Good somewhat neutral mixing headphones with good bass extension for max. 350€?

[FireEmblem]

The sound does come in different for everyone. HRTF is different for everyone. It's due to the shape of your head / body.

However, your brain expects sound that is filtered by your personal HRTF, and effectively removes it from what you hear.

When the sound bypasses your body and thereby your HTRF via headphones, your brain still expects to hear it WITH the HRTF. So the headphone must compensate for that.

Since everyone's HRTF is different, coming up with a target curve that works for everyone is hard. It's doubly hard because measuring headphones accurately compared to what people hear is very hard at high frequencies.

Speakers don't have this problem, which is why a flat speaker is flat for everybody, but a flat headphone isn't flat for everybody.

...

It's sort of like RIAA equalization. The studio adds RIAA EQ to a record. Your phono stage takes it back out again when you play the record. But if you plug a CD player into a phono stage, it won't sound right because the CD doesn't have RIAA EQ. That's sort of like what happens for speakers vs. headphones, except everyone has a different "EQ".

But HRTF is because sound must travel through the ear, and both, overhead headphones and speakers start externally before they reach the ear. That's why I never understood why headphones supposedly bypass the ear. The sound just has to go through all the curves of the earshape still. For IEMs I could understand the argument, but over ear I don't get.

 

[solderdude]

If after all, the only goal is to create a pre-EQ that sounds balanced as default, wouldn't it be possible to just create my own EQ and dial in everything by ear until I have the feeling that everything sounds balanced? In theory wouldn't that lead to a better result than a one-size-fits-almost-all curve, even tweaked?

This is extremely difficult to do and requires a lot of experience.

I thought these things mainly influence spatial information.. this is confusing. So basically headphones just alter a signal differently than speakers in a room? And because rooms can easier be made objectively flat than ears, they came up with the target curve idea, to find a common dominator amongst all ear shapes?

To make a headphone sound similar to a speaker in a room you have to use tuning or EQ.
This is because a speaker that measures flat in anechoic conditions on axis is not flat anymore at the listening position.

2nd reason is sounds coming from the front have a different tonality as sounds (with the same tonal balance) coming from the sides of the head.

This means that to make a headphone sound similar in tonal balance to speakers in a room you have to change what comes out of the headphone driver in that small 'chamber' around your ear.
So 'flat' speaker centimeters away from your ear does NOT sound anything like flat speaker a few meter away from in front of you.

But even from a perfectly flat room the sound has to go through the same ears.. why would it suddenly not matter anymore, if it's such a big deal with headphones?

There is no perfectly flat room.

I just found out about raw measurements. That's exactly what I meant earlier when I asked if they don't measure just the actual headphone's output:


View attachment 498920

This seems not to be compensated for any target curve.

The target + correction curve is the dotted line.
The blue lines is what the microphone in the fixture pics up.
That has been modified by by the fake ear, fake ear canal and microphone compensation.

Do I understand that the grey line in raw measurements is just how much dB a given frequency needs to be perceived as loud as others, e.g. mathematical objectivity, regardless of an individuums specific hearing or preference?

The grey line is an average of the measured response + target.
How that averaged response is obtained is not clear. Could be DF in an anechoic room, could be done using some known reference.

Can I not just take this, EQ my headphone until it's aligned with this 'target' curve to get rid off all colourations that come from my headphones?

Yes, that's what everyone does.

The thing is that grey target of that fixture is NOT the same as the grey target inside your head.
They will certainly differ above a few kHz.
That grey line on the plot is from the measurement fixture + target curve.

 

[solderdude]

But HRTF is because sound must travel through the ear, and both, overhead headphones and speakers start externally before they reach the ear. That's why I never understood why headphones supposedly bypass the ear. The sound just has to go through all the curves of the earshape still. For IEMs I could understand the argument, but over ear I don't get.

Headphones don't bypass the ear.

However, sounds coming from the front and the exact same sounds coming from the sides of the head have a different tonality at the ear drum.
Add to that the changes a room makes and the changes the shape of the head and the fact that we hear the left speaker (differently) with both ears (as well as the right) also makes the sound different.

Sound from the front thus not only interacts with the pinna but also with the head itself.
The pinna reacts differently with sounds from the front as from the side.
Because of this headphones interact different from speakers in the front.
You thus change perception and headphones bypass the head interaction and differ in pinna interaction.

This is not your problem to solve but the headphone manufacturer must solve this (by tuning the headphone response).
The latter is VERY difficult to do 'perfectly'.
The errors that the manufacturer cannot 'correct' is what must be corrected using parametric EQ.

