Fletcher Munson correction IR pack

Would you be willing to pay for a Fletcher-Munson correction IR pack?

  • Hell yes! $50 seems about right

    Votes: 5 7.5%
  • Sure! $20 is good for me

    Votes: 21 31.3%
  • No, but I want it for free (it's only software, right?)

    Votes: 6 9.0%
  • No. I'll wait until someone shares the IRs and steal them

    Votes: 1 1.5%
  • No, this doesn't sound like something I need / I've solved it another way

    Votes: 34 50.7%

  • Total voters
    67

fractalz

Power User
So, I had a crap rehearsal yesterday due to being distracted by my boomy / shrill patches that, of course, I created at low volume. This is one consequence of the the way we hear sounds at different SPL and is a familiar discussion on these forums, just search for "Fletcher-Munson".

I was annoyed enough to read through a number of research papers and have come up with a way to model equal loudness response curves and create IRs that would allow one to:

1. dial in tones at bedroom levels and then apply a correction to make them sound the same at stage volume
2. dial in tones at bedroom level previewing them as they would sound at stage volume
3. dial in tones at stage volume then play them at bedroom levels with an inverse correction to sound the same

Understanding that "bedroom level" and "stage level" mean different things to different people, a complete solution would include a set of IRs that represent a range of differences between these levels to apply the appropriate correction.

I believe I have the solution and am going to start working on this for my own use. However, I'm open to sharing the results and thought I'd see how much interest there is from the forum here. I figure these would sell for something like the price of a cab pack ($30 - $50 USD) and if enough people are willing to pay, it would make it worth the extra effort to build out the complete set of forward- and inverse- IRs, package them up and release them.

I'll post more info and examples as I make progress. In the meantime, please cast your vote on the included poll and/or register your interest by shooting me a dollar (or fifty) at paypal.me/zazula.
 
I’m not quite so sure the only issue is just Fletcher Munson that you were/are experiencing. Certainly that is one issue, but there is also another huge hurdle, and that’s dialing in your patches to fit within the context of your particular bands mix. And that’s not going to be the same for everyone. Different instruments and different Bass/Drums/keys and their respective timber’s and choice of tone/how they dial in their sound can dictate where you need to be & what you have left to work with within the confines of that particular mix.
Then you’re also going to be facing the physics of different gear and how different speakers react & change when pushed.
I guess ya never know until you try, it’d be fantastic if it worked, but I’m sure there would be a simple EQ plot that was figured out long ago that could be simply applied at the end of the chain or as a Global EQ. Hell, there’d probably even be an actual button that you could push to switch on/off, right on the front of units.
 
To be honest I'd vote no for the reason being I don't feel it's necessary. The EQ page in the amp block allows for setting a post power amp 3 band (passive I think I remember liking?) equalizer to adjust on the fly just in case it needs a little bit of tweaking to fit the location without drastically changing overall patch and balance I worked to create. Essentially exactly what the Seymour Duncan PowerStage amps allow you to do, but already included in your fractal package
 
I’m not quite so sure the only issue is just Fletcher Munson that you were/are experiencing. Certainly that is one issue, but there is also another huge hurdle, and that’s dialing in your patches to fit within the context of your particular bands mix. And that’s not going to be the same for everyone. Different instruments and different Bass/Drums/keys and their respective timber’s and choice of tone/how they dial in their sound can dictate where you need to be & what you have left to work with within the confines of that particular mix.
Then you’re also going to be facing the physics of different gear and how different speakers react & change when pushed.
I guess ya never know until you try, it’d be fantastic if it worked, but I’m sure there would be a simple EQ plot that was figured out long ago that could be simply applied at the end of the chain or as a Global EQ. Hell, there’d probably even be an actual button that you could push to switch on/off, right on the front of units.

I ported my Ax8 patches to the III and had an issue. I dialed in my Ax8 patches at high volume and used the global EQ to adjust. My amp/speaker were at the rehearsal room and I didn't have a chance to dial my new III patches at high volume. The EQ I ended up with at the end of rehearsal was basically the inverse FM curve. I think it would have saved me a ton of aggravation.
 
To be honest I'd vote no for the reason being I don't feel it's necessary. The EQ page in the amp block allows for setting a post power amp 3 band (passive I think I remember liking?) equalizer to adjust on the fly just in case it needs a little bit of tweaking to fit the location without drastically changing overall patch and balance I worked to create. Essentially exactly what the Seymour Duncan PowerStage amps allow you to do, but already included in your fractal package

Yes, you can absolutely use the global EQ for this. The challenge is exactly what EQ to use.

