Target Curve Correction for HPs

[ric982]

Hi,
REW allows you to enter a target curve and will provide PEQ parameters to correct a measured signal to match the target. I'd like to use that capability generate PEQ parameters to correct a target curve to be 'flat'. I probably could figure a way to 'record' a flat curve, but it would be nice if I could just ask REW to generate the parameters for a flat measurement.

What I'm trying to do:
I have some headphones (HD600 In this case). I'd like to EQ the low end response to extend the bass beyond what the drive does naturally. So I want to enter the HP's frequency response curve as my 'target curve' and have REW generate the PEQ parameters that would convert a flat curve to this response. I can then take the REW filter paramaters and invert the gain coefficients so that it would 'correct'' the target curve (which is my FR) back to a flat curve which is what I want.

Is there way to do that now? Is this a function that could be added? If it would be added as a feature, it might be a bit better if could enter two curves - the starting curve and the target curve and then do what you would normally do with REW (i.e. correct a measurement to match a target with PEQs). This amounts to giving me a way to enter a curve for what the 'measurement' is. For HPs, there is no way for me to measure what the head phone frequency response is at my eardrum (without buying a very expensive apparatus that emulates a head and ear).

 

[canavan]

For HPs, there is no way for me to measure what the head phone frequency response is at my eardrum (without buying a very expensive apparatus that emulates a head and ear).

You'd actually need a set of calibrated in-ear microphones to measure what's getting to your eardrums. The Smyth Realiser does everything you want in a box, but at a price, and it works really well (but mine got lost in the mail, so...)

Since you apparently would consider generic correction curves acceptable, AutoEq may be just what you're looking for.

 

[Mitchco]

Another approach: http://www.davidgriesinger.com/

 

[ric982]

Both interesting responses. So I have always been amused by the 'art of asking a question'. I tend to try to limit my question to specifically the question I'm trying to answer. A wise co-working once responded to one of these with "What are you trying to do?" because to him, the answer would vary with this additional bit of information. So let me give the back story and see if that helps.

I'm setting up an ATMOS mix room. I plan to use REW to measure my room response at the listening position and have it spit out filter parameters to correct my room response to match the Dobly target curve (which is not flat by the way). So this handles my speaker monitoring. I now want my headphones to also have the same target curve response. There are multiple sites that have used head measurement systems to measure the response of my HPs and their published reviews can be compared and are generally in agreement below about 6K. So I could use these measurements to program a target curve in REW and (with a flat measurement) have it provide the correct I need to correct the lower end of my HP response to a studio ideal HP response curve. My fall back here would be to use something like Sonarworks or NX plugins. My concern with these is that there's no guarantee they won't change their correction over time (I think Sonarworks already has once). In either case, I can then add the Dolby target curve on top of that So that's kind of my plan of attack right now. In the absense of what I'm asking for in the original post, I would just loop the REQ speaker line out to REW mic line in REW and that should give me a flat response and then proceed as described.

Perhaps another alternative would be to just use a EQ plugin and play with the PEQs till I get a curve that looks like the inverse of what I want to correct on my HPs. But I've seen some of what REW generates for PEQs and they seem to be more 'efficient' (fewer fitlers, smoother response, etc) and more accurate at accomplishing the desired target than I am doing it manually.

 

[canavan]

I now want my headphones to also have the same target curve response.

As measured how and where? You may want to ask Dolby what they recommend for headphones.

Otherwise, if you want to perfectly emulate a loudspeaker that has an in room response with the Dolby target curve at your listening position, my recommendation would still be a Smyth Realiser A16 - which would be indistinguishable from the speakers you already have, if you do everything correctly.

 

[JStewart]

Since you apparently would consider generic correction curves acceptable, AutoEq may be just what you're looking for.

@canavan have you seen this program that makes use of those GitHub files?

AutoEq

Automatic headphone equalization

autoeq.app

@ric982 the program will also EQ the HD600 to flat if that helps. From there, if you can create a text file of the response you’re looking for, it can be imported to REW and you can create a 2nd set of filters to EQ from flat to it. You may be limited by the number of filters your DSP can manage IDK.

