Speaker Equivalent SINAD Discussion

[JustIntonation]

I share your intuition - as speaker designers we have to use whatever measurements we have available and just shoot for the lowest possible THD. The idea that higher orders are less desirable than lower is somewhat sensible as well but the fact remains that THD does not correlate to sound quality! We use it as a guide to speaker design excellence, but I wouldn't use it to rank speakers in isolation. Like impedance response, it's mostly a graph which allows designers to see what's going wrong rather than a graph which allows us to evaluate a design.

Yes, I think it's the combination of these graphs / measurements including the room which determine soundquality. I don't think there's anything which can't be measured that affects soundquality. No magic. But knowledge on how to read measurements and their audibility which isn't that simple (and in my opinion includes knowledge of psychoacoustics and things like HRTF).

 

[MediumRare]

Yes, I think it's the combination of these graphs / measurements including the room which determine soundquality. I don't think there's anything which can't be measured that affects soundquality. No magic. But knowledge on how to read measurements and their audibility which isn't that simple (and in my opinion includes knowledge of psychoacoustics and things like HRTF).

This is very interesting. Your prior comment abut Toole perhaps reflects that we are an inquisitive and data-driven bunch, but can't individually be experts at a professional level in many different fields. So Toole does seem to show us one path, perhaps not the only path or a complete path. But, as you and some others show, as a group we include many individual experts. So, please help us understand further: what measurements CAN indicate audible SQ issues that are relevant? For example, how would clipping, blurriness, or that classic, the "veil", be reported in measurements?

 

[JustIntonation]

This is very interesting. Your prior comment abut Toole perhaps reflects that we are an inquisitive and data-driven bunch, but can't individually be experts at a professional level in many different fields. So Toole does seem to show us one path, perhaps not the only path or a complete path. But, as you and some others show, as a group we include many individual experts. So, please help us understand further: what measurements CAN indicate audible SQ issues that are relevant? For example, how would clipping, blurriness, or that classic, the "veil", be reported in measurements?

In my experience by far the most important metric is frequency response at the listening position. I'll repeat that: at the listening position so very including room effects! The next one is dynamic range, so a silent listening position (no computer fan etc) and enough loudness without high HD/IMD. A high true dynamic range is only realizable in a very dead room I think as reverb and reflections reduce short term dynamic range.
Then IMD which becomes quite audible when above conditions are met (especially in direct comparison with a speaker with lower IMD, it'll give you an "ahaa" moment).
And when all the above is done well enough you'll find HRTF becomes important and stereo is insufficient of a format for perfection. You'll find that the treble becomes too sharp under near anechoic conditions for instance. To remedy this do not place your speakers further apart than the equilateral triangle and furthermore put an eq slope on the treble to better match the average frequency response to the average frequency response of the ears if the sound were really to come from the center otherwise there's a mismatch between the phantom center and the frequency response of the ear which it gets from 30degrees left and right where the speakers actually are.
Overall I personally hold a very well treated big room as more important than the speaker. It is truly incredible how much one can hear and how real well recorded music can get with a truly good room and setup, even with half decent speakers, easily beating the most expensive headphones in detail and reality etc (which will never be audiophile in my opion due to single driver treble problems, bass feel problems and even moreso lack of natural HRTF). I think audiophiles should focus on treating their room first with 20cm thick very large absorbtion panels + large deep bass traps (for most self built for lwo costs) and only then focus on speakers + amps. If on the other hand one has very expensive speakers like Wilson or Revel or whatever (I used to have 20.000 euro speakers too) even in a big and considered "good" sounding room but not extensively treated, this won't hold a candle to a much cheaper speaker in a truly well treated room, not even close.

 

[Rick Sykora]

FYI I heard back from Sean Olive. Good news is that we can use his formula without worrying about licensing the patent. Bad news is that no one computed the value inside Harman so he has no code to give us.

I am starting to warm up to the idea of three or four scores for each speaker. One could be difficulty of drive per above. Another could be low frequency extension/power. Another smoothness of on-axis and another, off-axis. We could just give it scores A to F so that people don't zoom in too much on what they mean.

Am I correct that these scores would apply to any speaker tested regardless of what application space it is designed to address? So, actives, passives, consumer, pro, monopoles, dipoles, etc. (and all major permutations), would all be lumped together? This might be useful at some level, but with a large database, probably overwhelming.

Just like price, am hoping there would be some filtering based on intended application as well?

 

[MZKM]

Am I correct that these scores would apply to any speaker tested regardless of what application space it is designed to address? So, actives, passives, consumer, pro, monopoles, dipoles, etc. (and all major permutations), would all be lumped together? This might be useful at some level, but with a large database, probably overwhelming.

