Speakers are least important

[STC]

24 minutes ago, PeterSt said:

No. That could have been so, though

 

FM response is the response inside your ears which transforms the flat response that reaches your ears. 

 

We already taken the the FM curve into consideration since birth and adapted to the sound to interpret them as flat or bright or dull based on the FM curve. No reproduction system should ever try to recreate the FM curve because that would again transform based on the natural FM curve inside your ears. 

 

This is is not the first I have seen because I know of audiophiles who use paraEQ to change the output from speakers to represent the FM curve. I am not really following what you are saying because it looks like you have software which actually transforms the supposedly flat FR from a perfect system to be curve like FM response. 

 

Having said that, it is normal to have one preference over the other which is subjective and inconsistent. The actual correcting would have made them to bright. A common most act by recordists who uses binaural recording to play them with loudspeakers without correction. 

[Paul R]

7 hours ago, Abtr said:

Are you saying that cabinet resonances don't show up in FR measurements?

 

Turn that around - are you are saying that any tonal difference between speakers  (bright, soft, forward, reserved, any adjective you choose to use...) must be a result of differences in frequency response? 

 

While I would expect that to be mostly true,  would say it is far from true in all cases.

 

In other words, frequency response is not the only thing that defines the quality of what we hear. 

 

Do you disagree? 

[fas42]

2 hours ago, PeterSt said:

 

I don't agree, again. The better the system "is" as a whole, the more frequency response curves are directly audible.

 

 

That is from our Orelo MKII speaker.

 

Yes, our thinking is very different in many areas. You see, all I see there is a weak link, in the playback chain, right there in those set of switches - I don't care how fabulous the speaker is in its raw capabilities, how tremendous are the qualities of the parts used in the crossovers; to me, the fact that you can fiddle with the response of the speakers, and have indicator lights as well, sets off alarm bells for me ... if I had a pair of these to fool around with, the first thing I would do is completely disable, bypass all that fanciness.

 

2 hours ago, PeterSt said:

 

 

Possibly the moral should be: the better all is for technical quality, the more the smallest changes become vastly audible. The other moral could be that I don't work with your (Frank's) mind (thing). Just measurement up to capability and reason beyond that.

 

 

And how I interpret hearing the "smallest changes" is that the added complexity of the circuitry, and parts, introduces some distortion element which is audible - I have always removed the extra doodahs in a rig, because I can always hear the degrading impact of their presence. Once simplified, the level of 'purity' of the SQ lifts, and I never regret having dumped them - this has always been a one-way street for me, for decades.

[PeterSt]

50 minutes ago, STC said:

I am not really following what you are saying because it looks like you have software which actually transforms the supposedly flat FR from a perfect system to be curve like FM response.

 

Maybe it is harder to follow because I / Phasure design and build this stuff (everything in the audio chain) which is meant for customers to buy - which some do. 😑 Also, I am the user of my own (Phasure) system. So that can be confusing.

 

The FM curves are hardware in the speaker (say filters) and where actually meant as a "justified" means to offer 27 different FR curves. They are really FM, but only for a combination of one (3 fold) setting. The idea is that nobody can go wrong with this speaker because if the FR is not exactly as people want it, one can change it by means of simple switches (on the panel II showed). This is not related to software at all.

For the LF part it is DSP (also in the speaker).

 

1 hour ago, STC said:

No reproduction system should ever try to recreate the FM curve because that would again transform based on the natural FM curve inside your ears. 

 

I fully agree with that.

 

Keep in mind the gist of this all. About the +/- 3dB thing, which applies way more of the so-called FM curves, though at random places (random for you as the customer and a pain for the manufacturer who can't get rid of it). IOW a mess. So again my maybe controversial point of view: why bother discussing FR, while it is out of control anyway, WHILE we know (OK, I do) that the minute changes are vastly audible.

