• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Cheapest Full Range 20hz - 20khz Speakers?

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Note that the "Lokki" Griesinger quotes is Finnish acoustics researcher Tapio Lokki, rather than the step-brother and nemesis of Thor.

:D

My understanding is that one area of importance is the phase relationship between fundamentals and overtones. When the overtones arrive at the same instant as the fundamentals, there is a brief spike in SPL. When the overtones are smeared out in time, those spikes disappear.

The reason those spikes matter has to do with signal-to-noise ratio, which in turn has to do with clarity and intelligibility.

Woe is me to disagree with someone of Griesinger's status, but I'm aware of a lot of psychoacoustics literature that appears to contradict this position (if indeed I've understood correctly what his position is).

Of all the group delay audibility studies of which I'm aware, this is the study that found the lowest thresholds for group delay audibility. In this study, all-pass phase characteristics of numerous loudspeakers were convolved with impulses that were then played back to subjects over headphones. Subjects were tested to determine whether they were able to discern a difference between the convolved impulse and the un-convolved original (note that some low-pass filtering was applied to increase subjects' sensitivity - I won't go into the details here but it's in the linked paper).

Previous studies (e.g. the classic by Blauert) had already determined that subjects are far more sensitive to phase distortion when listening on headphones (vs. loudspeakers). Note also that previous studies had found that subjects are more sensitive to phase distortions when the stimuli are impulses than when the stimuli are music signals.

Keeping all that in mind, these graphs summarise the study's findings:

1605489035552.png


Even under highly contrived conditions (in comparison with normal loudspeaker listening and music), subjects were unable to detect phase differences of up to 1ms above 300Hz, and were unable to detect phase differences of well beyond 10ms across the entire audio band.

To put those results into perspective, a 4th order crossover at 500Hz or above will not exceed the threshold established in Graph (a). Even an 8th-order crossover at (for example) 2kHz will not exceed this threshold.

In other words, a typical loudspeaker is quite unlikely to produce enough phase distortion to be audible, even under the most contrived conditions (i.e. taking the speaker's phase characteristic, convolving it with an impulse, filtering it in such a way that sensitivity to phase distortion is increased, then playing it back through headphones).

Thus the idea that the group delay generated by normal loudspeaker crossovers could be a relevant factor in subjective sound quality does not seem plausible to me.

Anyway, I'm really not an expert on this (@j_j is who we would want to ask). But my understanding is that gradual shifts in phase such as those generally introduced by typical crossovers cannot result in audible group delay.

(IIUC, sharp changes in phase within a single critical band can be audible. This does not occur under normal circumstances in loudspeakers. It could occur with atypically high-order crossover filters, but we haven't seen that so far in the speakers measured.)
 
Last edited:

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
Woe is me to disagree with someone of Griesinger's status, but I'm aware of a lot of psychoacoustics literature that appears to contradict this position (if indeed I've understood correctly what his position is).

Thank you for your in-depth post.

I am well aware of the preponderance of evidence showing that phase shift of the order typical of loudspeakers is inaudible. Blaurt's book is on the shelf beside my desk, next to both copies of Toole's book. For many years I thought as you and almost everyone else does on this topic, and took comfort in the fact that the amount of phase shift in my passive crossovers had been proven to be inaudible.

Then I listened to Griesinger's lecture. And dangit, his argument seemed plausible. Rather than trying to mentally reconcile "dueling experts", I conducted an experiment. I introduced a relevant change into one of my crossover designs without saying anything to my beta-testers, and they heard it and described the difference. The old crossover may have measured better but the new one sounded better in a blind audition, and the improvement of the new one was in the phase domain.

Not that I expect anyone here to be convinced by my anecdote!

Thus far every time I have put Griesinger to the test via informal but controlled blind listening, his ideas have been validated, to me anyway.
 
Last edited:

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
Thank you for your in-depth post.

