Open303 - open source 303 emulation project - collaborators wanted

[mistertoast]

Careful there. Your filter is probably not analog, and therefore can possibly only attempt to model the analog behavior up to some finite frequency. By feeding it with something that isn't band-limited just means feeding it stuff (namely alias) that isn't there in the analog version.

Yes, this is my first non-bandlimited synth and I'm taking the problems as I go. But, honestly, I'm really enjoying doing this way and am loving the sound. I'm really just making what I want to make for my own enjoyment (but I would, of course, love to finish it and get it out for the purpose of making ass-kicking music).

You pick a road and you deal with the consequences. That's one reason the differences between approaches get so exciting. I chose this approach: "On the topic of oversampling, it is speculated that the Nord synths simply generate "naive" waveforms at a very high samplerate; then downsampling (with a proper LPF) is performed at the output. This has the added advantage that higher-order modulations will not alias noticably (for instance, ring mod and FM)." (from http://ccrma-www.stanford.edu/~dfl/220b/hw4.htm)

Sure, not all approaches will be successful. But I've done bandlimited before and I'm really learning a lot going down another path.

I do most certainly appreciate you pointing out some of the dangers I face.

[mistertoast]

Look at this from the MC-202 scoping. Top point curvy bottom point sharp.

[antto]

i'm not sure how the scope works..
this basicaly looks like a pure ramp, being a little lowpass filtered

btw, the tuning coefficient (for rv0's TB) is somewhere around 0.992.. gotta increase the precision a little

[mistertoast]

i'm not sure how the scope works..
this basicaly looks like a pure ramp, being a little lowpass filtered

btw, the tuning coefficient (for rv0's TB) is somewhere around 0.992.. gotta increase the precision a little

I wanted to point out the asymmetry. And in the pulse it's interesting, too. What I wonder is whether that's the actual raw osc (as it comes out of the circuit). Not sure what the guy did to capture it. There could be a bit of low-pass inherent in the circuit? Maybe all analog oscs act that way? Some implied LPF in them? I don't know. And I don't know if the LPF varies per note. I'm not a circuit guy and I'd love to know. Maybe that's what my "raw" osc should look like.

But if we're talking about asymmetry in the oscs--there's some right there. The square is similar. http://analogsweden.com/blog/wp-content ... c00850.jpg

Cool on the tuning coefficient. Is that per not or octave?

[valhallasound]

http://analogsweden.com/blog/?tag=mc-202

Look at the saw and its asymmetry. And look at the edges of the pulse and square. Just thought it was interesting, as I don't know the MC-202 well at all, but it has high ratings over at Vintage Synth. And it is a Roland analog design. http://www.vintagesynth.com/roland/mc202.php

The MC-202 is very close to the SH101. Curtis CEM3340 VCO with flip-flop sub oscillator, filter based around the IR3109 chip, VCA based around the BA662.

The asymmetry of the saw looks like some asymmetrical soft clipping, in conjunction with whatever the filter setting is at. I doubt that the probe was directly taken from the CEM3340 outputs. The filter and the VCA in the MC202/SH101 have some audible nonlinearities.

The MC-202 and SH-101 filter is a cascade of 4 one-pole filters with equal frequency (buffers between the stages), with back-to-back diodes in the feedback path. In this sense, it is pretty close to the filter in the TeeBeeFilter.h, which has 4 one-pole sections, and a single nonlinearity for the last filter stage, which results in a distorted feedback path.

The TB-303 filter would be better modeled by 4 one-pole filters in series, but with different frequencies for each filter, as the lack of buffering between filter stages in the diode ladder results in different cutoff frequencies for each stage. Such a system will have different migrations of the poles with feedback, which results in much higher feedback gain being required for oscillation. Since there are lots of nonlinearities in the filter, the higher feedback gain results in much more distortion, giving the nasty squelchy filter sound.

Sean Costello

[mystran]

Yes, but you seem to be doing an acid synth of some kind. Didn't mean to say what you were doing, but you have been posting some nice sounds.

Nah, it's a bass synth.

