The Etymotic Target (R.I.P. Harman)

[Newman]

Trying to untangle the OP's pretzel here.

The "dashed orange line" in post #28 is from an old (maybe 15+ years) bit of research where Harman were comparing several early auto-EQ systems to one another and to no-EQ. These auto-EQ systems all measured and equalized the ungated in-room frequency responses. Some of them actually made things subjectively worse than no EQ at all. The best of the bad bunch was the one that looked like the dashed orange line. That's why it is called "the preferred target" -- preferred to whatever target the others were working to, out of that bunch of products.

So Harman were saying, if you were going to equalize loudspeakers based only on in-room, at-the-listening-seat measurements that are ungated, then that dashed orange line is your best-bet target. Best of the set of targets they compared.

Then Olive makes the observation that, if you take a flat-anechoic-response speaker like the Revel F208, and measure it ungated at the listening seat, it already matches the orange dashed line, with no EQ needed, except in the bass region. Well, near enough to make the following point, namely, that except in the bass, using the orange dashed line as an averaged-in-room ungated EQ target is really equivalent to adjusting the speaker's direct-arriving sound to a more-or-less flat frequency response.

So, nowadays, with better in-home analytic software that can take gated measurements outside of the bass frequencies, we can be more direct. Instead of using the dashed orange line as a target as a proxy for roughly getting the gated (direct-arriving sound) response flat, we can more precisely get the direct-arriving response truly flat by measuring it directly.

As for the green curve, please realize that it is being measured inside the ear. Olive is trying to explain that when you take a speaker like the F208 with its flat anechoic (direct-sound) frequency response, and measure the response inside the ear, you will get that green line with the big bump in the HF. So, he says, working backwards, if you use that green line as a target response inside the ear when measuring headphones, you will be adjusting the headphones so they sound like flat-anechoic-response speakers.

 

[richard12511]

Preferred is not neutral. Why everyone gets it so wrong??

That's exactly what I'm saying, or are you agreeing with me? Unless you're talking about speakers. In the case of speakers, anechoic neutral speakers are the most neutral speakers(we "hear through rooms"), and they are also the most preferred. Neutral and preferred are one in the same for loudspeakers, but not for headphones. For headphones, neutral and preferred are different.

The Harman headphone target is a preference curve, not a neutrality curve.

 

[Newman]

This is distinctively different from the Harman target either for speakers or headphones.

No it's actually very similar, as I tried to untangle above.

Speaker with perfect flat anechoic 20-20kHz response <=> dashed orange line measured in-room ungated at the listener seat <=> headphone curve measured inside the ear. Roughly. More or less.

 

[JohnYang1997]

So Harman were saying, if you were going to equalize loudspeakers based only on in-room, at-the-listening-seat measurements that are ungated, then that dashed orange line is your best-bet target. Best of the set of targets they compared.

Nothing about neutral sound reproduction. Only preference from "people".

except in the bass region

The biggest problem IS the bass region. It's the core of the discussion.

The main difference between etymotic and harman target(OE) is in the bass region.

 

[richard12511]

Neutral is flat speakers in well treated room. Also neutral is basically what you eq to in speaker reviews.
This is distinctively different from the Harman target either for speakers or headphones.

Your first two sentences are correct. Your third sentence is incorrect, and shows a lack of understanding of the loudspeaker science of Toole/Olive, at least for loudspeakers(you are correct for headphones).

 

[JohnYang1997]

That's exactly what I'm saying, or are you agreeing with me? Unless you're talking about speakers. In the case of speakers, anechoic neutral speakers are the most neutral speakers(we "hear through rooms"), and they are also the most preferred. Neutral and preferred are one in the same for loudspeakers, but not for headphones. For headphones, neutral and preferred are different.

The Harman headphone target is a preference curve, not a neutrality curve.

No. You are contradicting yourself the whole thread. He pointed out the Harman targets are preferred targets. And the gree line is speakers eqed to flat in the room not room response with speakers eqed to flat anechoic response. He gets the whole thing and you didn't understand the point he's talking about.

 

[Newman]

Nothing about neutral sound reproduction. Only preference from "people".

It has been DBT established that it's the same thing. People prefer the direct-arrival sound to have the same FR as the original live instruments and voices. And it makes logical sense that they do so.

