KVR Audio

mistertoast · Post by **mistertoast** » Fri Sep 18, 2009 2:19 pm

mistertoast wrote:Interesting. I now slide the resonance down during the cycle. It may not be what the 303 does, but it's pretty cool.

Since I have a variable-duty pulse (and the 303 doesn't), I need to decide if it's cooler to have the rez drop-off be the same no matter what the duty cycle size is, or if I want it to be dependent on the phase offset.

The advantage I see to the latter approach is the assymetry. A duty cycle of 10% would have a very short period of resonance while a duty cycle of 90% would have tons of it.

Would make PWM more lively.

mistertoast · Post by **mistertoast** » Fri Sep 18, 2009 3:48 pm

Thoughts on cutoff of HP filter? I have mine at 25Hz.

Would there be any advantage to raising the HP cutoff with higher notes?

Can I implement or emphasize a subosc with the Q of a HP resonant filter?

antto · Post by **antto** » Fri Sep 18, 2009 4:38 pm

i guess you're talking about the HP filter in the mainfilter? (the one that pre-filters the oscillator, like a DC blocker or something)
(or maybe the HP-filter of the feedback?)

i'm just tweaking like 4-5 parameters looking very carefuly at a big FFT (VoxengoSPAN) trying to get the same shape+picture of the square wave

i'm tweaking my two HP Filters (the DC-Blocker and the HP on the feedback of the main filter)
the PW of the "square"
the LP filter of the negative saw of the square (phew)
and i'm thinking i need to add one other parameter here, i'll post pictures and details when (if) i nail it ;]

another thing that is probably not a secret about analog synths (specificaly VCOs) that the osc is detuned
should we emulate this too? (i guess.. not?)
what do you think?

mistertoast · Post by **mistertoast** » Fri Sep 18, 2009 4:43 pm

antto wrote:i guess you're talking about the HP filter in the mainfilter? (the one that pre-filters the oscillator, like a DC blocker or something)
(or maybe the HP-filter of the feedback?)

i'm just tweaking like 4-5 parameters looking very carefuly at a big FFT (VoxengoSPAN) trying to get the same shape+picture of the square wave

i'm tweaking my two HP Filters (the DC-Blocker and the HP on the feedback of the main filter)
the PW of the "square"
the LP filter of the negative saw of the square (phew)
and i'm thinking i need to add one other parameter here, i'll post pictures and details when (if) i nail it ;]

another thing that is probably not a secret about analog synths (specificaly VCOs) that the osc is detuned
should we emulate this too? (i guess.. not?)
what do you think?

Good clarification.

I have two single-pole HPs inside my LP and one outside-everything single pole HP for DC killing.

They are all at 25hz cutoff now. I'm not sure how to approach thinking about them.

antto · Post by **antto** » Fri Sep 18, 2009 5:25 pm

the third one (the one outside) if you ask me - it's too much
you can usualy mimmic a Mixer or A/D-convertor or distortion with that one (since it's behind everything
recently i was talking about "internal high-pass-filtered-hardclipper" which turns out to be external, but to get the effect i used a third HP filter (like you)

anyway
here's what i did now:
TB-303square_02.jpg
top: showing the harmonics in the DAW
bottom: the waveform in the wave editor...

mistertoast · Post by **mistertoast** » Fri Sep 18, 2009 10:25 pm

antto wrote:the third one (the one outside) if you ask me - it's too much
you can usualy mimmic a Mixer or A/D-convertor or distortion with that one (since it's behind everything
recently i was talking about "internal high-pass-filtered-hardclipper" which turns out to be external, but to get the effect i used a third HP filter (like you)

anyway
here's what i did now:
TB-303square_02.jpg
top: showing the harmonics in the DAW
bottom: the waveform in the wave editor...

Are you saying I shouldn't even have an outside HP? Or that I should lower the cutoff?

