The “Real” Speaker Preference Test

[Arindal]

When I have to record and mix live concerts which incl. large orchestra, rhythm section, vocals - one does what one must in order to get a cohesive end product, and usually this involves more than a few HP filters (among others).

So HP is necessary for the FOH mix solely in your case, and parallel multitrack recording for later mixdown is done without HP in the chain?

Other times, sure, there can be none - it's not something the consumer can know for sure.

No-one needs to know for sure what was in the recording chain. But the possibility that at least a decent number of recordings where made without HP (not so speak of artificially created sounds) and excess GD/long decay in bass bands might be audible, in my understanding is reason enough to keep potential reproduction errors low, and listen to the outcome.

There is one concert I did which for some reason on youtube has the option for "vocal presence" which is ON unless you turn it off.

If you want to listen to mixes untouched, use Qobuz or Apple Music for streaming. It is as easy as that.

 

[TimVG]

So HP is necessary for the FOH mix solely in your case, and parallel multitrack recording for later mixdown is done without HP in the chain?

No, this is not for the live mix, this is for the recording only. There are shared mics, but also mics used for the recording only.

No-one needs to know for sure what was in the recording chain. But the possibility that at least a decent number of recordings where made without HP (not so speak of artificially created sounds) and excess GD/long decay in bass bands might be audible, in my understanding is reason enough to keep potential reproduction errors low, and listen to the outcome.

As I said, one needs not excuse the other. In a small mixing room the multisub approach has served me well as amplitude can be evened out over a relatively wide area and it seems to translate well to a variety of circumstances. I have not noticed any timing issues, but I will admit I have never done my very best to detect them as well.

However, I am by no means dogmatic, and if there are even better ways, and if there are no detrimental effects to the created material, I am all ears.

If you want to listen to mixes untouched, use Qobuz or Apple Music for streaming. It is as easy as that.

We do not decide on which platform things go.

 

[kimmosto]

I am pretty cautious when it comes down to predicting an outcome or clear audibility thresholds.

I'm well aware that audibility and significance of phase distortion are very very low with normal 2-way speakers without sub. Ignorance of timing on ASR's reviews does not usually matter in practice. But all that changes with 4- and 3-ways with steep crossover slopes. Typical IIR XO for subwoofer damages timing and transient dynamics really badly. I have tested that few times with headphones only. Difference with continuous transient signal with low fundamental is radical even without haptic feedback. Haptic is quite totally missing with headphones without very high volume causing hits to skin around the ears.
So missing timing error as normal and excess GD (measured vs. minimum-phase) is mostly just theoretical flaw in the reviews, but it could turn to actual flaw/ignorance as soon as 3-4-way is tested. I'm confident that Tracy Chapman etc. does not reveal the problem, but "unfortunately" some of us might listen percussions, piano or electric where requirements for transient force are much higher in case listener knows possibilities or how instruments actually hit to the face and body when timing is flawless and other dynamic features such as thermal compression and radiator locations don't significantly reduce haptic feedback.

Not that one should necessarily be an excuse for the other, but production chain depending on the conditions can have many different HP filters in place.

Agree, but as we both said, it's very bad excuse to totally ignore audibility and significance without doing anything such as valid study if available studies are suspicious or clearly bad. Null hypotesis is default, but it is not meant for publications such as 'Sound reproduction' as probable "truth" without doing anything for it.

 

[thewas]

These might be the variables we should listen and evaluate from the system to be worthy and respected members here. I probably continue to ignore Harman's narrowed psychoacoustics and rely on objective measurement data. Distortion is distortion no matter how many doctors try to ignore it.

The thing is that Harmans research was based on real listening tests including several listeners while your theses on speculation with no proven relation to audibility. Toole's book is also by far not just an overview of the Harman research but the corresponding state of international literature. Also it doesn't say that only amplitude response matters but that it matters by far most.

I'm well aware that audibility and significance of phase distortion are very very low with normal 2-way speakers without sub. Ignorance of timing on ASR's reviews does not usually matter in practice. But all that changes with 4- and 3-ways with steep crossover slopes. Typical IIR XO for subwoofer damages timing and transient dynamics really badly. I have tested that few times with headphones only.

Listening with headphones as you know is very different to listening of loudspeakers in normal reflecting rooms. Talking about anecdotal experience, I had once the chance to listen to 4-way steep slope systems (K+H O500C plus Neumann subwoofers) in good room acoustics with direct switching between minimum and linear phase filtering and the difference was very small even with signals you mention and not really of a preference, just slightly different sounding at direct switching.

 

[kimmosto]

can have many different HP filters in place

This discussion reminded me about one interesting test. I scanned maybe 100 piano recordings and listened with some midsize or larger speakers with minimum-phase response. It was ridiculously easy to distinguish differences in transient force i.e. differences how much recordings had HP filtering or some other dynamic processing by sensing haptic feedback from the hammer and string.
Haptic feedback has been very low or missing with every multi-way speakers with very long excess GD at LF. Correct tonality is there, but otherwise just soft and powerless melody.