All these proxy curves are just trying to equate ungated responses to flat anechoic responses.

 

[richard12511]

For loudspeakers, neutral(anechoic) and preferred are one in the same, which is really a great convenience for us. Perhaps @Floyd Toole can expand, but it's all explained within his book.

Again, there are differences with headphones, and I think the preference science is likely still evolving.

 

[JohnYang1997]

Your first two sentences are correct. Your third sentence is incorrect, and shows a lack of understanding of the loudspeaker science of Toole/Olive, at least for loudspeakers(you are correct for headphones).

What about it? The "preferred" harman in room response shows the very similar relative response as the Harman OE target.

 

[JohnYang1997]

It has been DBT established that it's the same thing. People prefer the direct-arrival sound to have the same FR as the original live instruments and voices. And it makes logical senses that they do so.

All these proxy curves are just trying to equate ungated responses to flat anechoic responses.

People have very varied preference. It is not a matter of what's more sensible to do or what's better to reach for more audience.
The discussion here is what's the objective accurate response. There is one for speakers. So what's the equivalent for headphones.

 

[richard12511]

No. You are contradicting yourself the whole thread. He pointed out the Harman targets are preferred targets. And the gree line is speakers eqed to flat in the room not room response with speakers eqed to flat anechoic response. He gets the whole thing and you didn't understand the point he's talking about.

No, this shows you have a lack of understanding of loudspeaker science. You're a great engineer, but you're incorrect here. Anechoically neutral loudspeakers are the most preferred loudspeakers in double blind listening tests. It's explained very well in Toole's book.

 

[Newman]

The message being, "If all you can do is take in-room measurement without gating, then here, try this target curve and EQ to that, it's at least a guess that you will be somewhere near to having EQ'd the direct-arriving (anechoic) response to flat."

 

[JohnYang1997]

No, this shows you have a lack of understanding of loudspeaker science. You're a great engineer, but you're incorrect here. Anechoically neutral loudspeakers are the most preferred loudspeakers in double blind listening tests. It's explained very well in Toole's book.

It's in the grand scheme that it's the case. And that human can bypass the room hearing the speaker's anechoic response. Yes.
I was not arguing about what needs to be done in speakers. The current way to evaluating speakers in good and correct.
It's the difference here in headphones people are making confusion.

 

[JohnYang1997]

https://www.audiosciencereview.com/...e-why-is-it-represented-as-a-thin-line.19886/
Here's the thread to discuss about in room targets or stuff like that. Not this thread. That is a whole controversy on its own.

 

[Sharur]

@JohnYang1997 I think the split is in the amount of room treatment being used. If the X-Curve results from a well treated room, what exactly is the dashed orange line besides a useless line for poorly treated rooms?

 

[JohnYang1997]

The message being, "If all you can do is take in-room measurement without gating, then here, try this target curve and EQ to that, it's at least a guess that you will be somewhere near to having EQ'd the direct-arriving (anechoic) response to flat."

That's true. But there's definitely a big discrepancy as we all know.
The key is what's the equivalent of the eq to flat in anechoic response for headphones.

 

[JohnYang1997]

@JohnYang1997 I think the split is in the amount of room treatment being used. If the X-Curve results from a well treated room, what exactly is the dashed orange line besides a useless line for poorly treated rooms?

X curve doesn't necessarily represent a well treated room. It just ignores the biggest part of nonideal room aspect which is the bass. The high frequency is obviously changing with the size of the room. So that's not a fixed line.

 

[Sharur]

To be honest, I'm not very informed on loudspeaker science. I just wanted to point out in my video that the OliveToole target for in-room response is not representative of flat speakers in a well treated room.

 

[Newman]

If the X-Curve results from a well treated room, what exactly is the dashed orange line besides a useless line for poorly treated rooms?

The X curve has nothing to do with domestic rooms. It was for large spaces. And a poor proxy for that too.

The dashed orange line was at least for domestic rooms with normal treatment.

 

[Sharur]

Listeners who vote for such, consider it neutral. It is what sounds "right" to them.

Viewers preferring oversaturated screens all of a sudden makes oversaturated screens neutral? Neutral should be something that is measured accurate.