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

[Nouvruaght]

@Sean Olive you said your listeners preferred a neutral speaker target. But according to the Genelec survey study of 164 professional control rooms, the neutral speaker response inside a pro control room on average is actually pretty flat.

You pretty much confirm this flat response with the bold dashed line of this image as well. "Predicted steady-state room curve for highly rated loudspeakers in a typically-reflective room, with shaded area showing possible results in room having different amounts of reflectivity. Flat direct sound.

so why then, did the listeners prefer an obviously not neutral at all speaker target?

 

[pozz]

There are some weird/good characteristics of NS10 that made them that way. It never was claimed to sound like bigger systems. But when one mix on them and of course knowing how they sound one can make reliable decisions to sound good on bigger systems. If it's not the case no one would be using them.

Toole cited this paper in his book: https://www.aes.org/e-lib/browse.cfm?elib=3354 One of the conclusions drawn was that mixing engineers often just invert the frequency response of the monitors they use.

You're making the subject more mysterious than it really is.

The so-called "harman" curve is no different to a 1970s loudness button on a receiver or amplifier. Sounds big, fat and fun at low levels, but absolutely horrid at anything approaching medium levels. There's a reason loudness contours fell out of fashion, and so will this slavish adherence to an arbitrary brand's key selling point of difference

I don't believe you really read or try to understand the research before posting criticism.

Most of what you see in the curve for headphones is gain due to the ear coupler in the measurement system. It is not a characteristic of the driver, other than the bass shelf.

Best demonstration of accuracy is that for headphones that actually sound tubby, like Beats, there is a pronounced midrange dip. The bass does not show an unusual boost.

The curve was/is designed to find out what people "liked" and let's face, what people "like" and what's actually uncoloured accuracy are completely different things.

Not the case for speakers. What people like is accuracy. No reason it should be different for headphones.

 

[Sharur]

Not the case for speakers. What people like is accuracy. No reason it should be different for headphones.

Can you please explain this?

 

[pozz]

View attachment 155051
Can you please explain this?

Do you see any wild deviations that would hint at inaccuracy? Tilts and a bass shelf is all. Plus it's a steady state curve, not a loudspeaker.

 

[pozz]

@Sean Olive you said your listeners preferred a neutral speaker target. But according to the Genelec survey study of 250 professional control rooms, the neutral speaker response inside a pro control room on average is actually pretty flat.
View attachment 155040
You pretty much confirm this flat response with the bold dashed line of this image as well. "Predicted steady-state room curve for highly rated loudspeakers in a typically-reflective room, with shaded area showing possible results in room having different amounts of reflectivity. Flat direct sound.
View attachment 155044so why then, did the listeners prefer an obviously not neutral at all speaker target?

I realize your question is directed at Sean, but you're misreading the first graph. It shows how massive the deviation is inside studios and is unrelated to preference. Genelec use that graph to show the extent of the problem and to justify, for example, their approach, current product line and features, aimed at reducing the variability.

 

[Sharur]

Do you see any wild deviations that would hint at inaccuracy? Tilts and a bass shelf is all.

That is the inaccuracy

Plus it's a steady state curve, not a loudspeaker.

It says "predicted steady-state room curve for highly rated loudspeakers in a typically reflected room" with areas below 2 kHz changing depending on the amount of reflection. Isn't that the same thing?

 

[Thomas_A]

@Sean Olive you said your listeners preferred a neutral speaker target. But according to the Genelec survey study of 250 professional control rooms, the neutral speaker response inside a pro control room on average is actually pretty flat.
View attachment 155040
You pretty much confirm this flat response with the bold dashed line of this image as well. "Predicted steady-state room curve for highly rated loudspeakers in a typically-reflective room, with shaded area showing possible results in room having different amounts of reflectivity. Flat direct sound.
View attachment 155044so why then, did the listeners prefer an obviously not neutral at all speaker target?

What one can conclude is that the median seems rather flat, with the exception of room-related notches e.g. at 100 Hz and some boost below 100 Hz (but a limited response to ≈40 Hz). Due to the large variation above and below the median, I don't think there is any conclusion to be drawn regarding preference.

 

[pozz]

That is the inaccuracy

It says "predicted steady-state room curve for highly rated loudspeakers in a typically reflected room" with areas below 2 kHz changing depending on the amount of reflection. Isn't that the same thing?