 

[ric982]

@ric982 the program will also EQ the HD600 to flat if that helps. From there, if you can create a text file of the response you’re looking for, it can be imported to REW and you can create a 2nd set of filters to EQ from flat to it. You may be limited by the number of filters your DSP can manage IDK.

So my understanding of reading HP info, the response curve you need from a HP to make it sound like 'flat' speakers in the studio is not a flat EQ. The way they sort of figure this out is that make 'head models' that allow a mic to be placed at the position of the ear drum and have representative pinna. So they can see the response of the flat speakers transferred to the ear drum is. They then measure a given HPs on the head model and expect it to produce an equivalent response. The ideal HP FR would be something like flat below 1K but similar to what the HD600 has now (with some adjustments). Plugin's like Waves NX know how to correct for this (though I'm not sure they are all the same and that they don't change over time - so I was hoping to lock it in with some EQ so I know what I've got). I should be able to look at the plugin and see what correction they are going to apply and model that in an EQ so it's fixed.

As a second filter, you'd want to add the Dolby room curve on top of that, as you say. I need to create a filter that converts from a flat response to the dolby room curve. I know how to enter the target curve in REW. I know how to tell it to limit to the number of filters I have and other constrants. And I know that you can make a measurement (of the room's response which is not flat) and have REW provide the correction to get it to the target curve.

But how do you tell REW that the measurement is flat? (without taking a flat measurement)

Again, I could probably create the EQ manually, but the room curve has some slopes that are not integral multiples of 6 dB (-1.5 db/oct and -3 db/oct) so it's a little tricky to keep the slope constant with PEQs.

 

[ric982]

Ok - my bad.

I just checked the REW manual and there is a section on importing measurement files in various formats. I should be able to input the 'flat' response measurement I need and then have it provide the EQ to the target curve.

Again, sorry for lack of due dilligence on my part.

 

[ric982]

I guess to maybe take this one step further, if I have the HD600 response curve (like the one shown above), I can input that as my measurement. And if I pick an 'idealized' HP response curve (more than one of these running around, but for example see "Sound Guy's Studio Head Phone Curve"), I could input that curve as the REW target curve, and then have REW generate the PEQ filters to correct ithe HD600 from its response to the choosen ideal response.

It probably makes sense to keep the HP correction EQ separate from the room curve EQ if it's in plugins in the DAW. But if I I have enough channels in my hardware PEQ, I could also combine the two target curves and generate one EQ correction fto handle both.

I suspect some of the other programs suggested could accomplish the same thing, but I think I like keeping it in REW if it's possible. That you all for your suggestions and coments.

 

[AudiocRaver]

My advice for EQing headphones is DON'T. Interaction with the ear causes a lot of the HF variations and when EQed the result just ends up sounding weird and extremely sensitive to placement variations on the head.

 

[JStewart]

My advice for EQing headphones is DON'T. Interaction with the ear causes a lot of the HF variations and when EQed the result just ends up sounding weird and extremely sensitive to placement variations on the head.

Can we agree to disagree on that? My experience is the opposite, but you have to start from a measurement on a headphone rig and a target that works with said rig. With that some may prefer to adjust from there with broad tone adjustments.
@Mitchco uses a different method. His general process is outlined on his website. I use filters he did for a model of cans I have. Fabulous.

 

[ric982]

My advice for EQing headphones is DON'T. Interaction with the ear causes a lot of the HF variations and when EQed the result just ends up sounding weird and extremely sensitive to placement variations on the head.

Can we agree to disagree on that? My experience is the opposite, but you have to start from a measurement on a headphone rig and a target that works with said rig. With that some may prefer to adjust from there with broad tone adjustments.
@Mitchco uses a different method. His general process is outlined on his website. I use filters he did for a model of cans I have. Fabulous.

So maybe a few comments:

1) I was more focused on tweaking the low end response of the headphones - like extending the bass. There are a lot of discussions on forums that HP600s can be pushed to extend there bass extension. Other than that,, their response is pretty much in line with an idealized studio HP response. There may be a minor correction needed in the mid bass as well, but were talking maybe a dB or so over an octave or so.