Measurement wise using a fixture this can be achieved with almost perfect results.
BUT... that is ONLY true for that headphone, in that position on that fixture which has a target+correction belonging to that fixture. The target part is 'an average' and human preference based.

Your head is not a test fixture. So a 'perfect response' on a fixture = not perfect response on your head because of physical differences + taste preference.
It will be closer to 'the ideal' than without that EQ so will bring improvement.
You can use that EQ (between 100Hz and 6kHz) and can tweak it from there.
This is NOT easy to do IF the goal is to create a reference but is possible when just tuning to personal preference.

Personal preference thus is NOT the same a reference which is needed for recording.
Personal preference is fine for music enjoyment.
Learning how a headphone deviates from reference speakers in a room makes it possible to use headphones to create a good mix.
Experience AND having a good reference are crucial for recording.
For music enjoyment it is not needed and only enjoyment is important. No one cares if that is 'flat' or not.

 

[FireEmblem]

Thanks for the in depth info that was rellay insightful!

You can use that EQ (between 100Hz and 6kHz) and can tweak it from there.
This is NOT easy to do IF the goal is to create a reference but is possible when just tuning to personal preference.

I think if an EQ is supposed to be "flat" (or closest to flat) but I hear something sticking out in every song, that would be a good indicator that for my ears that particular frequency is louder and dialing it a bit down might make the signal flatter, no? As Kemmler suggested, EQing towards balance, not personal preference what sounds 'good'.

Now for the whole Harman subject..

RTings measurements are not based on the Harman research.

How important is it, that the target curve is Harman? As far as I understood now, people hear different, so you just try different curves and see which of those sound most balanced to you and then just adjust from there on. Is Harman doing anything better then let's say RTINGS?

They write that there's no clear winner:

Seems not to be drastically different at all. They then move on and explain that B&K would be a new standard with modern technology.

That's probably nothing new for you guys here, so can you give me a quick rundown on how well this played out? Is it nonsense and Harman is unbeaten, or is their "new" approach with modern takes on measurements legitimate?

Here's the article: https://www.rtings.com/research/frequency-response-1-8


According to their comparison graph, their new own graph doesn't stray much from Harman:

It just has a bit less low-end and a bit less treble. So shouldn't be terribly off, working with Rtings target, no?

Looks like something super easy to fix. If mixes end up having too much bass and too powerful heights, we'd just lower the EQ slightly at the bottom and top? But maybe they're really ahead of the curve (hehe) and this works even better in terms of translation.. I mean, doesn't sound too far fetched, since most people love a bright highend and a powerful lowend.

 

[kemmler3D]

But HRTF is because sound must travel through the ear, and both, overhead headphones and speakers start externally before they reach the ear. That's why I never understood why headphones supposedly bypass the ear. The sound just has to go through all the curves of the earshape still. For IEMs I could understand the argument, but over ear I don't get.

HRTF stands for "head-related transfer function" and its effect includes that of the upper torso, head, neck and ears of course. Headphone sound, depending on the style of headphone, obviously bypasses the entire head and body, but also at least part of the ear.

Think about it. Your speakers do not have a direct line of sight to your ear canal while you are listening to them... but any style of headphones you use will.

This is also leaving aside crosstalk, which as you know inherently makes things sound different on speakers vs. headphones.

How important is it, that the target curve is Harman?

Harman has done actual science-grade research on headphone preferences and attempted to get the target curve to match what you would hear from a neutral speaker. They're not the only ones that have put work into developing a curve, but their work is considered more robust than most.

As far as I understood now, people hear different, so you just try different curves and see which of those sound most balanced to you and then just adjust from there on.

It's pretty hard to judge by ear to dial in neutral / flat sound with good accuracy. Within +/- 2dB I would believe someone with good listening training could do it. Better than that, I think is hard to do without external references or measurements.

Is Harman doing anything better then let's say RTINGS?

RTINGs article is interesting, and they say they're mostly doing their own curve because Harman's curve isn't valid for the newer, fancier measurement rig they have. And, if you read how they created their target curve, they didn't do anything complicated. AFAIK it's similar to how Harman initially did their curve, but with a new setup.

If you want my opinion, there is nothing wrong with their curve, the interesting thing is really more whether their measurement rig actually correlates better with what people hear.

 

[FireEmblem]

Harman has done actual science-grade research on headphone preferences and attempted to get the target curve to match what you would hear from a neutral speaker. They're not the only ones that have put work into developing a curve, but their work is considered more robust than most.

Sonarworks is probably on pair with them given how popular they are?