I can imagine that this is harder to follow (this starts with my English), and it actually needs a reference graph of what I mean. Maybe you have one smoothed 1/12 or less (with proof of that please) so we have a chalkboard picture to talk about. But example:

 

 

That dip you see is at 17.8KHz. This is cancellation in the horn because of the too much on-axis the measurement as taken. For me this is a severe dip (though still 1dB or so). But take the measurement more off-axis and other anomalies show. So I chose to use this one.

IOW, when not careful we are hoaxed anywhere and only when we (can) measure ourselves we know for sure what the real outcome is (if we know how to measure of course). And at least I don't fool myself with it, plus it encourages for improvement again. This is on behalf of myself, and the customer at second place (yep) too.

 

Because you like to know: IIRC these measurements are taken at 118cm from the mouth opening, which is hardly really relevant for this. But now think what happens when the crossover is to be measured decently and how the height is totally crucial (woofers are at the bottom). This is a bit horn related of course, but I can help you out of all dreaming: this measurement is impossible because of the necessary distance, the sheer height of the woofer section itself (not aligned time arrival at the too short distance) and warbling and more. Oh, you can do it all right, if the norm would be +/-3dB (or even worse). And so this goes per section, and the remainder is theory (a crossover just has to match - period). Thus, the bass is set for its crossover (which is say 1000 measurements) and the mid is set separately from that. The connection can only be roughly checked with measurement (but usually does not make real sense because the whole FR looks like a mess because of the necessary distance - we talked about that).

 

Maybe the above was not all relevant - I just explain a little how I do it. It is no sinecure at all, which starts with a horn not being allowed to sound like a horn at all (take am empty horn, talk through it and listen to the most odd sound). The filter for the top horn (co-axial mid and high) took me a year of tuning. Actually two of us.

 

 

[semente]

1 hour ago, Paul R said:

 

Turn that around - are you are saying that any tonal difference between speakers  (bright, soft, forward, reserved, any adjective you choose to use...) must be a result of differences in frequency response? 

 

While I would expect that to be mostly true,  would say it is far from true in all cases.

 

In other words, frequency response is not the only thing that defines the quality of what we hear. 

 

Do you disagree? 

 

I think that cabinet or driver resonces will affect the frequency response but in most cases one cannot easily determine that they exist (or what is causing them) without resorting to other measurments like impedance, CSD or THD+N, or by listening if you're experienced.

 

Here's an example:

 

Monitor Audio Studio 15, anechoic response on tweeter axis at 44" (black), corrected for microphone response, with individual responses of the tweeter (red), woofer (blue), and port (green), measured in the nearfield below 300Hz.

Monitor Audio Studio 15, cumulative spectral-decay plot of woofer at 44" (0.15ms risetime).

Monitor Audio Studio 15, cumulative spectral-decay plot at 44" (0.15ms risetime).

 

Drive units: 1" (25mm) aluminum-alloy dome tweeter, 6.5" (170mm) aluminum-alloy cone woofer

Crossover: approximately 3kHz, 2nd-order

 

https://www.stereophile.com/content/monitor-audio-studio-15-loudspeaker-measurements

[semente]

2 hours ago, STC said:

 

So you believe that correcting the FR so that the curve is somewhat closer to Fletcher Munson curve?  

 

 

 

2 hours ago, PeterSt said:

 

No. That could have been so, though. Only flat works out best for me. I don't know about others (other customers).

 

In my view, using adjustable to FM/Contour curves compensation EQ settings allows the user to achieve flat FR at different listening/playback SPLs.

That is why some amplifiers used to have a Loudness/Countour circuit, which would boost frequency extremes to adjust to our lower sensitivity at these frequencies when listening at low SPLs.

Some speakers (with a fixed FR curve/balance) will only sound balanced (FR-wise) when playing at very loud SPLs (inverted smile curve) whilst others will sound balanced only at moderate levels (smile curve).

