The tweeter is the most critical part...

[mhardy6647]

I guess tonality is more important than anything else

What, like I'm gonna argue?

 

[napilopez]

Yes but still
If the highs feel rolled off, the sound is dull and clarity feels less good.
The KH120 may measure perfectly but sounded dark to me, on the other hand the Genelecs sounded more lively and clarity was better to me, I also liked Focal's Berillyum.

So the problem with your argument is that you're talking as if all these speakers measured the same. They don't. The Kh120 measures darker than your typical genelec of focal, so it does not support your argument that it sounds darker because of the tweeter material.

The KH120 is a "neutral-ish" speaker but obviously no speaker is perfectly neutral, especially not in a direct comparison. The KH120 specifically has broad scoop with two prominent dips in the off-axis response from ~1.5 to 7kHz. A region that is exactly where you'd associate 'brightness' and detail within a speaker. Behold, @dominikz measurements of his KH120:

Dip aside, there is also a slight downward trendline to the measurements overall (it's not much, but we can definitely be sensitive to small FR tilts).


Compare that with, say, a Focal Solo6 (current version, my measurements):

Notice the pronounced brightness off-axis in the presence region, and note that we are more sensitive to peaks than we are valleys. Amir also measured an older version of the solo6 which had somewhat different on-axis:

Here are erin's measurements of the Focal Twin6 Be:

Notice similar off-axis peaking in the presence region.

Now let's look at some Genelecs. 8341B:

Pretty flat throughout, although also notice the on-axis seems

8361B:

Or a genelec 8050B:

There is in the 8050B above, but it happens lower in the frequency response, not to much in the detail region.

All of this is to say that before we even consider what possible effect the tweeter material alone could have, the frequency response and directivity already suggest the neumann will be the darkest-sounding of these speakers.

 

[Larry B. Larabee]

With midrange drivers such as this available:https://daytonaudio.com/product/1203/rs100p-8-4-reference-paper-midwoofer-8-ohm down 3db at 10k, 6db at 15k it appears that tweeters are unnecessary or at least overrated in the absolute requirement for wideband reproduction related to your typical human's hi frequency response accuity.
The subjective advantages are debateable, the objective advantages, maybe staggering.

 

[fredstuhl]

We mainly hear direct sound frequency response, directivity and distortion (in this row of decreasing significance), here I had posted a research about the effect of different tweeter materials and types.

The effect of materials employed in drivers re loudspeaker performance

I noticed in one of Amir's speaker reviews a newer member made an allusion to driver materials having some salutary influence on the reproduced sound of a speaker. All things being equal, I wonder how true this assertion really is. I know it's a great marketing point for some of the companies...

www.audiosciencereview.com

Since I‘ve read that thesis and heared how different a speaker sounds just by sloping the fq response by half a db per octave downwards, I don‘t care about tweeter material anymore. Fully agree with that previous statement - it‘s all implementation. Don‘t believe in a specific sound quality/ signature of specific tweeter materials that supposedly can not be reached by other constructions.

 

[valerianf]

To me, if the tweeter is badly behaving, there is no way to have any attention to a speaker.
That said, once the tweeter behave properly , then the mid-range is the key point in my speaker choice.
Once in my life I listened to the Magnat plasma tweeter and it was a revelation.

 

[test1223]

So the problem with your argument is that you're talking as if all these speakers measured the same. They don't. The Kh120 measures darker than your typical genelec of focal, so it does not support your argument that it sounds darker because of the tweeter material.

The KH120 is a "neutral-ish" speaker but obviously no speaker is perfectly neutral, especially not in a direct comparison. The KH120 specifically has broad scoop with two prominent dips in the off-axis response from ~1.5 to 7kHz. A region that is exactly where you'd associate 'brightness' and detail within a speaker. Behold, @dominikz measurements of his KH120:

View attachment 182972

Dip aside, there is also a slight downward trendline to the measurements overall (it's not much, but we can definitely be sensitive to small FR tilts).


