Does Harman Target Curve for Headphones Respect Music Producer's Authorial intent?

[Eary]

This is the confusion I have after learning about Harman curve etc. I will write what I understand and where this question pops out from.

From my understanding, Harman curve came from the combination of two curves: Raw Harman Reference Room Measurement Curve (apparently in latest measurement with B&K5128 this curve is very similar to diffuse field target) + Largely Agreed Preference Filters.

So here is the first question: Do Harman reference room have the average room response of average professional music studios?

I tried to figure out the answer by searching for how the pro-studios are treating and calibrating their studio room responses, so I Googled that. The result I am seeing is, a lot of the related videos seems to target a flat room response, first by using acoustic treatments like panels, bass traps etc., and end the room treatment process using automatic EQ software llike Sonarworks SoundID.

I cannot find direct comfirmation on the room response curve of the Harman Reference Room. I read in different discussions that a flat speaker placed in an average room would have the bass boosted and that's likely what the Harman room is like. Well, my search seems to indicate that studios are using SoundID etc. software to calibrate the room to be flat, so the baseline seems to deviate from the author intent.

Or maybe some professional music producer can show up and suggest that most music studios are not calibrating their rooms to be flat, but rather, a similar state to the Harman room. This I can accept, but we need to explain the existance of software like SoundID that are used by studios to make it flat.

Ok move on, the preference filters. This is where stuff gets even more complicated.
Let's think of what the filters are depending on the answer earlier.
If the Harman room's room is flat and most professional production studios are also flat, original measurement meets authorial intent, adding the preference filters means we are deviating from the authorial intent.
If the Harman room is not flat, but most pro studios are flat, the baseline is already deviating from authorial intent, adding the filters just makes it go further away (natural room response bass boost + filter bass boost).
If both the Harman room and most studios are not flat with natural room response bass boost, then original measurement meets authorial intent, adding the preference filters means we are deviating from the authorial intent.
If the Harman room is flat and most pro studios are not flat and has natural room response's bass boost, the post-filter curve might be closer to the author intent.

If it's the last case, why don't we just measure different pro studios around the world and get an average studio room response first, and tune the Harman reference room to that? In that case we can just use the raw measurement data.

Also why the more accurate measurement by 5128 will turn out similar to diffuse field? Diffuse field is a extremey reflective room with ultimately flat response, is this an evidance that Harman room have flat response?

Basically, if I were to be a professional producer choosing a target for monitoring my mix, should I trust Harman? Or should I trust the raw measurement curve of the Harman reference room? Or some other target curve?
I care about this because I am a songwriting hobbyist. I write my own songs and arrange the songs with DAWs and VSTs, and I am learning to mix my stem tracks. I don't even know whether I can trust my headphone, if not, which target curve I should trust. I would like my songs to sound about the same when played in a professional music studio.

 

[Eary]

Found partial answer for the first part of the question

The in-room measurements of the 19 loudspeakers is close to the DF target because the large number of sources coming from all directions approximate a DF and the speakers are equalized from anechoic flat to in-room flat meaning the speakers will also sound thin and bright.

"speakers are equalized from anechoic flat to in-room flat", this comfirms the Harman reference room was indeed tuned flat during the measurement.

The rest of the question is, what are the professional studios out there doing with there room response? I am aware there are tilted room reference curve out there, but the wide use of software like Sonarworks SoundID seems to suggest studios are just tuning their room to a flat response.

What exactly are the music producion studios doing in general?

 

[Eary]

Not sure why nobody replies. I guess either people don't know the answer just like me, or people just don't care.

Anyhow, I am coming to my conclusion that Harman Target Curve is not meant to respect authorial intent, but instead, the different tuning preference clusters of different populations, including how the cluster of younger males generally prefer more bass, the cluster of elderly people generally prefer less bass, etc.

So I am going with this:

For reference of how the piece will sound like in a pro-studio:
1. Sean's B&K5128-measured flat-room 19-channel-average curve (simulate the FR of a tuned-to-flat studio)
2. B&K5128 DF with -1db/oct tilt (simulate the FR of a tuned-to-tilted-room-target studio)

For reference of how the piece will sound like to the general public:
Harman B&K 5128 translated Harman 2019 IE curve.

I guess I will have to keep multiple target curves' EQ profiles around and actively switch bewteen them during the mixing session.

 

[Bugal1998]

Does Harman Target Curve for Headphones Respect Music Producer's Authorial intent?​

No it does not. In the aggregate, no curve does. Read up on the circle of confusion.

To resolve this issue, I always assume the authorial intent is maximum enjoyment, and therefore, I use the response that gives me the most enjoyment; intent achieved.

Oh, and if their intent wasn't maximum enjoyment, well, then I don't care about their intent.

 

[Eary]

Oh, and if their intent wasn't maximum enjoyment, well, then I don't care about their intent.

The thing is, as I said, I am a songwriting hobbyist, if I use a headphone that doesn't match what the actual studios sounds at all, the piece sounds good to me but it sounds like garbage outside my own headphone.

