MT-45, MT-46, PT-7
  • Hitachi HD44140 (8 Sounds, 8-note polyphony, 3 octaves, two integrated 12-Bit ? 14-Bit ? DAC for each wave)

 
Were the instruments with what I call simplyfied consonant vowel synthesis generating their mix of two different square waves (and not more (staircase) waveforms?
So like (see consonant vowel01.jpg)?
How many and which waveforms were there for each of the two waves?
 
4. consonant vowel Synthesis (extended) ???

• digital sound synthesis (with D/A converters - DACs)
• dual (?) or more (?) rectangular (?) staricase / stair-step waveforms (?) with 16 (?) steps
• each with own digital ASDR (?) enevlope with ? steps
• eight notes polyphony
• overlay / mix of more consonant vowel Synthesis generated sound for one tone ?

Casiotones

• CASIO CT-6000
  • NEC D932G(A) + NEC D932G(B) + NEC D932G(C): 'Music LSI' (3 x melody or 2 x melody + 1 x accomp.)
  • 17-Bit (!) discrete DACs (x3)
  • (NEC D7811G: CPU, MSM6200 Key Touc Sensor), Hitachi HD6170: Rhythm generator)

 
Was this still consonant vowel sound synthesis?
What was the difference to the earlier consonant vowel synthesis?
Apparently a higher resolution?
Also more different waveforms?
 
5. SD synthesis

• digital sound synthesis (with D/A converters - DACs)
• dual more complex waveforms waveforms ?
• eight notes polyphony

Casiotones

• HT-700, HT-3000, HZ-600
  • NEC D935G, one filter
• HT-6000
  • NEC D935G x4, 8 filters

 
What differences in sound generation and filtering did these keyboards have to the earlier Casiotones and the CT-6000
 
6. All the smaller ones
 

• Casio VL-Tone 1, - Casio PT-1
  • NEC D1867G
• Casio VL-5
  • NEC D910G
• Casio PT-30, PT-50, PT-80, KX-101
  • NEC D1868G
• CASIO MT-88
  • Hitachi HD61702
• CASIO MT-18, PT-82, PT-87, EP-20
  • Hitachi HD61703
• Casio MT-36, MT-90, MT-200
  • Toshiba TMP8049P

...
 
With the earlier PTs and later MTs, the organ sound still sounds as if it is not a pure analog wave and could come from a digital synthesis.
Do these keyboards still use digital sound generation with DACs or are there several analog pulse or whatever waves simply mixed per tone?
 
 

Edited November 12, 2024 by horniger

[CYBERYOGI =CO=Windler]

HD43720 can be found in MT-36, MT-200, MT-90, CT-102. It was apparently Casio's cheapest polyphonic sound IC with percussion generator. Yamaha made a somewhat similar thing named DSG (digital sound generator, YM2163 and YM2142 (GE8)) which however employed simple pulsewave and sawtooth. But common is that both only have simple linear volume envelopes those make them sound artificial.

 

Although it can sound similar, the real Consonant-Vowel synthesis is not multipulse-/squarewave at all. The waveforms look like pixelated symmetric skyscrapers morphing into each others. The volume envelope switches through a a sequence of multiple linear sections turning gradually shallower to approximate logarithmic decay, and its clock rate is driven by the waveform clock, which makes (like playing a sample at different speeds) low notes decay slower, which mimics the natural behaviour of most plucked string, piano and mallet instruments.

 

The sinewave instruments (1000p, MT-70 etc.) sum 5 sines like a drawbar organ. Beside somewhat low resolution these were actual sines. The sound IC is very detailedly explained in the Casiotone 1000P patent (US patent 4538495) and further details in US patent 4453440, which describes digital envelope control by adding phase shifted copies of the same sine wave to avoid multiplication, and how overtones are produced by a kind of phase distortion predecessor. In this patent Casio also describes how to mix the outputs of multiple sound ICs digitally, which was never done in any other 1980th Casio keyboards I am aware of.

 

I think the main reason why Casio had so much secrecy about their sound generators was that to be first on the market, they often sold keyboards with a new sound source inside before the patenting process was finished, and so had to try hard to prevent competitors from figuring out what they did. But beside FM variants (phase distortion, using no exact sines to avoid the definition of "multiplication") they didn't copy much from Yamaha.

 

To learn more about the inner working, it is more useful to websearch for Casio patents with "priority date" close to the release date of the instrument you are looking for. Unfortunately patent names are often very generic or even grossly misleading (e.g. "System for generating sample tones on an electronic musical instrument" is about playing a demo note during preset sound selection, which has absolutely nothing to do with sound sampling) and the interesting parts are often not even that what is patented, but technical details about other parts of a described instrument. The disclosed reference implementation often also differs from the actual finished product, which can make it hard to identify the actual instrument and to figure out how it functions. But patent texts are the only officially published technical documents about Casio hardware those can be found online in large quantities, which makes them worth to look into even though they can be hard to understand. Because one purpose of patent texts is legally documenting the state of art of already known inventions (to prevent patenting twice), unlike websites they are not prone to suddenly disappear. So not every hobbyist website or forum talking about them is in the need of keeping backups online (telling a patent number is better than an URL). Patent texts can be hard to understand, but often they are the only traces that orphan hardware left in the noosphere. It is on us now to find out and index what they are about. There is are a huge online archive of unidentified technical information out there, that is useful far beyond the realms of companies and lawyers.

Edited November 13, 2024 by CYBERYOGI =CO=Windler