Just like price, am hoping there would be some filtering based on intended application as well?

I can't make filters i a published Google Sheet, but I can make sheets for select groups (e.g., only passives, only <$500, etc.)

And yes, dipoles can and have been scored by others using this algorithm. I probably wouldn't give it much weight, instead mainly focusing on the in-room response.

 

[Rick Sykora]

Between Amir suggesting multiple scores (some that may not apply to a particular speaker type) and others discussing the caveats of too much focus on a single score, it reminded about the need to focus on the consumer.

While having a thoughtful score is helpful, generating one that may be meaningful for an engineering team is not always useful for others (that may not understand the context or the science behind the metric(s)). Clearly the benefit of discussing in the larger forum here is to bring forward other potential solutions. While they are all speakers, it seems unfair to endorse a scoring system without proper context. If I built a stage speaker with hifi parts (or vice versa), the outcome would be wrong. One could say they are all speakers, but that is a bit like saying a taxi and a dragster are both cars!

So, was just suggesting that there is value in considering who and how the scoring might be applied. Thanks for encouraging the discussion.

 

[MZKM]

Between Amir suggesting multiple scores (some that may not apply to a particular speaker type) and others discussing the caveats of too much focus on a single score, it reminded about the need to focus on the consumer.

While having a thoughtful score is helpful, generating one that may be meaningful for an engineering team is not always useful for others (that may not understand the context or the science behind the metric(s)). Clearly the benefit of discussing in the larger forum here is to bring forward other potential solutions. While they are all speakers, it seems unfair to endorse a scoring system without proper context. If I built a stage speaker with hifi parts (or vice versa), the outcome would be wrong. One could say they are all speakers, but that is a bit like saying a taxi and a dragster are both cars!

So, was just suggesting that there is value in considering who and how the scoring might be applied. Thanks for encouraging the discussion.

Which is why I also include sensitivity specs under the SPL Specs tab. Sadly, as not wanting to risk damaging the speakers, Amir wont't be testing max SPL.
I have asked before, but I wish for the minimum distance for far-field conditions to be included in the review, this will let us know which speakers are not that suited for near-field usage.

These types of caveats are included in the Notes section of the database.

 

[jhaider]

In my experience by far the most important metric is frequency response at the listening position. I'll repeat that: at the listening position so very including room effects!

How do you reconcile that position with studies indicating listener perception of spectral balance in the statistical region track the anechoic data rather than the in room response?

In the modal and transition regions, no argument.

I think audiophiles should focus on treating their room first with 20cm thick very large absorbtion panels + large deep bass traps (for most self built for lwo costs) and only then focus on speakers + amps.

Preference is preference, but...yuck. Both in terms of looks and sound.

 

[amirm]

Am I correct that these scores would apply to any speaker tested regardless of what application space it is designed to address? So, actives, passives, consumer, pro, monopoles, dipoles, etc. (and all major permutations), would all be lumped together? This might be useful at some level, but with a large database, probably overwhelming.

Just like price, am hoping there would be some filtering based on intended application as well?

I don't think there were any powered speakers in there. Or Pro/nearfield monitors. There was only one dipole (Martin Logan).

 

[amirm]

From experience I can tell you that harmonic distortion and IMD meaurements do indeed not correlate substantially to subjective experience in normal listening rooms except in extreme cases, BUT.. they DO correlate to subjective experience in very well treated rooms / studios / anechoic rooms!
The thing is, room reflections and reverb mask most low level detail and HD and IMD up to moderate levels are easily burried in the much great linear and time distortions caused by room reflections / modes and reverb.

Research data disputes this. For one thing, i can tell you Pros are just as bad as audiophiles in detecting non-linear distortions. In one public blind test, one famous movie soundtrack mixer catastrophically failed a lossy compression test, saying two identical tracks sounded different! I was I think one of two people who found out the error in the test that resulted in identical files. So if there is an implication in your post that pros have better ears when it comes to distortion, they do not.

On the rest of the point, a very careful study was performed recently with headphones. Once again, frequency response differences dominated and distortion except for extreme level in one sample I think, made no difference. Headphones naturally remove the room, and provide far higher acuity for detecting non-linear distortions.

It is a simple thing really: in presence of frequency response differences, distortion has no prayer of overriding those differences.

Now, if one is designing a speaker, they should make it as low a distortion possible as in that case, it is the same frequency response more or less.

 

[amirm]

I have asked before, but I wish for the minimum distance for far-field conditions to be included in the review, this will let us know which speakers are not that suited for near-field usage.

That number is frequency dependent. The lower the frequency, the longer the distance. What do you want it at?