Compare that with, say, a Focal Solo6 (current version, my measurements):

Notice the pronounced brightness off-axis in the presence region, and note that we are more sensitive to peaks than we are valleys. Amir also measured an older version of the solo6 which had somewhat different on-axis:

Here are erin's measurements of the Focal Twin6 Be:

Notice similar off-axis peaking in the presence region.

Now let's look at some Genelecs. 8341B:

Pretty flat throughout, although also notice the on-axis seems

8361B:

Or a genelec 8050B:

There is in the 8050B above, but it happens lower in the frequency response, not to much in the detail region.

All of this is to say that before we even consider what possible effect the tweeter material alone could have, the frequency response and directivity already suggest the neumann will be the darkest-sounding of these speakers.

Yes you are absolutely right. The dark dull sounding is a problem of all small speaker with a (deeper) waveguide. As I wrote here just a minute ago.

Post in thread 'Geithein RL 906 vs Genelec 8030C' https://audiosciencereview.com/foru...in-rl-906-vs-genelec-8030c.25163/post-1069313

 

[peniku8]

No, youre right about that. The instruments caracters is defined by the upper frequency range. We dont listen to sine-sweeps when we listen to music. If the hights is not audible good /correct sounding, everything is gonna sound bad. Try to use a program or eq that can cut everything above 1 kHz and listen how bad everything sounds. Doing the same cutting at 100 Hz and below still gonna sound good if everything above 100 Hz is intact.

You're comparing cutting 2 octaves of low-end vs 4 octaves of top-end. Not a fair comparison imo.
Also varies with source content imo. If I had to choose giving up one octave of the spectrum, I'd choose to remove the lowest octave when listening to Classic Rock or a string quartet, while choosing to remove the top octave while listening to Techno or EDM.
Violin solo? Can live without a subwoofer. Electric guitar? Don't need a tweeter (guitar cabs don't even have tweeters).

 

[tuga]

We mainly hear direct sound frequency response, directivity and distortion (in this row of decreasing significance), here I had posted a research about the effect of different tweeter materials and types.

The effect of materials employed in drivers re loudspeaker performance

I noticed in one of Amir's speaker reviews a newer member made an allusion to driver materials having some salutary influence on the reproduced sound of a speaker. All things being equal, I wonder how true this assertion really is. I know it's a great marketing point for some of the companies...

www.audiosciencereview.com

In my view, that research about the effects of different tweeter materials and types is not fit for proving (or disproving) anything.
The sample is tiny (and the summary you translated makes no mention of the level of listening ability) and then there's a lot of processing to the recorded sound before the evaluation of the simulation of a stereo pair finally takes place on headphones.
Some researchers seem to live on another planet... Was this a Masters thesis? How did it pass peer scrutiny?

Summary
The present study examined the popular opinion that the subjective sound impressions of tweeters are clearly related to the drive principle, the diaphragm material used and the geometrical extension of the diaphragm. For this purpose, six typically in tweeters used in the current loudspeakers, which can be both in the transducer principle, its geometry and in the sound-generating material of the diaphragm.
In order to exclude typical discoloration of the sound image due to interaction with the surrounding room as well as with the baffle of a loudspeaker cabinet, all drivers were placed in a quasi infinite baffle in an anechoic environment. In order to avoid discoloration of the sound image, which solely from the driver-specific distortions of the amplitude and phase response all drivers were tested using a digital crossover based on FIR filters for one point in space equalized. The goal of the compensation was a linear amplitude response in the range from 2 kHz to 22 kHz with maximum deviations of ±1 dB and a linear phase response. After successful equalization of all drivers, they were used in combination with binaural impulse responses are recorded by a compensated bass-midrange driver. The Measurement of the BIR was performed in positions relative to the drivers, which are a left and right corresponded to a right speaker in a typical stereo arrangement. In order to enable a dynamic simulation of the loudspeaker pairs in a later comparison, the BIR were plotted in a radius of typical head movements.
In a final listening test, the two virtual pairs of loudspeakers were compared, which showed the greatest deviations from each other in the analysis of the binaural data. On the basis of the binaural, dynamic simulations of the loudspeaker pairs, ten test persons were asked to find out whether there was any difference at all, which could be determined by the subjective sound image exists between the tweeters. The result of the experiment allows the conclusion that there is no longer any perceptual difference between the tweeters.
In conclusion, it can be stated that in the case of excluded interaction with the room and the baffle, with drivers linearised axially in amplitude and phase, there is no longer any difference in linear operation between loudspeakers with different transducer principles, different diaphragm materials or diaphragm expansions exists. The common vernacular can thus, at least under these conditions be considered disproved.
Further binaural measurements and simulations can now be performed to evaluate the effects of other parameters such as the directional characteristic and the resulting driver-specific interaction with the space. Likewise the differences can be examined, which result from driver-specific interactions with the baffle or resulting from non-linear operation of the tweeters.