[semente]

On 7/13/2019 at 9:05 AM, STC said:

There was a disputed between Harbeth and Stereophile when the measurements of Stereophile did not agree. Measurements varies as the standard/method is not fixed.

 

That happened because Stereophile doesn't have an anechoic chamber and they use a "trick" to overcome this:

https://www.stereophile.com/content/measuring-loudspeakers-part-three-page-6

 

Still if you take the resulting 6dB boost in the bass artefact (usually around 100Hz with large speakers) then the listening-window measured responses aren't too different from those performed at the anechoic chamber of the NRC in Canada:

 

Revel Ultima Salon2

 

 

 

[PeterSt]

33 minutes ago, fas42 said:

if I had a pair of these to fool around with, the first thing I would do is completely disable, bypass all that fanciness.

 

Yes, good point. That is, for someone who does not measure during the design, thus does not know the difference, and maybe (for sure) does not know how all is set up. Still I'd agree with you from the distance. Now look:

 

Hey, here I am. I sell a speaker. The pair actually. They sound the very best because I say so. Shipping to your place down under is 2000 euros on 3 pallets. Go for it.

Don't like it ? just send them back. If you are lucky you can do that for 2000 euros too.

 

Or do you really think that you can change the lot to your liking ?

Of course you can't. That's why the provision for it. And it really is not detrimental (at least not by measurement).

 

So that's why. Internet selling is a bit tough on the auditioning part. There must be some kind of guarantee that it is "best".  Never mind this is not common thinking.

Frank, you can't participate. You're that guy who isn't even interested in all below 100Hz. Remember ?  Still I would agree with you. But spending a lot of money is not fun without some kind of guarantee it is "good". So you give in a little (if at all) and be more happy. At least that is the idea.

[STC]

28 minutes ago, PeterSt said:

 

Maybe it is harder to follow because I / Phasure design and build this stuff (everything in the audio chain) which is meant for customers to buy - which some do. 😑 Also, I am the user of my own (Phasure) system. So that can be confusing.

 

The FM curves are hardware in the speaker (say filters) and where actually meant as a "justified" means to offer 27 different FR curves. They are really FM, but only for a combination of one (3 fold) setting. The idea is that nobody can go wrong with this speaker because if the FR is not exactly as people want it, one can change it by means of simple switches (on the panel II showed). This is not related to software at all.

For the LF part it is DSP (also in the speaker).

 

 

I fully agree with that.

 

Keep in mind the gist of this all. About the +/- 3dB thing, which applies way more of the so-called FM curves, though at random places (random for you as the customer and a pain for the manufacturer who can't get rid of it). IOW a mess. So again my maybe controversial point of view: why bother discussing FR, while it is out of control anyway, WHILE we know (OK, I do) that the minute changes are vastly audible.

I can imagine that this is harder to follow (this starts with my English), and it actually needs a reference graph of what I mean. Maybe you have one smoothed 1/12 or less (with proof of that please) so we have a chalkboard picture to talk about. But example:

 

 

That dip you see is at 17.8KHz. This is cancellation in the horn because of the too much on-axis the measurement as taken. For me this is a severe dip (though still 1dB or so). But take the measurement more off-axis and other anomalies show. So I chose to use this one.

IOW, when not careful we are hoaxed anywhere and only when we (can) measure ourselves we know for sure what the real outcome is (if we know how to measure of course). And at least I don't fool myself with it, plus it encourages for improvement again. This is on behalf of myself, and the customer at second place (yep) too.

 

Because you like to know: IIRC these measurements are taken at 118cm from the mouth opening, which is hardly really relevant for this. But now think what happens when the crossover is to be measured decently and how the height is totally crucial (woofers are at the bottom). This is a bit horn related of course, but I can help you out of all dreaming: this measurement is impossible because of the necessary distance, the sheer height of the woofer section itself (not aligned time arrival at the too short distance) and warbling and more. Oh, you can do it all right, if the norm would be +/-3dB (or even worse). And so this goes per section, and the remainder is theory (a crossover just has to match - period). Thus, the bass is set for its crossover (which is say 1000 measurements) and the mid is set separately from that. The connection can only be roughly checked with measurement (but usually does not make real sense because the whole FR looks like a mess because of the necessary distance - we talked about that).

