Idea for product (Stringer Module)

[themilford]

SO, a sub-$500 desktop Stringer with Midi and Patch memory... selectable filter (para/poly). Ensemble effect (4xBBD) with flu control, I/O, and maybe memory as well. I've had RS-09s in the past so I'm thinking a similar control layout to that with the footages, envelope, tone, etc. Maybe an organ section as well. I also have a Crumar Trilogy which has a nice sounding string section.

looking into ways to do the oscillator(s) (VCO vs. DCO), divide-down, etc.

Should it be small, say MFB or DSI sized? Or should it have a nice comfortable layout with full-sized sliders and switches?

[garranimal]

I would have to agree with keeping the price down, and therefore the functions simple:
- Two simple sawtooth oscillators with one fixed octave below the other.
- An external input probably should be there with a line-in attenuator, and mixer pot to adjust the amount of mix between external/internal sounds.
- Ability to pitch bend the whole sound up/down to a selectable interval up to an octave like the VP-330.
- Simple fixed string filter circuit like that in the VP-330, RS-09, Solina. If there's budget I like how the VP-330 has a lowpass filter to brighten or darken the sound.
- Three BBD circuits are very good, but four are even better.
- Simple amount, rate controls for the chorus.
- A simple Attack/Release envelope scenario like the old string machines.
- The MIDI board would be the only digital circuit needed there: note on/off, pitch bend, attack/release amount, filter amount, chorus settings.
- I wouldn't put a phaser into this unit to keep price down. Let folks pick their own favorite phaser/flangers to use.

I think it would be really cool to be able to sweep the pitch of one of the oscillators against the other, or set it some interval other than octave below. I did this on my software clone of the VP-330 and it sounded pretty damn cool.
The good chips like SAD512, TCA350(z)/1022, and Roland MN300X chips are in very limited short supply. My JH Triple Chorus uses TCA1022s and it sounds superb. BBD availability could make-or-break how much units get produced.
Patch save features dictate an additional multiplexer/data-bus/memory-chip setup which may make it cost prohibitive when competing with good (and free) soft-synths like the Stringer VST. If the unit were simple enough like the old stringers then patch save wouldn't really be necessary IMHO. I would be willing to pay around $300-400 for such a unit, which is the trick. With surface mount technologies I have now idea how much it would cost to produce and still profit.

[Ashe37]

I'd skip the selectable filter. There are too many things you'd have to do to be able to switch all the voices from separate filters to a single filter...

I've been playing with the idea of a 1U rack module, not a keyboard or desktop module. Unless you're using direct pot control of the features on the unit, the circuitry will already be there to have patch memory. The 'memory' part itself is largely inconsequential as the 'MIDI board' has enough memory to save patches on it already.

Parts availability would be considered in the design. Since the filter is a 24 Db lowpass, the filter is likely to be a discrete filter, OTA-based.

[Mooger5]

Many if not most of the stringers had all the notes synced to one master oscillator. It was a key part of the sound as much as the chorus and paraphonic filter. Read this http://www.organservice.com/crm/topdividers.htm to have an idea of how the thing worked and availability of the old parts.

Maybe the best thing to do is to go DSP (at least for that specific section and VCA and EGs) for cost-effectiveness, analogue for the filter and BBD FX for purism...

[Ashe37]

why would that be the 'best thing to do'? it kinda defeats the purpose of making a stringer.

[Mooger5]

Because of the sub-500 price.
I don´t think there would be much objection regarding the use of DSP to emulate the divide-down techniques. When all notes in the keyboard are phase-locked to a master oscillator there isn´t the harmonic richness associated with analogue polys. That´s how stringers work, and why the complex choruses and filtering were added.
So it might be worth to use DSP for that instead of sourcing the chips. Then add analogue VCF and BBD FX section.

[Ashe37]

Sourcing and programmoing a small number of DSPs is not going to be any cheaper....

[Mooger5]

Yeah I don´t know about that, so I said "maybe"

[Stab Frenzy]

Sourcing and programmoing a small number of DSPs is not going to be any cheaper....

