Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Audible differences between neutral speakers (Read 16558 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Audible differences between neutral speakers

Between different speakers (or headphones) that are considered and tested neutral, you'll find people saying one sounds harsh and one doesn't. I was skeptical about this until I tried a pair of both the Yamaha HS8 and the Neumann KH 120; I found, as do several reviewers, that the HS8 is harsh and the KH 120 isn't. People use all sorts of weird language like "musical", "sweet", "warm", and I wonder how these could be measured. What I experienced with the HS8 is discomfort after extended listening, the same discomfort you'd find when increasing the upper-mids.

What causes this to happen if a speaker's frequency response is actually neutral?

Re: Audible differences between neutral speakers

Reply #1 –
Define 'neutral' in an objective way.

Re: Audible differences between neutral speakers

Reply #2 –
Quote from: krabapple on
Define 'neutral' in an objective way.

A sine sweep from the speaker reads flat on a perfect sound level meter in a perfect anechoic chamber. These are theoretical conditions; in reality measurement flaws exist, but "neutral" seems to mean this when used by people who publish measurements. In this context, I use neutral to refer to a flat signature, as opposed to a V, or something else.

Re: Audible differences between neutral speakers

Reply #3 –
Quote
Between different speakers (or headphones) that are considered and tested neutral, you'll find people saying one sounds harsh and one doesn't....

...A sine sweep from the speaker reads flat on a perfect sound level meter in a perfect anechoic chamber.
Neutral headphones should all sound similar, but perhaps not identical.   Different people have different ear shapes, and to some extent this can affect how headphones sound to different people.   In an anechoic chamber,  I assume neutral speakers also sound similar.    But in real world listening rooms, the off-axis response comes into play and it's unlikely that any two speakers will sound alike.    That is, there is usually enough difference that you can hear the difference in a blind listening test.

I don't see how a speaker or headphone could be both neutral and harsh.   But, different people have different preferences and different opinions, so one person may say neutral and another person may describe the same speaker as harsh...

Quote
People use all sorts of weird language like "musical", "sweet", "warm", and I wonder how these could be measured.
That drives me nuts!    Moulton Labs - Whaddya Mean The Sound Is Fluffy? Those words are mostly meaningless.    There's frequency response (off & on-axis) and there's distortion.    Other than power rating (related to distortion) and possibly angular dispersion, that's about it.

Re: Audible differences between neutral speakers

Reply #4 –
Quote from: DVDdoug on
I don't see how a speaker or headphone could be both neutral and harsh.
It's measured neutral, but perceived harsh. Going back and forth between the HS8 and the KH120, the former is noticeably harsher, but they're both measured to be neutral. At first I thought people were full of it when they said the HS8 was harsh, but I can personally verify it. If you're like me, you'd have to hear it and see the charts to be equally confused.

I guess those wishy-washy terms could be used to describe distortion. Maybe certain frequencies impact harder than others because of distortion.

Re: Audible differences between neutral speakers

Reply #5 –
Quote from: Lumify on
A sine sweep from the speaker reads flat on a perfect sound level meter in a perfect anechoic chamber.

That is a very very incomplete measurement and thus not a basis for determining that one or more speakers are 'neutral'.

Do yourself a favour and read this

http://www.amazon.co.uk/Sound-Reproduction-Psychoacoustics-Loudspeakers-Engineering/dp/0240520092/ref=sr_1_1?ie=UTF8&qid=1455863010&sr=8-1&keywords=floyd+toole

Re: Audible differences between neutral speakers

Reply #6 –
Quote from: WernerO on
That is a very very incomplete measurement and thus not a basis for determining that one or more speakers are 'neutral'.
That's what I thought. The frequency response is usually the only statistical data one can find besides "it sounds good and musical". In my search for good speakers, I found a bunch of "precise and clinical", so it's hard to find what's actually accurate.

Re: Audible differences between neutral speakers

Reply #7 –
Frequency response is only half of the equation. It is possible to create a neutral frequency response while ignoring the phase response.
http://www.bksv.com/doc/17-198.pdf

Re: Audible differences between neutral speakers

Reply #8 –
Quote from: Lumify on
Between different speakers (or headphones) that are considered and tested neutral
Only audiophiles and studiophiles use those type terms, since they are equally clueless and delusional.

Quote from: Lumify on
People use all sorts of weird language like "musical", "sweet", "warm", and I wonder how these could be measured.
Possibly with fMRI, but I'm no psychologist or psychiatrist. Are these people using their ears, or a whole lot else?

Quote from: Lumify on
What I experienced with the HS8 is discomfort after extended listening
What is your definition of "listening"?