I am well aware of the preponderance of evidence showing that phase shift of the order typical of loudspeakers is inaudible. Blaurt's book is on the shelf beside my desk, next to both copies of Toole's book. For many years I thought as you and almost everyone else does on this topic, and took comfort in the fact that the amount of phase shift in my passive crossovers had been proven to be inaudible.

Then I listened to Griesinger's lecture. And dangit, and his argument seemed plausible. Rather than trying to mentally reconcile "dueling experts", I conducted an experiment. I introduced a relevant change into one of my crossover designs without saying anything to my beta-testers, and they heard it and described the difference. The old crossover may have measured better but the new one sounded better in a blind audition, and the improvement of the new one was in the phase domain.

Not that I expect anyone here to be convinced by my anecdote!

Thus far every time I have put Griesinger to the test via informal but controlled blind listening, his ideas have been validated, to me anyway.
What are the specifics of a relevant change? Have you considered designing and running a controlled test to validate your results? In your own words you mention a "preponderance of evidence" directly against your own anecdotal claim. In your opinion, what did they do wrong in the AES study? Are you able to repeat the testing done in the AES study with differrent results? If you are convinced of this wouldn't it be worth the effort to publish evidence to the contrary and improve everyone else's understanding of audio science?
 
Last edited:

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
What are the specifics of a relevant change?

Increased perception of hearing a live performance.

Have you considered setting up a controlled test and validating your results?

Actually the test was double-blind because no one present even knew that a change had been made. I was twelve hundred miles away.

The results are as valid as I need them to be for in-house research.

Not every question needs Harman-research-level investigation and "validation" in order to be answered. Even at Harman.

If you are convinced of this wouldn't it be worth the effort to publish evidence and improve everyone else's understanding of audio science?

If are willing to pay for my time and for all associated expenses, I'd be more than happy to.
 
Last edited:

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
Increased perception of hearing a live performance.
I was asking what change did you make to your crossover. I don't understand this answer.
Actually the test was double-blind because no one present even knew that a change had been made. I was twelve hundred miles away.

The results are as valid as I need them to be for in-house research.

Not every question needs Harman-research-level investigation in order to be answered. Even at Harman.

If are willing to pay for my time and for all associated expenses, I'd be more than happy to.
That's unfortunate. Truthfully I don't see the value in paying you to perform testing that has already been done and published by AES. I am implying that if you were benevolent it would be of great benefit to the audio science community. If however profit is a main driver for you I would imagine that it might be in your interests considering that, from perusing your website, you appear to be in the business of at least partially profiting off of this claim. I say that in seriousness, not to poke fun at you.
 
Last edited:

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
I was asking what change did you make to your crossover. I don't understand this answer.

Sorry, my mistake. I thought you were asking what change my beta-testers perceived, but I did use the words "relevant change" exactly as you described. My bad.

Briefly, I went from fourth-order acoustic slopes to slopes that started out first-order and then accelerated to fourth-order after about an octave.

I don't see the need in paying you to perform testing that has already been done and published by AES. I am implying that if you were benevolent it would be of great benefit to the audio science community.

Of course I don't really expect you to pay me!

I was trying to make the point that asking someone to produce a paper worthy of peer review and publication is asking a LOT. Costs include test equipment, a test facility, the time of those administering the test, the time of those taking the test, and the time of those writing the paper. For a small company like me, you're asking that I stop production and make my customers wait while I do the testing and research and write the paper that you asked for. So to you it's nothing, but to the person you're asking, it's a LOT.

If you were benevolent, you would have offered up front to pay me to do the research that you asked me to do. You see, two can play the passive-aggressive "if you were benevolent" game.

If however profit is a main driver for you I would imagine that it might be in your interests considering that, from perusing your website, you appear to be in the business of at least partially profiting off of this claim. I say that in seriousness, not to poke fun at you.

I profit off of the knowledge gained IF it results in a better product and IF that results in more sales.

And yes I am in the audio business. Most of my work is in prosound. And it is for profit; it's not a charity, though I have donated speakers as well as design work on occasion.

What business are you in? Are you doing it for profit?
 