 

Maybe the above was not all relevant - I just explain a little how I do it. It is no sinecure at all, which starts with a horn not being allowed to sound like a horn at all (take am empty horn, talk through it and listen to the most odd sound). The filter for the top horn (co-axial mid and high) took me a year of tuning. Actually two of us.

 

 

 

Thanks. I am okay with correction and I think it is a necessary evil. I always believed it was the conspiracy of the high end propagandists to do away with the most basic correction tool, i.e the treble and bass, which would make  most equipment to sound almost alike. Of course, that’s not healthy for high end market. 

 

I need more more time to digest your approach but on the surface I agree to your approach.   

[STC]

25 minutes ago, semente said:

 

 

In my view, using adjustable to FM/Contour curves compensation EQ settings allows the user to achieve flat FR at different listening/playback SPLs.

That is why some amplifiers used to have a Loudness/Countour circuit, which would boost frequency extremes to adjust to our lower sensitivity at these frequencies when listening at low SPLs.

Some speakers (with a fixed FR curve/balance) will only sound balanced (FR-wise) when playing at very loud SPLs (inverted smile curve) whilst others will sound balanced only at moderate levels (smile curve).

 

Bass and treble will do the job just fine. 

[PeterSt]

14 minutes ago, semente said:

That happened because Stereophile doesn't have an anechoic chamber and they use a "trick" to overcome this:

https://www.stereophile.com/content/measuring-loudspeakers-part-three-page-6

 

See ? :

 

 

... this already shows + and - 3dB and this is smoothed at 1/3 octave. You would be shocked when that was done on 1/12th as the peaks and dips would be ... can't tell. But possibly +/- 10dB (that's a 20dB difference).

Please notice that this is not about good or worse speakers; It merely is about the nonsenical thing to talk about Frequency Response when first everybody smooths out ... well ... the response. So this is not only about the +/- 3dB spec, but also about how the proof of that is again smoothed or not. So for example, if I claim a +/- 0.5dB under 100Hz, I should add "@ 1/12 smoothing". Otherwise it does not tell a thing.

 

In addition - and that is what the Stereophile article is about - one should tell how this is measured. Or better, people must understand that it makes no sense at all to measure in-room when this room is not anechoic (and smooth until the mess has finally disappeared). Otoh, explaining this all is beyond what people want to know or comprehend (I suppose). In my view measuring a loudspeaker is extremely difficult and comparison with measurements of other manufacturers / other speakers, is sheer impossible. Maybe I am wrong.

[PeterSt]

11 minutes ago, STC said:

I always believed it was the conspiracy of the high end propagandists to do away with the most basic correction tool, i.e the treble and bass, which would make  most equipment to sound almost alike.

 

Now I am not sure what you are heading for. But indeed the bass and treble knobs were banned from my system since - I forgot how long ago. Also, there will never be anything of an equalizer in my software. And why the FM curves exist in the speaker, I just explained.

 

Maybe you recall me saying (stating) that any analogue attenuation means is the most detrimental to sound (the most easy to measure). Bass and treble "knobs" are even worse. Mind you please, this is a technical thing - thus regarding measurement again. How it would net work out in anyone's room for the better, is something else. But as we know, that part I stay as far as possible from. What measures (ultimately) poorly, will never work out for the better, net. No matter is boosts bass where bass is seemingly needed. If that is the case, a few other things will be wrong, with "wrong" between quotes because it could be the inherent response of the speaker. Boost that (bass) and you'd have even more distortion again. Boosting treble I can imagine to some degree, but looking at the unlinearity over the whole spectrum, you really don't want to go for that. And then again remember the +/- 3dB story. If that is the basis, who cares (I am serious). All this stuff only starts to count when all what you deem possible is attacked for the better. I'm almost Frank.  But then without brains.

