Doepfer

while trying to approximate my favourite bass station sounds on the a100,
i found the only thing i couldn't get near to was the adsr curves,
which seem less exponential (vis a vis pitch perception) than the a140.

does anyone have any tricks to flatten out the curve ?

how are the adsr's from other firms in comparison ?

dieter, would there be no demand for an alternative adsr ?  the way
the a140 slowly settles into the sustain level is brilliant for many
things, but i could certainly also use a curve that makes straight for
the bullseye, no messing.

cheers,

stu

Hi Stu,

Since I had a set of TB-303 schematics open in front of me on my desk 
(been studying the filter), I thought I'd ponder this on a slow 
Sunday afternoon - but do I assume correctly that the Bass Station is 
largely based on the '303?. From what I've spotted so far, there 
doesn't seem to be much of a concept of 'sustain' on the 303, but I 
don't know it that well. In any case, the envelope mechanism is the 
usual charge/discharge of a capacitor, and since this is the same as 
the A-140, it must be different relative voltage levels between the 
two machines that account for the different bend/linearity of the 
envelope curve.

I think there might be some room for manoeuvre in terms of a 
straightforward change to the A-140: during 'decay' the charge drains 
away via a transistor, which slowly turns itself off as the 'sustain' 
level is reached. If this transistor was kept hard on, and switched 
off via a separate means (say a comparator, or a simple differential 
pair), the charge would continue to drain off faster for longer, and 
hence the decay curve would not end up so steep, thus going 'straight 
for the jugular' (taking liberties with your metaphor!). The biggest 
problem is that such a scheme might open up a myriad of possibilities 
of what is and isn't a good curve, and it might be hard to remove 
this subjectivity if it came down to chosing component values etc.

I guess there is also the possibility that the Novation envelope is 
not generated in real hardware, but done by software? I looked around 
for a set of schematics, but from comments on other forums it looks 
like they don't exist, so without a unit to look at, we get no help 
in establishing just *what* contributes to its envelope shape.

Thats my few pennyworth - perhaps others and Dieter might have some 
more ideas?

Tim

--- In Doepfer_a100@yahoogroups.com, "Stu Grimshaw" <grimshaw@...> 
wrote:
>
> while trying to approximate my favourite bass station sounds on the 
a100,
> i found the only thing i couldn't get near to was the adsr curves,
> which seem less exponential (vis a vis pitch perception) than the 
a140.
> 
> does anyone have any tricks to flatten out the curve ?
> 
> how are the adsr's from other firms in comparison ?
> 
> dieter, would there be no demand for an alternative adsr ?  the way
> the a140 slowly settles into the sustain level is brilliant for many
> things, but i could certainly also use a curve that makes straight 
for

> the bullseye, no messing.
> 
> cheers,
> 
> stu
>

Hi Tim

> I guess there is also the possibility that the Novation envelope is 
> not generated in real hardware, but done by software? 
Thats a fact. I had a novation bass station for testing when it appeared 
on the market. The envelopes and the LFO are generated by software.

Florian

> while trying to approximate my favourite bass station sounds on the a100,
> i found the only thing i couldn't get near to was the adsr curves,
> which seem less exponential (vis a vis pitch perception) than the a140.
>
> does anyone have any tricks to flatten out the curve ?
>
> how are the adsr's from other firms in comparison ?
>
> dieter, would there be no demand for an alternative adsr ? the way
> the a140 slowly settles into the sustain level is brilliant for many
> things, but i could certainly also use a curve that makes straight for
> the bullseye, no messing.
>
> cheers,
>
> stu

ADSR envelopes are normally exponential. One would need a current source
(instead of the simple R/C circuit) to obtain linear slopes. The voltage
controlled version of the ADSR A-141 can be modified to obtain more linear
slopes by overdriving (clipping) the OTA that deliveres the charge/discharge
current for the capacitor. For the standard ADSR A-140 that's not possible.