It's pretty hard to judge by ear to dial in neutral / flat sound with good accuracy. Within +/- 2dB I would believe someone with good listening training could do it. Better than that, I think is hard to do without external references or measurements.

As far as I understand, the region below 100Hz and above 10kHz is anyways highly dependend on HRTF and should be tuned individually by ear, is that correct? Since Harman and RTINGs seem to mainly deviate there from each other, I think that would further diminish the differences.. furthermore, if many people use RTINGs for reference, and that's probably the case, that will perhaps as well lead to a sound that people out there are largely used to.

If you want my opinion, there is nothing wrong with their curve, the interesting thing is really more whether their measurement rig actually correlates better with what people hear.

Yes I'm curious about this too. But I'd like to compare. Is there any place that measures tons of headphones with harman as a target, so that I can see the measurement and try it with some local EQ plugin? Would like to see how it plays out versus Sonarworks for instance, and RTINGs.

Sonarworks is really good when it comes to precision regarding their target. If I calibrate my Superlux HD681 with it, and then switch to HD650 and calibrate the HD650 too, it will end up with the very same sound (98%, subjectively) that the Superlux has.

What I notice most between both that Superlux seem to be better suited for mixing as it feels more centerd and close, creating the phantom center image. HD 650 has difficulties with that because the soundstage is so wide and deep, that a lot of stuff sounds a bit diffuse and further away (with and without calibration), so it's a bit more difficult to really get the stage into the center of your "brain". The HD650 places the soundstage more around you while HD681 really sends it through you.

 

[kemmler3D]

Sonarworks is probably on pair with them given how popular they are?

I know of Sonarworks but not much about it, I imagine they're not far off, but doesn't it also do some spatial stuff, or am I thinking of something else?

As far as I understand, the region below 100Hz and above 10kHz is anyways highly dependend on HRTF and should be tuned individually by ear, is that correct?

I am not sure about that, I know that high frequencies are increasingly inaccurate, to the point that results from many headphone couplers should just be ignored above 6khz or so. I have not heard that it's due to HRTF variation so much as the simulated ears and mics aren't accurate compared to what a person hears, for various reasons.

Is there any place that measures tons of headphones with harman as a target, so that I can see the measurement and try it with some local EQ plugin? Would like to see how it plays out versus Sonarworks for instance, and RTINGs.

Not sure, I don't read a lot of headphone reviews other than at ASR.

The thing is, if your headphone is measured on any other setup than the B&K Type 5128-B, applying EQ to reach the RTINGs target will probably not sound right.

What I notice most between both that Superlux seem to be better suited for mixing as it feels more centerd and close, creating the phantom center image. HD 650 has difficulties with that because the soundstage is so wide and deep, that a lot of stuff sounds a bit diffuse and further away (with and without calibration), so it's a bit more difficult to really get the stage into the center of your "brain". The HD650 places the soundstage more around you while HD681 really sends it through you.

Kind of funny, as soundstage width like you're describing with the 650 is considered desirable and hard to find in headphones, but I tend to agree that it might be a little tricky when it comes to mixing.

 

[FireEmblem]

I know of Sonarworks but not much about it, I imagine they're not far off, but doesn't it also do some spatial stuff, or am I thinking of something else?

I don't think so, it's simply correcting the frequency response, you can also modify it with the custom tab, match it to Dolby Atmos or do translation checks with various places like cars, laptops, smartphones, studio speakers, TVs and such.

Simulated after-EQ curve:

I must say it sounds super balanced.

I am not sure about that, I know that high frequencies are increasingly inaccurate, to the point that results from many headphone couplers should just be ignored above 6khz or so. I have not heard that it's due to HRTF variation so much as the simulated ears and mics aren't accurate compared to what a person hears, for various reasons.

Yeah, RTINGS used to grey out everything above 9kHz for that reason. But the newer measurements don't have that greyed out area anymore, maybe their new more modern setup makes up a bit for these issues.

The thing is, if your headphone is measured on any other setup than the B&K Type 5128-B, applying EQ to reach the RTINGs target will probably not sound right.

You mean if I took the RTINGs EQ settings but with a measurement not made with B&K by RTINGS?

That's why I was asking for a place that measures headphones with harman as target, so I could compare.. or do I misunderstand you?

Kind of funny, as soundstage width like you're describing with the 650 is considered desirable and hard to find in headphones, but I tend to agree that it might be a little tricky when it comes to mixing.