Last edited:

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
Of course I don't really expect you to pay me!

I was trying to make the point that asking someone to produce a paper worthy of peer review and publication is asking a LOT. Costs include test equipment, a test facility, the time of those administering the test, the time of those taking the test, and the time of those writing the paper. For a small company like me, you're asking that I stop production and make my customers wait while I do the testing and research and write the paper that you asked for. So to you it's nothing, but to the person you're asking, it's a LOT.

If you were benevolent, you would have offered up front to pay me to do the research that you asked me to do. You see, two can play the passive-aggressive "if you were benevolent" game.
I didn't force you to put yourself out there firmly opposed to the "preponderance of evidence" against your anecdotal claims. I guess I shouldn't have tried to "passive-aggressively" ask you pointed questions to imply that your methodology isn't sound. So how about I just say it flat out then? I think that you're wrong. Evidence that you're clearly aware of again:
https://www.aes.org/e-lib/browse.cfm?elib=19404
I profit off of the knowledge that I gained if it results in a better product.

And yes I am in the audio business. Most of my work is in prosound. And it is for profit; it's not a charity, though I have donated speakers on occasion.

What business are you in? Are you doing it for profit?
I'm a degreed mechanical engineer and I currently work in furniture. Previously in medical device and aerospace. The only relevant thing about my business is that it doesn't profit from the audibility of group delay.

I guess we can leave it at an agreement to disagree, with the caveat that the existing evidence doesn't agree with you.
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
I guess I shouldn't have tried to "passive-aggressively" ask you pointed questions to imply that your methodology isn't sound. So how about I just say it flat out then? I think that you're wrong.

Thank you. I appreciate the courtesy of your telling me flat out that you think I'm wrong. This feels straightforward to me; the other felt like an attack on my character. I'd rather be wrong than be evil.

Evidence that you're clearly aware of again: https://www.aes.org/e-lib/browse.cfm?elib=19404

For the record, that paper is apparently about the audibility of group delay below 1 kHz, and Griesinger is talking about the audibility of phase (he doesn't use the term "group delay") north of 1 kHz.

I'm a degreed mechanical engineer and I currently work in furniture. Previously in medical device and aerospace. The only relevant thing about my business is that it doesn't profit from the audibility of group delay.

Well I was going to take a pot shot at you for working "for profit", but you took all the fun out of that. I wanted to be an aerospace engineer but kept flunking calculus, so I tip my virtual hat to you.

I guess we can leave it at an agreement to disagree...

That's the best I could have hoped for. Thank you.

... with the caveat that the existing evidence doesn't agree with you.

The preponderance of the existing evidence does not agree with me, but Griesinger might.
 
Last edited:

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
I agree about the limitations of the spin (see above). But I do think (other than the lack of IMD measurements) that ASR takes measurements beyond the spin that allow us to characterise the loudspeaker's performance reasonably well. Between the spin, the polar response measurements, and the distortion measurements, we get a very good (albeit not quite perfect) picture of the speaker's performance

You don't agree that measuring FR, HD, IMD, decay of individual drivers would provided any extra meaningful information? And measuring the port? Or the box (my BBC-style cabinets are quite resonant but they seem to measure flat on-axis)?
What about the speaker's step response?
What about an in-room response measurement?

But my understanding is that gradual shifts in phase such as those generally introduced by typical crossovers cannot result in audible group delay.

What is the possibility of such shifts combined with another albeit low-level type of distortion that also affects transients (cummulative effects) being audible?

You can back this up with current knowledge re audibility but to my ignorant eyes his idea that FR is enough looks overly simplistic.
I haven't read a lot of the literature so I'm using only a bit common-sense logic and anecdotal listening experience...

"Science" seems to indicate (very important words here) that high-Q resonances are inaudible yet the BBC Research Department research says otherwise, that if trigered by a fundamental with music programme they will be audible. So, not audible with pink-noise over a mono speaker but audible in real world conditions?

