KVR Audio

antto · Post by **antto** » Mon Sep 14, 2009 6:47 pm

i told you ;P~

mistertoast · Post by **mistertoast** » Mon Sep 14, 2009 6:49 pm

antto wrote:i told you ;P~

Remember you've been living this stuff a lot longer than we have. Give us some time to catch up.

antto · Post by **antto** » Mon Sep 14, 2009 6:52 pm

Robin from www.rs-met.com wrote:could it be that the standard RC envelope goes through some nonlinearity with a cubic component in the original circuit?

not sure, but since i got the same behaviour, i guess it's a linear decay of some sort, with a shaper that resembles my curve
<OR>
a decaying RC (as you said) with a shaper, um, that's entirely what you said

blah
but if it's the second one, whatever the Decay knob is set to, the whole thing must look like a waveshaped linear decay (since, the output is exactly this)

antto · Post by **antto** » Mon Sep 14, 2009 6:55 pm

mistertoast wrote:
antto wrote:i told you ;P~
Remember you've been living this stuff a lot longer than we have. Give us some time to catch up.

yeah
the discussion here looks like we're argueing (damn, how do you spell that word) with each other about most details

this is quite good actually, since we understand each other sooner or later and eliminate the weak spots ;]

Music Engineer · Post by **Music Engineer** » Mon Sep 14, 2009 7:00 pm

antto wrote:the discussion here looks like we're argueing (damn, how do you spell that word) with each other about most details

this is quite good actually, since we understand each other sooner or later and eliminate the weak spots ;]

yep, that was my intention with this project and thread. to discuss and debate such details and hopefully at the end arrive at a good implementation (or more than one

- i just recently became aware of your venom)

mistertoast · Post by **mistertoast** » Mon Sep 14, 2009 7:05 pm

I don't think we're arguing. We're just trying to figure out what we're talking about.

I think another worthy goal would be a document of statements about the 303 backed up with actual samples.

nollock · Post by **nollock** » Mon Sep 14, 2009 7:10 pm

antto wrote: PS: my FFT (sonogram) is linear
my filter has a 1V/KHz response

ah, ok

both waveforms look the same on the FFT (mine, and the original)
so, i'm right ;]

I dont doubt it having listed to the sample you posted earlier. I could hardly tell the two channels apart. And really wouldnt know which is which if you hadnt said so beforehand.

Excellent work!

Music Engineer · Post by **Music Engineer** » Mon Sep 14, 2009 7:15 pm

mistertoast wrote:I think another worthy goal would be a document of statements about the 303 backed up with actual samples.

yes, that would great

Music Engineer · Post by **Music Engineer** » Mon Sep 14, 2009 7:16 pm

nollock wrote: And really wouldnt know which is which if you hadnt said so beforehand.

Excellent work!

+1. the only cue for me was the (slight) noisefloor in the original.

antto · Post by **antto** » Mon Sep 14, 2009 8:25 pm

nah, there is difference, but it's fully tweakable
left channel (my synth) has little bit more HP filter (was calibrated over another audio sample)
this new audio i got has almost no (external) HP filtering, you can see it clearly
now i just have to turn a (hidden) knob to match this pretty easy (was too lazy)

also, mine has a little too much resonance, you can even see some noise (that's a bug in the filter that i haven't yet fixed)

btw i just wrote myself a little tool that helps me convert patterns from audio files
actually, just by looking at the waveform, i can tell if there is a slide, accent and so on, the only harder thing is to distinguish the exact notes played
so now i just open the audio file in SoundForge, i zoom in on a note, to see the waveform, then i select 1 cycle of the waveform at that note, and watch the selection length in samples (for example ~604 samples at 44100 SR)
then i enter 604 into my tool and it says: D-2
i tested it on some odd patterns i got from here and there, it works ;]

valhallasound · Post by **valhallasound** » Mon Sep 14, 2009 9:34 pm

antto wrote:
Robin from www.rs-met.com wrote:could it be that the standard RC envelope goes through some nonlinearity with a cubic component in the original circuit?
not sure, but since i got the same behaviour, i guess it's a linear decay of some sort, with a shaper that resembles my curve
<OR>
a decaying RC (as you said) with a shaper, um, that's entirely what you said

I haven't followed the thread this closely, but I would presume that any envelopes in the 303 go through an exponential convertor before modulating the filter freq, as this is how most VCFs I have seen worked. Could the curve be an exponential envelope, going through an exponential convertor?

Sean Costello

antto · Post by **antto** » Mon Sep 14, 2009 9:49 pm

valhallasound wrote:
antto wrote:
Robin from www.rs-met.com wrote:could it be that the standard RC envelope goes through some nonlinearity with a cubic component in the original circuit?
not sure, but since i got the same behaviour, i guess it's a linear decay of some sort, with a shaper that resembles my curve
<OR>
a decaying RC (as you said) with a shaper, um, that's entirely what you said
I haven't followed the thread this closely, but I would presume that any envelopes in the 303 go through an exponential convertor before modulating the filter freq, as this is how most VCFs I have seen worked. Could the curve be an exponential envelope, going through an exponential convertor?

Sean Costello

yes, it could be (that's what i meant when i said a RC decay thing with a shaper on it)

mystran · Post by **mystran** » Mon Sep 14, 2009 10:38 pm

valhallasound wrote:
antto wrote:
Robin from www.rs-met.com wrote:could it be that the standard RC envelope goes through some nonlinearity with a cubic component in the original circuit?
not sure, but since i got the same behaviour, i guess it's a linear decay of some sort, with a shaper that resembles my curve
<OR>
a decaying RC (as you said) with a shaper, um, that's entirely what you said
I haven't followed the thread this closely, but I would presume that any envelopes in the 303 go through an exponential convertor before modulating the filter freq, as this is how most VCFs I have seen worked. Could the curve be an exponential envelope, going through an exponential convertor?

Yeah. In addition to that the antilog is biased depending on the position of the env-mod knob such that reference cutoff drops when modulation is increased. This is actually explained in the service manual thingie that you find as the first hit when you google for "303 schematics"

mistertoast · Post by **mistertoast** » Mon Sep 14, 2009 11:40 pm

I'm going to start on a document that's a discussion of the TB-303. My current thinking is that it will be in FAQ form (Q&A), and will be evidence based.

By "evidence based," I mean that assertions will be backed up with data such as samples from the 303.

Let's get started.

Q. Do accented notes have the same filter resonance (Q) as unaccented notes? Background: Some references say that one of the characteristics of accented notes is that they have higher resonance than unaccented notes.

Proposed Test: Compare two sets of notes. One set is accented and one is unaccented. ACCENT and DECAY should be turned down completely (fully counter-clockwise).

Now, before anyone creates such a file with their TB-303, does anyone have any comments on the test?

Music Engineer · Post by **Music Engineer** » Tue Sep 15, 2009 12:01 am

@mistertoast: great iniatitive. looking forward to see this grow. one of my questions would be:

Q: is the filter envelope added (by some amount) to the amplitude envelope?

proposed test:
-set Cutoff to 100%, Reso to 0%, EnvMod to 0%, Decay to 0%, Accent to 0%
-play some (long) unaccented notes and watch the amp-env
-set Decay to some higher value, play notes and watch if the amp-env has changed (longer initial decay)

repeat the same test with accent set to 100% and with accented notes (in order to spot whether accent influences this)

i have observed the described behavior in the ABL and autodafe's samples.

Open303 - open source 303 emulation project - collaborators wanted