KVR Audio

wetdentist · Post by **wetdentist** » Wed Sep 12, 2018 11:26 am

weird that this is the only synth i have that acts this way. i remember mentioning this issue a couple/few years ago and had hoped that this issue had been fixed by now, but alas, it has not. oh well.

Local Man · Post by **Local Man** » Wed Sep 12, 2018 3:30 pm

Yeah, this synth also handles automation and midi learn in a strange way also (at least in Logic). When I try to write automation in latch mode (or any other real time automation mode) I get nothing. But you can select parameters from the drop down menu on the automation lane and draw in automation. I don't think I've come across another synth that works like that. Usually if the parameters are mapped for automation, you can draw it in or do it real time. I don't know what the hang up for this one is.

zzz00m · Post by **zzz00m** » Wed Sep 12, 2018 4:21 pm

It's a cool piece of freeware, and a decent Oberheim emulation. It is what it is.

Local Man · Post by **Local Man** » Thu Sep 13, 2018 1:56 am

zzz00m wrote:It's a cool piece of freeware, and a decent Oberheim emulation. It is what it is.

I agree. Which is why I still use it extensively, despite its little quirks.

wetdentist · Post by **wetdentist** » Thu Sep 13, 2018 11:07 am

yeah, no one uses Maschine anyway.
i can, if i must, use it in Cakewalk, though i do most of my composing in Maschine.

george · Post by **george** » Wed Jan 16, 2019 11:10 am

I have some words to share from 2Dat regarding the old vs new OB-Xd filters:

Yes, they are different. Latest version is more physically correct. It's based on a "mystran" tangent method and does good with lightly non-linear filters such as Oberheim Ob-x or SEM.
I also tried non-iterative method based on a non-elementary omega function to solve the non-linear equation with exponents.

https://en.wikipedia.org/wiki/Wright_Omega_function

Double the cpu cost and barely any difference in sound.

The non-linearities in the latest filter version are not an exact approximation to a diode pair, but rather like a "soft" saturation curve. I thought it sounded better at the moment.

So the reality about old filters sounding better is not true

Cheers!

fmr · Post by **fmr** » Wed Jan 16, 2019 11:49 am

discoDSP wrote: Wed Jan 16, 2019 11:10 am So the reality about old filters sounding better is not true Cheers!

Since this is an entirely subjective perception and opinion, it will always be debatable

It's like the old debate between analog and digital, or hardware vs software.

Anyway, now we have it clear that they are different. But it's a take it or leave it option.

george · Post by **george** » Wed Jan 16, 2019 11:53 am

Old filters using double the CPU is not subjective

if the person who designed the filters decided the new ones were better at the time is something to consider. But yes, each to his own.

mcbpete · Post by **mcbpete** » Wed Jan 16, 2019 12:00 pm

Of course there's always the option of them being switchable a la the 3 engine modes on Dexed, leave it up to the user as to what they think sounds 'best'

aciddose · Post by **aciddose** » Wed Jan 16, 2019 4:41 pm

Enabling the switching can cost a lot. Either you need to template the entire process function (block processing) and switch "outside the loop" or there is some potential for significant overhead to using function pointers, switches or other methods. Optimization issues aside (which are likely minor today anyway, although not always) there is also the overhead of managing the GUI stuff, extra parameters, how is this saved in presets, is it compatible and numerous other issues that don't seem obvious at a glance.

Plus you'd need to write the code. That is usually a significant "but it isn't MY time" overlooked cost.

mcbpete · Post by **mcbpete** » Fri Jan 18, 2019 11:23 am

As the source seems to suggest the filter is indeed a separate function ( https://github.com/reales/OB-Xd/blob/ma ... e/Filter.h ) with case switching for the filter modes (eg HP, LP etc.) couldn't you just have the 'old' and 'new' versions in that same header file, but switch which version is called depending on a global user setting.

I don't think it'd need to be stored in presets, like the dexed engine it'd just be a global setting (possibly saved to an .ini ?) and the GUI front panel wouldn't need tweaking either - Just add the filter-type option to the right-click menu

aciddose · Post by **aciddose** » Sat Jan 19, 2019 12:07 am

Are you going to program all that stuff? Feel free, it's open source.

mcbpete · Post by **mcbpete** » Sat Jan 19, 2019 10:43 am

I don't have Juce, Visual Studio, access to the earlier filter code, VST coding skills, and hardly any free time (currently working 6 days a week, 9am-8pm shifts) so ... nope.

Vortifex · Post by **Vortifex** » Fri Mar 29, 2019 4:50 pm

albertodream · Post by **albertodream** » Tue Apr 02, 2019 3:47 pm

Vortifex wrote: Fri Mar 29, 2019 4:50 pm Can someone explain why the pitch modulation works on some patches and not others? In the following screenshots the first example works and the second example doesn't. What's the really obvious thing that I'm missing here?

The Sync in the oscillators?

Obxd synthesizer