 

[JustIntonation]

How do you reconcile that position with studies indicating listener perception of spectral balance in the statistical region track the anechoic data rather than the in room response?

In the modal and transition regions, no argument.

I don't think that's true in detail. Room reflections do alter perception of spectral balance even with speakers with a completely smooth off-axis.
It's simply not true that we can seperate the direct sound from the reflected sound to a great degree (though one can argue what defines "a great degree" here). And even if we were able to seperate the spectral balance of direct vs reflected sound (which is not the case I believe) then still there is the matter of rooms completely changing the balance of transients vs sustained part of sound by smearing the transients in time through reflections and reverb which alone completely changes the balance of most music and which is completely room dependent.
Btw, I've read plenty of studies in the past and while I'm not an expert and have always taken such studies with a big grain of salt as to the conditions (most) such studies take place etc I do believe the studies you're referencing to indicated that the direct sound is the most important factor, not the only one.

Preference is preference, but...yuck. Both in terms of looks and sound.

I don't care about looks much, though I think my own panels are pretty I've attached a picture. (panels 245x125x20cm they cost only 70 euro a piece to make + bigger deeper bass traps, speaker in the picture is the Kali LP-6 which I'm using temporarily while I build my own high-end speakers.)
As far as sound goes, I'm not sure you've heard the kind of sound I'm referring to. It's the difference between listening "inside" the recording and listening to a recording "in your own room". I greatly prefer the first and while some people think it would be somewhat like listening to headphones it is not comparable (and I hate headphones, have tried to like them several times but I just can't).

 

[JustIntonation]

Research data disputes this. For one thing, i can tell you Pros are just as bad as audiophiles in detecting non-linear distortions. In one public blind test, one famous movie soundtrack mixer catastrophically failed a lossy compression test, saying two identical tracks sounded different! I was I think one of two people who found out the error in the test that resulted in identical files. So if there is an implication in your post that pros have better ears when it comes to distortion, they do not.

On the rest of the point, a very careful study was performed recently with headphones. Once again, frequency response differences dominated and distortion except for extreme level in one sample I think, made no difference. Headphones naturally remove the room, and provide far higher acuity for detecting non-linear distortions.

It is a simple thing really: in presence of frequency response differences, distortion has no prayer of overriding those differences.

Now, if one is designing a speaker, they should make it as low a distortion possible as in that case, it is the same frequency response more or less.

I wasn't talking about the ears of pro's vs the general public. (though I think trained ears/brains that know what to focus on when listening are better at such things). I was talking about the difference in audibility of nonlinear distortion when listening in (near) anechoic conditions vs listening in a normal listening room. This difference is huge I fully stand behind my statement.

As for headphone listening. I think headphones are very poor for listening to such things. The main reason why is that they lack natural HRTF (and a few other very bad quirks) and simply won't give you a natural transparent view into the music which causes in my opinion nonlinear distortions to be nowhere near as audible as with good speakers properly set up in (near) anechoic conditions. Again, this is not a small difference. Headphones are just a very bad comparison to speakers without a room.

I DO agree with you statement that frequency response is king! And that non-linear distortion does not override the importance of frequency response (unless extreme).
I however do not agree with views that comb filter effect of room reflections and room reverb are somehow even preferrable to (near) anechoic conditions. I think room reflections, modes and reverb are incredibly detrimental to sound quality and ability to detect non-linear distortions and ability to be transported to the acoustics of the recording, etc etc. And frankly I can't believe that Toole's setup and research on this particular subject (while obviously doing great work in other areas) are taken by many as some kind of reference while most of todays pro's clearly have vastly superior setups and rooms (measureably so unless one uses in my opinion selective half assed bad "research" to argue away the massive nonlinear distortions created by rooms at listening position).

 

[Blumlein 88]

I don't think that's true in detail. Room reflections do alter perception of spectral balance even with speakers with a completely smooth off-axis.
It's simply not true that we can seperate the direct sound from the reflected sound to a great degree (though one can argue what defines "a great degree" here). And even if we were able to seperate the spectral balance of direct vs reflected sound (which is not the case I believe) then still there is the matter of rooms completely changing the balance of transients vs sustained part of sound by smearing the transients in time through reflections and reverb which alone completely changes the balance of most music and which is completely room dependent.
Btw, I've read plenty of studies in the past and while I'm not an expert and have always taken such studies with a big grain of salt as to the conditions (most) such studies take place etc I do believe the studies you're referencing to indicated that the direct sound is the most important factor, not the only one.