[antto]

..Cool on the tuning coefficient. Is that per not or octave?

you can scale your pitch with this coefficient, when the pitch is in the form of 1Volt/Octave
i have this natively in my synth

it's sort of like this:
+5V = A4 key
+5.25V = C5
+6V = A5

then i multiply this with the coefficient
then i can convert this 1V/Oct pitch to frequency in Hz (so i can use it for the Osc) with the formula:
freq = 440 * pow(2,(volts - 5.0));

EDIT: the coefficient was actually 0.9917

[rv0]

I wonder if the tuning quirks are per-instrument (different among different hardware 303s) or whether there's some sort of distinctive tuning "errors" caused by the circuit design.

i guess a little of both ;]

exactly..

there's trimpots inside
TM 6 controls the CV..

you have to adjust it carefully because otherwise the scaling wont be as correct for different tunings.
there's a procedure for it in the service notes.
i've noticed low tuning always gives terrible error
i just checked btw, there was quite an explanation about VCF envelope modulation & the gimmick circuit.

Checking the service notes will give you an idea of the variable factors in the 303 design.

It would be a little insane, yet cool, if you could somehow flip the lid of the software clone and adjust those trimpots yourself .. so several instances of the instrument would sound different (add user-triggerable random function for the positions)



About that tuning.. Maybe I can record another sample...
I can do the full range with my midified 303? With different tuning settings.. (i tune my 303 with a guitar tuner: korg ca-20)

[antto]

...Checking the service notes will give you an idea of the variable factors in the 303 design.

damn, is that in the docs that can be found on the net?
i think i don't have it, grrr..

[mistertoast]

"It would be a little insane, yet cool, if you could somehow flip the lid of the software clone and adjust those trimpots yourself .. so several instances of the instrument would sound different (add user-triggerable random function for the positions)"

Yeah, not out of the question. Some soft synths have "internal" panels.

Also, I remember Morgana advertised that each sold copy was unique. I was never sure how much I liked that idea. What if mine sounded the worst of all of them?

[mistertoast]

http://analogsweden.com/blog/?tag=mc-202

Look at the saw and its asymmetry. And look at the edges of the pulse and square. Just thought it was interesting, as I don't know the MC-202 well at all, but it has high ratings over at Vintage Synth. And it is a Roland analog design. http://www.vintagesynth.com/roland/mc202.php

The MC-202 is very close to the SH101. Curtis CEM3340 VCO with flip-flop sub oscillator, filter based around the IR3109 chip, VCA based around the BA662.

The asymmetry of the saw looks like some asymmetrical soft clipping, in conjunction with whatever the filter setting is at. I doubt that the probe was directly taken from the CEM3340 outputs. The filter and the VCA in the MC202/SH101 have some audible nonlinearities.

The MC-202 and SH-101 filter is a cascade of 4 one-pole filters with equal frequency (buffers between the stages), with back-to-back diodes in the feedback path. In this sense, it is pretty close to the filter in the TeeBeeFilter.h, which has 4 one-pole sections, and a single nonlinearity for the last filter stage, which results in a distorted feedback path.

The TB-303 filter would be better modeled by 4 one-pole filters in series, but with different frequencies for each filter, as the lack of buffering between filter stages in the diode ladder results in different cutoff frequencies for each stage. Such a system will have different migrations of the poles with feedback, which results in much higher feedback gain being required for oscillation. Since there are lots of nonlinearities in the filter, the higher feedback gain results in much more distortion, giving the nasty squelchy filter sound.

Sean Costello

Great info. Thanks.

[rv0]

...Checking the service notes will give you an idea of the variable factors in the 303 design.

damn, is that in the docs that can be found on the net?
i think i don't have it, grrr..

then get it while you can:
http://users.pandora.be/darffader/rv0te ... eNotes.pdf

[mistertoast]

I bet you're right. that MC-202 scope shot is probably just hooked up on the output. I would love to see some raw oscillator shots.

[Kritical Audio]

@rv0 Are you going to add shuffle to your sync unit plug ? That would be something!

[antto]

"It would be a little insane, yet cool, if you could somehow flip the lid of the software clone and adjust those trimpots yourself .. so several instances of the instrument would sound different (add user-triggerable random function for the positions)"

Yeah, not out of the question. Some soft synths have "internal" panels.

Also, I remember Morgana advertised that each sold copy was unique. I was never sure how much I liked that idea. What if mine sounded the worst of all of them?

yeah, this is why i have a "back-panel" of my synth
while in the beta version, it's pretty empty, now there will be such tuning controls, like 2 controls for tweaking the two HP filters (which change the sound & shape of the filter a lot)