Best wishes
Dieter Doepfer

Hi Stu

> does anyone have any tricks to flatten out the curve ?
If you speak about loudness envelope: Are you aware that you should use 
a linear VCA for the envelope modulation, because the envelope _IS_ 
exponential already. Exponential VCAs are required only for modulations 
by linear voltages, e. g. LFOs.

Florian

interesting stuff, my thanks.

@tim: i'm not sure how close the bass station is modelled on the 303,
since the last time i had a 303 in my hands was about 1984 :). it
seems that software is responsible for the curve - cheers florian -
which i assume to be quite different to the 303. however, the sounds
i'm aiming at are less the squelchy acid stuff than the round and sexy
sawtooth things, which the novation does well enough that it gets
noticed by those suitable disinterested in the technical details. the
fact that i don't have a moog lying around is probably self evident. i
had to stifle a giggle when i read this;

The biggest
problem is that such a scheme might open up a myriad of possibilities
of what is and isn't a good curve, and it might be hard to remove
this subjectivity

like, "i got this synth from a firm called doepfer, but it's crap -
you even have to decide in what order you want the components". just
kidding.

@dieter: overdriving the OTA (what on earth is that?) sounds like what
i've been trying to do since my post, i.e. clip the envelope. although
putting it through the extreme filter produced an in interesting
envelope in its own right, this wasn't exactly the yoke of the egg.
the filter is dc coupled or something ? anyway, is there a module that
would serve as a cv compressor ? used to do this a lot with max/msp,
and it yielded superb results - i read somewhere that the moog
envelopes need a few milliseconds to start dropping to the sustain
level. anyone know more ?

@florian: this was all about the filter envelope, but now i know why i
like the cheapo a132 vca's better than the a131 exp. vca :) thanks for
that!

just out of curiosity, if someone has time: is the slew limiter not
much more linear ? am i imagining it ? i can't afford to, but if i
were to replace the adsr with, say, a sequencer (clocked suitably
quickly) that would send the appropriate cv's at appropriate times to
a slew limiter, would i not then have my linar curves ?
duh! if i put the adsr through the slew limiter ?? (back in a minute)

find me a warship and a silly jacket and line up the sailors, mission
accomplished.

slewing it down, so to speak, is pretty much what i was after. i'm not
saying that one is better than the other, but varying the lengths of
the decay at the adsr and the slew rate introduces a lot of
flexibility. can't offer any comparisons with moogs, oberheims et al,
don't have any,  but i'd recommend trying it out to anybody interested
in that more sluggish george duke thang as opposed to the percussive
kraftwerk ding.

thanks again for the input,

stu

> Hi Stu
>
> > does anyone have any tricks to flatten out the curve ?
> If you speak about loudness envelope: Are you aware that you should use
> a linear VCA for the envelope modulation, because the envelope _IS_
> exponential already. Exponential VCAs are required only for modulations
> by linear voltages, e. g. LFOs.
>
> Florian

Hi Florian,

I think there are no such strict rules. I personally like the double
exponential behaviour (i.e. exp. ADSR + exp. VCA) for punchy, percussive
sounds. On the other hand I very often use a linear VCA for LFO based
modulations.

Dieter

Hi Dieter,

 > I think there are no such strict rules. I personally like the double
 > exponential behaviour (i.e. exp. ADSR + exp. VCA) for punchy,
 > percussive sounds. On the other hand I very often use a linear VCA
 > for LFO based modulations.
as I wrote in my book: "There is no rule how to use a modular system. If 
you like the sound of a synth crashing on the boardwalk, then do let you 
urge by the owners manual, open the window and throw it out" ;-)

But regarding amplitude modulation by linear voltages (like LFOS) and 
linear/exponential VCAs there is a big difference in the sound of audio 
crossmods! I'd always use the exponential VCAs for these (principle: 
more bang for bucks).

Regarding envelope modulation of VCAs: for live Music (as you do it) the 
combination exp. ADSR + exp. VCA is ok and works more like the famous 
"icepick in the forehead", but if you want them to use it in a recording 
the sound will be mostly _thin_.

Florian

You find my comments written inline.

