Preference curve for speakers and headphones....I still don't get it....

[Pdxwayne]

I have been listening to music mostly with recommended headphones (ak371 and he400se) for the last couple months.

This is one song that I like because it is kind of funny:

With k371, the cymbals are noticeable and sounded sharp. With he400se, the cymbals sounded even sharper, with bites!

So, yesterday I decided to listen to same song with my AV setup in living room. First impression was: where are those cymbal's sharpness??

So, I checked again today using stereo
mode (2.2), with Audyssey xt32 room eq on and no other enhancements. I still missing that sharp cymbals.

I then checked my other AV setup in my family room, using stereo mode (2.1), with Pioneers MCACC Pro room eq on, cymbals still missing that sharpness and bites....

Assuming that Audyssey xt32 and MCACC pro both are somewhat following the Harman curve for speakers, I am now confused about which curve is "right"...

How can two distinctly different curves being "preference" curve? I don't get it....

Since I am used to the headphones sound, suddenly hearing my speakers setup made me think that my speaker systems sounded dull. Naturally, I kind of prefer the headphones sound.

On the other hand, if I have been listening mostly to my speakers setup for the last few months, listening to headphones now would make me think headphones sound way too bright/sharp!

This got me thinking:

What if those teenagers and younger generations who use headphones (that mostly comform to headphones preference curve) exclusively for music are invited to Harman blind speaker preference listening test today?

If they are given a choice of two speakers, one conforms to the normal Harman speakers curve, and the other conforms to the headphones preference curve, which speaker will they prefer today?

Logically, I will say majority of them will select the one that conforms to headphones curve.

On the other hand, if we invite older generations, who listen to music exclusively with speakers (that mostly conform to Harman speakers curve) for a blind headphones preference test today, will most of them prefer a headphones that conform to Harman speaker curve over the one that conforms to headphones curve?

So, yeah, I still don't get these preference curve...

 

[dasdoing]

the 2 curves actualy kind of match when you subtract the difuse field curve from the headphone curve.
you room setup probably deviates too much from the harman room setup. if you have more direct sound then them it will obviously sound dull

 

[Eetu]

Another aspect to consider is that if you listen to your speaker setups at lower-than-reference levels you're missing out on some treble bite (and bass slam) unless you have loudness compensation. But yeah, likely your room. Maybe try a flatter target curve?

 

[abdo123]

Room acoustics are a bitch. Nothing unusual imo.

 

[Pdxwayne]

Thanks all for comments.

Yeah, for both speakers setups, I sit about 10 ft or more from my speakers. Rooms are large size and mid size and heavily carpeted. As I recall, when I last used REW to measure from my seat in my living room, the measurements looked OK (before applying any eq). There is a steep downward slope after 10khz, a bit more than I would like, but I can't really sense over 15khz anyway.

When I did my listening yesterday, I listened loud (should be over 90db max, but most likely not over 100db max) and can "feel" the bass fine. Just that the cymbal is much more subdued....as compared to what I am used to with headphones.

 

[Pdxwayne]

the 2 curves actualy kind of match when you subtract the difuse field curve from the headphone curve.
.....

Can you show examples with pictures? Not so sure I follow....

Thanks!

 

[Coin3]

Another possibility could be that your HRTF deviates significantly from average in the cymbals' frequency range. How you hear sounds in free space sets the reference point for what your brain considers neutral. If the frequency response of your headphones on your ears boosts those frequencies higher than your HRTF, then the cymbals will sound sharper. Of course, if you listen to those headphones for extended periods of time, it's possible that immediately switching over to neutral speakers would sound dull in comparison.

About the preference targets, I'm just going to reiterate what I've read on these forums many times before. For a pair of speakers in a normal room, there is no real preference target. Good speakers should have a flat anechoic response. When they are placed into a room, the response will curve down, and there will be deviations caused by many factors in the room. A lot of people seem to agree that as long as you have good speakers, the only deviations you should try to correct with EQ are those in the bass frequencies.

The Harman preference target for headphones is essentially a smoothed, averaged HRTF meant to make headphones sound like good speakers in a good room. Everyone has different HRTFs, so it's quite likely that any one person's truly neutral headphone preference curve will deviate from Harman's target. If you are confident that you have good speakers in a good room, I would recommend using that as a reference from which to tune the frequency response of your headpones.

 

[Pdxwayne]

Another possibility could be that your HRTF deviates significantly from average in the cymbals' frequency range. How you hear sounds in free space sets the reference point for what your brain considers neutral. If the frequency response of your headphones on your ears boosts those frequencies higher than your HRTF, then the cymbals will sound sharper. Of course, if you listen to those headphones for extended periods of time, it's possible that immediately switching over to neutral speakers would sound dull in comparison.

About the preference targets, I'm just going to reiterate what I've read on these forums many times before. For a pair of speakers in a normal room, there is no real preference target. Good speakers should have a flat anechoic response. When they are placed into a room, the response will curve down, and there will be deviations caused by many factors in the room. A lot of people seem to agree that as long as you have good speakers, the only deviations you should try to correct with EQ are those in the bass frequencies.

