Real-time dynamic convolution

DSP, Plug-in and Host development discussion.
tunca
KVRist
474 posts since 26 Mar, 2015

Post Thu Jan 11, 2018 4:36 am

Hi,

Title says everything.Is this possible?

I have an idea something like Waves Q-Clone but dynamic.It can capture saturation and dynamics of eq.

For example i have vintage Harrison channelstrip.I can use it for every channel with Q-Clone but only impulse response...

Possible to make plugin like this or already exist?

Thanks.
http://analogobsession.com/ VST, AU, AAX for WIN & MAC

User avatar
whyterabbyt
Beware the Quoth
26115 posts since 4 Sep, 2001 from R'lyeh Oceanic Amusement Park and Funfair

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 5:22 am

Responded to your other thread of this same question....
"The bearer of this signature is a genuine and authorised pope."

tunca
KVRist
474 posts since 26 Mar, 2015

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 5:38 am

whyterabbyt wrote:Responded to your other thread of this same question....
Just saw.

I opened another topic in DSP section cause i'm plugin developer,too.Maybe i can find someone who working on this type of project.And maybe i can get some advice to make my own...

Thanks.
http://analogobsession.com/ VST, AU, AAX for WIN & MAC

stratum
KVRAF
1840 posts since 29 May, 2012

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 6:38 am

sintefex.com›docs/appnotes/dynaconv.PDF states:


In a non-linear case we make an assumption that the non-linear process can be modelled
by a linear summation of responses to impulses that are appropriate to the magnitude
of each individual input sample.


And apparently it's a patented algorithm.

How many people can actually hear this effect is an interesting question.
~stratum~

tunca
KVRist
474 posts since 26 Mar, 2015

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 7:54 am

stratum wrote:sintefex.com›docs/appnotes/dynaconv.PDF states:


In a non-linear case we make an assumption that the non-linear process can be modelled
by a linear summation of responses to impulses that are appropriate to the magnitude
of each individual input sample.


And apparently it's a patented algorithm.

How many people can actually hear this effect is an interesting question.
Thanks for reply.

I still can't understand that how IRs capturing harmonic distortion...
http://analogobsession.com/ VST, AU, AAX for WIN & MAC

stratum
KVRAF
1840 posts since 29 May, 2012

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 8:20 am

I still can't understand that how IRs capturing harmonic distortion...
Well, not that I have carefully read the paper but nonlinearity and distortion are not the same thing and dynamic convolution does not look like an algorithm intended to capture harmonic distortion. It's just that impulse responses of certain analog circuits have been observed to vary with respect to the signal level, so why not measure it at several different levels and pick the one that matches the level for each sample? As far as I can see this is the idea. Implementing it using the formula that calculates convolution directly does not seem hard, I have no idea if that would be suitable for realtime usage.
~stratum~

matt42
KVRian
1057 posts since 9 Jan, 2006

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 2:28 pm

Not convolution, but Nebula does that kind of thing with volterra kernels.

camsr
KVRAF
6858 posts since 17 Feb, 2005

Re: Real-time dynamic convolution

Post Thu Jan 11, 2018 3:45 pm

Building a simulation of the actual circuit would be more efficient, if the intended result is to be near-identical.
Image

Return to “DSP and Plug-in Development”