Good morning everyone
in a rather interesting video by GR Research Company they talk about a blind comparison test between two speakers with the same frequency response and sound judged different
Admitting the validity of the test I think can we can that the FR alone is not enough to characterize the sound of a speaker ?
The discussion on instrumental analysis protocols unfortunately ended here
The aim of the video was to establish the importance of the xover design and construction mostly Same values different parts different sound
The somewhat banal consideration is that for example speakers showing a very similar FR can have a very different CSD and distortion behaviour
At this point I ask myself but why not make further efforts to understand the dynamic behavior of a speaker ?
instead I see that important and perhaps decisive tests to establish the reproduction quality of a speaker such as CSD and distortion measurements are extremely rare in the various tests
Why? because they are difficult to perform ?
to be honest some datasheets report them But very few unfortunately
i have an idea but i do not want to provoke a storm
i am here to learn
Thank You very much
in a rather interesting video by GR Research Company they talk about a blind comparison test between two speakers with the same frequency response and sound judged different
Admitting the validity of the test I think can we can that the FR alone is not enough to characterize the sound of a speaker ?
The discussion on instrumental analysis protocols unfortunately ended here
The aim of the video was to establish the importance of the xover design and construction mostly Same values different parts different sound
The somewhat banal consideration is that for example speakers showing a very similar FR can have a very different CSD and distortion behaviour
At this point I ask myself but why not make further efforts to understand the dynamic behavior of a speaker ?
instead I see that important and perhaps decisive tests to establish the reproduction quality of a speaker such as CSD and distortion measurements are extremely rare in the various tests
Why? because they are difficult to perform ?
to be honest some datasheets report them But very few unfortunately
i have an idea but i do not want to provoke a storm
i am here to learn
Thank You very much
You may be interested in following Erin's Audio Corner over at Youtube, where he auditions loudspeakers, measures them, and then gives his impression of how the measurements correlate with what he hears.
https://www.youtube.com/@ErinsAudioCorner
https://www.youtube.com/@ErinsAudioCorner
Hi ginetto61, i do have one reason why speakers with the same fr response sound different. There may be more reasons, but one important parameter is directionality. Or beaming, in other word.
Here is why i think so. I made many speakers. Dozens. Typically i make bass box, and separate top with mid and tweeter. Bass stays, and many tops are being swapped and compared. Every time i replace the mids/tweeters, i do equalize for flat response in the listening position. (behringer 31 band ultracurve)
Inspite the same flat fr response, each speaker has slightly different sound. Each has slightly different directionality. Each 'sprays' the room with different width. This you can not eq. Each type of tweeter has directionality index. Selenium horn has completely different directionality vs hiquphon small dome. Different directionality energizes room in different manner with different amount of reflections. Hence slightly different sound.
You can easily measure directionality index. But how you evaluate final sound in room, when it comes to multiple speakers comparison, especially when it comes to soundstage? We can not measure soundstage. Subjective listening test for now is the only way.
Here is why i think so. I made many speakers. Dozens. Typically i make bass box, and separate top with mid and tweeter. Bass stays, and many tops are being swapped and compared. Every time i replace the mids/tweeters, i do equalize for flat response in the listening position. (behringer 31 band ultracurve)
Inspite the same flat fr response, each speaker has slightly different sound. Each has slightly different directionality. Each 'sprays' the room with different width. This you can not eq. Each type of tweeter has directionality index. Selenium horn has completely different directionality vs hiquphon small dome. Different directionality energizes room in different manner with different amount of reflections. Hence slightly different sound.
You can easily measure directionality index. But how you evaluate final sound in room, when it comes to multiple speakers comparison, especially when it comes to soundstage? We can not measure soundstage. Subjective listening test for now is the only way.
