KVR Audio

mistertoast · Post by **mistertoast** » Sun Sep 13, 2009 5:56 pm

I see, so you have a tool that lets you plot the peak through time. Whew! I thought I'd never understand what the hell you were doing.

antto · Post by **antto** » Sun Sep 13, 2009 5:56 pm

Kriminal wrote:
antto wrote:
you know, some say "logarithmic envelope" or "exponential" or "linear"
this is exactly the same, only it doesn't seem to be any of these

in other words, if you look how the cutoff (with a lot of resonance) travels on the FFT (when you're looking at a TB-303 sample) you'll see that it is not linear, it falls down fast in the begining, then slower and slower

isnt that what a logarithmic curve does?

could be, try to approximate the curve with a log() and tell me if you were succesful ;]

Kriminal · Post by **Kriminal** » Sun Sep 13, 2009 6:02 pm

antto wrote:
Kriminal wrote:
antto wrote:
you know, some say "logarithmic envelope" or "exponential" or "linear"
this is exactly the same, only it doesn't seem to be any of these

in other words, if you look how the cutoff (with a lot of resonance) travels on the FFT (when you're looking at a TB-303 sample) you'll see that it is not linear, it falls down fast in the begining, then slower and slower

isnt that what a logarithmic curve does?
could be...

well, from your description, thats what it sounds like...have you tried a log curve, just out of interest?

(its pretty much all i use these days)

antto · Post by **antto** » Sun Sep 13, 2009 6:17 pm

no, i haven't explicitly tryied a logarithmic curve, tho, CurveExpert i think has such, but the closest match was a logistic curve (which had exp() and division)

Kriminal · Post by **Kriminal** » Sun Sep 13, 2009 6:22 pm

antto wrote:no, i haven't explicitly tryied a logarithmic curve, tho, CurveExpert i think has such, but the closest match was a logistic curve (which had exp() and division)

ok, im not an expert at anything like this, but after some advice i got the DH stuff and made my own ADSR using his modules...and used the log curves (for a more accurate emulation of something i did)

http://www.dehaupt.com/SynthEdit/DH_EnvSeg_FAQ.htm

apologies if you already know all this, just trying to throw some info your way

antto · Post by **antto** » Sun Sep 13, 2009 6:22 pm

BOZA_65_01_DEMO.mp3 <- my synth "in the mix" (two instances)

antto · Post by **antto** » Sun Sep 13, 2009 6:28 pm

Kriminal, i know about DH_EnvSegment, i don't like it, i've used it before to make a 6 stage envelope, but i figured it had a "bug" .. you can't change the time of a segment that is currently runing which is exactly what i need for this project, so the standard SE ADSR works fine ;]

i tested a Logarithmic curve in CurveExpert, it failed, division by zero ;]
btw, the curve is very specific, don't be fooled by how simple i explained it ;]

Kriminal · Post by **Kriminal** » Sun Sep 13, 2009 6:30 pm

antto wrote:Kriminal, i know about DH_EnvSegment, i don't like it, i've used it before to make a 6 stage envelope, but i figured it had a "bug" .. you can't change the time of a segment that is currently runing which is exactly what i need for this project, so the standard SE ADSR works fine ;]

i tested a Logarithmic curve in CurveExpert, it failed, division by zero ;]
btw, the curve is very specific, don't be fooled by how simple i explained it ;]

ok, fair enough, just thought i'd mention it

mistertoast · Post by **mistertoast** » Sun Sep 13, 2009 6:45 pm

In the 303, what changes based on tempo (BPM)? Anything? Any curves?

Also, do we know if Q changes dynamically (independent of the knob)? Or is it just at one level for accented notes and another level for unaccented notes?

antto · Post by **antto** » Sun Sep 13, 2009 6:49 pm

mistertoast wrote:In the 303, what changes based on tempo (BPM)? Anything? Any curves?

nothing..

Also, do we know if Q changes dynamically (independent of the knob)? Or is it just at one level for accented notes and another level for unaccented notes?

i've seen a lot of audio samples (including good ones)
i've never seen such a thing, there might be some current going to the resonance input of the filter when an accented note is playing (thus, increasing the resonance level a bit) but it must be sooooo small that i can't see/hear it in any audio sample
so i assume there isn't such a thing

the VCA is louder for accented notes (depending on the Accent knob) but this has nothing to do with resonance ;]

mistertoast · Post by **mistertoast** » Sun Sep 13, 2009 8:04 pm

"the VCA is louder for accented notes (depending on the Accent knob) but this has nothing to do with resonance ;]"

I'm certainly willing to believe you, given the amount of misinformation and contradictory stories out there. But there are sources that say res is higher on accented notes.

antto · Post by **antto** » Sun Sep 13, 2009 8:11 pm

in the begining, i read all of these articles too, and believed them
just like the 200ms decay time and the "attack" (wtf?) and even wikipedia with the "-18dB filter" or even that i read somewhere about "the TB-303 has a excelent 2-pole -18dB Resonant filter" (holly crap!)

now take some good samples and see for yourself, i don't see any raise in resonance on accented notes (even when accent knob is at MAX)
the maximum resonance level of the Resonance knob on the TB-303 is quite close to self oscilation
you know what'll happen if the resonance was increased on accented notes?
i know, i've done that to see, it's bad ;]

mistertoast · Post by **mistertoast** » Sun Sep 13, 2009 8:13 pm

Well, antto, you've paid your dues, so you should be able to help us clear up the misconceptions out there.

What tools are you using for your detective work, besides SoundForge? Just that and some mIRC scripts?

antto · Post by **antto** » Sun Sep 13, 2009 8:32 pm

nice question ;]
well, i can plot math expressions in my mIRC and see curves, i can also plot 2 different expressions and see them both plotted
i also use Winamp's AVS (visualisation) it has this SuperScope effect, which is scriptable, i've made myself a scrolling spectrograph there too (notice: not just a spectroscope)
CurveExpert
Spectroman (FLStudio) (another scrolling FFT, but darn small)
MS Calculator

Code::Blocks (IDE) + GCC (mingw32)
SynthEdit 1.0170 (ofcourse)
MS Paint (infact, i measured the the FEnv curve amplitude with Paint and transformed the coordinates in mIRC to be able to paste them into CurveExpert, pretty wierd, eh?)

mIRC helps me alot too, since i know the scripting like my fingers, i can do pretty much anything, a downside is that it's interpreted and runs slow (but this also means that you can change the code and it's updated instantly, no compile & wait)
i think every decent developer has his own good tools which will probably differe from mine (i got different background than you might think)
probably some guys would use Python for what i use mIRC for, or C++ itself, or MATLAB, whatever ;]
i just still can't make a proper win32 GUI application (in C++), for example to plot something on the screen
i hate the winapi, it's so confusing
this is why i use mIRC a lot

antto · Post by **antto** » Sun Sep 13, 2009 8:36 pm

ok, you probably wanted to know how i can tell from an audio sample so much about what's the resonance level, what were the knob values..
i can, because i've stared at oscilloscopes for about 7 years maybe, i have an idea how sounds "look like" as a waveform

i guess my brain works in an odd way ;] (don't get me wrong, i'm no genious)

Open303 - open source 303 emulation project - collaborators wanted