Maybe I'm just complicating things, maybe I'm fantasising and biased from all those audio magazine reviews (I didn't read), perhaps all that needs to be done is to slap any couple of drivers into a box, correct directivity and cabinet diffraction et voilà. As long as it Spins right it is right. Right?
Maybe people aren't taking the current knowledge too literaly or as the be-all and end-all, nor Toole bias for wide-directivity.
Kef are undoubtebly overingineering their Reference series because a simpler entry-level box will Spin identically and thus perform and sound identically.
(sorry for the sarcasm Andreas, you're actually quite sober-minded and don't deserve it; I really enjoy discussing these things with you and have learned a lot)


As a side note I don't find any usefulness in the preference ratings, they're but fancier star ratings.
Preference depends on many factors including listener experience, as shown by the albeit minuscule sample of listeners in Olive's target in-room curve study.
Some types or distortion are widely recognised as euphonic, and even some unpleasing distortions may sound pleasing to a few people (if what almost 15 years of observing a huge sample of audiophiles in webforums from several countries seems to indicate counts for anything).
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Thank you for your in-depth post.

I am well aware of the preponderance of evidence showing that phase shift of the order typical of loudspeakers is inaudible. Blaurt's book is on the shelf beside my desk, next to both copies of Toole's book. For many years I thought as you and almost everyone else does on this topic, and took comfort in the fact that the amount of phase shift in my passive crossovers had been proven to be inaudible.

Then I listened to Griesinger's lecture. And dangit, and his argument seemed plausible. Rather than trying to mentally reconcile "dueling experts", I conducted an experiment. I introduced a relevant change into one of my crossover designs without saying anything to my beta-testers, and they heard it and described the difference. The old crossover may have measured better but the new one sounded better in a blind audition, and the improvement of the new one was in the phase domain.

Not that I expect anyone here to be convinced by my anecdote!

Thus far every time I have put Griesinger to the test via informal but controlled blind listening, his ideas have been validated, to me anyway.

Well, kudos for sticking your neck out on this forum ;)

The first thing I'm curious about is the crossover topology itself. What you describe sounds pretty similar to a 4th-order Bessel. Is that in the ballpark?

Secondly, it seems obvious that phase was not your only variable. Switching from a 4th-order XO to something like the XO you describe is clearly going to produce side-effects in the amplitude response, both on-axis and off-axis, as well as changes in the profile of the nonlinear distortion (I know you're aware of all that ofc).

Thirdly, I understand you don't want to halt production to run a full suite of blind tests culminating in a peer-reviewed AES paper, haha. But maybe I could suggest a relatively straightforward sanity check?

What you could do is simply measure the phase response of the speaker with the original 4th-order XO in place, and then with the new crossover in place. Perhaps you already have these measurements on hand? Then, you could calculate the relative excess phase of the 4th-order XO and model an all-pass FIR filter based on it. Convolve some audio samples through that filter (tracks you're certain the difference is most audible on), and then compare (initially sighted, then if you hear a difference, under controlled conditions) the convolved audio with the unconvolved audio played back through the new (lower phase-distortion) speaker. If phase is the decisive factor, the excess phase should be quite audible. If no difference is discernible, then some factor other than phase must account for the perceived difference.

Anyway, I realise that probably isn't of much interest to you - you've designed a speaker that you like, and then done something that improves its sound: understanding the reason underlying it may not be a priority. But I guess that would be a relatively easy check to perform, if you have interest in doing so.

For the record, that paper is apparently about the audibility of group delay below 1 kHz, and Griesinger is talking about the audibility of phase (he doesn't use the term "group delay") north of 1 kHz.

That's true in terms of the study's stated goals, but if you look at the samples in Group A of the Finnish study (i.e. the samples in respect of which no audible difference was discerned), you'll notice that many of them exhibited levels of group delay above 1kHz that imply 360°+++ of phase shift. In fact, one of those group delay curves is up at almost 1ms at 3kHz, which implies >1000° of phase shift. And, as I know you know already, Blauert quite extensively investigated phase shift above 1kHz (as well as below it...) :)
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
That's unfortunate. Truthfully I don't see the value in paying you to perform testing that has already been done and published by AES.

