Harman curves sound different? (room x headphones x IEMs )

[Matias]

The other day I tried comparing my main system with my in ear monitors. The main system is in my living room and I applied EQ to a target of a linear tilt following the slope simimlar to the most preferred response according to Harman's research. I was sitting in the sweet spot listening to a song, and quickly switched to the same song playing through my IEMs, which were connected to my phone using the Wavelet app to corrected their response with AutoEQ database, which in turn aims at the 2019 Harman target curve for IEMs. Switched between them back and forth a few times to be sure.

I was shocked to find that they sounded nothing alike! The IEMs were a lot brighter, with treble and bass a lot more pronounced than my main system, which in comparison had a lot more mid bass and mids. I find my main system to be right and the IEM highly V shaped. This was not subtle!

This experience left me thinking...

Assuming that recorded acoustic instruments sound right and natural in one, the other necessarily is going to sound wrong and artificial.

Shouldn't the target curves try to be psychoacoustically similar? I mean, I think that Harman got both target curves from what most people preferred in each case, OK, but apparently they had no intention that each remotely tried to match the other.

Did I do anything wrong with my testing? Anyone tried this too and had different thoughts?

 

[dc655321]

Without seeing measurements, particularly of your speaker system, it's difficult to comment constructively.

I will say that headphones often have a much different sonic character from loudspeakers, ime...

 

[Matias]

@dc655321 my main system response is here, and the IEMs were Moondrop Aria corrected to Moondrop Starfield as they measured identically.

I agree speakers and IEMs and headphones will sound different because of their construction, position and the sound directivity. But they should at least be in the same ballpark in tonality, and in my experience they were not.

Have you tested this yourself too in your system? Does anyone have a different experience like "in my system they sound very similarly"?

 

[pma]

Headphones always exaggerate details in a non-realistic manner and are overbright. Your impressions are correct.

 

[Matias]

I had to apply the manual EQ below for the IEMs to be somewhat close to the speakers.

 

[Matias]

Pinging @jaakkopasanen any idea?

 

[dc655321]

@Matias I don't think it is realistic to expect your iems and speakers to have similar tonality.
I understand that is how the Harman curve is portrayed, but still...
The Harman curve for headphones is a best-fit guess estimate, and sensitive to anatomical variance.
Plus, iems have nothing to attenuate higher frequencies or produce reflections - it's a direct, sonic injection to the eardrum!

I suspect that listening to anechoically-flat speakers, near-field, in an anechoic chamber may produce the tonal similarities you thought you might find...

 

[Matias]

@dc655321 Yes, that is what I wanted to check.
1. Did I test something wrong? If not,
2. Do others agree with their testing? If so,
3. Should the headphone and IEMs targets try to meet the speakers target?

The last one, in my opinion, they should. The music should sound natural in all mediums. For example, a piano should sound like a piano when playing through speakers, over the ear headphones or IEMs, and not have its tonality altered depending on which one we listen.

 

[Tim Link]

My experience with headphones vs my main speakers has been similar. The midrange is more prominent on the main speakers with the headphones sounding brighter up top and with more bass if adjusted to the target curve. I can adjust the speakers to match the bass and still have it sound good. This has caused me some concern at times because once I've adjusted to the headphones it makes the speakers seem dull at first. I've had to learn to give it some adjustment time. If I adjust to the speakers the headphones can sound overly crispy at first. I don't see how a speaker firing sound across your ear can ever compete at the highest frequencies with a headphone that is firing straight into your ear canal unless you boost the treble considerably on the speakers. For some reason though my ears find the extra treble from the headphones acceptable in a way that I don't from boosting the treble on the speakers.
Lately I've been using the curve rtings.com came up with when they measured the Grado SR325e. Using that curve adjustment I'm finding my headphone and main speaker listening experience has become much more similar. It would seem that the dummy head they used to measure there matches my ears fairly well.