I beg to differ.

[Ashe37]

i just see using a dsp in the design as defeating the purpose for the kind of small production run i'm thinking about. maybe an fpga...

[redchapterjubilee]

Price would be a large factor, at least for me. It would have to offer something that you couldn't just spend $500 to get vintage or $150 to get close enough to with the G-Media stringer VST. Something in-between there. Digital oscillator wouldn't bother me so much, but you'd pretty much HAVE to have BBD chips for the effects. Like a Shruthi or Meeblip kind of box.

[Sir Ruff]

Price would be a large factor, at least for me. It would have to offer something that you couldn't just spend $500 to get vintage or $150 to get close enough to with the G-Media stringer VST. Something in-between there. Digital oscillator wouldn't bother me so much, but you'd pretty much HAVE to have BBD chips for the effects. Like a Shruthi or Meeblip kind of box.

Yeah, I think the osc is probably the least important aspect, and using a digital osc would give you more options. Shruthi is a good example for a low-cost unit though I keep wondering why they didn't make it polyphonic. Surely one voice is the same as many on a DSP? Perhaps it was since then you would then need multiple filters. But in the case of a paraphonic stringer, that would be "normal".

[redchapterjubilee]

Right. Which is why I think a paraphonic Shruthi box with one filter but with BBD chips would be easily possible. But it might push the price a bit. But if you could provide your own choice of filter, like those do now...that could actually be really interesting.

[Ashe37]

Price would be a large factor, at least for me. It would have to offer something that you couldn't just spend $500 to get vintage or $150 to get close enough to with the G-Media stringer VST. Something in-between there. Digital oscillator wouldn't bother me so much, but you'd pretty much HAVE to have BBD chips for the effects. Like a Shruthi or Meeblip kind of box.

Yeah, I think the osc is probably the least important aspect, and using a digital osc would give you more options. Shruthi is a good example for a low-cost unit though I keep wondering why they didn't make it polyphonic. Surely one voice is the same as many on a DSP? Perhaps it was since then you would then need multiple filters. But in the case of a paraphonic stringer, that would be "normal".

unless you want each voice to sound the same, say, one voice is x dsp load, two voices is going to be between 1.75 and 2x the amount of dsp, 3 voices would be 2.5-3x, etc hence why adding an oscillator to a patch on the Blofeld, Virus, etc, cuts into your polyphony.

The Shruthi really doesn't have a DSP. There's a reason why so many people compare it to an ESQ- because its the 'lo fi samples plus analog filters' method of synthesis. The Shruthi is basically an arduino...

[HideawayStudio]

Alot of people would have fits if the master oscillator was a DCO...

Why so? String machines oscillators are about as lifeless as you get - they're almost all divide down so you might as well be using a DCO.

NO!!!!!!!

Well ok, sort of yes!... here's the deal - this is only true if the stringer/synth makes use of a top octave generator phased locked to a single master HF oscillator and, worse still, when it is then modulated by a single vibrato LFO - eg. ARP Omni.

If a divide down architecture makes use of 12 discrete free-running top octave oscillators (and better still with independant vibrato LFOs, and even better still, the dividers are discrete) then the story is -very- different... the following divide-down instruments are anything but lifeless... Novachord, Philips Philicorda, Eminent 310U, PS3100 - all use the techniques I've cited to gain wide, detuned, free-running analog goodness - the Novachord went one stage further and utilised monostable dividers (the bistable hadn't been invented in 1938!) and hence no two cycles were quite the same!

Divide down can be really cool (believe me it can!) but it must be implemented in a musically sympathetic manner and this is often very expensive. Very sadly when the bean counters came along most divide down synths suffered terribly - and the very last thing you want is a totally phase locked stringer!

Incidentally, you might be wondering why the Omni-2 therefore sounds half decent - well the secret is that it's triple parallel BBD chorus unit makes use of 3 free running LFOs, each running at half the frequency of the other and hence there is a ton of movement in the FX section which sort of compensates for the sterility of it's tone generation stages!!