So the testing has been done and dusted then. It can't be challenged. It can't possibly be wrong. Burn the witches.

Wait, that sounds like dogma, not science. Though I can understand the appeal. A problem solved is one less thing to be worried about... Let's tick a few boxes and then we can concentrate on the Spins.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
When pressed on these statements with the fact that FR and directivity have been shown in controlled studies to be much more important that group delay, reasonable levels of distortion, other audio magic; which by the way, are also generally accounted for in the measurements taken at ASR, there was just additional hand waiving.

I love hand-waving and audio magic. Maybe I am in the wrong forum, Audio Dogma Review is not the place for a herectic or a fool like miself.
The Sun still spins around the Earth I believe and thus...
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
you've designed a speaker that you like, and then done something that improves its sound: understanding the reason underlying it may not be a priority. But I guess that would be a relatively easy check to perform, if you have interest in doing so

This is a very good point.
It's actually what I try to do with listening assessments: find faults and then try to determine possible causes (easier said than done).
For that one needs measurements of as many parameters as possible and (I admit to my crime of ignorance) some ideia of the currently accepted audibility thresholds (just in case I've been "hearing" things).
Some idea of the mechanics as well as mechanisms which could cause particular audible artefacts is also important.
 
Last edited:

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
Hopefully I've cleared up any confusion regarding my interpretation. I agree that distortion, sensitivity, CSD should be looked at in addition to the spinorama when categorizing and comparing loudspeakers. I believe that is why we do it here.

There's hardly ever a CSD being shown unless one ask for it (apparently because Toole says it don't matter).
FR of individual drivers and port are also something that one requested (not shown in earlier reviews) and I don't thing that they're being measured independently (when bi- or tri-wiring is available to make this possible).
What about an in-room response measurement to cross-check with the estimated PIR, do you think it's worthless?
The step response won't tell you anything that you don't already know from a Spin and the Impedance plot?

What counts to you as distortion? HD and IMD probably, anything else?
 

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
There's hardly ever a CSD being shown unless one ask for it (apparently because Toole says it don't matter).
FR of individual drivers and port are also something that one requested (not shown in earlier reviews) and I don't thing that they're being measured independently (when bi- or tri-wiring is available to make this possible).
What about an in-room response measurement to cross-check with the estimated PIR, do you think it's worthless?
The step response won't tell you anything that you don't already know from a Spin and the Impedance plot?

What counts to you as distortion? HD and IMD probably, anything else?
See the problem here is that I'm asking you to prove to me what else is more important than what's already being measured. You're the one making the claims. You can't keep trying to flip it over to me to answer for you.

As far as I can tell Amir does make measurements of the individual drivers and port when possible. He seems to be pretty keen on checking how port noise does or doesn't affect the overall frequency response. Ask David Fabrikant if you don't believe me! So, what exactly are you complaining about? Also for my own knowledge, if you know the overall summed frequency response, directivity, sound power, and harmonic distortion through the audible frequency range what do the individual driver frequency responses tell you that would be of extreme importance? As for CSD, I think Amir has made it clear that's its hard to graph CSD in a way that adds much value due to the nature of the manipulation possible to make it look "good" or "bad". In the end it's usually posted anyway and it's another way of looking at the frequency response. In room response could be useful as a sanity check if you don't believe that the Klippel produces real anechoic results, I could see how that's something you would really push for. Yes, I don't think the step response will tell you anything you don't already know. Are you able to explain to me how it would? What other distortions do you think are more important than harmonic and possibly intermodulation? Do you know of any audibility studies for them? I'm genuinely curious.
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
You don't agree that measuring FR, HD, IMD, decay of individual drivers would provided any extra meaningful information?

I am in favour of measuring IMD and compression, but I see no reason (other than curiosity) to measure an individual driver as opposed to the whole system. My reasoning is this: We listen to speakers, not to individual drivers. What would individual driver measurements reveal that measurements of the loudspeaker as a system do not?
And measuring the port?