 

Maybe a small speaking example: If a loudspeaker is designed (build/tuned), the manufacturer will run into anomalies he likes to get rid of. Could be the bump just under 100Hz (I think @sementereferred to that). To some degree this can be attacked by fairly normal means (a notch filter) but since this bump exists because of e.g. a resonance to begin with, the result is a bit unpredictable when all is drawn out of balance because the bass is boosted (and the remainder is not). The manufacturer will never take this into account (he just can't predict what you want, in your room). And so the result will be worse (you may not recognize it per se, but worse it will be).

 

The even better example could be the molesting of the cross over(s). I wonder whether people even think about this;

When you boost the bass, your pre-amplifier really won't know anything about where your crossover to the mid resides. It will just boost up to where the preamp designer thought it was good. And so you molest your crossover without realizing it. You will perceive a different sound anyway (more bassey) - may like the more bass, and the downside (if recognizable) you take for granted. 

Ever thought about that ?

[semente]

On 7/13/2019 at 9:05 AM, STC said:

And how do the measurements made at “the fully developed” spot is useful to the one who believe that measurements of speakers determine the preferred sound? Are the listeners going to sit at the exact distance and turn his room to an anechoic chamber?  You do not know the variables to have any meaningful interpretation of the speakers measurements. 

 

There was a disputed between Harbeth and Stereophile when the measurements of Stereophile did not agree. Measurements varies as the standard/method is not fixed.

 

 I don’t even want to get why there need a specific distance for the sound to be fully developed or what happens outside the particular spot. Or why the speakers sensitivity is quoted at 1 meter distance if that’s not the spot where the sound is fully developed. 

 

In the end, whatever information you get from reading measurements of the speakers provides no constructive information that will be useful. If I am not mistaken, once Toole said that you get better information information from a tyre specs than the manufacturers speakers measurements.  Not to mention about AES papers on objective and subjective evaluation of loudspeakers to address how to get useful measurements to judge speakers correctly. 

 

I have stated my position and nothing further.

 

A single on-axis frequency response measurement is incapable of characterising the sound of a speaker. It will only tell you about the tonal balance on axis.

But if you add horizontal and vertical dispersion characteristics you can get a more significant idea of what the speaker will sound like in a room.

 

As you've mentioned, cabinet and driver resonances as well enclosure characteristics and how the speaker deals with floor-bounce will also affect sound quality. CSD and THD plots will provide information about the former and a room response plot about the latter. Using that information in conjuction with listening experiences of different topologies, enclosure types and driver materials etc. is very important to achieve a more accurate picture of the sonic character.

Other aspects like "clarity" at low SPL, distortion at high SPL and driver/crossover integration are a bit trickier to extrapolate from measurements, though CSD and THD plots may help here too.

[fas42]

2 hours ago, PeterSt said:

 

 

Frank, you can't participate. You're that guy who isn't even interested in all below 100Hz. Remember ?  Still I would agree with you. But spending a lot of money is not fun without some kind of guarantee it is "good". So you give in a little (if at all) and be more happy. At least that is the idea.

 

100Hz? Try 60, most speakers can do something reasonable down to this point - and most importantly, make music replay sound like the real thing. In turn, you should remember I have heard endless rigs that would be capable of rumbling rooms down to 20 Hz or so - but, oh dear!, they don't produce sounds that correlate that well with music from acoustic instruments ...

 

So, if I can choose between a unit specialising in bass effects, or, a setup that gets 99.9% of the sound picture right ... this is an easy one for me, .

[semente]

2 hours ago, PeterSt said:

 

See ? :

 

 

... this already shows + and - 3dB and this is smoothed at 1/3 octave. You would be shocked when that was done on 1/12th as the peaks and dips would be ... can't tell. But possibly +/- 10dB (that's a 20dB difference).