These IRs will preserve the relationships between frequencies as you adjust volume up or down. You would still have to adjust for a different room or any other EQ needs to fit the band mix.
 
How do you know the frequency response of your home monitor system at a given intensity level matches the frequency response of a PA system ?

Your not only dealing with equal loudness contours but different systems, different acoustical environments, boundary effects of monitors being near a wall at home etc

I spent half of grad school working on this and it’s not nearly as simple as just making an inverse loudness contour EQ curve. If it was, it would already be common place
 
How do you know the frequency response of your home monitor system at a given intensity level matches the frequency response of a PA system ?

This is not intended to correct for differences between different loudspeaker systems. This is to correct for the same speaker system at different volumes. I think this covers two very common use-cases:

1. using a powered speaker at home and at rehearsal / performances
- e.g., using a powered wedge to practice at home and as a monitor at a gig
2. building patches using a PA-like speaker as a best-guess for your PA sound
- e.g., using, like I do, a QSC K12 to dial in patches knowing that it is representative of a flat PA system

Your not only dealing with equal loudness contours but different systems, different acoustical environments, boundary effects of monitors being near a wall at home etc

I spent half of grad school working on this and it’s not nearly as simple as just making an inverse loudness contour EQ curve. If it was, it would already be common place

You are correct, but I think there is still value in removing one deterministic change in sound, that of the effect of different volumes. As I said above, you'd still have to tweak for room resonances and other things, but maybe/hopefully those would be less drastic tweaks once the FM effects are neutralized.
 
It would be worth paying for I suppose if it wasn't too much fuss. But currently I only use 1 or 2 presets and dial those in pretty quick when we do sound. It's probably quicker for me to fiddle with BMT when we set levels than mess with IRs. Maybe worth it if I had more presets to deal with though.
 
Would you change among IRs during the gig to accommodate the room and level?

This wouldn't be to help with room resonances / boundary effects, etc. as noted above. This is just to deal with the change in frequency relationships as you change volume.

You would need to have a sense for the loudness difference between your reference (tweaking) level and the final level. You could select the IR when building presets and then turn it off when you turn up for the gig.

I think this could work really well for a home v. rehearsal use-case where your home and rehearsal volumes are pretty static. Get a cheap SPL meter, measure your @home and @rehearsal levels and choose the appropriate IR to apply. You need only apply one and I'd do that where you build/tweak your patches to account for the volume level at the other location.

In the case of a gig where you now have a third level, you would be able to apply a less drastic IR for a first-pass correction and then tune to taste. The amount of correction goes down as the difference from the reference level decreases. So, if stage volume is > rehearsal >> home, you have a large difference from home->stage but a smaller difference from rehearsal -> stage.

I'm open to people testing this out and letting me know how it works... I have yet to build the full compliment of IRs but I know how to do it.
 
It would be worth paying for I suppose if it wasn't too much fuss. But currently I only use 1 or 2 presets and dial those in pretty quick when we do sound. It's probably quicker for me to fiddle with BMT when we set levels than mess with IRs. Maybe worth it if I had more presets to deal with though.

I wouldn't use BMT as much as a Filter block at the end of the chain that makes the changes you need.

IMHO, this would ideally be a block in the AxeFX itself. We'd just have an Equal-Loudness block with a couple of knobs to dial in the amount of forward or reverse correction. You could try to be exact with it, or use it as an effect.

In the absence of being a block w/knob, I'll have to create a matrix of IRs which means just a few in each direction so it is manageable.
 
No flaming intended, but there is another option missing from your voting criteria: No - Because I don't need them.

I do not have the same problem you have. I fine tune my EQ, if I ever need to, at rehearsal where levels are high enough not to make any difference with stage use. Once through this procedure, I (almost) never have issues on stage. And if I do, it's usually time to change my strings!
 
If you look at the equal loudness contours, you will notice a pretty flat response from around 100-6000ish hertz. The real lows and highs are regions where the ear has less sensitivity, but, guitar isn’t really in that range. I high and low cut at 100 and 6000 hz with guitar as is, as I think many do.

Sub bass at 35hz at 30dB is different than 90dB in terms of perception, but 1000hz is pretty linear overall, and guitar being a midrange instrument falls into the spectrum where our ears are the most linear. Assuming of course that one has normal hearing to start with.
 
If you look at the equal loudness contours, you will notice a pretty flat response from around 100-6000ish hertz. The real lows and highs are regions where the ear has less sensitivity, but, guitar isn’t really in that range. I high and low cut at 100 and 6000 hz with guitar as is, as I think many do.

