Dynamic IR Block

GM Arts

GM Arts

Fractal Fanatic
Not really a desperate Wishlist item, just some fanciful meanderings of my mind ...

IRs are a combination of a general broad EQ shape with many very high Q peaks and notches. What I have in mind is a Dynamic IR block that can generate typical IRs that are easy to dial in & fine-tune. I'm not thinking this would be used to emulate real-world cabs exactly; it's too generic to be able to do that. It's more of a tool to easily dial in something we like that has the general character of a real cab.

So I wonder if it's possible to have a Dynamic IR bock where we choose a general EQ shape, something like the PEQ block, and also choose peak/notch density & severity.

For the sake of this post and as an attempt at clarity with the peaks/notches, I'll call peaks/notches "ripple", density I'll call "spread" and severity I'll call "depth".

It would need to be repeatable, not using true randomisation, but I think it needs a pseudo-random element in conjunction with both ripple spread and depth settings to avoid any evenly spaced notches which might sound like a dense unmodulated phaser or flanger.

Here are the controls I have in mind for a Dynamic IR block, which would be used instead of a cab block:

The PEQ section has 5 frequencies like the current PEQ, with a restriction that Frequencies 1 to 5 must be in order.
Band 1: Low blocking Freq, Slope (eg -3db/oct to -24dB/oct)
Band 2, Band 3, Band 4 all have: Freq, Q, Level
Band 5: High blocking Freq, Slope (eg -6dB/oct to -48dB/oct)

Then Ripple controls for zones between & outside frequencies 1 to 5, for these 6 zones:
< Freq 1, Band 1-2, Band 2-3, Band 3-4, Band 4-5, > Band 5
The ripple controls for each of these 6 zones are: Spread, Depth
As above, the application of both spread and depth should be further affected by a pseudo random calculation.

Finally, the usual block mixer controls (level, panning), plus possibly the cab Drive parameter.

No need for Air (simply use less ripple depth in lower zones).

Thoughts, anyone?
 
Is this similiar to the Nebula IR system where there are multiple IR's in a single plugin (would be IR User Cab block in the AxeFx) that are selectively applied based on the input signal or other parameters?

Richard
 
This doesn't sound like an IR block at all, but rather a "cab sim" using eq and comb filtering. Try using PEQ and Filter blocks with non-sweeping phase shifters and/or flangers (micro delays) adjusted judiciously. Before I had the Axe, I used this technique with some success, but IMO it is no match for true IRs.
 
I think maybe he wants to EQ the IRs themselves before they're ran through the convolution reverb
 
Perhaps some pictures will make this clearer. This is way more sophisticated than a couple of PEQs - my hope is that it will produce flexible and very real sounding IR responses. 1st pic shows a typical IR response. The 2nd pic shows the broad EQ shape in green - just a gentle low rolloff, some mid cut, and a more severe high end rolloff. This could be achieved with just 3 bands.

On top of that broad EQ shape are the peak/notch ripples. In this example, they are more dense and severe at higher frequencies. This is typical, and might be a simpler way to implement a control for these ripples - maybe we'd just need 2 parameters for ripple spread and depth that automaically increase severity with frequency.
 
If you created an IR with just the ripple response (no rise/drop on a broad scale) you could try that in series with a PEQ. You'd need a different IR for each notch density but you could adjust severity by mixing non-IR signal (with a parallel path or cab AIR).
 
Bakerman said:
If you created an IR with just the ripple response (no rise/drop on a broad scale) you could try that in series with a PEQ. You'd need a different IR for each notch density but you could adjust severity by mixing non-IR signal (with a parallel path or cab AIR).
Click to expand...

Yes that could work well, Bakerman. The PEQ might need the ability for more severe high blocking without a resonant peak, but otherwise it suits well. I'm thinking of a generated ripple response, not from using any actual IR. What I'm hoping for is a way to create a "real sounding" cab IR and the ability to fine tune it to taste.

I guess this is borne from some frustration in loading multiple user IRs to find something close, then using corrective EQ to get it closer.
 
GM Arts said:
So I wonder if it's possible to have a Dynamic IR bock where we choose a general EQ shape, something like the PEQ block, and also choose peak/notch density & severity.
Click to expand...
It's called a "workstation." With one of those, you can start with any IR you like filter it in any way you see fit, and convert and upload it to the Axe-Fx.

Thoughts, anyone?
Click to expand...
You just described a PEQ.
 
GM Arts said:
On top of that broad EQ shape are the peak/notch ripples.
Click to expand...
Multipath phenomena. What you need to replicate that is a multitap delay with extremely fine increments of delay adjustment (25uS or so). And you'll play hell tweaking that by ear to get anything resembling the "ripple" in the curves above. If you've got the means to analyze the response of an IR, you have the means to do what you want outboard of the Axe-Fx, which is where you should be playing this kind of game. Once you're done, you can convert and upload back to the Axe-Fx.
 
Thanks Jay, I'm aware of those options. They allow for generation of a user IR, one at a time. I'm wondering whether there might be a way to dynamically generate typical IRs within an effects block for a more flexible approach.

I suspect there are multiple delay paths with different and very short times that cause this "ripple", so we'd want to limit user controls to things that are clearly audible. My suggestion is just ripple density and depth, and maybe a rate of increasing intensity from low to high frequencies. The algorithm would generate a response from this, maybe with a simple delay filter and pseudo-random variations, or perhaps from a detailed calculation of multiple delay paths.

I think it's doable, but I'm only hopeful that the end result would sound realistic, when used with a PEQ for broad spectrum shaping.

This is not about what should be done where. I've posted this in the wishlist forum, for discussion about innovation that maybe could be done in the AxeFX. With ideas from many, even an impractical idea can spark something new, exciting and achieveable. Do you have any thoughts about what might be practical in the AxeFX to allow creation of dynamic IRs?
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom