A 'synth' with no knobs or sliders?

[k brown]

In researching the Trautonium, I came across a whole class of pre-synth electronic instruments that were very popular in the late 40's through the early 60s that were designed to be attached below a piano (or organ) keyboard to offer an alternative 'voice' for the right hand. The simplest ones had no knobs, just push buttons for selecting pitch ranges and tone colors.

One of the earliest was the Hammond Solovox. The similar, but much lighter Clavioline became much more popular because of it's greater prortability. I chose to model this on the Solovox simply because it was the only one I could find a service manual for with a good deal of tech specs in it (in English, anyway ) .

Volosox_nk copy.png (38.45 KiB) Viewed 18814 times

https://ln.sync.com/dl/62c9b9f70/bbjdaw ... 3439840004

Would love to have done as full and accurate emulation as I cold muster of the most sophisticated instrument of this type, the Odioline, but I wasn't able find any details about it's voicing filters. I suppose the only option would be buy Soniccouture's sampled Ondioline and try to 'voice' it by ear.
http://www.soniccouture.com/en/products ... ondioline/

I did include two features of the Ondioline - volume can be increased by applying aftertouch, and the 'Fast Attack' setting has an alternate mode - 'Percussive', similar to the Ondiline's touch-wire.

UPDATED to v1.002 7/20/20.

[Spogg]

I’ve grown a fascination for these ancient precursors to modern synths and this is great, and of course looks good too!

One question or observation (I only tried the plugin)…
When I turn Vibrato off I can still hear some Vibrato. Is this authentic and therefore intentional?

[k brown]

Authentic - pretty good idea, I thought too; the service manual actually stated this was to prevent the 'too-perfect' sound of electronics - sound familiar? Of course, I had to guess at how much residual vibrato is left when it's off.

With your electronics background, you might find the schems interesting - I was impressed with how much tonal flexibility was achieved by so few passive parts on the input to the preamp section. You might also find that I've bollocksed something up in my attempt to emulate them.

Solovox-technical.pdf.zip

(3.76 MiB) Downloaded 1159 times

[Spogg]

It was great to be able to see the circuit diagram!

Some comments on the Model L:

The highest (Master) frequency was a sine wave. When “Mute” was selected it would change to a half-wave rectified version (positive half-cycle only).

The Volume control was actually a 7-way switch (horrible idea).

Vibrato was a soft square wave (longer rise and fall times than a perfect square).
I can see no evidence that when Vibrato was off there would still be some present. Cleverly, they introduced a tuning capacitor, so when Vibrato was on, the pitch would be bipolar around nominal and when off the capacitor would cause a static shift of pitch. This meant you should tune the instrument with Vibrato off and the centre pitch would still be correct when Vibrato is turned on.

On the slower attack there is likely to be some distortion at the start of the attack. This is because the VCA works by changing the bias on the control valves. This might be nice because there would be some higher harmonics present to maybe simulate bowing, like on a cello.

To properly simulate the tone settings would require a very complex simulation of the total reactance caused by the L-C networks and transformers. So without impulse responses and a convolver we’re always guessing. But to my ears your solution sounded nice.

I couldn’t resist looking at the Model J circuit too. It’s very different! It uses a master sinewave oscillator and 5 synchronised sinewave sub oscillators. The system works more like an additive synth in that respect, to get the different mixtures for the registers. I bet it sounded nothing like the Model L. What I thought was a remarkable but crazy thing was the electro-mechanical vibrato mechanism. It used a resonating vibrating reed energised by a low frequency oscillator.

[k brown]

Wow! You able to determine all that just from the schematic? I figured it would lend some insight, but not this much.

If I'm interpreting what you're saying correctly, the top octave on it's keyboard would have produced sine waves, that switch to half-rect sine with Mute on? Somehow octave-dividing this master for the lower octaves converts it to something closer to square? Perhaps I could get a little closer to this with a little inverse key-tracked filtering on the final output?

From the text description of the vibrato I figured it might be something like a 'rounded' square - maybe a slightly filtered trapezoid would be closer (I've got such a wave in my wavetable collection).

In my cursory look at the schem, I didn't notice the 7-way switch for volume - pre-digital zippering!

Is it possible the reference to some vibrato still present with Vibrato off was actually a bit of detuning going on?

Interesting about the attack distortion - maybe it sounded a bit like the 'chiff' of the Hammond organ note attacks?

Also interesting that the earlier model was basically sine-additive, since that's what their organs were.

Great info - thanks for 'digging in'.

[Spogg]

It was fun for me! Really took me back.

The top notes were sine or rectified sine (Mute). The rectified sine was then overdriven to create suitable switching pulses for the first divider. Subsequent dividers had a nice square pulse to work with.

The sine was not just for the top octave, it played at whatever pitch was played on the keyboard. The dividers then provided square waves for the sub harmonics (which unlike Sala’s were always octave intervals). So you get a mixture of sine/half wave and octave square waves to mix together over the whole key span, before filtering. So no need for key tracking stuff. Just provide the sine/half wave for the “master” oscillator; the one with the highest fundamental.

The “Vibrato when off” can’t have been beating because all the sound generators were phase-locked to the master oscillator. Maybe it was an unintended “feature” due to wiring or earth impedance issues. Certainly the circuit itself doesn’t provide for it. It only takes a very small signal to notice Vibrato.

Regarding the distortion during initial attack it’s hard to imagine the resulting sound. To simulate it you’d have to soft-clip the signal a bit at the start of the attack and remove it by the time sustain level was reached. And it would vary from tube to tube and with vacuum tube age because operation would be in the non-linear range. Also the attack sound would depend on the mixture of course. You wouldn’t notice it with a single square wave selected for example, but for a sine/half-wave it would be very apparent due to the wave shape changing dynamically.

[k brown]

Just provide the sine/half wave for the “master” oscillator; the one with the highest fundamental

Sorry to be so thick, but how would this translate to FS?

[Spogg]

Use one of Martin’s lovely sine oscillators, and for the half-wave pass the sine signal through a stream max prim (one input set to zero) so it acts like a diode. You’ll get some aliasing on the half wave of course.

[k brown]

No, I know how to do that, I didn't what you meant by 'master oscillator'; there's no octave division going on here. Do you mean use the sine/rect sine for the Soprano register?

[Spogg]

Oh sorry!

Yes I meant for the Soprano register.