Yes the soundstage on HD650 is wider, and it's easier to distinguish elements from left to right. A little brass note playing on the right side while a female choir chimes in from the left is much more noticeable on the HD650. But at the cost of that phantom center image and the feeling of being close to the sound. I think when mixing with headphones the most important thing is that it feels >zoomed in<, so that we can really spot tiny details with ease. HD650's soundstage is great for listening, but I think a bit disadvantageous for mixing.

And since spatial decisions should be made on monitors anways, I think it's smarter to look for other traits (such as clarity / audio imaging size) in headphones.

The T7V offer both, they have that very wide L/R separation and depth, but in the same time create a powerful phantom center image inbetween. My plan is to use the monitors for everything spatial (panning, reverb) and as well ensuring that a mix sounds tight in an average untreated room, while I will fall back on headphones for making precise EQing decision when it comes to tonal balance and colouring, so that I don't have to worry about room influences.

 

[FireEmblem]

I tried something now. I'm new to this:

I downloaded APO EQ, then found Superlux HD681 measured to harman 2018 OE linear:

Is this a reliable source?

I then added 10 bands and clicked auto-EQ, exported that one and imported it in APO EQ.

This harman curve sounds different to Sonarworks, it leaves more of the 'shimmering' in the treble and sounds a bit more unbalanced to me. It leaves more of the typical HD681 character with the harsh top that sticks out vs. everything else.

I didn't look into the exact values or tuned anything yet, cuz it's super late and bedtime for me. Just a rough first impression with the lazy auto-EQ button.

 

[kemmler3D]

You mean if I took the RTINGs EQ settings but with a measurement not made with B&K by RTINGS?

That's why I was asking for a place that measures headphones with harman as target, so I could compare.. or do I misunderstand you?

Targets are used for reviews, (Harman, RTINGS, and others) and most of the time they give you the raw response on top of the target.

The targets are also specified as being for a specific test setup/ mic. So basically you just can't mix and match across different test setups.

The RTINGs target is only valid when the headphone has been measured on the B&K. Harman 2018 is only valid on certain rigs. Etc.

The problem is that headphone measurement rigs are known to vary considerably, this is not because the mics are inaccurate but more about the shape of the fake ears and ear canals. The B&K RTINGS is using is newer and potentially more accurate than what most reviews are and have been done with.

So when you EQ to a target you just need to be sure that the measurement and target are both done with the same measurement equipment.

All that said FWIW amir measures against Harman on this site, otherwise not sure.

 

[solderdude]

I think if an EQ is supposed to be "flat" (or closest to flat) but I hear something sticking out in every song, that would be a good indicator that for my ears that particular frequency is louder and dialing it a bit down might make the signal flatter, no? As Kemmler suggested, EQing towards balance, not personal preference what sounds 'good'.

Yes, if something sticks out or you feel is lacking you can/should EQ that.
Here's the thing....
A: you would need to know what frequency is sticking out, how many dB and which bandwidth. Very hard to determine by ear. DON'T use a sweep or test tones for that.
B: you would need to use white or pink noise to do that and know how that noise sounds on a reference. Or use well recorded music (reference recording) which you know how it (should) sound on reference monitors.
C: the listening level should be correct otherwise the tonal balance might be wrong (equal loudness contours)

When that is fixed you will prefer that sound and thus reference should be the same as 'sounds good'.

Below the Superlux HD681 measured without a pinna nor ear canal. Horizontal trace perceived flat nearfield monitor target (mine, is a bit Harman-ish).
This is what the driver actually does at your ERP (where the sound hits your ear).
The linked measurement shows what is measured at the EDRP but ONLY for that test fixture which will differ from your ear.

That's why you need to use narrow-band noise 'tones' or white /pink noise if doing the tuning by ear.

This suggests changing the 5kHz PK filter should be 8kHz and 15kHz should be -5dB instead of -8.2dB.
Your ears may have peaks/dips that differs from your linked plot yet again and those should not be compensated (your brain does that for you).
As can be seen my measurement is closer to that from Sonarworks.
The squiglink EQ will leave 'sharpness' that Sonarworks won't. Different test fixture, different target, different sound.

Below is how my (modified so not EQed) HD681 measures/sounds.

I suggest NOT to fill in the 4kHz 'dip' as that is a 'null' which can't and should not be EQ'ed. Should you want to no more than +5dB.

and how my HD650 measures/sounds:

 

[Tito]

that is interesting, oratory did use 3826 hz +3.2db with q 5.6 on autoeq for peace apo, and filled in as most others around this region.
headphones.com Legacy did not fill in the gap in 4000hz like solderdude.

well, i guess teh best thing with superlux is to use someones preset and modify bass and work around that, and put toilet paper inside ( solderdude hack suggestion) to lower treble.

ony my peace apo setting for hd681 (this i guess is not harman bass, slightly lower i guess, but harman bass is bit too much for me, i like lower 1.5 db in general it appears)
15 hz 3.9db ,Q 0.9
55 ,-0.8DB, q1.7
174. -1.1, q 1.1



and i did bump 4000 hz, but this is before i knew that one should not do that.

i did not use toilet paper-

.they are not here right now, but when i get hd681 back, i will try toilet paper hack without 4k null fill and compare.