Quote from: Lumify on
What causes this to happen if a speaker's frequency response is actually neutral?
Speakers radiate 3 dimensionally. Millions of frequency responses. Which particular "frequency response" are you referring to, where and what is "neutral"?

cheers,

AJ
Loudspeaker manufacturer

Re: Audible differences between neutral speakers

Reply #9 –
Quote from: Lumify on
That's what I thought. The frequency response is usually the only statistical data one can find besides "it sounds good and musical". In my search for good speakers, I found a bunch of "precise and clinical", so it's hard to find what's actually accurate.
You mentioned the KH120. If you look on their web page at the measurements section on the RHS you will see a a set of measurements to help judge the sound of speakers. It is not complete but it is a reasonable set to get started. Now you need the same for the HS8.

Re: Audible differences between neutral speakers

Reply #10 –
STEP #1: Take a large, thick bed pillow and directly cover/smother the drivers, tweeter and woofer, of a conventional, small bookshelf speaker and what do you hear from your living room chair position? Answer: a profound reduction in sound, especially in the mids and highs. What does this prove? That the pillow successfully blocks a good portion of the sound.

STEP #2: Sit back down and now hold the same pillow up, at arms length, such that the direct path from the speaker to your ears is blocked by the pillow. What do you hear? Only a comparatively mild change, but the mids are still there at almost full force and high frequency content is still readily apparent too, only slightly subdued compared to step #1. What does this prove? Answer: the direct, on axis output of the speaker is actually less important than the sound coming out at all the other millions of angles, which bounce around various room surfaces and then are summed together when they finally hit your ear/brain combo.

The on-axis sound becomes more  important to the overall sound the closer we are to a speaker (think desk top monitor) and less important for farther away use, say in a typical living room/home theater use application. The reason the direct on axis frequency response measurements we see published for speakers don't give us a good picture for the overall sound is because it may not be representative of the total sound output in all directions which when summed together is called the acoustical power response.

The typical magazine/manufacturer published, on-axis, 1-meter response curves are only a tiny slice of what's important and don't represent what we actually hear in non-anechoic chamber use.

Re: Audible differences between neutral speakers

Reply #11 –
Quote from: Lumify on
What causes this to happen if a speaker's frequency response is actually neutral?

Hi,
I can't believe that a speaker may have a neutral frequency response. I've looked at the measurements published by Neumann for the KH120, and I am very surprised.
When the same speaker is measured by different people, the differences between the curves can be up to 5 dB. See for example these curves for the Kef R300 speaker (purple line in the first graph; go to next page to see a third measurement) : http://www.homecinema-fr.com/forum/post177852450.html#p177852450

Moreover, it is impossible to reliably measure the frequency reponse below 100 Hz, even in an anechoic chamber. The curves published are only an rough estimation. To get a true measurement, the speaker and microphone would have to be hung in the air at least 10 meters above the ground with a crane to avoid any early reflection.

My guess is that the curve published by Neumann has been artificially neutralized using various consideration such as "removing artifacts caused by the microphone placement, by the room, or by the measurement process", assuming that it is the speaker that should be correct, and not the room / setup / process.

I would be happy to be proven wrong, though. Have these speakers been measured by independent reviewers ?

Re: Audible differences between neutral speakers

Reply #12 –
I've found an independent measurement of the Neumann KH 120 : http://kenrockwell.com/audio/neumann/kh-120-a.htm#meas

It's much more realistic. The frequency response measured here is excellent, but not neutral to the point that it should sound identical compared to another excellent monitor.
The measured deviations are enough to explain the difference in sound.

Moreover, as already stated above, the frequency response out of axis is very important too.

Re: Audible differences between neutral speakers

Reply #13 –
Quote from: Pio2001 on
I can't believe that a speaker may have a neutral frequency response.
As mentioned above, a speaker has many frequency responses and not all of them are supposed to be flat.

Quote from: Pio2001 on
I've looked at the measurements published by Neumann for the KH120, and I am very surprised.
You do not say what you find surprising?

Quote from: Pio2001 on
When the same speaker is measured by different people, the differences between the curves can be up to 5 dB.
Not if it is a competent speaker measured by competent people. There are speaker to speaker differences and there are measurement errors but these do not add up to anything like 5 dB. I would suggest ten times smaller would be a more realistic figure with factors both increasing or decreasing it from there.

Quote from: Pio2001 on
See for example these curves for the Kef R300 speaker (purple line in the first graph; go to next page to see a third measurement) : http://www.homecinema-fr.com/forum/post177852450.html#p177852450
Chortle. How accurate do you think measurements made by hi-fi enthusiasts presented on forums are likely to be? Some are likely to be reasonable but many are going to be far from it. Neumann on the other hand sell their speakers to professional institutions and will back their specification/measurement to a fair extent although a +/- dB is not part of the Tender Description just below the measurements.

