KVR Audio

brambos · Post by **brambos** » Fri Jul 31, 2009 7:24 am

No real TB-303 here to check... does anyone have a pointer to a description of the TB's envelope behavior (notably the attack)?

The EG's attack seems to be relatively short, but I get the feeling the ENVMod alters the attack duration.

antto · Post by **antto** » Sun Aug 09, 2009 9:14 pm

erm, no
there's no attack, the TB-303 envelope has only decay
EnvMod doesn't alter the "attack" (since there ain't such) ..it doesn't alter the decay neither..
EDIT: um.. i ain't got a TB-303 neither, but i'm working on a something that has something to do with a TB-303, and i've analyzed it a lot..

mistertoast · Post by **mistertoast** » Sun Aug 09, 2009 9:33 pm

Obviously there is some attack, even if you can't set it.

I've read that it's 3 milliseconds. I don't know if any controls modify that. I don't know if it's different with ENV changed or if it's affected by ACCENT.

antto · Post by **antto** » Sun Aug 09, 2009 9:43 pm

ok, so we're talking about the VCA?
the GATE itself has a little bit of filter or something - it's not "instant" otherwise it would "click"
but that's not an "attack" in the ADSR-sense..
"3 milliseconds" ... yeah, that's like a "de-clicker"

the Cutoff Envelope doesn't have any attack

ACCENT actually adds a tail on the VCA on accented notes, i realised this just a few days ago..
..and the "gimmick" effect, which i think we ain't talking about, are we?

mystran · Post by **mystran** » Sun Aug 09, 2009 9:56 pm

brambos wrote: The EG's attack seems to be relatively short, but I get the feeling the ENVMod alters the attack duration.

I got this with quick Google: Devilfish page says "TB-303 attack was ~3 msec." in relation to the adjustable attack mod, which based on quick empirical test seems to sound reasonable.

For really clean and easy to read schematic, you can look here:
http://machines.hyperreal.org/manufactu ... chematics/ (see number 5 if you wanna try to make sense of it)

Other than potential current limitations, env mod doesn't indeed seem to affect anything but cutoff which is lowered for larger modulation depth (except is that dot a connection to resonance or just dirt in scan?). That's what I got out of it in 15 second with my absolutely terrible circuit analysis skillz (so please don't trust me, hehe).

antto · Post by **antto** » Sun Aug 09, 2009 10:13 pm

envmod > resonance?
wasn't that statement "accent > resonance" (on accented notes only)
tho i don't believe it'll have a huge effect anyway
and i haven't observed any stronger resonance on accented notes (with accent knob turned high) in the audio materials i have..

infact, looking at the audio, this "attack" you're saying is ~3ms (i'm talking about the VCA) is actually 20-35 samples long (at 44100Hz) so guess what miliseconds that is ;]
and i think it's not intended in any way to be an "attack" .. i guess the GATE is meant to be instant, but this "smoothing" is due to the way analog work (excuse me, i can't express myself good)
i mean.. it was a cheap synth, why would they put an attack anyways..

mistertoast · Post by **mistertoast** » Sun Aug 09, 2009 10:43 pm

So by 20-35 samples in at 44100, it's full volume? Great info. Thanks! Does the ramp up look linear? I've read that the decay is exponential, but you can never trust anything you read about the 303.

mystran · Post by **mystran** » Sun Aug 09, 2009 11:00 pm

Just so we're talking about the same thing: there's two EGs in a 303. I was talking about the one that controls the filter, since env-mod was mentioned.

antto · Post by **antto** » Sun Aug 09, 2009 11:03 pm

it's really hard to tell if this short 20-35sample long VCA "fade-in" is linear or not.. i think such a small detail is insignificant
the "fade-in" itself is a good thing (de-clicker) linear should be fine
my calculator says like 0.000795 seconds for the longer case - 35samples
that's like less than 1 milisecond..
it'll sure be heard tho if you have all C3 notes on a pattern and you "push the tempo" (like a fast trance gate, so fast that it's not usable in acid music) then the clicking might be heard if it was instant, so even tho that's like 35 samples, they have effect, depends on you..

exponential decay? well, again, you have to specify whether you're talking about the VCA decay or VCF..
it's hard to tell about the VCA decay, i really need to put my hands on a real one (or have a specific pattern recorded by someone who has one) to analyze it (no luck yet)
the VCF decay is exponential, yeah..

"but you can never trust anything you read about the 303."
true, and this i think is because not only that two tb303 are not 1:1
..but also, if you remove the front panel, there are "adjustment" knobs or something, that can change (or tune) the most important knobs on the front panel (tune, cutoff range, envmod range, accent amount, and i think even the resonance)
and some people have messed with these (or bought a machine that was "modified" by the previous owners)
so it's hard to tell, and there will never be a perfect clone, because everyone who has a TB-303 would say "they all have little defects, but mine is the perfect one, all clones should be modeled after my TB-303, not the others"

antto · Post by **antto** » Sun Aug 09, 2009 11:08 pm

mystran wrote:Just so we're talking about the same thing: there's two EGs in a 303. I was talking about the one that controls the filter, since env-mod was mentioned.

exactly
one EG for the VCA, and the other for the filter cutoff ;]
and things get complicated, since when there's accent (and an accented note), some of the VCF EG goes to this so-called "gimmick" circuit..
and, hm.. it also goes to the VCA (adds the punch)