 

[FireEmblem]

The targets are also specified as being for a specific test setup/ mic. So basically you just can't mix and match across different test setups.

The RTINGs target is only valid when the headphone has been measured on the B&K. Harman 2018 is only valid on certain rigs. Etc.

So when you EQ to a target you just need to be sure that the measurement and target are both done with the same measurement equipment.

But isn't the measurement always tied to the equipment? Or do you mean that when I grab a harman curve from a website like I did yesterday, that maybe the person measured their Superlux HD681 differently than the harman curve he uses was measured with, which leads to problems?

In that case this would mean that very transparent and professional places like Sonarworks or RTINGS would be better sources than websites like SQUIGLINK?

Yes, if something sticks out or you feel is lacking you can/should EQ that.
Here's the thing....
A: you would need to know what frequency is sticking out, how many dB and which bandwidth. Very hard to determine by ear. DON'T use a sweep or test tones for that.
B: you would need to use white or pink noise to do that and know how that noise sounds on a reference. Or use well recorded music (reference recording) which you know how it (should) sound on reference monitors.
C: the listening level should be correct otherwise the tonal balance might be wrong (equal loudness contours)

When that is fixed you will prefer that sound and thus reference should be the same as 'sounds good'.

Hmm, that sounds like an impossible task. How should one have access to a flat room with flat studio monitors to know what pink or white noise sounds there in order to replicate that at home with headphones.

Why using sweeps not sufficient to do that?

Let's say there's a particular frequency that always sticks out. You could download the 3 different tracks, sweep, and see if it's always the same region that gets highlighted, which is the one you felt is too prominent. That should give enough certainty to know it's the one you want to lower, or not?

The squiglink EQ will leave 'sharpness' that Sonarworks won't. Different test fixture, different target, different sound.

Okay but how do we know which one is closer to "neutral"? If you ask me to judge by ear, I'd say Sonarworks is more close, because it just sounds more balanced without the treble sticking out. This could lead to dark mixes. On the other hand, maybe Harman is better because it would make you add more treble, and maybe that's favoured by more people.

I suggest NOT to fill in the 4kHz 'dip' as that is a 'null' which can't and should not be EQ'ed. Should you want to no more than +5dB.

How do we know that this is a null and not just an intentional dip in the headphone because Superlux thought this sounds good?

 

[solderdude]

But isn't the measurement always tied to the equipment?

The measurement is, the chosen target isn't.
The 'curve' you see is always the correction (for that fixture) + target combined. And that is highly smoothed.

Or do you mean that when I grab a harman curve from a website like I did yesterday, that maybe the person measured their Superlux HD681 differently than the harman curve he uses was measured with, which leads to problems?

In that case this would mean that very transparent and professional places like Sonarworks or RTINGS would be better sources than websites like SQUIGLINK?

That is the responsibility of the website hosting squiglink which has to depend on the info received from the measurebator and have to rely on their info.
Sonarworks and Rtings have different targets.

Hmm, that sounds like an impossible task. How should one have access to a flat room with flat studio monitors to know what pink or white noise sounds there in order to replicate that at home with headphones.

Yes, in order to do that properly you need access to calibrated monitors.
Most 'home studios' should at least have that. Does not need to cost that much when no high SPL is expected/demanded.

Why using sweeps not sufficient to do that?

Because you cannot distinguish between variances caused by the ear and by the headphone and above all your ears aren't analyzers that can gauge dB's nor bandwidth.

Let's say there's a particular frequency that always sticks out. You could download the 3 different tracks, sweep, and see if it's always the same region that gets highlighted, which is the one you felt is too prominent. That should give enough certainty to know it's the one you want to lower, or not?

That will work for you but is not the basis for a reference. For one thing that what 'sticks out' may even be as it should but you could have been used to that part of the frequency range always being subdued.

Okay but how do we know which one is closer to "neutral"? If you ask me to judge by ear, I'd say Sonarworks is more close, because it just sounds more balanced without the treble sticking out. This could lead to dark mixes. On the other hand, maybe Harman is better because it would make you add more treble, and maybe that's favoured by more people.

