[sdiy] Pole Mixing (with Resonance)
tom at electricdruid.net
Sat Apr 10 12:07:43 CEST 2021
It occurs to me that using bandpass feedback as described in the SSI application note might fix the problem with highpass modes. They propose it as an alternative way of doing passband droop compensation, but it might be a better way of doing resonance in a pole-mixing filter too.
Clearly when you use feedback from the LP4 output, you’re feeding a lot of low-end signal back into the filter which then screws up the balance required for good stop-band performance in the highpass modes.
I’ll give it a try one day!
Synth & Stompbox DIY
> On 9 Apr 2021, at 21:47, David Moylan via Synth-diy <synth-diy at synth-diy.org> wrote:
> Got a resonance slider added. I had to attack the math a few different ways, but finally got the proper cancellations and looks very simple in retrospect. Definitely ruins the HP modes. :)
> On 4/9/21 2:03 PM, David Moylan via Synth-diy wrote:
>> Hi Brian. This was inspired by the Xpander/Matrix filter but is more of a general exploratory tool at this point. I think most of the names overlap but there are definitely some not implemented on the Obies, such as HP4 and the double notch curves. Those require all 4 poles plus a dry signal and the Obies don't have that capability. The curves are calculated on the fly from the theoretical transfer functions.
>> I'm in algebra hell right now trying to work the math out to handle the feedback signal. It would be a lot easier with no dry signal but once you start mixing that in you get different denominators that have to be resolved and suddenly I'm looking at 8th order polynomials :(
>> On 4/9/21 1:53 PM, Brian Willoughby wrote:
>>> This is excellent! Thank you for sharing this online.
>>> As a Matrix-12 owner, I appreciate the list of presets that cover the available modes. It looks like you've even included some modes that aren't available in the Oberheim. Is that correct?
>>> Did you use impulse response measurements to obtain the frequency response?
>>> The reason I ask is that I wrote a macOS CoreAudio application that can load various AudioUnit plugins and measure their frequency response by simply feeding a short, 1-sample impulse into the plugin and doing a Fourier transform on the result. This only works for pure LTI (Linear Time-Invariant) plugins, but it's spot-on for those that fall into the category (i.e. no distortion or envelope-controlled dynamics).
>>> On Apr 8, 2021, at 15:24, David Moylan wrote:
>>>> Hi All. I banged together a little web app to play around with filter pole mixing, of the Oberheim Xpander type. You can mix poles in varying amounts and see the output magnitude shape as well as the transfer function. Y axis is Db and X axis is log scale based on normalized frequency (so basically 1 equals the cutoff frequency). Haven't done phase plot yet.
>>>> If you have an interest in this sort of thing check it out:
> David Moylan
> Expedition Electronics
> sonic adventures!
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> Selling or trading? Use marketplace at synth-diy.org
More information about the Synth-diy