Best wishes
Dieter Doepfer

> @dieter: overdriving the OTA (what on earth is that?)

an OTA (operational transconductance amplifier) is similar to an operational
amplifier (I hope you know this circuit) but with current in/out (instead of
voltage in/out) and variable current amplification (so called
"transconductance"). OTAs are used in many voltage controlled applications
like filters or VCAs.

> sounds like what i've been trying to do since my post, i.e. clip the
envelope. although
> putting it through the extreme filter produced an in interesting
> envelope in its own right, this wasn't exactly the yoke of the egg.
> the filter is dc coupled or something ?

yes and no: the input is DC coupled, but the output is AC coupled because of
the DC offset of the filter output.

> anyway, is there a module that would serve as a cv compressor ?

a log->lin converter could do this job. Such a module would be no problem
(very similar to the exponential converters in the VCOs, VCFs and VCAs) but
so far nobody asked for such a module.

> just out of curiosity, if someone has time: is the slew limiter not
> much more linear ?

The slew limiter is nothing but a variable resistor and a capacitor followed
by a buffer. Consequently the charge/discharge behaviour of the slew limiter
is exponential too. Only the voltage controlled version of the slew limiter
(A-171) with an OTA could be modified by overdriving/clipping the OTA in the
same way as the VCADSR A-141.

would it then not make a lot of sense to have a vca with a linear cv
input without attenuation for the envelope, and an exponential one (or
2?) with attenuation for lfo/kbd-tracking/am ?

or even put a switch linear/exp next to each input ?

stu

p.s. what's the difference between exponential and logarithmic, and
which one do i mean überhaupt ???

hi florian,

so a synth voice that won't shut up - interesting for guitar emulation
i suppose but i take your point.

stu

--- In Doepfer_a100@yahoogroups.com, Florian Anwander
<Florian.Anwander@...> wrote:

>
> Hi Stu
> 
> > would it then not make a lot of sense to have a vca with a linear cv
> > input without attenuation for the envelope, and an exponential one (or
> > 2?) with attenuation for lfo/kbd-tracking/am ?
> Of course one could do it like that, but then you can use it _either_ 
> for loudness contour _or_ for amplitude cross modulation. But both 
> together won't work, because the LFO will still open the VCA where the 
> envelope would like to make it shut up.
> 
> 
> Florian
>

Hi Stu

> would it then not make a lot of sense to have a vca with a linear cv
> input without attenuation for the envelope, and an exponential one (or
> 2?) with attenuation for lfo/kbd-tracking/am ?
Of course one could do it like that, but then you can use it _either_ 
for loudness contour _or_ for amplitude cross modulation. But both 
together won't work, because the LFO will still open the VCA where the 
envelope would like to make it shut up.


Florian

> would it then not make a lot of sense to have a vca with a linear cv
> input without attenuation for the envelope, and an exponential one (or
> 2?) with attenuation for lfo/kbd-tracking/am ?

For this appliaction usually two daisy chained VCAs are used: one with
linear and one with logarithmic response. This works as a multiplication of
both the linear and logarithmic amplification (i.e. if one of both is zero
the output is zero too).

> or even put a switch linear/exp next to each input ?

This feature will be available at the planned dual VCA (module number will
be probably A-132-3)

> stu
>
> p.s. what's the difference between exponential and logarithmic

That's mathematics. The exponential function is the reverse function of the
logarithm (like multiplication <-> division, or addition <-> subtraction).
Sometimes the terms are used as a synonym but that's wrong (even by me).
More details in your mathematics book.

Best wishes
Dieter Doepfer

> > or even put a switch linear/exp next to each input ?
> 
> This feature will be available at the planned dual VCA (module
number will
> be probably A-132-3)

excellent. does planned mean this year ?

stu

> excellent. does planned mean this year ?
>
> stu

I cannot promise that. It will be probably included with the next news batch
(i.e. after the coming batch that includes the Crossfader/FX Insert A-135d,
Manual Gate A-164-1, Switched Multiple A-182, Precision Adder A-185-2,
Gate/Trigger Combiner A-186) but I'm not sure if will be available still
this year.

Best wishes
Dieter Doepfer