No. My point was about accurate loudspeakers, not a room curve.

If your point is that preferences vary, then all that graphs shows is that they don't vary much.

 

[MaxBuck]

I can see how debating preference curves makes sense for a speaker manufacturer trying to sell more product.

I can't for the life of me figure out why anyone else would want to do so. It's not as though it's rational to purchase transducers on the basis of a graph. One likes what one likes, regardless of what an AES PhD says.

 

[Sharur]

It's not as though it's rational to purchase transducers on the basis of a graph. One likes what one likes, regardless of what an AES PhD says.

I will not hesitate to buy a headphone without trying it if it measures well enough

 

[MaxBuck]

I will not hesitate to buy a headphone without trying it if it measures well enough

Fair enough. I'm not a headphone guy.

 

[Nouvruaght]

No. My point was about accurate loudspeakers, not a room curve.

If your point is that preferences vary, then all that graphs shows is that they don't vary much.

I can understand the boost below 100hz by the trained listeners, it's within the "shaded area showing possible results in rooms having different amounts of reflectivity." But that -5db deviation from neutral past 2khz for trained listeners is insane. Not to mention the untrained listeners preference. Massive 10db differences throughout. How is that not a huge difference in preference?

 

[Sharur]

Fair enough. I'm not a headphone guy.

same with speakers

 

[MaxBuck]

same with speakers

Are you saying you'd buy speakers based only on measurements, without first listening to them?

 

[richard12511]

@Sean Olive you said your listeners preferred a neutral speaker target. But according to the Genelec survey study of 164 professional control rooms, the neutral speaker response inside a pro control room on average is actually pretty flat.
View attachment 155040
You pretty much confirm this flat response with the bold dashed line of this image as well. "Predicted steady-state room curve for highly rated loudspeakers in a typically-reflective room, with shaded area showing possible results in room having different amounts of reflectivity. Flat direct sound.
View attachment 155044so why then, did the listeners prefer an obviously not neutral at all speaker target?

You’re getting anechoic and in room mixed up. The Hartman loudspeaker target is flat in an anechoic chamber. The graphs you posted are in room measurements, so therefore there is no conflict. It’s a very common confusion.

 

[richard12511]

View attachment 155051
Can you please explain this?

Yes.

That’s a loudspeaker in steady state room measurement. The Harman loudspeaker target is neutral.

Neutral means flat direct sound in an anechoic chamber, not in room. In room will look different because of the difference in how different frequencies radiate and reflect in room.

 

[pozz]

I can understand the boost below 100hz by the trained listeners, it's within the "shaded area showing possible results in rooms having different amounts of reflectivity." But that -5db deviation from neutral past 2khz for trained listeners is insane. Not to mention the untrained listeners preference. Massive 10db differences throughout. How is that not differences in preference?

It's to do with how loudspeakers radiate sound in rooms, which combines to form the steady state curve. Note that there aren't random deviations. Just a question of tilt and a bass shelf.

Flat isn't neutral, by the way, it's equal magnitude per frequency. If you look up the spectra of natural sounds, they all exhibit tilts. You want your electronics to be flat, since they only transmit a signal, but loudspeakers transform that signal into acoustical waves. Neutrality there is a complex question partially answered by how we hear, like our response to direct sound vs. reflections, and how loudspeakers influence that tumult through their radiation patterns.

This graph has little to do with the point I made originally, which is that accurate sound is by and large the same as good sound. The conditions for accuracy as far as headphones go are pretty similar to speakers: no jaggedness or abrupt discontinuity, a tendency towards more bass, and some attempt to match the response to our hearing, which for headphones means accounting for direct coupling to our ears.

Are you saying you'd buy speakers based only on measurements, without first listening to them?

I would too, if the measurements were comprehensive.

 

[MaxBuck]

I would too, if the measurements were comprehensive.

OK. I don't regard measurements of transducers to be definitive, but each to his or her own.

 

[pozz]

OK. I don't regard measurements of transducers to be definitive, but each to his or her own.

What do you think is missing? Real question.

 

[richard12511]

@amirm, I think it could be very valuable to do a video about the difference between in room measurements and anechoic measurements. I've lost track of how many times I've seen someone post a downward sloping in room figure to make the point that "see, Harman target is not flat". We even have very good engineers that seem to have this same misunderstanding. It's obviously a very common confusion.