 

[abdo123]

I don't know about IEMs but headphones tuned to the Harman curve are one of the most clinical ways to listen to music. I find it really difficult to find a bad sounding track on my headphones.

I did the same experiment you did, and my headphones felt way more accurate because of the sublime bass response. having the frequencies decay exactly when you need them to decay is a marvelous feat and experience on headphones. it is something that i haven't been able to create with my budget limited main speaker setup.

 

[Tim Link]

I don't know about IEMs but headphones tuned to the Harman curve are one of the most clinical ways to listen to music. I find it really difficult to find a bad sounding track on my headphones.

I did the same experiment you did, and my headphones felt way more accurate because of the sublime bass response. having the frequencies decay exactly when you need them to decay is a marvelous feat and experience on headphones. it is something that i haven't been able to create with my budget limited main speaker setup.

For years my speakers and room couldn't even get close to headphones for bass clarity. Lately I'm surprisingly close. The headphones still have the edge but I don't feel like I'm missing much detail or hearing excessive coloration from the speakers and room. If you use REW you can run a sweep through your speakers and measure it at the listening position. Look at the clarity tab and the C50 measurement. I'm currently getting around 10db on the C50 chart in my front room through most of the bass and lower midrange. It'd be interesting to see what other people are getting and how they perceive the sound. Here at ASC we are thinking about 10db is a good C50 goal, with 15db or higher being excellent for bass articulation.

 

[abdo123]

For years my speakers and room couldn't even get close to headphones for bass clarity. Lately I'm surprisingly close. The headphones still have the edge but I don't feel like I'm missing much detail or hearing excessive coloration from the speakers and room. If you use REW you can run a sweep through your speakers and measure it at the listening position. Look at the clarity tab and the C50 measurement. I'm currently getting around 10db on the C50 chart in my front room through most of the bass and lower midrange. It'd be interesting to see what other people are getting and how they perceive the sound. Here at ASC we are thinking about 10db is a good C50 goal, with 15db or higher being excellent for bass articulation.

I guess I would be pretty close too if I manufactured bass traps .

thank you for the hint, I'm out of fresh REW measurments (been using Dirac lately) so i will check it out tomorrow

 

[Tim Link]

I guess I would be pretty close too if I manufactured bass traps .

thank you for the hint, I'm out of fresh REW measurments (been using Dirac lately) so i will check it out tomorrow

I have to confess - I don't currently have any bass traps in my main room! I'm using woofers stacked floor to ceiling to take care of vertical bass storage. Side to side I don't have as much pattern control but each corner is disrupted because it has a huge stack of horn woofers in it so the bass can't concentrate so well there. I've also got windows and sliding glass doors slightly ahead of the speakers and they tend to leak bass so that helps a bit. My biggest problem is front to back reflections but that's broken up partially because the back of the room is staggered rather than having a solid back wall, and it's vented in to other rooms. Still, I do think I would benefit greatly from bass traps in the back of the room but I need to make space for them somehow. It's a multi purpose room and I don't live alone so I can't just do anything I want in there. I've already hogged the two front corners with my speakers.

TubeTraps traps have worked great for me in my bedroom system with it's bookshelf speakers and small subwoofers. That was in heavy use when I was living with my mom, helping her take care of my dad. Now that I'm back at my place I have yet to put the bedroom system back together. I'm too busy listening to the main system, and there's too much junk in my room.

 

[Pdxwayne]

For years my speakers and room couldn't even get close to headphones for bass clarity. Lately I'm surprisingly close. The headphones still have the edge but I don't feel like I'm missing much detail or hearing excessive coloration from the speakers and room. If you use REW you can run a sweep through your speakers and measure it at the listening position. Look at the clarity tab and the C50 measurement. I'm currently getting around 10db on the C50 chart in my front room through most of the bass and lower midrange. It'd be interesting to see what other people are getting and how they perceive the sound. Here at ASC we are thinking about 10db is a good C50 goal, with 15db or higher being excellent for bass articulation.

