• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

EQ Correction/Room Emulation Paradox

aspillane789

Member
Joined
Oct 19, 2022
Messages
25
Likes
13
Hi there i have this thought. As you guys know we have a beautiful list of Impulse responses and Parametric EQ settings for almost every pair of headphones from AutoEQ. This is incredible; high praise for the dev who accomplished this.

Anyways There seems to be this anomaly when using eq correction with a speaker simulation. And that discrepancy is the Harman Curve. As we know the harman curve allows headphones to sound more natural and authentic to represent what real monitors in a control room would sound like. Okay Fair enough. But what if we want to follow this EQ correction with a room emulation software? Ie. AA Sienna, Waves NX oceanway, realphones, etc.

Well there lies the paradox o_O. Having A harman curve correction along with Speaker Emulation. Wouldn't this be doubling down on the "authentic" room sound, in a sense that The harman Curve's purpose is to sound more like speakers, then you have the room emulation, "Which also sounds like speakers."

My theory is that if one chooses to use a room emulation program/plugin its best to EQ your Cans to a flat response, then adding the room emulation would be more accurate/appropriate.

Thoughts?

Btw... I have the hifiman sundara 2020 revised earpads. If anyone knows where i can find the "Flat" eq settings for this pair of cans i would highly appreciate it.
 
OP
A

aspillane789

Member
Joined
Oct 19, 2022
Messages
25
Likes
13
EQ corrects frequency response, room emulation adds reverb. I don't see how using both at the same time would be problematic.
I dont think thats necessarily true. As an example. Realphones from dsoniq has a eq curve for the room, along with the spatial effects. not just for the headphone correction.
 

staticV3

Master Contributor
Joined
Aug 29, 2019
Messages
7,531
Likes
11,997
I dont think thats necessarily true. As an example. Realphones from dsoniq has a eq curve for the room, along with the spatial effects. not just for the headphone correction.
Yeah that still checks out imo.
If your room emulation emulates the frequency response colorations of a specific room as well, then you want to first EQ your headphones to neutral so that the colorations are reproduced as intended.
 
OP
A

aspillane789

Member
Joined
Oct 19, 2022
Messages
25
Likes
13
Yeah that still checks out imo.
If your room emulation emulates the frequency response colorations of a specific room as well, then you want to first EQ your headphones to neutral so that the colorations are reproduced as intended.

i hear you. i guess it matters how the Room emulation program is designed to be used as well. I think my question still provokes some thought however because i would assume the eq correction from the room itslef is closer to harman than flat.. Curious on other ppls opinion :)
 

ADU

Major Contributor
Joined
Jul 29, 2021
Messages
1,584
Likes
1,086
To EQ or not to EQ, that is the Q. :)

I have never used any of the above applications, so wouldn't really have a clue. I think I'd just ask the developers of the simulations what they recommend though, and whether adding your own software EQ would interfere with their simulation.
 
OP
A

aspillane789

Member
Joined
Oct 19, 2022
Messages
25
Likes
13
To EQ or not to EQ, that is the Q. :)

I have never used any of the above applications, so wouldn't really have a clue. I think I'd just ask the developers of the simulations what they recommend though, and whether adding your own software EQ would interfere with their simulation.
Your point is very valid. Infact I would argue that’s the best approach. We as audio geeks love to pick apart things. So I’ll say this.

What if they didn’t know what is best for their software?

The original question does remain. So supposedly we start with a harmon curve, which sounds speaker like.. then we have a speaker simulation which also sounds speaker like. Hmm a room within a room perhaps?.

Some could easily argue… Make good music, be an effective engineer, do your job and get it done without worrying about the minute details. True

But if I can get even %1 better with technical advances I would do so.
 

Cars-N-Cans

Addicted to Fun and Learning
Joined
May 19, 2022
Messages
819
Likes
1,009
Location
Dirty Jerzey
Its worth noting that the purpose of the Harman curve was to imperially determine what FR was necessary from headphones to make them be perceived as tonally neutral. Furthermore, its only valid with the fixture they used, which is why Amir has a GRAS fixture as well, so that the measurements correlate. The "room target" for speakers that we are all familiar with is really a function of the speakers radiation pattern and the room characteristics. The characteristic 1 dB/octave downward tilt originates from the fact that your typical box speakers have a 4 pi radiation pattern at low frequencies but this narrows to the listening window by the time it hits the cross-over region. Since a portion of the sound will be the direct sound, it introduces a slight tilt. The rest of it, I believe, also originates from measurements done in their own listening rooms as well. Again, the goal was to make them perceptually neutral tonally.