Hi just a first quick answer I will get back with comments after i have checked all the very kind and valuable advice
I just say now that i have the feeling to be one of the very few obsessed by distortion
Just think to the definition of distortion
the act of twisting or altering something out of its true, natural, or original state
see you later
I just say now that i have the feeling to be one of the very few obsessed by distortion
Just think to the definition of distortion
the act of twisting or altering something out of its true, natural, or original state
see you later
Hello professor G, nice to meet you again, adding a new conundrum to the already gazillion ones... or as the mosquito said when taking a pee while flying over the atlantic ocean... what does it matter? :)
Well... it's said the room reflections accounts for nn% (can't remember the figures but relatively quite high, usually) vs the direct sound from the speakers, so the listening venue and its acoustic treatment, or lack of, accounts for maybe the same nn% of the sound coloration... so why not throw out the room and put on a headphone on the head, problem solved... partially, no? Orr.. make the speakers beaming lake lasers towards the ears, projecting by means of some exotic technology... microwaves maybe? ::)
Now the CSD measurements is something I miss too, but at least Stereophile add them and have done so for ages, look here for instance in their latest loudspeaker review in September reviewing a B&W, they are measuring both the box CSD attaching an accelerometer to it, and making a CSD measurement usually aligned with the tweeter axis, good? :::)
https://www.stereophile.com/content/bowers-wilkins-805-d4-signature-loudspeaker-measurements
Well... it's said the room reflections accounts for nn% (can't remember the figures but relatively quite high, usually) vs the direct sound from the speakers, so the listening venue and its acoustic treatment, or lack of, accounts for maybe the same nn% of the sound coloration... so why not throw out the room and put on a headphone on the head, problem solved... partially, no? Orr.. make the speakers beaming lake lasers towards the ears, projecting by means of some exotic technology... microwaves maybe? ::)
Now the CSD measurements is something I miss too, but at least Stereophile add them and have done so for ages, look here for instance in their latest loudspeaker review in September reviewing a B&W, they are measuring both the box CSD attaching an accelerometer to it, and making a CSD measurement usually aligned with the tweeter axis, good? :::)
https://www.stereophile.com/content/bowers-wilkins-805-d4-signature-loudspeaker-measurements
The difficulty with that assumption is: what is the "original state" or how would you define the relevant parameters?
Is it how music was reproduced and shaped in the mastering room?
In the electrical domain it's easier to compare two signals to the original ones. But in a real acoustic situation?
Would you need an anechoic room?
Headphones?
The only tools available that are objective and repeatable are measurements
When Dynaudio for mention one name check their products before shipping they measure them
The lazy people listen from a sofa
The good guy puts up microphones to measure
The ideal driver or speaker has perfect measurements
Flat FR clean CSD zero distortion at the wanted SPLs
In the video I mentioned they were relying on listening to assess the sonic superiority of one speaker vs the other
One important step is missing
That's to perform a full set of measurements on both speakers and compare the results
I am pretty confident that the better sounding one had lower thd imd and a cleaner CSD
But I can't carry out the needed instrumental check
There's ideality
The more close the reality to ideality the better
But a testing protocol is needed
When Dynaudio for mention one name check their products before shipping they measure them
The lazy people listen from a sofa
The good guy puts up microphones to measure
The ideal driver or speaker has perfect measurements
Flat FR clean CSD zero distortion at the wanted SPLs
In the video I mentioned they were relying on listening to assess the sonic superiority of one speaker vs the other
One important step is missing
That's to perform a full set of measurements on both speakers and compare the results
I am pretty confident that the better sounding one had lower thd imd and a cleaner CSD
But I can't carry out the needed instrumental check
There's ideality
The more close the reality to ideality the better
But a testing protocol is needed
but who defines the measurement parameters?
you seem to exclude the room. that's perfectly fine for isolated measurements,
but you explicitly stated "listening impression" in your title.
you absolutely need to include the room (and the off-axis radiation/energy) in your evaluation.
it's one of the main influences for "listening impression".
and then you need to clarify what is "correct" in the "listening impression" category. that's where you enter the circle of confusion.
https://www.audioholics.com/room-acoustics/sound-reproduction
Most measurements you can find will tend to focus on the frequency response on axis and maybe the HD peformance. That is only the tip of the iceberg in terms of acoustic outputs and acoustic related parameters that can and will effect the perceived sound quality in a number of ways, including:
Intermodular distortion
How loud should the loudspeaker play, on average?
Radiation pattern (what's the output of the loudspeaker off axis, on every non-listener axis)
How is the loudspeaker's output (on all axes) interacting with the ROOM. E.g. what is the RT of the room, are there hard or soft surfaces and where are these located, shape of room, diffusing surfaces vs reflecting surfaces, etc.
How large is the room (W/D/H and total volume) and where in the room are the loudspeakers and the listener located? Does the room have openings into other spaces?
The bottom line is that this very question @ginetto61 is asking has been asked by others, and professionally investigated (e.g. see Sean Olive's work) for decades, and we have many clues but few definitive answers.
Intermodular distortion
How loud should the loudspeaker play, on average?
Radiation pattern (what's the output of the loudspeaker off axis, on every non-listener axis)
How is the loudspeaker's output (on all axes) interacting with the ROOM. E.g. what is the RT of the room, are there hard or soft surfaces and where are these located, shape of room, diffusing surfaces vs reflecting surfaces, etc.
