Anti-cab Block

Rajnar

Inspired
While it seems impossible for a normal speaker cabinet such as a Marshall 4X12 to emulate a full-range cabinet no matter how cleverly any standard EQ, including global EQ, is applied, I wonder if it might be possible to develop an anti-cab block.

My concept would allow the user to type in whatever actual cab they are using and this new block would apply an inverted IR for that specific cab to neutralize, as much as possible, the native responses of the cab. If that could be done, any cab IR could then be applied to emulate a different cab.
 
If you have cab-lab (full version) you can create an inverted IR by yourself. I think you can do it in REW too.

But it's not that simple..
If an IR has null response on some range of frequencies (and a guitar cab IR has them for sure) you can't correctly invert it. Furthermore the resulting IR will be much lower in volume.

Fact is I tried that some time ago and it's not worth the effort imho
 
What I have in mind is, as much as possible, creating a flat response, which means adding missing frequencies (more accurately, trivial frequencies since none are totally absent due to the nature of harmonics). I doubt that Cab-lab is capable of doing that although I could be wrong. We can easily compensate for any drop in volume, and I also had in mind being able to apply any normal IR rather than leave the result as neutral.
 
I wonder if it might be possible to develop an anti-cab block.

In "audiophile" circles this is known as speaker calibration/correction. It's very similar to generating an IR (or freq response) and "inverting it". However, the situation there is already rather different:
  1. Most hi-fi speakers are designed to be full-range unlike guitar speakers which roll-off
  2. They are usually not near-field and never close mic'd
  3. Calibration includes room effects (resonances/peaks, anti-resonances/nulls)
  4. The correction (inverse) can only be applied for one listening location wrt the speakers+room
  5. If the speaker+room (system) has true nulls or significant LF/HF roll-off, these cannot be corrected for or would cause unwanted distortion or would otherwise overload speaker
I did this with various speaker systems and the result to my ears was "meh" and I tended to prefer uncalibrated speakers. IRRC it also introduces phase shifts so it is not win-win.

If that could be done, any cab IR could then be applied to emulate a different cab.
Guitar cabs are non-invertible. There are frequencies at which the inverse would be infinity.

Cliff's right. However one approach would be to use a calibration tool (like REW) to measure the FR where you normally stand/sit to play guitar and determine a "smoothed inverse" EQ curve to try to roughly compensate for the general EQ of your cab+room at your listening location. As @DLC86 says it might not be worth the effort.
 
Guitar cabs are non-invertible. There are frequencies at which the inverse would be infinity.
You may be taking my suggestion too literally. Rather than a perfectly inverted algorithm, I had in mind a compensatory algorithm, bringing the real cab into neutrality as much as possible.
 
In "audiophile" circles this is known as speaker calibration/correction. It's very similar to generating an IR (or freq response) and "inverting it". However, the situation there is already rather different:
  1. Most hi-fi speakers are designed to be full-range unlike guitar speakers which roll-off
  2. They are usually not near-field and never close mic'd
  3. Calibration includes room effects (resonances/peaks, anti-resonances/nulls)
  4. The correction (inverse) can only be applied for one listening location wrt the speakers+room
  5. If the speaker+room (system) has true nulls or significant LF/HF roll-off, these cannot be corrected for or would cause unwanted distortion or would otherwise overload speaker
I did this with various speaker systems and the result to my ears was "meh" and I tended to prefer uncalibrated speakers. IRRC it also introduces phase shifts so it is not win-win.




Cliff's right. However one approach would be to use a calibration tool (like REW) to measure the FR where you normally stand/sit to play guitar and determine a "smoothed inverse" EQ curve to try to roughly compensate for the general EQ of your cab+room at your listening location. As @DLC86 says it might not be worth the effort.
I've been trying exactly that and have had some success, and some success in correcting phasing errors. But using global or individual EQ ruins my studio patches built for direct recording, and isn't precise enough. While I could and have created specific patches for each application, having a correction block available would be far more convenient and, I'd hope, accurate.
 
You may be taking my suggestion too literally. Rather than a perfectly inverted algorithm, I had in mind a compensatory algorithm, bringing the real cab into neutrality as much as possible.
No, this is an old idea. It's been tried many times and always fails. You can't make a given cabinet sound like another cabinet using filters. The invertability problem, even using all kinds of tricks, is too severe.
 
Besides the basic non-invertibility problem, there is no "universal IR" (filter) for a cab/speaker as all (IR) measurements are in context: mic, location, time window, room/reflections, monaural vs binaural, etc. Each position/context can give radically different IRs (peaks/nulls), so there is no universal IR for a given cab to invert anyway.