How would you suggest measuring the port exactly? What would you be trying to find out?
What is the possibility of such shifts combined with another albeit low-level type of distortion that also affects transients (cummulative effects) being audible?

The established audibility thresholds for typical time domain issues are quite high relative to distortions caused by typical speakers (as per my earlier post). This is actually an area that has been extensively researched, with reasonably consistent findings across the board.

A group delay curve or step response could, I guess, put potential issues in this department beyond doubt. It's also an interesting measurement to have for the technically-minded.
"Science" seems to indicate (very important words here) that high-Q resonances are inaudible yet the BBC Research Department research says otherwise, that if trigered by a fundamental with music programme they will be audible. So, not audible with pink-noise over a mono speaker but audible in real world conditions?

What "science" are you referring to that says high-Q resonances should be inaudible? That doesn't fit with my understanding of the science at all. The threshold of audibility should be higher for high-Q resonances than for low-Q resonances. Roughly speaking, the area under the curve of a resonance is a good indicator of its audibility. The narrower a resonance is, the higher its amplitude must be before it's likely to be audible.

Anyway, I don't understand why you're raising this point in the first place: The spin is an excellent tool for identifying both narrow- and wide-band resonances. And even the Olive preference score takes both wide-band resonances (SM) and narrow-band resonances (NBD) into account.
Maybe people aren't taking the current knowledge too literaly or as the be-all and end-all, nor Toole bias for wide-directivity.

I agree with this to an extent. I think I've said this a number of times in the thread and elsewhere. Firstly, I don't think the science is settled as to the relationship between preference and off-axis radiation, and secondly, I think the Olive preference rating tends to be accepted too uncritically.
Kef are undoubtebly overingineering their Reference series because a simpler entry-level box will Spin identically and thus perform and sound identically.
(sorry for the sarcasm Andreas, you're actually quite sober-minded and don't deserve it)

Are you certain that models in the Ref series don't "spin" better than the R series? All the measurements I've seen suggest that they do....
Some types or distortion are widely recognised as euphonic, and even some unpleasing distortions may sound pleasing to a few people (as almost 15 years of observing audiophiles in forums from several countries seems to indicate).

What loudspeaker distortions are widely recognised as euphonic? Where's the science to demonstrate that listeners prefer these?
 
Last edited:

vavan

Senior Member
Joined
Apr 29, 2019
Messages
341
Likes
212
Location
Kazan, Russia
"Science" seems to indicate (very important words here) that high-Q resonances are inaudible
From the book: "The probability of any high-Q resonance being heard is low because exact frequency matches with some component of the program are required. However, it does happen. Anyone who holds negative opinions about resonances in metal diaphragms has heard very high-Q, very-narrow-bandwidth effects. Superb metal diaphragms exist, and high-resolution measurements confirm their excellence"
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
The first thing I'm curious about is the crossover topology itself. What you describe sounds pretty similar to a 4th-order Bessel. Is that in the ballpark?

The topology and component values of my crossover are "what works".... what delivers my target response (or comes close enough). I have no idea how closely it resembles any named filter. At first glance the crossover board looks like a second order lowpass and a third order highpass, but there are other filters present which do some response shaping.

Secondly, it seems obvious that phase was not your only variable. Switching from a 4th-order XO to something like the XO you describe is clearly going to produce side-effects in the amplitude response, both on-axis and off-axis, as well as changes in the profile of the nonlinear distortion (I know you're aware of all that of).

Yes there were small changes in the frequency response on paper, which were not commented on by my beta-testers. The improvement my beta-testers noticed was consistent with Griesinger's description of "proximity", though they didn't use that word. Non-linear distortion was extremely unlikely to be an issue in this case (I use prosound drivers).

Anyway, I realise that [your proposed "sanity check"] probably isn't of much interest to you - you've designed a speaker that you like, and then done something that improves its sound: understanding the reason underlying it may not be a priority.