2) I'd be suspicious of doing anything on the high end unless it was a vary broad adjustment on the order of what a 'treble knob might do on a 3 band EQ. Exceptiions on a per case basis maybe.

3) I'm using these in an ATMOS mix room and Dolby has a target room curve they expect you to apply to your speaker monitoring in the room. I haven't heard back from them yet whether I should apply this to the HPs. Part of the room curve they want on speakers has to do with having a consistent room response between different ATMOS mix rooms, but HPs have a different situation (e.g. no room). So they may say no. But, I was sorta gearing up for the possibility that they would say yes. Their target curve has something like a -1.5 dB/oct roll off from 1.6K to 10K and a -3.0 db/Oct above 10K - so its very gradual and broad. This would go on top of whatever correction you might make to modify the HPs inherent response. And it was this curve that I was interested in getting REW to spit out some PEQs for in particular. If I can make the measured curve 'flat' and use the Dolby room curve as the target, the generated EQ would be just add the room curve on top the HP's natural response.

So In some ways I agree with both of you.

 

[ddude003]

I also use a layered approach using Mitch's Hang Loose Convolver along with REW's Trace Arithmetic to achieve a more or less flat FIR DSP room correction... Then layer, to taste, a Parallel Parametric EQ plugin like the GML 8200 or Tone Projects Michelangelo... For multiple channel work you might take a look at Liquid Sonics Cinematic Rooms...

 

[JStewart]

1) I was more focused on tweaking the low end response of the headphones - like extending the bass. There are a lot of discussions on forums that HP600s can be pushed to extend there bass extension. Other than that,, their response is pretty much in line with an idealized studio HP response. There may be a minor correction needed in the mid bass as well, but were talking maybe a dB or so over an octave or so.

It would seem the case that the headphones are neutral sounding when their response is compared to the Harman over ear headphone target.

Not REW, but Using AutoEq and adjusting the bass target to 0dB vs +3dB for the Harman target looks like this

and yields these PEQs:
Preamp: -8.12 dB
Filter 1: ON LSC Fc 105.0 Hz Gain 2.9 dB Q 0.70
Filter 2: ON PK Fc 20.7 Hz Gain 3.0 dB Q 5.82
Filter 3: ON PK Fc 24.7 Hz Gain 3.1 dB Q 1.87
Filter 4: ON PK Fc 32.4 Hz Gain 0.3 dB Q 2.26
Filter 5: ON PK Fc 102.5 Hz Gain -2.0 dB Q 0.89
Filter 6: ON PK Fc 211.7 Hz Gain -1.1 dB Q 1.74
Filter 7: ON PK Fc 1267.5 Hz Gain -2.0 dB Q 2.69
Filter 8: ON PK Fc 3113.6 Hz Gain -2.9 dB Q 3.73
Filter 9: ON PK Fc 9075.5 Hz Gain 6.5 dB Q 1.18
Filter 10: ON HSC Fc 10000.0 Hz Gain -3.9 dB Q 0.70

I don’t see why you cannot enter only the 1st 4 or 5 if you only wish to provide bass extension to “flat”.

3) I'm using these in an ATMOS mix room and Dolby has a target room curve they expect you to apply to your speaker monitoring in the room. I haven't heard back from them yet whether I should apply this to the HPs. Part of the room curve they want on speakers has to do with having a consistent room response between different ATMOS mix rooms, but HPs have a different situation (e.g. no room). So they may say no. But, I was sorta gearing up for the possibility that they would say yes. Their target curve has something like a -1.5 dB/oct roll off from 1.6K to 10K and a -3.0 db/Oct above 10K - so its very gradual and broad. This would go on top of whatever correction you might make to modify the HPs inherent response.

The Dolby Atmos music mix EQ was found in this Document.
If you use them also the headphones should be EQ’d as you wanted. Whether or not it then sounds as expected TBD.

 

[ric982]

It would seem the case that the headphones are neutral sounding when their response is compared to the Harman over ear headphone target.

