Exponential Envelope Option?

[Naive Teen Idol]

I was wondering: is it possible to create an option to use exponential envelopes? My understanding is that PFM2 uses linear envelopes, but my sense is that exponential would help with some of the percussion sounds where a slightly faster attack would be appropriate.

Not sure if this would be a global parameter or per sound. But is something like that doable?

[Toltekradiation]

i fear the envelopes are already exponential, at least the useful parts :
https://github.com/Ixox/preenfm2/blob/f359b92195af6a0370a08d279eb9081354f93f67/src/synth/Env.h#L76

with an attack set to zero and a short decay, you can get snappy envelopes.
maybe add some click part to the sound to make it more percussive ?

[Xavier]

Yes, but i think what he meant was for the attack which would require an exponential in the opposite direction.
I have no idea whether it would really impact the sound.

For the moment attack is linear.

[Naive Teen Idol]

Thanks to you both. I am hesitant to even mention this because I hate when people compare the DX7 to the Preen. But here is an example of my Preen playing a single operator at the fastest attack and the DX7 (please ignore the hiss from the Preen -- this was a quick example so I didn't plug in directly to my output ... also, this is from Dexed, a DX7 emulator, but this is consistent w my experience with the actual hardware as well):

https://www.dropbox.com/s/j3sj5oljoej9526/No%20Click.wav?dl=0

There is a bright "clickiness" to every note on the DX that is not present on the Preen. You can even get it when you turn all the EG levels on a DX to 0 but leave the attack EG level at 99. A lot of the time, sound designers will add a little attack time to get rid of the sound where it's not appropriate -- but percussion sounds often utilize it to their advantage.

Again, I hate to mention this because I think people spend too much time comparing the two synths and the Preen is wonderful and a million times more flexible. And I'm not sure exactly what is in the DX7 code that accounts for it -- a little research suggests that Yamaha "faked" an exponential shape. But regardless, this would be a great feature to have given how full-bodied this synthesizer already sounds.

[Emesstwenty]

Hi please implement this! The attack stage has to be exponential in order for humans to perceive the response as linear. This thread supports this https://www.muffwiggler.com/forum/viewtopic.php?t=102357 and I'm sure there is other data out there. A linear envelope won't be as snappy and or do modulation sweeps the same. I'm on the fence about buying one and this issue is really the deal breaker

[Emesstwenty]

Just wanted to add that volume pots on synth and guitars are logarithmic/exponential in order to create the perception of a linear response.

[Xavier]

Hi please implement this!
I'm on the fence about buying one and this issue is really the deal breaker

That made me smile

You know it's open source right ?
I will be more convinced by hearing it than by 2 posts and an email
And how it's done makes this very easy to do (see Toltekradiation link in message #2 above)
For the final version It would need to be behind a menu option because it will change the existing patch sound.

[Xavier]

There is a bright "clickiness" to every note on the DX that is not present on the Preen. You can even get it when you turn all the EG levels on a DX to 0 but leave the attack EG level at 99. A lot of the time, sound designers will add a little attack time to get rid of the sound where it's not appropriate -- but percussion sounds often utilize it to their advantage.

I noticed that. The "clickiness" attack of the DX.
I don't say it's impossible but I would be surprised if that came from the exponential attack.

[matrix12x]

I always thought that click meant the oscillator does not start at a zero crossing.

[Emesstwenty]

You know it's open source right ?

Hi yes but I am not a programmer or engineer and i do not know how to code, and I am planning on getting a prebuilt unit.

[Naive Teen Idol]

Hi please implement this! The attack stage has to be exponential in order for humans to perceive the response as linear. This thread supports this https://www.muffwiggler.com/forum/viewtopic.php?t=102357 and I'm sure there is other data out there. A linear envelope won't be as snappy and or do modulation sweeps the same. I'm on the fence about buying one and this issue is really the deal breaker

Hi. Thanks for your support of this idea and link to the MW thread – but this shouldn’t be a deal breaker for buying the PFM2. It’s an absolutely incredible and flexible box that continues to inspire me most every time I plug it in. There are so many features packed into it – and Xavier has continued to add new ones (such as the random voice selection via CV gates) that only add to the value.

The envelope issue is simply something I noticed when converting DX-7 patches and think it would be a nice add.

[SMF]

I always thought that click meant the oscillator does not start at a zero crossing.

All for all I recall, this is correct.

BTW: An attack-time of 0 is neither exponential or linear. It's a step... always. 

best,
Stefan

PS: The main problem arising with exponential/logarithmic volume/envelopes is mapping them to a useful range of control-values... Say you want to map a range of 0...-60dB to a range of control-values of 0...+1,0. Where in a range of say 0 to +1 would you expect to hear the signal half as loud as at +1? At +0.9 (exponential) or at +0.5(linear)? You want a control range of 0...-96dB? Well, this would make things even worse... "Half as loud" would then be at a control-value of 0.9375. Even if you reduce the control-range to something like 0...-40dB the control-value would sit at 0.8, then.

This is the main reason why most synthesizers do *NOT* use exponential envelopes (even if some of the manufacturers perpetually keep on repeating to state that they do...).