^^This.^^

I'm much more interested in other aspects of loudspeaker design, which (if/when the day comes) will almost certainly be vastly more controversial around here.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
I am in favour of measuring IMD and compression, but I see no reason (other than curiosity) to measure an individual driver as opposed to the whole system. My reasoning is this: We listen to speakers, not to individual drivers. What would individual driver measurements reveal that measurements of the loudspeaker as a system do not?

Stereophile measures the response of individual drivers and port, have you never found it useful?
Even here at ASR we have seen that can sometimes help clarify some issues.
For one because unless you are using 8th-order filter slopes there will be some to considerable overlaping which can mask resonances.
High-Q resonances may show in on- and of-axis measurements as a blip but the extent in the time domain is not quantified, for that you need a CSD (you can give me a Toole and I'll respond with a BBC).

We listen to DACs, what is the purpose of measuring the effectiveness of a filter or the amount of jitter or the noise-floor level "cleanliness"?
Surely S/N, FR, HD and IMD are enough.
Channel bleeding is not important because we will be using a single speaker for the listening assessment. :cool:

How would you suggest measuring the port exactly? What would you be trying to find out?

It is important, for me at least, to learn if the port tuning is over-, under- or critically-damped as this will have an impact on the audible quality of the sub-bass. It may be difficult to determine with a FR plot.
The port response is often a peak but sometimes a plateau.
Port often produce resonance peaks which may not be obvious in the FR plot or inversely it may prove difficult to locate the cause of a peak in the FR.
fr_listeningwindow.gif

1113PSBT2fig3.jpg

thd_90db.gif

https://www.stereophile.com/content/psb-imagine-t2-tower-loudspeaker-measurements
https://www.soundstagenetwork.com/i...&catid=77:loudspeaker-measurements&Itemid=153

What "science" are you referring to that says high-Q resonances should be inaudible? That doesn't fit with my understanding of the science at all. The threshold of audibility should be higher for high-Q resonances than for low-Q resonances. Roughly speaking, the area under the curve of a resonance is a good indicator of its audibility. The narrower a resonance is, the higher its amplitude must be before it becomes audible.

Anyway, I don't understand why you're raising this point in the first place: The spin is an excellent tool for identifying both narrow- and wide-band resonances. And even the Olive preference score takes both wide-band resonances (SM) and narrow-band resonances (NBD) into account.

Sorry, I didn't that mean to say that high-Q resonances were inaudible but less audible for identical level than low-Q resonances. I think I read it in Toole's "Toole-Modification of Timbre by Resonances".

The Spin helps you locate the resonances but it doesn't provide information on the time domain nor is it discriminating enough to in certain instances identify the cause.
This is why we've seen measurements here where it was necessary to produce a CSD, FR of the individual drivers, or go elsewhere for those and other measurements.

I understand Toole's concern regarding the inadequacy of CSD's performed at low res but he also raises the same issue in regard to FR measurements.

Are you certain that models in the Ref series don't "spin" better than the R series? All the measurements I've seen suggest that they do....

It was meant as a hyperbole. The F208 seems to Spin just as well as the F228Be.
You are missing the point about driver accuracy and measurements.

What loudspeaker distortions are widely recognised as euphonic? Where's the science to demonstrate that listeners prefer these?

For example the BBC dip. Most manufacturers use it, many of them top the sales charts among audiophiles of normal income.
Most buyers were not trained by Harman to identify deviations on FR response and just buy what they like best.

I've often seen at forums people say that they sold their Benchmark DACs, which they boght on hype, because they didn't like the sound. Do you (really) think that they were mislead by not listening under controlled conditions?

If the research were right no one would prefer narrow-dispersion speakers like horns or dipole panels or even D&Ds. Unless they think that they prefer it but they don't...
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Stereophile measures the response of individual drivers and port, have you never found it useful.
Even here at ASR we have seen that can sometimes help clarify some issues.

I certainly find such measurements interesting, as I like thinking about the decisions the designer has made and whether I would have done things differently.