Please notice that this is not about good or worse speakers; It merely is about the nonsenical thing to talk about Frequency Response when first everybody smooths out ... well ... the response. So this is not only about the +/- 3dB spec, but also about how the proof of that is again smoothed or not. So for example, if I claim a +/- 0.5dB under 100Hz, I should add "@ 1/12 smoothing". Otherwise it does not tell a thing.

 

In addition - and that is what the Stereophile article is about - one should tell how this is measured. Or better, people must understand that it makes no sense at all to measure in-room when this room is not anechoic (and smooth until the mess has finally disappeared). Otoh, explaining this all is beyond what people want to know or comprehend (I suppose). In my view measuring a loudspeaker is extremely difficult and comparison with measurements of other manufacturers / other speakers, is sheer impossible. Maybe I am wrong.

 

JA no longer smoothes to 1/3 octave, he's been using 1/6 octave smoothing for many years.

 

For many years 1/3 octave was somewaht of a standard for in-room measurements, perhaps because it was close to how humans perceive tonal balance.

 

This is from the REW manual:

 

Psychoacoustic smoothing uses 1/3 octave below 100Hz, 1/6 octave above 1 kHz and varies from 1/3 octave to 1/6 octave between 100 Hz and 1 kHz. It also applies more weighting to peaks by using a cubic mean (cube root of the average of the cubed values) to produce a plot that more closely corresponds to the perceived frequency response.

[PeterSt]

1 minute ago, fas42 said:

this is an easy one for me, .

 

I agree. But music without under 100Hz (or under 60Hz) is not music. Try a kick drum with that limitation. Maybe it is good for Middle of The Road, which even is doubtful.

 

An other thing is that when you never heard "undistorted" music in that regions, you don't miss it either. And a promise: listening to it like I coincidentally do, requires a lot of re-learning (to listen). It is to-tal-ly different.

So again I must give you right, but actually only when you ever listening to something like I have - be that a million costing system or 200. It just is completely different and once used to it, there is no going back.

But would I rather cut off at 100Hz ? No. I'd pick up model railroading again.

[Summit]

The frequency response plot describes the range of frequencies the speaker can reproduce between 20 Hz and 20 kHz. The plot matches the sound signature of the speakers, as in when does the bass roll off, is there any peaks or big dips etc. Together with other measurements we can get some ideas about the sound signature and the overall tone balance. However the frequency response plot is a measurement of the amount/quantity at any given frequency and says nothing about the SQ of those frequencies.

[Abtr]

13 hours ago, Paul R said:

 

Turn that around - are you are saying that any tonal difference between speakers  (bright, soft, forward, reserved, any adjective you choose to use...) must be a result of differences in frequency response? 

 

While I would expect that to be mostly true,  would say it is far from true in all cases.

 

In other words, frequency response is not the only thing that defines the quality of what we hear. 

 

Do you disagree? 

 

In my experience audible tonality differences of speakers (e.g., resulting from crossover adjustments or cabinet changes) generally affect FR proportionately. The tonality change causes the FR change. I agree that in many cases the reverse is not possible, i.e., predicting the (tonal) character of a speaker based on its measured FR. Tonal character and FR are far from one-on-one related. Nevertheless, any consistent audible sound difference will be represented as a difference in FR.

 

Possibly some dynamic aspects of a loudspeaker or mechanical intermodulation effects are only minimally represented in a measured FR, but I can't imagine an audible difference that will completely elude FR. I might be wrong though.

[semente]

1 hour ago, Abtr said:

 

In my experience audible tonality differences of speakers (e.g., resulting from crossover adjustments or cabinet changes) generally affect FR proportionately. The tonality change causes the FR change. I agree that in many cases the reverse is not possible, i.e., predicting the (tonal) character of a speaker based on its measured FR. Tonal character and FR are far from one-on-one related. Nevertheless, any consistent audible sound difference will be represented as a difference in FR.