What you might be after is more of a "speaker/cab morphing" function which would be more physically-based than measurement-based. That is a whole other beast.
 
Last edited:
BTW @Rajnar, if you have determined a smoothed "cab correction EQ" for yourself that you like, the Axe has various EQ blocks and the IRP block that you could use for that purpose.
 
See it this way: if the signal at, let's say, 50Hz is completely filtered out so that you have no signal at all at that frequency, you can boost it all you want afterwards but you'd still get no signal.
1000*0=0

I've been trying exactly that and have had some success, and some success in correcting phasing errors. But using global or individual EQ ruins my studio patches built for direct recording, and isn't precise enough. While I could and have created specific patches for each application, having a correction block available would be far more convenient and, I'd hope, accurate.

Despite what I said above, I second this idea to have a built-in correction block for frfr monitors or headphones, I'm currently using a second cab block with IRs captured from sonarworks plugin exactly for that and it works pretty well imho.

PS: i actually made a wish for that some time ago
https://forum.fractalaudio.com/threads/correction-ir-loader.137037/
 
Last edited:
built-in correction block for frfr monitors or headphones

How would that be different from an IRP (besides in name)? I could see the benefit of having an optional "global" FR correction on each output but then that would consume CPU (therefore it should be optional).

EDIT: I see @DLC86's wish is for that...
 
How would that be different from an IRP (besides in name)? I could see the benefit of having an optional "global" FR correction on each output but then that would consume CPU (therefore it should be optional).
Yep that, "global". And stereo (see my wish above)
 
Ideal would allow any 'correction profile' to be selected from cab/IR slots as desired. (Thus utilizing already existing functionality)
 
Ideal would allow any 'correction profile' to be selected from cab/IR slots as desired. (Thus utilizing already existing functionality)
Yes, and my idea was to apply it only to physical outputs (analog/spdif/aes), so that you can record without correction baked into the signal via usb while still monitoring the corrected signal via those outs
 
Last edited:
Guitar cabs as a rule are not full range, actually having quite a steep drop off in the highs. Once a recorded frequency is not present due to this fact, how do you simply “add it back”?
 
Once a recorded frequency is not present due to this fact, how do you simply “add it back”

Well, in that case, I always multiply by infinity. Sure, 0*∞ is indeterminate you might say, but "something" is better than nothing (as some questionable psycho-logics go). ;)
 
I've been trying exactly that and have had some success, and some success in correcting phasing errors. But using global or individual EQ ruins my studio patches built for direct recording, and isn't precise enough. While I could and have created specific patches for each application, having a correction block available would be far more convenient and, I'd hope, accurate.
I'm slow to catch on sometimes, so I have some questions:
1. How would the Axe III measure the actual frequency response of your cab/monitors/headphones? Can it do that already through Tone Matching?
2. What end goal are you trying to achieve by doing this? True flat response? As others have stated, that's not possible with a guitar cab, but is approachable with FRFR monitors or headphones.

I've done quite a bit of home theater calibration, where this is done a lot. Most really in-depth calibration software runs independently from the home theater processor (for instance, DIRAC). Once measurements, and corrections are done, those filters are uploaded to the head unit. Phasing can be a problem with lesser calibration software - last time I used Audyssey it would detect an absolute phase error - such as a speaker wired out of phase - but couldn't correct it... it may now). However, most phase correction is typically to correct speaker placement... ultimately to correct time alignment at the listening positions. The point is, you could probably do what your asking for a FRFR at a single listening position that never changes, but not for a club or arena as the requirements change with each added listening position.
 
No, this is an old idea. It's been tried many times and always fails. You can't make a given cabinet sound like another cabinet using filters. The invertability problem, even using all kinds of tricks, is too severe.

Is this a case of taking the comment too literally? I think the poster was talking about making a cabinet more neutral, not completely neutral. I could be wrong. "Attenuate some of its more severe tendencies"
 
Is this a case of taking the comment too literally?
The OP literally said his comment is being taken too literally :).

If it's a guitar cabinet, then it will obviously have a severe high end rolloff and that will make it difficult to get a meaningful inverse IR. But, assuming the cabinet being emulated has a similar high end rolloff, that might not matter much. In other words, artificially adjust the coefficients so the high end of the IR is flat. I share some of the skepticism about whether it would sound good, but it would certainly be an interesting experiment.
 
Back
Top Bottom