Quote from: Pio2001 on
Moreover, it is impossible to reliably measure the frequency reponse below 100 Hz, even in an anechoic chamber. The curves published are only an rough estimation. To get a true measurement, the speaker and microphone would have to be hung in the air at least 10 meters above the ground with a crane to avoid any early reflection.
Measuring at low frequency is not difficult because the wavelength is long. In a room you would position the microphone close to the driver so that the direct sound is a lot louder than the reflected sound and so the latter only changes the measurement a small amount. Avoid severe room resonances. This is a near field measurement which needs modifying to be representative of a far field measurement in a full, half or whatever space but is going to be fairly accurate. Given a reasonable cabinet there will be negligible short circuiting of sound in the near field and so the sound from the monopole source will end up in the far field.

If you want to measure low frequency outside then simply lie the speaker on its back in an open space or better flush in a hole if you want to measure to higher frequencies.

Re: Audible differences between neutral speakers

Reply #14 –
Here is a set of thrid party measurements for the KH120 I stumbled across looking for something else. I would suggest they represent about the level of difference to expect between different (competent) measurements of different examples of (competent) modern speakers. The Neumann KH120 is of course an example of a well designed and manufactured speaker. Typical home audio speakers are not likely to be either as consistent from unit to unit or to have such a well sorted frequency response as can be gleaned from the measurement sections of Stereophile and the like.
 

Re: Audible differences between neutral speakers

Reply #15 –
As far as I'm aware, the measurements from the Neumann website are done by Anselm Goertz. This seems to be the semi-anechoic chamber where he(and his team) usually measures equipment. He also does the measurements for Sound&Recording magazine.

To get back to the difference between HS8 and KH120 and the "discomfort you'd find when increasing the upper-mids". If you look at the horizontal isobars of the HS8, you can see a widening in the horizontal directivity between ~2.5KHz - 9KHz. If you look at the vertical isobars, you can see something similar. The KH120 on the other hand stays quiet constant. Compared to the KH120 the HS8 is not as neutral on-axis and off-axis.

You can download a (less compressed) german version of the KH120 review from S&R for free at Thomann. It contains isobar graphs which are comparable to the HS8 graphs.


Re: Audible differences between neutral speakers

Reply #16 –
Quote from: ErnestS on
As far as I'm aware, the measurements from the Neumann website are done by Anselm Goertz. This seems to be the semi-anechoic chamber where he(and his team) usually measures equipment. He also does the measurements for Sound&Recording magazine.
Interesting. Thanks. Weakens my point a bit though.

Re: Audible differences between neutral speakers

Reply #17 –
Quote from: ErnestS on
If you look at the horizontal isobars of the HS8, you can see a widening in the horizontal directivity between ~2.5KHz - 9KHz.
That could account for Lumifys "fatigue", as excess power in that region tends to emphasize sibilant sounds in some recordings, but of course would also factor the room/position etc.
Then again so could that white cone. ;)

cheers,

AJ
Loudspeaker manufacturer

Re: Audible differences between neutral speakers

Reply #18 –
Quote from: honestguv on
Quote from: Pio2001 on
Moreover, it is impossible to reliably measure the frequency reponse below 100 Hz, even in an anechoic chamber. The curves published are only an rough estimation. To get a true measurement, the speaker and microphone would have to be hung in the air at least 10 meters above the ground with a crane to avoid any early reflection.
Measuring at low frequency is not difficult because the wavelength is long. In a room you would position the microphone close to the driver so that the direct sound is a lot louder than the reflected sound and so the latter only changes the measurement a small amount.

I don't think that solves the problem.  Yes, putting it close means that you measure a larger signal leaving the speaker, but the amount of energy coupled from the speaker into the room depends on the boundary conditions imposed on the room.  If a given room couples less energy, than you will measure a lower amplitude no matter how close you get to the driver. 

The reason you use an anechoic chamber is that it better approximates the free space response of the device.

Re: Audible differences between neutral speakers

Reply #19 –
Quote from: saratoga on
The reason you use an anechoic chamber is that it better approximates the free space response of the device.
An anechoic chamber is no different to any other room at low frequencies. Even if Prof. Goertz walked a few hundred metres to use the large "proper" ITA anechoic chamber it would only get him down to 63 Hz. Since the implosion of ITA I presume this would be an option if he thought it worthwhile rather than using a near field measurement at low frequencies.