That's why hearing it compared to a known reference is important in the production side.
Or ... you have to learn how that headphone sounds to calibrated studio monitors or reality (while recording).
This is what's meant with 'translates'.
Only important on the production/recording side.

How do we know that this is a null and not just an intentional dip in the headphone because Superlux thought this sounds good?

No sane manufacturer will create that type of (sharp/narronw) dip on purpose. It is caused by acoustic effects and is the result of tuning side effects or the driver (break-up, or with large drivers could be caused by different ear-membrane differences).
Manufacturers often use a trick to make it seem less problematic by applying 'heavy smoothing'. Just like 'target curves' are always smoothed.

Partly this is defendable because of the way the human hearing works (ERB) and the reason why one should not listen to test tones but narrow band noise instead of tones.
Warble tone (basically afrequency modulated tone) is also an option for measuring but less so for listening.


This all has to do with desiring/needing the headphone to be a reference.
This differs from music enjoyment.

Using Sonarworks, Oratory, Rtings or whatever measurements and basing EQ based on that is 'close enough' but as you already found out that when using headphones that need a lot of correction, especially in the bass and treble the choice of 'trusted measurements/EQ' can lead to substantially different results.
I can't say which one you can trust or which 'target' is the best for you.

 

[FireEmblem]

The measurement is, the chosen target isn't.
The 'curve' you see is always the correction (for that fixture) + target combined. And that is highly smoothed.

Yes I know, but I don't understand what you mean with mixing things. Is how a target curve also dependend on what it was measured with?

Like, does Harman just exist, published by harman, and is always the same, or do people have to measure harman first with their measurement rig and then measure a device against it?

Yes, in order to do that properly you need access to calibrated monitors.
Most 'home studios' should at least have that. Does not need to cost that much when no high SPL is expected/demanded.

I gave up on that. All measurement microphones suggested here are not available on Amazon or Thomann, except for the cheap Behringer one, which is only available in a couple of weeks. But then, it's uncalibrated and you need another professional environment to calibrate it.

The only thing I could do is getting Sonarworks which comes with a measurement mic for monitors too. In that case they offer out of the box to align headphones and your monitors (measured in your room) so that you've got the same sound on both. I'm thinking about doing this.

But I heard that room correction EQ can introduce new problems, and that it only works in the listening spot. If you ever move the table or the speakers or sit differently, it could be messed up already. That doesn't sound so convenient as we aren't static machines doing everything the same way every day..

On the other hand I heard, that Sonarworks wants 29 different measurements or so across the entire room. Maybe they do it a bit differently so that it's not too much limited to how you sit and minor changes?

No sane manufacturer will create that type of (sharp/narronw) dip on purpose.

Is the dip in the HD681 problematic or something that's not much audible? I remember Adam speakers having such a dip too and Adam said it's not audible.

Also, why is it not adviced to boost that region? As far as I understand, with headphones this is different because it's not room cancellation and it's as well measured in every unit, so boosting it shouldn't do harm as it's not the same as fighting a real null that just gets boosted without getting much up.

 

[kemmler3D]

Like, does Harman just exist, published by harman, and is always the same, or do people have to measure harman first with their measurement rig and then measure a device against it?

The harman curve basically says "this is how a good headphone measures, on this specific test fixture". So if someone is not using that same measurement rig, like you say, thy would need to calibrate it somehow.

I heard that room correction EQ can introduce new problems, and that it only works in the listening spot. If you ever move the table or the speakers or sit differently, it could be messed up already. That doesn't sound so convenient as we aren't static machines doing everything the same way every day..

This is true to an extent. Above a certain frequency (around 300hz) EQ only works for an area corresponding to the wavelength you are correcting. 5khz is only about 7cm, so you see why you can't really EQ small frequency ranges higher up. But, wide filters can be OK and correcting low frequencies is OK. This is especially important when it comes to bass and room modes. Lots of reading on that topic to be found on this site.

How are we supposed to work from a 'linear' basis if that one is lying to us and makes us overcompensate for its flatness? I even gave myself time to listen to the track 2-3 times on repeat to align my ears, and yesterday I spent hours listening to music just with the calibration on, but there's no effect taking place that shifts my perception. I can at any time clearly notice that there's this veil over everything caused by the linearity.

You would need to adjust your hearing to the new frequency response / tonality to use it properly for mixing, same as any other speaker or headphone. At some point "the linearity" will sound normal and your old sources will sound wrong. I think it can take 1-3 weeks of regular listening to get used to a new speaker or headphone.