Is there a thread that talk more about REW Clarity tab? May I know how loud you measure affect the clarity score?

I just did some adjustment of my computer room (using minidsp to high passed speakers and add a sub) and did some REW measurements.

Sweep is around 80db for bass region. Here are the chart for 20hz to 200hz measured db and the clarity chart for the same measurement. Looks good?

 

[Tim Link]

Is there a thread that talk more about REW Clarity tab? May I know how loud you measure affect the clarity score?

I just did some adjustment of my computer room (using minidsp to high passed speakers and add a sub) and did some REW measurements.

Sweep is around 80db for bass region. Here are the chart for 20hz to 200hz measured db and the clarity chart for the same measurement. Looks good?

View attachment 124465


View attachment 124466

Thanks for posting that. It looks very good. I wouldn't expect it to change much with volume unless perhaps you manage to structurally excite the room somehow. How does it sound? Do you feel you are getting good bass clarity in there compared to headphones? I don't think we have a thread related to REW clarity. Maybe I should start one?

 

[Pdxwayne]

Thanks for posting that. It looks very good. I wouldn't expect it to change much with volume unless perhaps you manage to structurally excite the room somehow. How does it sound? Do you feel you are getting good bass clarity in there compared to headphones? I don't think we have a thread related to REW clarity. Maybe I should start one?

My headphone can't really compete in term of bass feel. Headphone feels too bass light after listening to my stereo setup. So, hard to do apple to apple comparison. But, I do think my stereo setup is still clean, much cleaner than before I use minidsp. Do start a thread for REW clarity. I am interested in seeing people's charts and clarity comment. Thx

 

[Matias]

I was not even talking about clarity and details, just overall tonality between target curves. Way off in my experience.

 

[thewas]

Several reasons why they usually don't sound the same:

There is no real Harman generic target for loudspeakers, as the target of a loudspeaker depends on its directivity, the room reverberation and the listening distance, ideally a linear direct sound above transition frequency is preferred, see more here.

The Harman headphone target was not created by having as reference neutral loudspeakers in a good room (that was only the base line) but by listeners adjusting bass and treble to their taste listening to some existing recordings, a quite flawed approach continuing Audio's Circle of Confusion.

The tonal perception of headphones is different for every human as it depends on his individual HRTFs which are partially bypassed by the headphones, a reason why generic EQs and binaural localisation usually do not work on most listeners optimally.

 

[Pdxwayne]

My headphones is AKG 371. Amir said pretty much follow the Harman curve.

My computer room setup measurements pretty much following the curve Amir likes for speakers.

For me, my experience is also headphones sounded a lot brighter....

 

[Matias]

Several reasons why they usually don't sound the same:

There is no real Harman generic target for loudspeakers, as the target of a loudspeaker depends on its directivity, the room reverberation and the listening distance, ideally a linear direct sound above transition frequency is preferred, see more here.

I understood that in the listening position, it does not matter if the speaker has wide or narrow dispersion, dipole or front firing, the preferred curve was the linear curve with the downward slope no matter how it was created and reflected in the room, as long as it arrived at the sweet spot in that curve. Did I get that wrong?

The Harman headphone target was not created by having as reference neutral loudspeakers in a good room (that was only the base line) but by listeners adjusting bass and treble to their taste listening to some existing recordings, a quite flawed approach continuing Audio's Circle of Confusion.

Agreed!

The tonal perception of headphones is different for every human as it depends on his individual HRTFs which are partially bypassed by the headphones, a reason why generic EQs and binaural localisation usually do not work on most listeners optimally.

Shouldn't each ERTF also affect how the subject perceives a speaker in the room? So if the idea was a headphone/IEM target that resembles as close as possible a the linear downward slope in a room, then ERTF affect all curves equal (speaker, headphone, IEM).
If they were at least trying to match those curves, then with a few dozen subjects the average ERTF would be used and the end result would not be so different sounding imo.