Sub bass at 35hz at 30dB is different than 90dB in terms of perception, but 1000hz is pretty linear overall, and guitar being a midrange instrument falls into the spectrum where our ears are the most linear. Assuming of course that one has normal hearing to start with.

That is all correct. What you also see is that the relative level (relation between frequencies) compresses as volume gets louder.

So, bass and treble that seemed under control at bedroom level can dominate your sound as you turn up.

I high pass at 125 and low pass at 6500 and still have issues. I think those filters may not be high enough order to account for my level differences?

In any case, I'll know next weekend if my plan works!
 
This wouldn't be to help with room resonances / boundary effects, etc. as noted above. This is just to deal with the change in frequency relationships as you change volume.

You would need to have a sense for the loudness difference between your reference (tweaking) level and the final level. You could select the IR when building presets and then turn it off when you turn up for the gig.

I think this could work really well for a home v. rehearsal use-case where your home and rehearsal volumes are pretty static. Get a cheap SPL meter, measure your @home and @rehearsal levels and choose the appropriate IR to apply. You need only apply one and I'd do that where you build/tweak your patches to account for the volume level at the other location.

In the case of a gig where you now have a third level, you would be able to apply a less drastic IR for a first-pass correction and then tune to taste. The amount of correction goes down as the difference from the reference level decreases. So, if stage volume is > rehearsal >> home, you have a large difference from home->stage but a smaller difference from rehearsal -> stage.

I'm open to people testing this out and letting me know how it works... I have yet to build the full compliment of IRs but I know how to do it.
Well I would think most of the room EQing could be done at the board anyway.. if your dialed in at rehearsal & the whole band is dialed in together, you should be fairly good to go, and I would think & hope that the entire band isn’t trying to rework their individual tones to fit the room! That should all be done at the board during sound check & depending upon time constraints and how attentive the sound tech is, minor adjustments/fine tuning while you’re playing.
 
Well I would think most of the room EQing could be done at the board anyway.. if your dialed in at rehearsal & the whole band is dialed in together, you should be fairly good to go, and I would think & hope that the entire band isn’t trying to rework their individual tones to fit the room! That should all be done at the board during sound check & depending upon time constraints and how attentive the sound tech is, minor adjustments/fine tuning while you’re playing.

Exactly. If every musician has done his homework, room corrections shall be for the mixdown of all the band, not for each individual signal going to the desk.
 
That is all correct. What you also see is that the relative level (relation between frequencies) compresses as volume gets louder.

So, bass and treble that seemed under control at bedroom level can dominate your sound as you turn up.

I high pass at 125 and low pass at 6500 and still have issues. I think those filters may not be high enough order to account for my level differences?

In any case, I'll know next weekend if my plan works!


Don’t forget though that the equal loudness contours were created using an average of normal hearing listeners, which are actually pretty difficult to find.

When we were doing psychoacoustics work with volunteers we had to reject probably 75% or more of people due to degrees of hearing loss.

If someone has some degree of high frequency hearing loss they may have a nonlinearity of a 40dB dip at 3khz, however due to the nonlinear nature of sensorineural hearing loss, louder sounds will be perceived equally loud as someone with normal hearing.

Then there is also conductive losses, such as if there is a cold or sinus issue which can dampen the lows, and this is a more linear dampening independent of volume.

Point being, hearing can vary quite a bit between subjects and also day to day with the same subject.

When you factor in nonlonearity of auditory thresholds, hearing loss, room acoustics, nonlinearity of monitor systems, you end up with so many variables that any adjustment you try to make can end up compounding matters.

The basic premise is fine, but your far from the first person to think of this idea. It’s been attempted lots of times over the years, but it just doesn’t really work that well.

If it was as easy and transferable as you believe, and more so it worked, it would be widely used. Cliff would have it in the Axe, all monitor companies would add an intensity dependent processing to make their monitors more linear, your cablebox or tv would implement it etc.

The concept has been around for like 75+ years, it’s nothing new, it’s just recently been picked up by guitar players.

Certainly doesn’t mean you can try your hand at it, but it’s not going to work quite as well as I think you think it does, nor will it address many issues, so I could foresee folks saying they paid you $50 and they tweaked at home at 40dB and then it didn’t sound good at a gig, or vice versus, and it really isn’t your fault, but they thought they paid you to make their tone sound good anywhere, and it still doesn’t, so who’s going to get the blame and bad review ?
 
FM.png OUT-EQ.png
Don’t forget though that the equal loudness contours were created using an average of normal hearing listeners, which are actually pretty difficult to find.

When we were doing psychoacoustics work with volunteers we had to reject probably 75% or more of people due to degrees of hearing loss.