My take has been like this, when we go out for dinner, we don't like how the food taste, we can put more sauce into it (EQ settings in the music player APP), but saying stuff like "this spoon gives it extra salty taste" (the headphone has extra bass boost) just doesn't make sense to me.

 

[Bugal1998]

You aren't getting better answers because there is no such thing as what "the actual studios sound" like. Even if you average the responses of a variety of studios the result may not be representative of any actual studio that exists in reality. That's why Harman--I believe--starts with anechoicly flat speakers, which in a typical room results in a low frequency gain and a modest mid and high frequency rolloff (i.e. neutral).

Among the variety of studios I've been in (TV production/broadcast, radio production/broadcast, and music production) there have been significantly different frequency responses.

Dolby provides frequency response guidelines for production, but that in no way assures a truly consistent experience.

I'm not a mixer, but I have mixed, and in my experience even supposedly neutral headphones give different results compared to supposedly neutral speakers, all within the Harman family of products. YMMV.

So I'll again reiterate that understanding the circle of confusion may help. Floyde Toole's book Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms provides measurements of numerous different facilities, which can be informative.

 

[Eary]

That's why Harman--I believe--starts with anechoicly flat speakers, which in a typical room results in a low frequency gain and a modest mid and high frequency rolloff (i.e. neutral).

That was one of my confusions but I think I found the answer from Sean that though the speakers themselves were anechoicly flat at first, they were then EQed to in-room-flat before doing the measurement. Harman starting point is an in-room flat measurement. That was one of the stuff that was not cleared when I read from various sources. So no, Harman starting point doesn't have low frequency gain, and that's why the 5128 19 channel average measurement looks similar to the 5128 diffuse field curve.

The definition of "neutral" in my perspective is "faithful to what the audio file encodes", not "room low frequency boost", again, it doesn't make sense to use an extra salty spoon and call it "neutral".

I guess I need to look into some of those PHRTF binaural monitoring tools, I heard that Embody Immerse Virtual Studio is quite good, I haven't prepared the money to purchase it though.

 

[Bugal1998]

That was one of my confusions but I think I found the answer from Sean that though the speakers themselves were anechoicly flat at first, they were then EQed to in-room-flat before doing the measurement. Harman starting point is an in-room flat measurement. That was one of the stuff that was not cleared when I read from various sources. So no, Harman starting point doesn't have low frequency gain, and that's why the 5128 19 channel average measurement looks similar to the 5128 diffuse field curve.

The definition of "neutral" in my perspective is "faithful to what the audio file encodes", not "room low frequency boost", again, it doesn't make sense to use an extra salty spoon and call it "neutral".

Yes, the 19-channel and the DF response are nearly identical; but where are you getting the idea that studios or their headphones are tuned to a flat in-room response?!

Some mixers may use DF head phones, others use Harman curve with low end boost, others are using who knows what... I have all of the above, plus DF headphones EQ'd to the Harman over-ear headphone response.

A neutral speaker has a flat on-axis anechoic response and a smooth off-axis response that yields a neutral in-room response with a low-end rise and a high frequency rolloff. I believe this resulting in-room response was the basis for Harman's older over-ear headphone target.

If the mixing/mastering room had a low frequency boost, then playback also needs the same boost because that's how the content was created. But not every mixing room has the same low end response. So again, I will refer you to the circle of confusion.

 

[Bugal1998]

Thread 'What is the target frequency response/curve of a well treated room?' https://audiosciencereview.com/foru...-response-curve-of-a-well-treated-room.54773/

Thread 'Is there a rough response you should target for music?' https://audiosciencereview.com/foru...h-response-you-should-target-for-music.57329/

Thread 'Please help me understand the basic principles and get me up to a good start in understanding this comment by Floyd Toole.' https://audiosciencereview.com/foru...erstanding-this-comment-by-floyd-toole.54949/

Thread 'Headphones, Sonarworks and Harman Curves' https://audiosciencereview.com/foru...eadphones-sonarworks-and-harman-curves.55820/

 

[Eary]

Thread 'What is the target frequency response/curve of a well treated room?' https://audiosciencereview.com/foru...-response-curve-of-a-well-treated-room.54773/

Thread 'Is there a rough response you should target for music?' https://audiosciencereview.com/foru...h-response-you-should-target-for-music.57329/

Thread 'Please help me understand the basic principles and get me up to a good start in understanding this comment by Floyd Toole.' https://audiosciencereview.com/foru...erstanding-this-comment-by-floyd-toole.54949/

Thread 'Headphones, Sonarworks and Harman Curves' https://audiosciencereview.com/foru...eadphones-sonarworks-and-harman-curves.55820/

I have already read all those posts before. But when it comes down to my headphone, I still need to pick the target curve(s) to use, so simply saying that "no curve would work" is kind of a useless answer.