The Harman preference target for headphones is essentially a smoothed, averaged HRTF meant to make headphones sound like good speakers in a good room. Everyone has different HRTFs, so it's quite likely that any one person's truly neutral headphone preference curve will deviate from Harman's target. If you are confident that you have good speakers in a good room, I would recommend using that as a reference from which to tune the frequency response of your headpones.

Thanks. So I guess I should assume the headphones are giving me too much highs?

The pair of front speakers used in living room AV setup is Paradigm Signature S2. Should be decent in a room, especially after room eq with Audyssey xt32.....

From https://www.soundstagenetwork.com/measurements/paradigm_signature_s2/

Here is one plot:

 

[whazzup]

Alternatively, eq your speakers to match your headphones' highs? Just to see how it sounds.

Just to add, that sultans of swing cover is lovely!

 

[Pdxwayne]

Alternatively, eq your speakers to match your headphones' highs? Just to see how it sounds.

Both of my AV preamp/avr are old and manual eq option is very basic fix bands adjustments, so will need something like minidsp in my chain.

But, interesting idea. May try it when I have more time. Thanks!

 

[Pdxwayne]

For members who own both headphones setup that conforms to headphone preference curve and speakers setup that conforms to Harman speakers preference curve (speakers more than 6 ft away from you), it would be great if you can check with the same song.

Cymbals sound the same to you with both setups? Or different?

Thanks!

 

[JRS]

With the Sundara's and no EQ vs any of the speakers I have owned with the exception of the Thiel CS3.5's I had for 6 months, before giving up on them, the highs sizzle, and cymbals depending on recording can provoke the sibilance to almost painful levels. Using EQ for Harmon curve, things are much better, but with residual sibilance that at high volumes quickly fatigues. Keep in mind that I am very fond of them, but they will never be particularly natural in the same way that a good tweeter at 2 or 3 meters sounds. Just my experience--I'm not a huge headphone fan.

 

[Sancus]

You might want to look into having your own personal HRTF measured, there was an old thread on that here. Not sure where it ended up.

Genelec also has a new Aural ID thing where they use 360 degree video of your head to calculate the HRTF. I have no idea how well it works, but I wouldn't expect Genelec to do something if it doesn't at least work somewhat decently.

FWIW, the cymbals aren't harsh for me on my QC35s vs my Vanatoo 0 desktop speakers. They sound about the same. I don't love wearing headphones in general so I haven't done a lot of work to establish my preferences for particular ones.

 

[Pdxwayne]

You might want to look into having your own personal HRTF measured, there was an old thread on that here. Not sure where it ended up.

Genelec also has a new Aural ID thing where they use 360 degree video of your head to calculate the HRTF. I have no idea how well it works, but I wouldn't expect Genelec to do something if it doesn't at least work somewhat decently.

FWIW, the cymbals aren't harsh for me on my QC35s vs my Vanatoo 0 desktop speakers. They sound about the same. I don't love wearing headphones in general so I haven't done a lot of work to establish my preferences for particular ones.

Thanks for checking!

If I get your headphones correct, it is the green line one from measurements in

Bose QC35ii measurements and thoughts

[IMG] From this fall, my (real) life changed slightly, so I needed something that can protect me from noisy labs and library. After several auditions,...

www.superbestaudiofriends.org

It measured differently than the preference curve used by Amir. For example, here is my K371:

I can see why you don't hear sharp cymbals.

Thanks again!

 

[Sancus]

Amir reviewed the QC35 as well: https://www.audiosciencereview.com/...5-ii-review-noise-cancelling-headphone.20584/

I could try to bump up 4khz to see if there's any difference. Though I will say that it's normal to EQ treble and bass to taste anyhow, and the "harman curve" includes a range of different treble and bass preferences. Not just 1 fixed one.

Also, Dr. Olive found that headphone preferences varies from the harman curve(more treble, less bass) in 36% of cases, which is a good amount.

Considering that headphones don't have directivity and so EQ will always perform perfectly on them(unlike with speakers) if you're trying to optimize headphone sound I would consider personal EQ to be a basic requirement.

 

[Pdxwayne]

Amir reviewed the QC35 as well: https://www.audiosciencereview.com/...5-ii-review-noise-cancelling-headphone.20584/

I could try to bump up 4khz to see if there's any difference. Though I will say that it's normal to EQ treble and bass to taste anyhow, and the "harman curve" includes a range of different treble and bass preferences. Not just 1 fixed one.

Also, Dr. Olive found that headphone preferences varies from the harman curve(more treble, less bass) in 36% of cases, which is a good amount.

Considering that headphones don't have directivity and so EQ will always perform perfectly on them(unlike with speakers) if you're trying to optimize headphone sound I would consider personal EQ to be a basic requirement.

Interesting.....

Amir's version: 250hz to ~4kh, it is ~10db increase.

Your chart: same range, minus 2db. Even including the max variation, the up treand is only ~3db up for the same range.

Big difference.....So which version is the correct one for headphones preference?