However, I don't believe they are necessary to characterise a speaker's performance from the point of view of the end-listener. ASR measurements give us extremely accurate 360° frequency response data, distortion data down to below the noise floor of the room, etc. If only we also had IMD and compression (aka linearity) data, that would be close enough to the full picture IMO.

High-Q resonances may show in on- and of-axis measurements as a blip but the extent in the time domain is not quantified, for that you need a CSD (you can give me a Toole and I'll respond with a BBC).

Haha, I think I know which BBC paper you're talking about, but just to make sure we're on the same page, could you tell me which it is, please?

We listen to DACs, what is the purpose of measuring the effectiveness of a filter or the amount of jitter or the noise-floor level "cleanliness"?
Surely S/N, FR, HD and IMD are enough.
Channel bleeding is not important because we will be using a single speaker for the listening assessment. :cool:

I barely look at DAC measurements as I don't believe DACs affect the sound unless there is something terribly wrong with them. I certainly don't look at the jitter measurements ;)

I believe this is consistent with my pragmatic attitude towards speaker measurement (for the purposes of the end-listener): No need to measure what is not audible.

It is important, for me at least, to learn if the port tuning is over-, under- or critically-damped as this will have an impact on the audible quality of the sub-bass. It may be difficult to determine with a FR plot.
The port response is often a peak but sometimes a plateau.
Port often produce resonance peaks which may not be obvious in the FR plot or inversely it may prove difficult to locate the cause of a peak in the FR.

This point I completely disagree with you on ;)

The impedance and frequency response measurements fully characterise the behaviour of the port (within its linear range of operation - to go beyond that, we need compression testing). Assuming that peak in the Soundtsage! measurement is not a measurement artefact, we see the problem in the frequency response, anyway. Why would we (assuming we're not the designer) need to know what was causing it? A peak in the FR is a peak in the FR... right?

Re: your comments about damping, how will this impact the audible quality of the low-bass, other than by affecting its frequency response (which we see in the FR graphs)?

The Spin helps you locate the resonances but it doesn't provide information on the time domain nor is it discriminating enough to .
This is why we've seen measurements here where it was necessary to produce a CSD, FR of the individual drivers, or go elsewhere for those and other measurements.

Do you have an example of such an issue that you could share so I better understand what you're talking about? And do you have any evidence that such an issue could be audible even when there is no measurable anomaly in the amplitude response?

It was meant as a hyperbole. The F208 seems to Spin just as well as the F228Be.

Actually, bass distortion aside, my opinion is that the F208 is a better speaker than the F328Be :eek:

In general, based on the measurements I've seen, the Be series appears to be a slight step backwards from the original Performa series, which I suspect is mostly a marketing-driven exercise. But that's another story...

You are missing the point about driver accuracy and measurements.

Sorry, yes, apparently I am :) Would you mind explaining it?

For example the BBC dip. Most manufacturers use it, many of them top the sales charts among audiophiles of normal income.
Most buyers were not trained by Harman to identify deviations on FR response and just buy what they like best.

Ok, but let's say we consider a BBC-style dip to be desirable. The spin is the perfect set of data to tell us whether and to what extent it's there, is it not?

I've often seen at forums people say that they sold their Benchmark DACs, which they boght on hype, because they didn't like the sound. Do you (really) think that they were mislead by not listening under controlled conditions?

My guess is that they expected the DAC to have a special sound and then they listened to it and it sounded like every other DAC they'd heard, which they interpreted as sounding cold or boring. Or any of numerous other psychological factors that could be influencing their perceptions at the time...

If the research were right no one would prefer narrow-sispersion speakers like horns or dipole panels or even D&Ds. Unless they think that they prefer it but they don't...

I mostly agree. I don't think there's any direct research showing that people prefer wider-dispersion speakers. However, I acknowledge that there is a lot of indirect evidence that many or even most people tend to prefer them. It's certainly Toole's informed opinion based on a lot of different data, and it's a reasonable opinion for him to have. But I can see good arguments against this position in the data, too.
 
Top Bottom