Quote from: saratoga on
I don't think that solves the problem.  Yes, putting it close means that you measure a larger signal leaving the speaker, but the amount of energy coupled from the speaker into the room depends on the boundary conditions imposed on the room.  If a given room couples less energy, than you will measure a lower amplitude no matter how close you get to the driver. 
What would be required to directly change the measurement significantly is to change the impedance of the air. That is the force experienced by the cone pushing on the air. What is going on at the room boundary is an indirect influence. In order to do this the indirect sound waves will have to have a significant effect either by being large or by being a sequence of small ones which always nudge the cone in the same direction and add up to something significant.

Now the first acoustic resonance of a speaker enclosure will certainly put a blip in the response of the speaker if driven without stuffing in the enclosure but a very large room (e.g. outside) will not. So where is the line?

Re: Audible differences between neutral speakers

Reply #20 –
Quote from: saratoga on
Quote from: honestguv on
Quote from: Pio2001 on
Moreover, it is impossible to reliably measure the frequency reponse below 100 Hz, even in an anechoic chamber. The curves published are only an rough estimation. To get a true measurement, the speaker and microphone would have to be hung in the air at least 10 meters above the ground with a crane to avoid any early reflection.
Measuring at low frequency is not difficult because the wavelength is long. In a room you would position the microphone close to the driver so that the direct sound is a lot louder than the reflected sound and so the latter only changes the measurement a small amount.

I don't think that solves the problem.  Yes, putting it close means that you measure a larger signal leaving the speaker, but the amount of energy coupled from the speaker into the room depends on the boundary conditions imposed on the room.  If a given room couples less energy, than you will measure a lower amplitude no matter how close you get to the driver. 

The reason you use an anechoic chamber is that it better approximates the free space response of the device.

If you're interested in the onset response at LF, as Andy indicated, a very NF measurement is suffice. Of course it gets a bit more complicated if you have a reflex system.
And now back to Lumify....

cheers,

AJ
Loudspeaker manufacturer

Re: Audible differences between neutral speakers

Reply #21 –
Quote from: honestguv on
Quote from: saratoga on
I don't think that solves the problem.  Yes, putting it close means that you measure a larger signal leaving the speaker, but the amount of energy coupled from the speaker into the room depends on the boundary conditions imposed on the room.  If a given room couples less energy, than you will measure a lower amplitude no matter how close you get to the driver. 
What would be required to directly change the measurement significantly is to change the impedance of the air. That is the force experienced by the cone pushing on the air.

In freespace, yes. But in a room there is also the coherent sum of all other acoustic energy at that point, a sum which can be zero or even negative.  Its entirely possible for instance to have no energy whatsoever leave a speaker, and then to have a lot of energy leave a speaker without changing the air or the electrical drive waveform at all. 

Re: Audible differences between neutral speakers

Reply #22 –
Quote from: saratoga on
Quote from: honestguv on
What would be required to directly change the measurement significantly is to change the impedance of the air. That is the force experienced by the cone pushing on the air.
In freespace, yes. But in a room there is also the coherent sum of all other acoustic energy at that point, a sum which can be zero or even negative.  Its entirely possible for instance to have no energy whatsoever leave a speaker, and then to have a lot of energy leave a speaker without changing the air or the electrical drive waveform at all. 
The rate of work done by the cone on the air is the product of the cone velocity and the air pressure. This holds at all times and follows from the definition of work and Newton's third law. Impedance is the ratio. Are you claiming this product can be changed from large to zero without changing either the cone velocity or the air pressure? Or something else given electricity does not seem to have a role?

Re: Audible differences between neutral speakers

Reply #23 –
Take a woofer and operate it without any baffle. Take the same woofer and mount it in a wall...
"I hear it when I see it."

Re: Audible differences between neutral speakers

Reply #24 –
Quote from: honestguv on
Quote from: saratoga on
Quote from: honestguv on
What would be required to directly change the measurement significantly is to change the impedance of the air. That is the force experienced by the cone pushing on the air.
In freespace, yes. But in a room there is also the coherent sum of all other acoustic energy at that point, a sum which can be zero or even negative.  Its entirely possible for instance to have no energy whatsoever leave a speaker, and then to have a lot of energy leave a speaker without changing the air or the electrical drive waveform at all. 
The rate of work done by the cone on the air is the product of the cone velocity and the air pressure.

This is only true if the air is completely stationary, which is equivalent to assuming you are in freespace.  If you are not in freespace, it is not true because the air can be driven by more than just the cone. 

Consider an example:  the cone is hooked up to a resistor, and then placed near a second speaker producing an output sound.  The cone and air will move in response to the first speaker.  Do you really think work being extracted from the resistor?  No, of course not.  As I said, you must consider the coherent sum of all the acoustic waves at a point unless you really are in freespace.