Page 1 of 1

EQ Based on multiple IRs

PostPosted: Wed Nov 04, 2020 10:25 am
by StereoSpace
Hello!
I know this may sound silly, but I'll share the idea anyway.
I recently experimented with Martin's IR Cabinet.
I recorded the impulse of another EQ and after loading it into the cabinet, I got exactly the same curve as in the original EQ.
Then the idea came to my mind, to sample several knobs of one equalizer and place a copy of the cabinet with the loaded impulse on each band.

For example: there is a good EQ with an interesting treble boost curve. I took a picture at maximum frequency gain and transmitted it to the IR cabinet. Now when I mix in to the dry signal, the signal processed by the cabinet, I get what I need.

But, now there is a problem, the solution to which I hope to find with your help.

The equalizer has interesting behavior. When a certain frequency is raised, it behaves differently than when the same frequency is cut. The curve changes in different ways.
As a result, I got the idea to load two separate pulses for attenuation and for amplification of the desired frequency. But I have no idea how to make such a scheme:

If the knob is in position 0, then nothing happens, none of the two pulses (A and B) are mixed with the original sound.
If the knob is in a position, say -1 to 0, this gives you the ability to influence the dry / wet ratio of Cabinet A.
If the knob is in the 0 to 1 position, it naturally affects the mixing of Cabinet B.

I hope I was able to explain with the help of google translate))

Even if this is solved through two separate knobs (boost and cut), like some devices, there are still a couple of questions.

For each specific position of the frequency selection knob, a pulse should be created, and I'm ready to spend a lot of time to take IR shots, but how to switch between pulses seamlessly?
When there are fixed frequencies, everything is simpler, but in the same devices from Acoustica Audio, which I was inspired by, everything works smoothly there.

Is there a solution for at least one of the questions?

If you need to lay out the cabinet diagram that I used, please write here.
Thanks to everyone who creates these wonderful schemes here.

Re: EQ Based on multiple IRs

PostPosted: Thu Nov 05, 2020 4:22 am
by ChrisHooker
Hi StereoSpace,
This idea could lead to a very CPU intensive plug-in, but that aside, it can be done.
Attached is an example of using 1 knob to control the EQ "Gain"... or more accurately: a balance between the dry signal, a "fully boosted filter" IR, and a "fully cut filter" IR.

This schematic is all primitive-based, for ease of understanding how it works. I'm sure it could be made more efficient by converting it into code/assembly.

Hope it helps ya.

Re: EQ Based on multiple IRs

PostPosted: Thu Nov 05, 2020 8:07 am
by StereoSpace
ChrisHooker wrote:Hi StereoSpace,
This idea could lead to a very CPU intensive plug-in, but that aside, it can be done.
Attached is an example of using 1 knob to control the EQ "Gain"... or more accurately: a balance between the dry signal, a "fully boosted filter" IR, and a "fully cut filter" IR.

This schematic is all primitive-based, for ease of understanding how it works. I'm sure it could be made more efficient by converting it into code/assembly.

Hope it helps ya.


Wow! Thanks for help! I think, high cpu eating is normal for this type of eq. :roll:

Re: EQ Based on multiple IRs

PostPosted: Mon Nov 09, 2020 9:35 pm
by martinvicanek
That is a somewhat expensive way to build an EQ. Usually you would use a convolver to implement a complex transfer function with many resonances, like the hollow body of a cello. Peaking EQ filters are quite straight-forward to implement and light on CPU.