If someone has some degree of high frequency hearing loss they may have a nonlinearity of a 40dB dip at 3khz, however due to the nonlinear nature of sensorineural hearing loss, louder sounds will be perceived equally loud as someone with normal hearing.

Then there is also conductive losses, such as if there is a cold or sinus issue which can dampen the lows, and this is a more linear dampening independent of volume.

Point being, hearing can vary quite a bit between subjects and also day to day with the same subject.

When you factor in nonlonearity of auditory thresholds, hearing loss, room acoustics, nonlinearity of monitor systems, you end up with so many variables that any adjustment you try to make can end up compounding matters.

The basic premise is fine, but your far from the first person to think of this idea. It’s been attempted lots of times over the years, but it just doesn’t really work that well.

If it was as easy and transferable as you believe, and more so it worked, it would be widely used. Cliff would have it in the Axe, all monitor companies would add an intensity dependent processing to make their monitors more linear, your cablebox or tv would implement it etc.

The concept has been around for like 75+ years, it’s nothing new, it’s just recently been picked up by guitar players.

Certainly doesn’t mean you can try your hand at it, but it’s not going to work quite as well as I think you think it does, nor will it address many issues, so I could foresee folks saying they paid you $50 and they tweaked at home at 40dB and then it didn’t sound good at a gig, or vice versus, and it really isn’t your fault, but they thought they paid you to make their tone sound good anywhere, and it still doesn’t, so who’s going to get the blame and bad review ?

I think a few of the things you mention are static in the given situation. For example, my hearing profile is the same at home, at rehearsal and at the gig. So, if I'm using these IRs for my convenience so I can hear the "same" frequency response at home and at rehearsal, that should work. If my hearing at 3k is shot, it'll always be shot and someone will have to clue me in to boost that frequency if it is missing.

I don't think I've come up with anything novel at all - I'm just taking the data and making IRs that match it. My motivation is what my Out EQ looks like compared to the equal loudness contours (see attached). Looks a heck of a lot like the FM curve, or this IR.

But, if the consensus is "don't bother", as it seems to be, I'll go about my business in silence (no harm done).
 
I've always been curious if something like this would work. In my research though, It looks like it may not solve the problem(s). Seems like too many variables affect the end product and therefore correcting only 1 variable doesn't solve the problem. Not to mention that the end product would most likely be interpreted differently by each individual listening.
That being said, I would be interested in hearing what you come up with.
 
View attachment 54245 View attachment 54244

I think a few of the things you mention are static in the given situation. For example, my hearing profile is the same at home, at rehearsal and at the gig. So, if I'm using these IRs for my convenience so I can hear the "same" frequency response at home and at rehearsal, that should work. If my hearing at 3k is shot, it'll always be shot and someone will have to clue me in to boost that frequency if it is missing.

I don't think I've come up with anything novel at all - I'm just taking the data and making IRs that match it. My motivation is what my Out EQ looks like compared to the equal loudness contours (see attached). Looks a heck of a lot like the FM curve, or this IR.

But, if the consensus is "don't bother", as it seems to be, I'll go about my business in silence (no harm done).


Thing is though that your hearing response isn’t linear, especially if you have some degree of noise induced hearing loss/age related “wear and tear”.

An earplug is essentially linear, it makes everything like 30dB softer. Soft sounds are inaudible, louder sounds are loud but reduced, safer.

Neural hearing loss doesn’t work like that though, softer sounds can be inaudible, but louder sounds are perceived equally loud. It’s essentially a reduction of dynamic range.

90dB to you can sound the same as 90dB to someone with normal hearing, but, maybe you can’t detect the presence of sound at a given frequency until 50dB while someone else can detect that at 10dB.

It’s a very tough variable to account for without knowing your specific hearing thresholds, and then you still do have loudness contour, room acoustics, what sounds musically pleasing etc on top of that, tons of variables

I did some sound installation for a while, long before getting a doctorate in audiology, and when we’d do a club system, we didn’t try to make it linear, we tried to tune it for the venue acoustics but also to sound musically pleasing. We didn’t want an accurate frequency response, we wanted the system to thump lol.

Most people don’t like the sound of studio monitors after all. they aren’t pleasing, they’d prefer more lows, muted highs that aren’t icepicks etc

A linear system is good for making objective mixing decisions, but you don’t really encounter them that often except for home monitors and the recent popularity of “frfr” monitors.

Everyone thinks they need this linear response in everything, but that alone doesn’t really make a mix sound any better, it just can help things translate better to a different system, assuming you know the differences between them and what your trying to get out of it.
 
Back
Top Bottom