Cool picture. My square finally looks right, too. And I'm loving the resonance dropping during the waveform. Looks just like the TB303. I'll post my waveform pic sometime.

antto · Post by **antto** » Fri Sep 18, 2009 11:00 pm

well, if you got the looks and behavior of the TB-303 square - keep the third HP
in my case - i will add a third HP only for the distortion (as part of it)

mistertoast · Post by **mistertoast** » Fri Sep 18, 2009 11:54 pm

antto wrote:well, if you got the looks and behavior of the TB-303 square - keep the third HP
in my case - i will add a third HP only for the distortion (as part of it)

My problem is that in I need the DC removal even without the distortion in a few situations (hard sinc, double saw osc). It's only a one pole hp. But I'll keep thinking about it.

Music Engineer · Post by **Music Engineer** » Sat Sep 19, 2009 12:01 am

antto wrote:anyway
here's what i did now:
TB-303square_02.jpg
top: showing the harmonics in the DAW
bottom: the waveform in the wave editor...

this is really damn close! and this is only the input HP, pulse-width and the half-wave lowpass? it seems, i can't get that dip in the waveform without introducing an allpass filter. how do you do that?

or wait - i just discovered that i can get that shape by using a second order highpass - is that what you are doing? i was using first order.

Music Engineer · Post by **Music Engineer** » Sat Sep 19, 2009 12:42 am

kunn wrote:i think this could explain the behavior:
Code: Select all
|      /|    |    /  |
|     / |    |   |   |
|    /  |    |   |   |
|   /   | => |   |   |
|  /    |    |   |   |
| /     |    |   |   |
|/      |    |__/    |
tanh(x*saw)

aha - so the 303 obtains the square by heavily overdriving some tanh-shaped nonlinearity, yes? this would indeed explain the behavior.

Music Engineer · Post by **Music Engineer** » Sat Sep 19, 2009 12:45 am

rv0 wrote:I know someone who has the schematics on his wall, he knows exactly what is going on in there. He's too busy to help out here, but a small, clear question won't hurt.

hmm...maybe we should then save that question for later

mistertoast · Post by **mistertoast** » Sat Sep 19, 2009 12:49 am

Antto, I have the feeling that we're going to have some people failing A/B taste tests with your synth soon.

Robin, my osc is perfectly square (not even bandlimited). I'm getting the characteristic TB-303 shape, but not as close as Antto's. What "behavior" in particular are you talking about regarding the tanh? The shape of the square or the resonance drop on one side?

mistertoast · Post by **mistertoast** » Sat Sep 19, 2009 12:50 am

Robin from www.rs-met.com wrote:
rv0 wrote:I know someone who has the schematics on his wall, he knows exactly what is going on in there. He's too busy to help out here, but a small, clear question won't hurt.
hmm...maybe we should then save that question for later

It's like having a genie and one wish.

rv0 · Post by **rv0** » Sat Sep 19, 2009 1:05 am

mistertoast wrote: It's like having a genie and one wish.

but yeah, keep this one as a helpline, in case you guys are really puzzled about something

mistertoast wrote:Antto, I have the feeling that we're going to have some people failing A/B taste tests with your synth soon.

2nd that

very impressive.

antto · Post by **antto** » Sat Sep 19, 2009 3:46 am

i actually tweaked it a little more after that image i posted, now it's a little closer
but that's only compared to 1 note, when the TB-303 goes higher for example, the difference starts growing a little bit ;]

Robin: nah, my first attempt at 303 waveform was exactly the same - allpass (long long ago) but it didn't work, so i never ever put an allpass filter there ;]
i achieved this waveform by tweaking 5 things:
erm, first, my original waveform isn't a perfect sawtooth (it has all the harmonics with "perfect" amplitudes tho) .. the phases are a little changed at generation time (which might be doing me a bad favour now, haven't checked that)
1. MainFilter: 6dB HPF (on the input signal, DC blocker)
2. MainFilter: 6dB HPF (filtering the feedback)
3. Pulse Width (phase difference between both sawtooths)
4. Osc: 6dB LPF filtering the Negative Saw (introduces additional phase marmalade on the Pulse Width, naturaly)
5. Osc: Positive Saw: Amplitude (it's less than 1.0)

gonna post a 1:1 compare audio with rv0's "slide" audio, where i calibrated this, you'll see how close the square looks on the first low tone, then not so close on the high tones (and also, the TB-303 gets detuned)

Open303 - open source 303 emulation project - collaborators wanted