With that in mind, the goal is not to EQ to some curve, but to EQ to what makes the speakers/Headphones/etc. tonally neutral. This can be radically different for differing setups. An example would be my nearfield system where the room measurement follows the baffle step compensation of the woofers and has a nearly 4 dB/octave downward tilt until it hits the crossover region and the tweeter takes over. The reason for this paradox is that what is being measured is the direct sound of the speakers, and measurement microphones basically integrates the overall sound power that is emitted by the speakers. Even though the SPL levels at one point are constant with frequency, the total sound power emitted is not due to the changing spatial load of the speakers as the baffle takes over and they become more directional. This system, however, has the SAME tonality as headphones on the Harman target when A/B testing is done. This is no surprise given that tonally neutral speakers were used as a reference as part of the development of the Harman target, AFAIK.

In my humble opinion, having some program to try and emulate a room is a moot point unless you like the 2D soundstage along with reflections and all the other things that speakers do in less than optimal environments. We need to remember that these are not "features" but defects and shortcomings in the sound that arise from issues with sound reproduction in a room. What would be better would be some form of general purpose processor that applies some form of HRTFs to overcome the fact that headphones cant really interact with the pinnae, and you end up with things like IHL. That is my biggest complaint is the soundstage is always along the axis of the ears, and that is what makes them somewhat unnatural to me. With binaural audio (e.g. Sennheiser's Ambeo material and similar things), its much better, and more in line with the more natural imaging I can get from speakers in nearfield setups, but with added granularity and detail. I think this is what you are really after since speakers can fully interact with our heads and torsos and the headphones and IEMs cannot. This requires things like HRTF libraries and some why to match those to the user, or measure the users HRTFs. I believe Apples spatial audio is shooting for doing this now as well. Maybe at some point a company will make a box with SOTA audio performance that we can use in our own systems. But I think room emulation is a step in the wrong direction.
 
OP
A

aspillane789

Member
Joined
Oct 19, 2022
Messages
25
Likes
13
Its worth noting that the purpose of the Harman curve was to imperially determine what FR was necessary from headphones to make them be perceived as tonally neutral. Furthermore, its only valid with the fixture they used, which is why Amir has a GRAS fixture as well, so that the measurements correlate. The "room target" for speakers that we are all familiar with is really a function of the speakers radiation pattern and the room characteristics. The characteristic 1 dB/octave downward tilt originates from the fact that your typical box speakers have a 4 pi radiation pattern at low frequencies but this narrows to the listening window by the time it hits the cross-over region. Since a portion of the sound will be the direct sound, it introduces a slight tilt. The rest of it, I believe, also originates from measurements done in their own listening rooms as well. Again, the goal was to make them perceptually neutral tonally.

With that in mind, the goal is not to EQ to some curve, but to EQ to what makes the speakers/Headphones/etc. tonally neutral. This can be radically different for differing setups. An example would be my nearfield system where the room measurement follows the baffle step compensation of the woofers and has a nearly 4 dB/octave downward tilt until it hits the crossover region and the tweeter takes over. The reason for this paradox is that what is being measured is the direct sound of the speakers, and measurement microphones basically integrates the overall sound power that is emitted by the speakers. Even though the SPL levels at one point are constant with frequency, the total sound power emitted is not due to the changing spatial load of the speakers as the baffle takes over and they become more directional. This system, however, has the SAME tonality as headphones on the Harman target when A/B testing is done. This is no surprise given that tonally neutral speakers were used as a reference as part of the development of the Harman target, AFAIK.

In my humble opinion, having some program to try and emulate a room is a moot point unless you like the 2D soundstage along with reflections and all the other things that speakers do in less than optimal environments. We need to remember that these are not "features" but defects and shortcomings in the sound that arise from issues with sound reproduction in a room. What would be better would be some form of general purpose processor that applies some form of HRTFs to overcome the fact that headphones cant really interact with the pinnae, and you end up with things like IHL. That is my biggest complaint is the soundstage is always along the axis of the ears, and that is what makes them somewhat unnatural to me. With binaural audio (e.g. Sennheiser's Ambeo material and similar things), its much better, and more in line with the more natural imaging I can get from speakers in nearfield setups, but with added granularity and detail. I think this is what you are really after since speakers can fully interact with our heads and torsos and the headphones and IEMs cannot. This requires things like HRTF libraries and some why to match those to the user, or measure the users HRTFs. I believe Apples spatial audio is shooting for doing this now as well. Maybe at some point a company will make a box with SOTA audio performance that we can use in our own systems. But I think room emulation is a step in the wrong direction.
Thank you for your long and descriptive response. Sorry for the delayed reply in advance.

I brought up the idea of room emulations in software to multiple veteran mastering Enginner’s, including my personal mentor. They came to the same conclusion that all this will do is emulate the potential flaws from a control room, in a 2d soundscape like you said.

So I’ve come to my own conclusion that what I should do in the meantime, is get a good pair of neutral headphones, paired with my smsl m500 dac, use good eq correction from auto eq and pick the best profile, then finally in my monitoring chain use a plugin like goodhertz can opener for the crossfeed application. Thanks man super informative
 
Top Bottom