 

[kimmosto]

The thing is that Harmans research was based on real listening tests including several listeners while your theses on speculation with no proven relation to audibility.

Why should I prove anything as long as anyone can repeat that kind of tests. How helpless you really are here? What have you done for this? I don't see any value or facts in you messages. Just text bubbles praising lazy speculations including references to poor ancient studies.

You are ignored.

 

[thewas]

Why should I prove anything as long as anyone can repeat that kind of tests. How helpless you really are here? What have you done for this? I don't see any value or facts in you messages. Just text bubbles praising lazy speculations including references to poor ancient studies.

You are ignored.

Because as I had written your tests are flawed, for example while using headphones making assumption on the audibility of multiway loudspeakers in normal reflective rooms. If you have anything of scientific significance write a journal paper on it before making condescending comments on such.

 

[Arindal]

There are shared mics, but also mics used for the recording only.

So, recording all mic signals untouched, and applying FIR-based high-pass in the DAW, would be an option?

I have not noticed any timing issues, but I will admit I have never done my very best to detect them as well.

Can absolutely confirm @kimmosto ´s approach here. I found the best approach being listening tests with ultra-transient artificial signals (lots of ambient techno, deep house out there) as well as rather dry yet acoustically recorded drumkits (can recommend some jazz recordings or retro-funk/post-disco stuff), which you both know well from headphones, as alternating tracks. Over time, you get a pretty good feeling what sounds like impulses are ´time-coherent´, and what shows either GD or overly long decay, or a combination thereof.

I'm confident that Tracy Chapman etc. does not reveal the problem, but "unfortunately" some of us might listen percussions, piano or electric where requirements for transient force are much higher in case listener knows possibilities

Exactly. Some groovy Jazz with kicking bass drum, by Jacques Loussier or Retro-Disco like L´Impératrice would do.

I had once the chance to listen to 4-way steep slope systems (K+H O500C plus Neumann subwoofers) in good room acoustics with direct switching between minimum and linear phase filtering and the difference was very small even with signals you mention

I have undergone the ´LLL´ vs ´MLL´mode test with O500C as well, years ago. With the right transient material in the right control room (well-treated in the bass range), it was pretty pronounced and not just subtle. Of course, long decay, both originating from the room or the (vented) speaker, severely mask GD problems in the same band, according to my experience.

 

[thewas]

I have undergone the ´LLL´ vs ´MLL´mode test with O500C as well, years ago. With the right transient material in the right control room (well-treated in the bass range), it was pretty pronounced and not just subtle. Of course, long decay, both originating from the room or the (vented) speaker, severely mask GD problems in the same band, according to my experience.

In my case there were no long decay bass problems as it was room corrected.

 

[olieb]

It's not "preference test" with Tracy Chapman etc. and single variable (spectral balance/tonality), though phase distortion can transform also to clear tonality change with transient signal.

Can you please name some well executed (as opposed to poor) and recent (as opposed to old) studies that show the impact of low frequency phase distortion on tonality for sound reproduction with speakers in small rooms.
I am sure you know of such studies as otherwise, how could older studies be "outdated"?

This ´carved in stone´ is a very good point. What is oftentimes referred to as ´scientific´ on this board, setting extremely high thresholds how controlled, blind listening tests should look like, is in my understanding a measure to prevent any other entity than Harman, to ever produce or publish any test result that would be accepted as ´scientific´, due to high costs of blind test facilities and staff to run it. While on the other hand disparaging all other tests conducted ever since as ´non-scientific´, ´biased´.

This is a funny way of putting things upside down.
The way I understand it, one major result of Toole's research has been that to get ANY meaningful data about sound quality/preference, one must conduct CONTROLLED tests. Otherwise all one gets is bias.
Now, instead of honouring that result, people call the research either "outdated" and "poor" ("measurement techniques of the days"), and at the same time, exaggerated and overly rigid [Some people apparently CAN have it both ways].
To me, this seems like an attempt to give one’s own anecdotal experiences more weight. Understandable, but is that really a reason to lower the bar and allow bias to creep back in?
I wonder whether you would advocate for the same in medical research, lowering the thresholds for testing new medication or treatment.

I have undergone the ´LLL´ vs ´MLL´mode test with O500C as well, years ago. With the right transient material in the right control room (well-treated in the bass range), it was pretty pronounced and not just subtle. Of course, long decay, both originating from the room or the (vented) speaker, severely mask GD problems in the same band, according to my experience.

When all this is so obvious and easy to hear and show, why are there no robust scientific results about the claims of enormous GD effects in LF?

 

[Audionaut]

Does anyone have any actual facts and figures to show if speakers that have been designed to apply the Harman research findings have actually made any significant sales inroads with the buying public?