Again I already came to the conclusion that I will need to keep multiple curves around and keep in mind what their goals are. I might need to look into PHRTF binaural monitoring in the future when I get the money. I heard when people using those binaural tools they thought their speakers were on and they needed to take off their headphone to double check, that just sounds intriguing.

 

[solderdude]

A speaker will sound different in different rooms at different distances.
Different types of room conditioning will have different results.
Different listening SPL will have different results.
Different target curves will have different results.
All headphones sound/measure different.
There is no single 'audible flat' headphone.
Headphones can sound different on different heads (interaction with pinnae and seal issues).
Headphones EQ'ed to a specific target on fixture A will sound different to headphones EQ'ed to the same target using another test fixture.
Then there are different targets, different types of headphones (in/on/over-ear).
Then there is personal taste which simply does not come in one flavor.
Then there are different recording techniques which also highly influence the recorded sound.

No matter how one twists and turns this and believes target A, B or C is the way to go and measurement method D, E, or F is the right way to go there is no single target and there is no single 'best method'.
There may well be a 'best method' (which seem to agree with one's hearing) for different persons and different circumstances but you will still have to deal with:
A: the recording (techniques) used.
B: the choice of speakers/room/headphone/target/SPL of the recording studio including the taste(s) of the people involved.
C: the choice of speakers/'room'/headphone/'target'/SPL in your listening environment.
D: your taste.

Circle of confusion.

I would say... pick a method/target that 'works' for you and accept there is no 'single true reference' or 'best reference' but you will have to pick one if you are looking for consistency.

At ASR Amir chose Harman for headphones on a specific test fixture and Klippel for speakers which are sound choices but what comes out of those tests may not be how the same copy of the tested headphone or those speakers in your room at a certain listening SPL will sound to you (even with EQ).

For recording... try to get it right on your monitoring... use known good monitoring gear/techniques. Pick your target.
For reproduction... go for maximum enjoyment which may or may not be 'perfectly flat and lowest possible distortion' but could well be what is needed and pick your target (if aiming for that).

 

[goat76]

The thing is, as I said, I am a songwriting hobbyist, if I use a headphone that doesn't match what the actual studios sounds at all, the piece sounds good to me but it sounds like garbage outside my own headphone.

Don't mess with different target curves, just make sure that well-balanced and well-known reference tracks also sound well-balanced to you in your system, and you are set to go. That will break the "circle of confusion" as the well-balanced reference tracks will have a known reference point.

As you have probably figured out by now, there is no single "target curve" that suits everyone's taste so you need to find what sounds well-balanced to you personally. To do this you should start by putting together a collection of reference tracks in the same style as the music you make, tracks that you over the years have found sounding great to you (and probably most others too) on most stereo systems. Everyone has a collection of tracks they find exceptionally great-sounding and I'm sure you already know what those reference tracks are to you.



My suggestion is that you create a new project in your DAW where you have all your reference tracks collected, and make sure they are all level-matched to the same LUFS. I have my reference tracks set to -18 LUFS as that level fairly well matches the "raw" mixing level of the type of music I mix.
Now listen to these tracks for hours and days and find out what sounds most correct to you tonality-wise. It may be a specific target curve that comes closest which you can at least use as a starting point, and then adjust the overall curve further until those reference tracks sound correct to you.

My next suggestion is that you always have this project open as a separate tab in your DAW for fast reality checks, constant comparisons to your mixes, and for re-calibrating your hearing to the correct-sounding tonality (that you know those well-known reference tracks have), as there is a big risk that your hearing has adjusted itself to a sometimes wacky-sounding tonality and "got lost" by focusing on details in the mix for a long time.

Whenever you reach the point where your mixes sounds similar tonality-wise to the reference tracks, you know you are on the right path.



As already said, don't mess with different target curves or try to find out what the average Joe finds most well-balanced sounding, because if that target curve doesn't sound well-balanced to you personally, you will still end up over-compensating for things in your mixes until it sounds well-balanced to you, but that will only lead to the result that the mix will no longer sound well-balanced when you finally hear it in a sound system that sounds correctly-balanced to you.

 

[Bugal1998]

I have already read all those posts before. But when it comes down to my headphone, I still need to pick the target curve(s) to use, so simply saying that "no curve would work" is kind of a useless answer.

A neutral response may be your best bet. Many listeners will have deviations from neutral in their own playback systems, but overall their deviations will cancel out in terms of any changes you should make to the frequency response of your mixing gear.

We have offered a definition of a neutral response above, (and you have created your own definition) as well as noted which headphone curves attemp to reproduce it (hint: it's not DF or the 19 speakers set to in-room flat... as Sean said, they don't sound pleasing).

As others have said, no matter what response you use, frequently 'calibrate you ears' with good recordings while you work.

Helpful or not, the simple fact remains that no curve satisfies the answer to the question you posed in this thread. However, a curve that includes the typical in-room boost of low frequencies and a gentle rolloff of higher frequencies will get you in the right ballpark. The rest is up to the listener to also have a neutral playback system, and thats out of your control.