I don't care about looks much, though I think my own panels are pretty I've attached a picture. (panels 245x125x20cm they cost only 70 euro a piece to make + bigger deeper bass traps, speaker in the picture is the Kali LP-6 which I'm using temporarily while I build my own high-end speakers.)
As far as sound goes, I'm not sure you've heard the kind of sound I'm referring to. It's the difference between listening "inside" the recording and listening to a recording "in your own room". I greatly prefer the first and while some people think it would be somewhat like listening to headphones it is not comparable (and I hate headphones, have tried to like them several times but I just can't).

It would seem you are attempting to approach listening to speakers outdoors with no reflections in one sense. Is that listening inside the recording?

 

[JustIntonation]

It would seem you are attempting to approach listening to speakers outdoors with no reflections in one sense. Is that listening inside the recording?

Yes. Except that outdoors one almost always has a very high noisefloor.
Look at for instance the rooms made by www.northwardacoustics.com
Realize that it has very thick (often 1m+) absorption behind fabric in ceiling and all walls except the front wall which acts as infinite baffle to the speakers and does not cause reflections. The few diffusion panels you see are there for added "creature comfort" from listening position and do not give audibly significant reflections from sound coming from the speakers.
Many if not most current studio designers use a similar approach (and otherwise something half in that direction according to older designs like LEDE).
Those pros with enough money and having compared different designs by actually visiting many mastering studios etc most often prefer these near-anechoic rooms and their immense soundquality.
I think it would be correct and very fitting for ASR to push "room, room, room!" before anything else as it's the most important factor for soundquality, only after that you get speakers, amps and DACs. This is not just my personal opinion, go to for instance gearslutz.com and virtually all experienced pros will tell you this exactly even for nearfield monitoring.

 

[Blumlein 88]

Yes. Except that outdoors one almost always has a very high noisefloor.
Look at for instance the rooms made by www.northwardacoustics.com
Realize that it has very thick (often 1m+) absorption behind fabric in ceiling and all walls except the front wall which acts as infinite baffle to the speakers and does not cause reflections. The few diffusion panels you see are there for added "creature comfort" from listening position and do not give audibly significant reflections from sound coming from the speakers.
Many if not most current studio designers use a similar approach (and otherwise something half in that direction according to older designs like LEDE).
Those pros with enough money and having compared different designs by actually visiting many mastering studios etc most often prefer these near-anechoic rooms and their immense soundquality.
I think it would be correct and very fitting for ASR to push "room, room, room!" before anything else as it's the most important factor for soundquality, only after that you get speakers, amps and DACs. This is not just my personal opinion, go to for instance gearslutz.com and virtually all experienced pros will tell you this exactly even for nearfield monitoring.

Maybe, maybe not. Some of the Harman findings are that with good speaker metrics the room is less important. We undoubtedly do listen thru the room to a very significant amount. Not totally. But it is quite significant. So should we push speakers, speakers, speakers on speaker quality or rooms, rooms, rooms on good room quality? I'm currently in the speaker camp. The room has some effect for sure, and one can find rooms that just eliminate the chance for any speaker to sound good at all.

Domestic rooms.........should it be great speaker and a little room help or does one get a better result with lesser speaker and heroic room help? It is a very interesting question. Of course everyone will say go way out on both.

I've certainly not hidden my opinion transducers are where money should go as everything between is pretty much a solved problem. I wouldn't argue with room treatment, but still don't feel it goes before quality speakers in terms of where to spend money.

I wouldn't mind if ASR started having some measurements of effects of various room treatments. It is somewhere one can make their own with not too much money and have it fit their customized needs. But are such things highly effective or not?

 

[JustIntonation]

Research data disputes this. For one thing, i can tell you Pros are just as bad as audiophiles in detecting non-linear distortions. In one public blind test, one famous movie soundtrack mixer catastrophically failed a lossy compression test, saying two identical tracks sounded different! I was I think one of two people who found out the error in the test that resulted in identical files. So if there is an implication in your post that pros have better ears when it comes to distortion, they do not.

On the rest of the point, a very careful study was performed recently with headphones. Once again, frequency response differences dominated and distortion except for extreme level in one sample I think, made no difference. Headphones naturally remove the room, and provide far higher acuity for detecting non-linear distortions.

It is a simple thing really: in presence of frequency response differences, distortion has no prayer of overriding those differences.

Now, if one is designing a speaker, they should make it as low a distortion possible as in that case, it is the same frequency response more or less.