 

[tuga]

If the highs feel rolled off, the sound is dull and clarity feels less good.
The KH120 may measure perfectly but sounded dark to me, on the other hand the Genelecs sounded more lively and clarity was better to me, I also liked Focal's Berillyum.

The problem with a listening assessment is that you are comparing what you listen to your references.

In other words, your references must produce a neutral tonal balance (frequency response) or your impressions will be skewed/biased. To make matters worse, the partnering electronics, the room and even the recordings will also affect your perception.

For this reason, measurements are a good starting point but cannot replace listening; a speaker may produce stellar measured performance but not sound good to you.

 

[thewas]

In my view, that research about the effects of different tweeter materials and types is not fit for proving (or disproving) anything.
The sample is tiny (and the summary you translated makes no mention of the level of listening ability) and then there's a lot of processing to the recorded sound before the evaluation of the simulation of a stereo pair finally takes place on headphones.

Very widely different types of tweeters were used

Name Diaphragm Drive Diaphragm Diaphragm Sound pressure
shape principle material expansion [dB]
A.D.A.M planar, magneto- plastic 2,8 x 3,5cm 92
X-type folded static
Expolinear planar magneto- aluminium 4 x 0,5cm 92,5
ARL 90 static
Vifa dome electro- aluminium 2,5cm 90
D25 AG35 dynamic
Vifa ring radiator electro- fabric 2,6 cm 91,5
XT 25 GT 30 dynamic
Expolinear planar magneto- plastic 11.6 x 2.4 cm 100
RT7-Pro static (polyimide)
SEAS dome electro- fabric 2.6 cm 91
Excel T25 dynamic
Eton cone electro- HexaCone 14 cm 87
5-880/25 dynamic (polyaramide)

and FIR filtering was done to EQ them to the same response. If you know a different/better study were such a test was done, please post it, otherwise it is what we have to live with at this time of point until proven otherwise.

 

[tuga]

Very widely different types of tweeters were used

Name Diaphragm Drive Diaphragm Diaphragm Sound pressure
shape principle material expansion [dB]
A.D.A.M planar, magneto- plastic 2,8 x 3,5cm 92
X-type folded static
Expolinear planar magneto- aluminium 4 x 0,5cm 92,5
ARL 90 static
Vifa dome electro- aluminium 2,5cm 90
D25 AG35 dynamic
Vifa ring radiator electro- fabric 2,6 cm 91,5
XT 25 GT 30 dynamic
Expolinear planar magneto- plastic 11.6 x 2.4 cm 100
RT7-Pro static (polyimide)
SEAS dome electro- fabric 2.6 cm 91
Excel T25 dynamic
Eton cone electro- HexaCone 14 cm 87
5-880/25 dynamic (polyaramide)

and FIR filtering was done to EQ them to the same response. If you know a different/better study were such a test was done, please post it, otherwise it is what we have to live with at this time of point until proven otherwise.

I don't know of any other study. But this one is unfit for purpose and thus proves nothing.

Sorry but I don't see why "it is what we have to live with" when the study methodology is rubbish and the sample is minuscule...
I prefer to live with "we don't know".

Eagerness to have some kind of proof no mater how feeble is bad science.