How large is the room (W/D/H and total volume) and where in the room are the loudspeakers and the listener located? Does the room have openings into other spaces?
The bottom line is that this very question @ginetto61 is asking has been asked by others, and professionally investigated (e.g. see Sean Olive's work) for decades, and we have many clues but few definitive answers.
Hi thank you very much for the link
He says that distortion is overrated
https://en.wikipedia.org/wiki/George_Clifford_Sziklai
if i remember well he contributed to the signal transfer technology reducing distortion on phone lines
lowering the distortion helped in phone communications a lot
another great read has been the Tom Danley speakers testing using successive recording of a human talk played back by speakers
after some recordings is impossible to understand the talk
this is the result of adding distortion to the original signal
that is the goal of the research Some are already defined
if you want to get insults just propose to use a square wave to test a passive component
i personally love SW Because is squared Precise
because it is difficult to sim a room That is why the measurements are performed in a standard anechoic room
nobody exclude the room influence
My guessing is that the better sounding speaker should have better measurements also in an anechoic room
yes but remember that both speakers had the same FR in the same room
fwiu the only difference was in the xover And i accept that different grade parts same value can contribute differently to the overall sound
I would be curious to see some measurements on the crossovers
If they sound different some measurement must be different
Just to add confusion i was watching a test on high end capacitors
They measured esr some dissipation factor i do not know
I asked ... since you've tried everything why not pass a square wave through it?
answer Great idea !
They didn't do it
I felt a bit like an idiot I must have missed something
Good luck with that! 😊
It's revealing that you used the wording "my guessing is" to relate "better sounding" to a measurement.
You also did not define what "better measurement" means. so even defining "better measurement" is your personal choice.
maybe it's square wave reproduction for you.
maybe i could change terms and saying a better behaviour A better response to impulse A lower IMD
i spent many years listening to small speakers thinking they were sounding great
i was missing a lot of the musical instruments
I understood this the 1st time i heard a pair of JBL L166 With a powerful 12"
a set of measurements could have predicted that listening impression
When i read of evaluation of passive parts by ear i am sincerely puzzled
How close a real part is to ideal can be measured
The same can apply to a complete speakers
Sometimes i think that even the taste of food could be measured With some chemical analysis
i spent many years listening to small speakers thinking they were sounding great
i was missing a lot of the musical instruments
I understood this the 1st time i heard a pair of JBL L166 With a powerful 12"
a set of measurements could have predicted that listening impression
When i read of evaluation of passive parts by ear i am sincerely puzzled
How close a real part is to ideal can be measured
The same can apply to a complete speakers
Sometimes i think that even the taste of food could be measured With some chemical analysis
I would be suspicious of the results from any test by GR Research of their competitors' products. Most, if not all, of those tests end up with them trying to sell you something to improve what you bought from someone else. I find it difficult to believe that all of those other manufactures have missed something important in designing and building their products.
And at the same time GR Research is trying to sell you expensive, overdesigned speaker cables of their own.
What really strikes me also as a matter of snake oil is their suggestion to get cable risers to improve the performance of your system. That's enough for me to discount just about everything else they have to say.
Last edited:
Yes i agree
I would be extremely curious to see on a lab bench one of his huge coils against a just decent one of the same value
They have to measure different on some parameters
Being able to listen only up to 13k I don't trust my ears anymore
I want to see the difference more than to hear it
I would be extremely curious to see on a lab bench one of his huge coils against a just decent one of the same value
They have to measure different on some parameters
Being able to listen only up to 13k I don't trust my ears anymore
I want to see the difference more than to hear it
However I had the impression that he has the ability to find very good and cheap drivers
He was showing a nice 6" woofer sourced from some Indian company
Plastic frame but very nice
I am sure that if I knew the price I would be shocked
That's a great gift
To find parts that perform much better than the price would tell
But again I'm sure that drivers selection can be carried out also by a deaf person
With the right equipment of course
He was showing a nice 6" woofer sourced from some Indian company
Plastic frame but very nice
I am sure that if I knew the price I would be shocked
That's a great gift
To find parts that perform much better than the price would tell
But again I'm sure that drivers selection can be carried out also by a deaf person
With the right equipment of course
Do you ever visit live performances? Experiencing any (classical) live concert will probably tell you more about sound reproduction in a room than a thousand loudspeaker listening tests. Just my 2ct.
How would you interpret a measurement without having experienced the bigger speakers?