Btw, I though of another thing to make my point.
With the JBL LSR305mkII if you play for instance a 1kHz sine and at the same time a sine sweep from for instance 500Hz to 2kHz, you'll clearly hear the IMD as sum and difference tones which very audibly sweep counter and parallel to the main sine sweep. Even at low SPL. (guess it's mostly caused by the cheap class-D amps). Same goes for the Kali LP-6. You can hear this in a normal room too.
When however listening to these speakers in a normal room with music you'd be hard pressed to put your finger on the effect of this massive amount of IMD. If you compare directly with a low IMD speaker switching A/B you'll probably be able to hear the difference if you know what to listen for.
In a (near) anechoic room however there's no sublety about this IMD, it'll hit you in the face with a sledgehammer you'll hear it destroy medium and low level detail, transparancy and put a big blanket of distortion over everything. Much much moreso than any headphone simulation will show you (with headphones you can't hear through the massive HRTF related linear distortions + single driver breakup linear distortions in the treble).

 

[JustIntonation]

Maybe, maybe not. Some of the Harman findings are that with good speaker metrics the room is less important. We undoubtedly do listen thru the room to a very significant amount. Not totally. But it is quite significant. So should we push speakers, speakers, speakers on speaker quality or rooms, rooms, rooms on good room quality? I'm currently in the speaker camp. The room has some effect for sure, and one can find rooms that just eliminate the chance for any speaker to sound good at all.

Domestic rooms.........should it be great speaker and a little room help or does one get a better result with lesser speaker and heroic room help? It is a very interesting question. Of course everyone will say go way out on both.

I've certainly not hidden my opinion transducers are where money should go as everything between is pretty much a solved problem. I wouldn't argue with room treatment, but still don't feel it goes before quality speakers in terms of where to spend money.

I wouldn't mind if ASR started having some measurements of effects of various room treatments. It is somewhere one can make their own with not too much money and have it fit their customized needs. But are such things highly effective or not?

Yes, agreed that good speakers make the room less important (by radiating the room more evenly). But even when less important than with a bad speaker the room is in my opinion still by far the most important.
I mean, in a normal room no matter the speaker you will not get the realistic impression that sound can be so incredibly tight and fast and close it actually gives the impression like it's located at the skin of your face while other sounds can give a realistic impression of being a hundred meters far in the distance, etc etc.
It's really something to experience, great music in a (near) anechoic environment with great speakers set up properly for that environment (including treble HRTF correction).
Downside is that you'll hear all imperfections in the music / recordings.

As for domestic rooms, I'm definately in the heroic room help category combined with a measurement mic and speaker/room correction.
If you look at commercial room treatment prices it's very expensive for most often products that don't cut it (foam or too shallow absorption panels etc). But if you build it yourself it can be done cheap. I'll be happy to start a thread on it with the science of absorption and practice of making cheap panels. Often a 1000 euro or so and some time and effort are all it takes if you already have a fairly good dedicated room (or are willing to put many big panels in your livingroom like me )
And yes, well made panels are highly effective and make a huge audible difference. And I mean truly rediculously huge audible difference.

 

[MZKM]

That number is frequency dependent. The lower the frequency, the longer the distance. What do you want it at?

Ah, ok; I asked about that here, as I wasn’t fully comprehending it (I do see it in the text you posted, guess I didn’t see that).

Why does it vary with frequency though? If it’s not in the crossover region, wouldn’t 500Hz or 5kHz result in the same distance regardless of speaker? I would have assumed the crossover region/frequency is where the minimum distance varies, not all frequencies. If that is truly the case, then I don’t have a suggestion for how to include it.

 

[JustIntonation]

Ah, ok; I asked about that here, as I wasn’t fully comprehending it (I do see it in the text you posted, guess I didn’t see that).

Why does it vary with frequency though? If it’s not in the crossover region, wouldn’t 500Hz or 5kHz result in the same distance regardless of speaker? I would have assumed the crossover region/frequency is where the minimum distance varies, not all frequencies. If that is truly the case, then I don’t have a suggestion for how to include it.

I must say I don't understand it either.
There are a few things related to distance. The first is baffle size, if you get close to the speaker the baffle step will be reduced which influences frequency response. The question is how much in dB will you allow baffle step reduction when setting minimal distance? Since its a broad bump in often most or all of the bass the audible difference starts at 0.5dB or so or less.
The second may be if one takes the tweeter axis as the on-axis then middriver-tweeter distance matters and their crossover freq too as the angle changes as we get closer so does the crossover sum (and angle to the mid driver). Again, at which amount of effect of this in dB on the fr is minimal distance?

One other general rule though not related to any specific speaker is that as you get really close the crossfeed between ears changes in volume as there'll be more reduction because sound drops at 6dB in doubling distance. Our brains can detect this. I believe we start speaking of "far field" in this sense when the soundsource is >1m away (don't quote me on the exact distance, from memory and not sure, could be that it was 60cm..).