Yes. HD600 has been a 'classic' studio HP for decades. HD650 is probably a bit warmer (mid/lower boost) and more to the tastes of music lovers. They are lean a bit more in the direction of the Harman curve than the HP600, but not quite that far I think. The two HPs are pretty close and you might not notice the differences without some A/B comparison.

I don’t see why you cannot enter only the 1st 4 or 5 if you only wish to provide bass extension to “flat”.

Exactly.

The Dolby Atmos music mix EQ was found in this Document.

Yes I"ve seen that and that would probably work fine. I guess I like to be able to test in house with my existing tools to confirm anything I find on the internet (or confirm my tools too) - but in general I would trust Grace. There monitor controller is mentioned in Dolby ATMOS information.

If you use them also the headphones should be EQ’d as you wanted. Whether or not it then sounds as expected TBD.

Well the idea is that you want your HPs to sound like your speakers because if you make mix some decisions on the HPs, you want that to translate back to your speakers (i.e. It sounded right on the HPs and it still sounds like I thought now that I hear it on my speakers). Whether the Dolby room curve is the right "room" curve to add to a HP to achieve that, I'm not positive, but I'm guessing it probably is. Smarter and more experienced minds may know better. Hopefully my ears will know when they hear it.

 

[ddude003]

Aren't Sonarworks or NX plugins basically an Impulse Response of some "well known" mix rooms, their monitors and HPs? Why not create your own IRs of your mix room including your current room corrections... And once you get something you consider flat for your HPs add in/layer your room IRs/FIRs...

 

[JStewart]

Smarter and more experienced minds may know better. Hopefully my ears will know when they hear it.

Maybe use pink noise to compare speakers to phones.
Good luck

 

[ric982]

OK. So I've got REW to generate EQ filters to correct from flat to the Dolby curve.

On the high end it used 4 filters compared to Graces two. The Grace's looked smoother where the REWs have some waving above and below target - both still very close to targe and withing the Dolby +/1 1dB spec. On the Low End, REW ignored the +1 dB correction. I though this was because you cn only tell it at mimium to stay within 1 dB of the target and the correction is +1 dB, but I set the target to +2dB and it still ignored it. It has some statement about not correcting below some freq and above some freq based on some criteria that I haven't quite decoded and says you can manually add EQs to 'drive beyond' your speakers rolloffs" with warnings about over drive. The REWs curves on the hight end all had Q's of "1" to a few decimal places. Graces were a mix. I'm not sure of the phase implications of that, but probably fewer filters and smoother response seems like the better choice. Unlike Grace's solution, I'll probably use a shelf filter plus some PKs for the +1 dB at the low end. (because my EQ has a shelft), and leave the +1dB all the way down. The speaker should roll off naturatlly where it does and the room response (with perhaps room gain) will do what it does.

In playing with REW, it seems that hand tuning the EQs works fine. And their auto correct is an easy button, but you can tweek what you get. I'll definitely need it to help correct my room speakers to target - the room response of the speakers is not flat to start with.

For the HPs, I'm finding that I don't see a lot of review measuments that make the data available (e.g. a .csv file). I'll probably just trust the HD600 measurement and EQ suggestions that someting like Wave's Abbey Road Studio 3 (which I have) generates for the HD600 add the Dobly correction on top. If the plugin doesn't extend the bass, I can add more EQ to attempt to do that, though I'm sure there's some limit to what the driver will take before cause too much distoration or damaging itself.

Anyway, I think we are done here. My original objective of trying to get REW to translate to a specified input curve to a specified output curve has been achieved. It's been a fun discussion as well. Thanks everyone for your input and advice!

 

[ric982]

One update. Dolby came back on the HP EQ question:

"In terms of headphones, when mixing for the binaural headphone experience we would not recommend applying equalisation (i.e., match them to the music curve on the speakers). In addition, when mixing on headphones as a workflow step before using speakers, we have found it is not necessary to apply music curve."

I'll qualify that in it is saying not to apply the Dolby Target Curve on top of the HP response. I doesn't mean that you shouldn't EQ the HPs if their natural response is different thant some target you feel is better (e.g "ideal" HP response curve, bass extension, Harman, more like your speakers sound, whatever).