 

Possibly some dynamic aspects of a loudspeaker or mechanical intermodulation effects are only minimally represented in a measured FR, but I can't imagine an audible difference that will completely elude FR. I might be wrong though.

 

I disagree. Frequency response is directly related to tonal balance but as @Summit said it will only provide information about this particular aspect of the speaker's performance, it won't be enough and should not be expected to describe the global sound quality.

[Ralf11]

global ??

[fas42]

11 hours ago, PeterSt said:

 

I agree. But music without under 100Hz (or under 60Hz) is not music. Try a kick drum with that limitation. Maybe it is good for Middle of The Road, which even is doubtful.

 

Yes, a kick drum is a good one - but what most people don't realise is that the intense, in your guts, impact of that sound comes from the transients, and all the harmonics - I have heard vastly superior renditions of that instrument's qualities from my 'tiny', bookshelf speakers, than all the megalithic speakers I have run across in the decades. There's a whole suite of recorded music making that sounds quite pathetic on conventional rigs, irrespective of how impressive their FR curves look - key aspects of the accuracy of the sound are just wrong, and it takes mere seconds of listening, to dismiss them as being competitive.

 

11 hours ago, PeterSt said:

So again I must give you right, but actually only when you ever listening to something like I have - be that a million costing system or 200. It just is completely different and once used to it, there is no going back.

But would I rather cut off at 100Hz ? No. I'd pick up model railroading again.

 

No, we're forgetting 100Hz - remember, 60Hz is the dividing line ... .

[fas42]

8 hours ago, Summit said:

The frequency response plot describes the range of frequencies the speaker can reproduce between 20 Hz and 20 kHz. The plot matches the sound signature of the speakers, as in when does the bass roll off, is there any peaks or big dips etc. Together with other measurements we can get some ideas about the sound signature and the overall tone balance. However the frequency response plot is a measurement of the amount/quantity at any given frequency and says nothing about the SQ of those frequencies.

 

The best example of the silliness of thinking that FR says it all, that I've come across, was a rig of an enthusiast some townships away from me, that I heard on a number of occasions. Massively heavy, sealed twin subwoofers, best quality drivers, a DEQX unit that has repeatedly had calibration runs done on it, including by senior people from the DEQX factory; running a frequency sweep up and down, from 20 to 20,000Hz showed very clean response, with excellently low levels of obvious distortion in the very low bass - did it deliver, Big Sound? Ummm, no ... key recordings that I tried on it, that had subjectively very intense bass lines, failed to deliver - they were quite ho hum; fell in the category of "Why did you bother asking me to play this" ...

[Abtr]

9 hours ago, semente said:

I disagree. Frequency response is directly related to tonal balance but as @Summit said it will only provide information about this particular aspect of the speaker's performance, it won't be enough and should not be expected to describe the global sound quality.

Well, that is basically what I said. Can you give an example of an audible aspect of speaker performance that has no FR component at all (apart from max SPL)?

[PeterSt]

1 hour ago, Abtr said:

Can you give an example of an audible aspect of speaker performance that has no FR component at all (apart from max SPL)?

 

All what is distortion ?

(I am making this up, so maybe it is rubbish)

[PeterSt]

So suppose you sweep the speaker; suppose it shows relatively severe THD in the lower regions, like 30Hz showing at 60Hz as well (or even more than the 30Hz itself). 31Hz shows as 62Hz. 32 as 64. Etc., until the distortion is so low that the base tone overwhelms the 2nd order sufficiently and we would not regard it audible distortion (like 20Hz should be inaudible because no audible 2nd order 40Hz shows).

 

I never tried it, but unless you look at the sweep in real time, the result will just be a following of the swept frequencies and the distortion products will have been "overwritten" by the base tone in order (like 60Hz base will make the 60Hz 2nd of 30Hz invisible).

 

Wrong ?