 

[FireEmblem]

The harman curve basically says "this is how a good headphone measures, on this specific test fixture". So if someone is not using that same measurement rig, like you say, thy would need to calibrate it somehow.

So if there is one person with test fixture A and the other with test fixture B, they both need to first evaluate the harman curve on that fixture each, before they measure the headphones against the curve?

Or do they just take the same pre-made harman curve published by harman and simply measure a device against it? That's the point where I got confused because I don't see how one could mix up things if the latter is the case.

This is true to an extent. Above a certain frequency (around 300hz) EQ only works for an area corresponding to the wavelength you are correcting. 5khz is only about 7cm, so you see why you can't really EQ small frequency ranges higher up. But, wide filters can be OK and correcting low frequencies is OK. This is especially important when it comes to bass and room modes. Lots of reading on that topic to be found on this site.

I don't have bass problems and room modes at my listening positions, luckily. So maybe I can get away without room calibration if everything else is anyways difficult to tune..

You would need to adjust your hearing to the new frequency response / tonality to use it properly for mixing, same as any other speaker or headphone. At some point "the linearity" will sound normal and your old sources will sound wrong. I think it can take 1-3 weeks of regular listening to get used to a new speaker or headphone.

Deleted that post so people don't have to read through nonsense. I must correct myself: I think APO EQ, although not being visible in the taskbar, somehow worked in the background. Because after shutting it down specifically, Sonarworks started to sound normal again. Now the HD681, calibrated, also plays back the strings very sharp and present, so that I wouldn't change it.

Only APO EQ with the harman curve creates that underwater-effect that would mislead me to overcompensate for the strings.

I was wondering because I thought, no way Superlux HD 681 sounded like this yesterday with Sonarworks. I don't remember such a strong veil. But when I checked for APO it wasn't visible. Still, the effect was gone after starting and shutting down APO. Weird.. or maybe Sonarworks glitched out.

But you know how it is, if you can't find an objective reason you start to question yourself.. "maybe I didn't spot this yesterday and I adapted over night" or something of that sort. But this felt almost unbelievable to me because no way my ears would deceive me that much.

I mean, the point still stands for the harman curve for HD681 with APO EQ, but who knows if the guy from the website where I got that measurement from didn't mess something up. I don't want to blame the harman curve for it. But Sonarworks works like a charm again. Harman I wouldn't trust as it stands.

 

[kemmler3D]

So if there is one person with test fixture A and the other with test fixture B, they both need to first evaluate the harman curve on that fixture each, before they measure the headphones against the curve? E.g, does harman not equal harman based on the test fixture, or is it always the same harman curve and they only need to test a device against it on their fixture?

Basically, the same headphone will measure differently on different fixtures. The Harman curve says what the measurements ought to be, but they are only 100% valid for one measurement fixture - more or less. So if you are evaluating a headphone for compliance with Harman (or any other target curve) you need to be sure it was measured with the same setup used to develop the target curve.

I don't have bass problems and room modes at my listening positions, luckily. So maybe I can get away without room calibration if everything else is anways difficult to tune..

Hmm, if so, you are very lucky, it would be the first time I've seen someone with a nice flat bass response at their listening position...

I must correct myself: I think APO EQ, although not being visible in the taskbar, somehow worked in the background. Because after shutting it down specifically, Sonarworks started to sound different. Now the HD681, calibrated, also plays the strings very sharp and present so that I wouldn't change it.

Only APO EQ with the harman curve creates that underwater-effect that would mislead me to overcompensate for the strings.

Makes sense, if you apply EQ correction twice (one with EQAPO, once with Sonarworks) it's probably gonna sound bad.

I was wondering because I thought, no way Superlux HD 681 sounded like this yesterday with Sonarworks. I don't remember such a strong veil. But when I checked for APO it wasn't visible. The effect was still gone after starting and shutting it. Weird.. or maybe Sonarworks glitched out. Anyways, now Superlux calibrated sound as they should again, as I remember them from yesterday.

But you know how it is, if you can't find an objective reason you start to question yourself.. "maybe I didn't spot this yesterday and I adapted over night" or something of that sort.

Definitely... you start to doubt your ears...

Once it's set up, EQAPO will always be running in the background whether or not you open the Editor. You basically have to remember if you turned it on or not. Normally that is not a problem because I only have 3 settings:

Headphone EQ
Speaker EQ with bass
Speaker EQ with no bass (for night time)

It's not subtle which setting is on.

If you are switching between different headphone EQs it might be hard to tell immediately which setting is in use, understandable.