 

[ddude003]

You have mentioned bass extension a few times as something your looking to enhance... I would suggest looking at the plugin emulation some of the classic Parallel Paremetic EQs like the Manley Massive Passive, GML 8200 and the Hendyamps Michelangelo... Tokyo Dawn Labs has just released a bundle of EQs that might be interesting... TDR Special Filters Bundle...

Good luck in your search for your sweetest signal chain and work flow...

BTY Tone Projects is having a black friday sale... It would be hard to beat a signal chain composed of Baseane Pro, Kelvin Tone Shaper, Unisum Mastering Compressor and Michelangelo tube EQ...

 

[ric982]

Aren't Sonarworks or NX plugins basically an Impulse Response of some "well known" mix rooms, their monitors and HPs? Why not create your own IRs of your mix room including your current room corrections... And once you get something you consider flat for your HPs add in/layer your room IRs/FIRs...

I just looked at Studio 3 and it provides several functions:

1) It allows you to listen to either a near-filed (2.0?), mid-field (7.1?), or far field (2.0?) speakers with the Studio 3 acoustics.
2) It allows you to correct the EQ for a given model HP (270 models). It says it provides a 'neutral' response that should let you expoerience the studio 3 speakers in Studio 3. There are some customizations for entering your personal head dimensions that affect the HP rendering - or the default is based on the average human head. Whether this flattens or extents the bass of the HP, I don't know yet - haven't done any listening. I was hoping it would let you see the correction or even tweak it, but it is a black box that you can turn on or off. The headphone measuremesnts were obtained from headphones.com. I'll assume that Abbey Road studios more or less controls the plugin and I'd trust them to know what they are doing.
3) It allows you to rotate the sound image as if you are turning your head. You can set the roation angle manually. If you have certain Blue Tooth head tracking devices, it will do it automatically - can turn on or off. Picture in the plugin of the room also rotates around the room - though if you're heads turned away from the screen, you won't see it.
4) It does binaural rendering. The width of the track it is in determines the number of channels that are rendered into 3D immersive sound - at least up to 7.1 width (I"ll tried that) - maybe 7.1.4?. On the left side of the plugin control window - there are levels for the input channels with solo and mute buttons. On the right side there are levels (2) for the binaural output. Binaural is supposed to make the HPs sound like the speakers would - even if it's stereo input. 2.0 on headphones is different that binaral rendering of 2.0 stereo on headphones.

One thing that this technology provides is a consistent monitoring environment between different locations - at least over HPs. The Dolby room target curve and the other specs they require also provides this consistency in the case of speakers in a studio. FWIW - it's on sale for Black Friday for $30.

As far as my room's IR, I can get that from REW. At least I can get the IR (or frequency response - they are just forier transforms of each other) of one speaker at a given mic position. If I have good room treatment so that the room matches the Dobly specs and I've calibrated the speakers to the Dolby House currve, then the result should be that my room should sound like Studio 3 if it's tuned like Dolby says. That's the point of the Dolby specs - make all their mix rooms sound the same. Given Dolby's response of don't bother adding their room curve to HP chains, I think the answer is don't bother to do that for HPs, but I think that is what NX is intended to do.

You have mentioned bass extension a few times as something your looking to enhance... I would suggest looking at the plugin emulation some of the classic Parallel Paremetic EQs like the Manley Massive Passive, GML 8200 and the Hendyamps Michelangelo... Tokyo Dawn Labs has just released a bundle of EQs that might be interesting... TDR Special Filters Bundle...

Good luck in your search for your sweetest signal chain and work flow...

BTY Tone Projects is having a black friday sale... It would be hard to beat a signal chain composed of Baseane Pro, Kelvin Tone Shaper, Unisum Mastering Compressor and Michelangelo tube EQ...

I have a lot of EQ plug-ins, some emulate hardware which provide some 'analog warmth' and 'non-linear saturation". I don't think having EQs of a particular type is the issue as much as knowing what filtering you need to achieve your objective. It's actually important not to have any coloring in your monitoring chain in a mix room. You want whatever coloring you choose to add in your recording chain. If it's in your monitoring chain, what you hear on your monitors is not what you get in the recording.