 

[thewas]

In other words, your references must produce a neutral tonal balance (frequency response) or your impressions will be skewed/biased. To make matters worse, the partnering electronics, the room and even the recordings will also affect your perception.

Would put "partnering electronics" with distance on the last place of significance though and typically unfortunately only the "audiophile grade" ones.

For this reason, measurements are a good starting point but cannot replace listening; a speaker may produce stellar measured performance but not sound good to you.

Can be to some people, although personally I have never had the case till now for myself, if of course well placed and bass corrected in a decent acoustics room and most people I know feel the same.

 

[thewas]

I don't know of any other study. But this one is unfit for purpose and thus proves nothing.

Sorry but I don't see why "it is what we have to live with" when the study methodology is rubbish and the sample is minuscule...
I prefer to live with "we don't know".

Eagerness to have some kind of proof no mater how feeble is bad science.

Sorry, but that is just your own subjective perception and not really tangible critic of the study, which I rather see as "rubbish".

 

[Thomas_A]

Sensitivity is frequency dependent, and JNDs vary over the spectra. So any deviations may have to be weighted in terms of audibility. 3-4 kHz sensitivity is one example in the tweeter range.

 

[thewas]

By the way to see how much the frequency response dominates the listening impression, I can only recommend below videos and recordings were some very different headphones were EQed to the same response and recorded binaurally:

Download of the ABX 5 headphones equalized to the same frequency response: FLAC 96/24 - https://cutt.ly/2YNcfSA

Yes, with careful listening some very small differences can be heard, but these are mainly due to high distortions and small FR irregularities like channel imbalance.

 

[tuga]

Can be to some people, although personally I have never had the case till now for myself, if of course well placed and bass corrected in a decent acoustics room and most people I know feel the same.

I use measurements for shortlisting but would not do without the listening assessment. It is how I find out whether or not I can enjoy listening to music when reproduced by a particular piece of equipment of system.

But many (perhaps most) audiophiles will spend their lives trying different equipment in search for their favourite presentation. And quite a few enjoy the journey more than the destination (not me though).

 

[thewas]

I use measurements for shortlisting but would not do without the listening assessment. It is how I find out whether or not I can enjoy listening to music when reproduced by a particular piece of equipment of system.

But many (perhaps most) audiophiles will spend their lives trying different equipment in search for their favourite presentation. And quite a few enjoy the journey more than the destination (not me though).

I also enjoy the listening assessment part and my final choice is usually done from that as there exist different implementations of very good measuring loudspeakers (for example wider vs narrower directivity), just wanted to say with the above post though that I had never had a loudspeaker with a full set of great measurements (on-axis is not enough) that I didn't enjoy its sound.

 

[DJBonoBobo]

So the problem with your argument is that you're talking as if all these speakers measured the same. They don't. The Kh120 measures darker than your typical genelec of focal, so it does not support your argument that it sounds darker because of the tweeter material.

The KH120 is a "neutral-ish" speaker but obviously no speaker is perfectly neutral, especially not in a direct comparison. The KH120 specifically has broad scoop with two prominent dips in the off-axis response from ~1.5 to 7kHz. A region that is exactly where you'd associate 'brightness' and detail within a speaker. Behold, @dominikz measurements of his KH120:

View attachment 182972

Dip aside, there is also a slight downward trendline to the measurements overall (it's not much, but we can definitely be sensitive to small FR tilts).


Compare that with, say, a Focal Solo6 (current version, my measurements):

Notice the pronounced brightness off-axis in the presence region, and note that we are more sensitive to peaks than we are valleys. Amir also measured an older version of the solo6 which had somewhat different on-axis:

Here are erin's measurements of the Focal Twin6 Be:

Notice similar off-axis peaking in the presence region.

Now let's look at some Genelecs. 8341B:

Pretty flat throughout, although also notice the on-axis seems

8361B:

Or a genelec 8050B:

There is in the 8050B above, but it happens lower in the frequency response, not to much in the detail region.

All of this is to say that before we even consider what possible effect the tweeter material alone could have, the frequency response and directivity already suggest the neumann will be the darkest-sounding of these speakers.