 

[FireEmblem]

Hmm, if so, you are very lucky, it would be the first time I've seen someone with a nice flat bass response at their listening position...

I have one or two modes, but they're 3m behind me. Sometimes I can hear a mode where I sit, but only rarely and located more in the background. So I can separate the actual bass from the "mode-echo" that is happening more far away, it feels very disjunkt and is impossible to mix up. Also I don't think I have a null where I sit since the lowend is really thick and nice and nothing changes when I move my sitting position or head.

Don't know how flat it actually is but it's not something that concerns me. I will be doing headphone + spectrum analyzer checks anyways.

Now that's with the T7V by Adam Audio. I also once measured my room, but with my consumer speakers (2.1 with sub that goes down to 50Hz), with an uncalibrated cheap Behringer mic:

I think this doesn't look too bad. This was with the sub under the table. With the T7V as they're on stands and satellites aren't disconnected from the sub, it will probably be a better and more even bass distribution?

Maybe I will order Sonarworks with measurement mic and measure out my room with it.. currently it's at 40% sale at Thomann. It's 299€ at Sonarworks, but only 179€ at Thomann (mic included). The software for headphone EQ alone would cost 99€, so that is really just a small leap for adding speaker/room calibration.

 

[solderdude]

Yes I know, but I don't understand what you mean with mixing things. Is how a target curve also dependend on what it was measured with?

There is a measurement fixture and that has its own correction curves... just not for headphones.

Then there are 'room correction curves' for speakers as 'perfect speaker in a somewhat reverberant listening room' is how one is supposed to listen to music.

With headphones you need to correct that what the fixture alters... AND on top of that you also need to 'simulate' the response speakers would have in a room when the goal is to have headphones sound like speakers in a room.
2 different factors that are combined and form 'the curve you see in raw plots'.

Harman combined these corrections so the Harman curve you see on all those 'raw' plots (which is the signal the microphone picks up but is NOT how you hear it) and made 1 single 'target curve'. Basically that consists of the correction needed for sounds coming from a headphone + the 'speaker in room simulation' with some extra bass to compensate for the lack of tactile feel. That part is based on experiments with people playing with bass controls they could set how they preferred it.

So it consists of measurements + required corrections + preference of the majority of listeners (64%) within a tolerance band.

Like, does Harman just exist, published by harman, and is always the same, or do people have to measure harman first with their measurement rig and then measure a device against it?

Harman target is ONLY valid for measurements taken on GRAS 45CA with the modified pinna from Harman.
It is pretty close to the pinna used by Harman and what Amir uses.
There is no Harman curve for other fixtures BUT as it is known what part is the Harman preference part and the fixture part you can basically take another fixture with its correction (usually Diffuse Field intended for acoustic measurements) and 'add' the Harman preference part to it and then you have a 'Harman' target for that fixture.
BUT ... again a but... it will NOT give the same results because the fixture (+ its correction) differs.

But I heard that room correction EQ can introduce new problems, and that it only works in the listening spot.

Basically yes.. In theory (and usually also in practice) you can get better sound at the listening spot.

If you ever move the table or the speakers or sit differently, it could be messed up already.

Yes, the same with headphones. If you move the positioning or seal you can change tonality.

That doesn't sound so convenient as we aren't static machines doing everything the same way every day..

It works for sound engineering when one sits in one position behind the console.
It also works for audiophiles sitting in their perfectly positioned chair.

On the other hand I heard, that Sonarworks wants 29 different measurements or so across the entire room. Maybe they do it a bit differently so that it's not too much limited to how you sit and minor changes?

Yes, this is done to come to a solution where the 'average' tonal balance in an area is 'decent to good'.

Is the dip in the HD681 problematic or something that's not much audible?

I would say EQ a similar dip into the HD650 and switch between with and without EQ and you will know.
It changes the sound a little.
In general peaks are MUCH more audible (in a negative way) than dips are.
In case of the HD681 the dip is less obvious because there are small peaks on either side which 'mask' it a bit.

I remember Adam speakers having such a dip too and Adam said it's not audible.

That dip usually is positioning dependent and caused by XO phase shift and sound origin point at the CO frequency where both drivers put out a signal but not in phase.
Different cause, different effect.

Also, why is it not adviced to boost that region? As far as I understand, with headphones this is different because it's not room cancellation and it's as well measured in every unit, so boosting it shouldn't do harm as it's not the same as fighting a real null that just gets boosted without getting much up.

A true 'null' has no signal and you can't boost what isn't there. You can boost the frequencies around it that are also affected.

Again... used the HD681 and boost that area 5dB and once do it 12dB and see what sounds correct to you.