Good try explaining that to him one more time. I was obviously not successful in March 2021... ;-)

 

[tuga]

Sorry, but that is just your own subjective perception and not really tangible critic of the study, which I rather see as "rubbish".

We've been through this when discussing Harman's research methodolgy. I am not only sceptical of subjective reports but also very demanding when it comes to research.

 

[fredstuhl]

In my view, that research about the effects of different tweeter materials and types is not fit for proving (or disproving) anything.
The sample is tiny (and the summary you translated makes no mention of the level of listening ability) and then there's a lot of processing to the recorded sound before the evaluation of the simulation of a stereo pair finally takes place on headphones.
Some researchers seem to live on another planet... Was this a Masters thesis? How did it pass peer scrutiny?

Summary
The present study examined the popular opinion that the subjective sound impressions of tweeters are clearly related to the drive principle, the diaphragm material used and the geometrical extension of the diaphragm. For this purpose, six typically in tweeters used in the current loudspeakers, which can be both in the transducer principle, its geometry and in the sound-generating material of the diaphragm.
In order to exclude typical discoloration of the sound image due to interaction with the surrounding room as well as with the baffle of a loudspeaker cabinet, all drivers were placed in a quasi infinite baffle in an anechoic environment. In order to avoid discoloration of the sound image, which solely from the driver-specific distortions of the amplitude and phase response all drivers were tested using a digital crossover based on FIR filters for one point in space equalized. The goal of the compensation was a linear amplitude response in the range from 2 kHz to 22 kHz with maximum deviations of ±1 dB and a linear phase response. After successful equalization of all drivers, they were used in combination with binaural impulse responses are recorded by a compensated bass-midrange driver. The Measurement of the BIR was performed in positions relative to the drivers, which are a left and right corresponded to a right speaker in a typical stereo arrangement. In order to enable a dynamic simulation of the loudspeaker pairs in a later comparison, the BIR were plotted in a radius of typical head movements.
In a final listening test, the two virtual pairs of loudspeakers were compared, which showed the greatest deviations from each other in the analysis of the binaural data. On the basis of the binaural, dynamic simulations of the loudspeaker pairs, ten test persons were asked to find out whether there was any difference at all, which could be determined by the subjective sound image exists between the tweeters. The result of the experiment allows the conclusion that there is no longer any perceptual difference between the tweeters.
In conclusion, it can be stated that in the case of excluded interaction with the room and the baffle, with drivers linearised axially in amplitude and phase, there is no longer any difference in linear operation between loudspeakers with different transducer principles, different diaphragm materials or diaphragm expansions exists. The common vernacular can thus, at least under these conditions be considered disproved.
Further binaural measurements and simulations can now be performed to evaluate the effects of other parameters such as the directional characteristic and the resulting driver-specific interaction with the space. Likewise the differences can be examined, which result from driver-specific interactions with the baffle or resulting from non-linear operation of the tweeters.

Yes, it‘s a masters thesis. There is some more info on the test subjects, directly copied and autotranslated from the main text:

A total of ten subjects took part in the listening test, with one person repeating the listening test (see evaluation). The average age of all persons was 31 years. Seven of the people were either employees or master students in the department of audio communication or came from their environment and were familiar with conducting listening tests and evaluating differences in acoustic simulations. Three of the people had no experience with listening tests, but had a musical education.

I don't know of any other study. But this one is unfit for purpose and thus proves nothing.

Sorry but I don't see why "it is what we have to live with" when the study methodology is rubbish and the sample is minuscule...
I prefer to live with "we don't know".

Eagerness to have some kind of proof no mater how feeble is bad science.

I agree that this study is not fit to unmistakably proof that tweeter material or construction has no audible effect after eqed to the same frequency response and with reflexions taken out of the picture.

At the same time, I don't see any technical flaws that would justify a complete invalidation of their data. So for me personally, unless disproven by a more thorough study with a higher level of scrutiny, this evidence definitely helps to put things in perspective.