Answering this question implicitly assumes that one knows exactly which companies tune their speakers precisely to this model.
Do you know this, and are you familiar with the companies and specific models?
And: Do you also include manufacturers who are developing with Klippel NFC for example in this category or is this an own field of Engineering?

 

[TimVG]

So, recording all mic signals untouched, and applying FIR-based high-pass in the DAW, would be an option?

For sure.

It just occurred to me, but we should also consider reverb which uses an enormous amount of all pass filters with modulated cutoffs and diffusion networks (think Bricasti M7 - one of the best and most popular analog reverb units). There can be a a lot of phase manipulation even in classical recordings.

 

[Kevin1956]

Answering this question implicitly assumes that one knows exactly which companies tune their speakers precisely to this model.
Do you know this, and are you familiar with the companies and specific models?
And: Do you also include manufacturers who are developing with Klippel NFC for example in this category or is this an own field of Engineering?

Thanks for the comment. I am satisfied that the answer to my question was: No there isn’t.

 

[kimmosto]

If I recall it correctly, that product was named O500C, and included instantaneous implementation of GD for listening tests. The ´fullrange linear-phase mode´ was particularly interesting.

Mid looks almost filler driver with probably quite high XO with the woofer. I did not find specs telling XO frequencies. Assuming that lower XO is ca. 600 Hz - just like with KH 420, and not much steeper than 4th order, I would not expect radical difference between normal and phase-linearized options. Subs and others with much lower XO point are totally different animals. Group delay can jump to 6-10x longer with equal filter order.
8260 had lower XO, 6th order or so at ca. 400 Hz. That produced (originally) so long GD that woofer's wave didn't fit much into time window of 4 ms while measuring quasi anechoic in a living room. 8260 wasn't able to reproduce piano's hammer with proper haptic feedback imo. 8351A was probably as bad.
Passive 4-ways such as Elac FS 509 have also been lame without much perceivable force in wide-band transients. I think Revel Ultima Salon 2 is not far from the same category. It may partly hint why Harman avoided the topic. Commercial science could have some biases depending on situation with their products.

 

[TimVG]

Mid looks almost filler driver with probably quite high XO with the woofer. I did not find specs telling XO frequencies. Assuming that lower XO is ca. 600 Hz - just like with KH 420, and not much steeper than 4th order, I would not expect radical difference between normal and phase-linearized options. Subs and others with much lower XO point are totally different animals. Group delay can jump to 6-10x longer with equal filter order.
8260 had lower XO, 6th order or so at ca. 400 Hz. That produced (originally) so long GD that woofer's wave didn't fit much into time window of 4 ms while measuring quasi anechoic in a living room. 8260 wasn't able to reproduce piano's hammer with proper haptic feedback imo. 8351A was probably as bad.
Passive 4-ways such as Elac FS 509 have also been lame without much perceivable force in wide-band transients. I think Revel Ultima Salon 2 is not far from the same category. It may partly hint why Harman avoided the topic. Commercial science could have some biases depending on situation with their products.

So the question begs what a good (but also practical) solution would be to smooth out bass response over a wider area in a small room (which multisubs can do) but without having to resort to large drivers in a resistance enclosure. FIR filtering? Or is pre-ringing also a potential issue. Or just take the loss and accept a compromise in dynamic reproduction of transients.

 

[kimmosto]

Can you please name some well executed (as opposed to poor) and recent (as opposed to old) studies that show the impact of low frequency phase distortion on tonality for sound reproduction with speakers in small rooms.

I haven't tried to find more than these. This list was collected quite a while ago. Toole lists few others in his book, but I don't recall are they all in AES e-Library where I have continuous access (for quite high price). Probably not.

As you can see, it's mostly just mid-range GD with headphones so official science does not offer much for us or professionals. Nothing really valid and plausible related to delay at LF, in my opinion. Logical conclusion is that phase distortion should be measured and documented as an obvious distortion - that data could be used for further investigations and analyses.

 

[TimVG]

I haven't tried to find more than these. This list was collected quite a while ago. Toole lists few others in his book, but I don't recall are they all in AES e-Library where I have continuous access (for quite high price). Probably not.
View attachment 521398
As you can see, it's mostly just mid-range GD with headphones so official science does not offer much for us or professionals. Nothing really valid and plausible related to delay at LF, in my opinion. Logical conclusion is that phase distortion should measured and document as an obvious distortion - that data could be used for further investigations and analyses.

There was a already a thread as well from @Kvalsvoll with interesting information in his blog.

 

[mj30250]

To address the original topic, anytime an attempt is made to directly link product sales performance to quality raises a multitude of flags for me. Does McDonald's make the best tasting hamburger? Per their sales number alone, the answer would be yes.

And of course I realize people don't go to a "hamburger store" to taste test a multitude of brands, but the sheer amount of variables and potential issues involved in a typical store audition of speakers renders the activity problematic at best. "Showroom sound" is a thing for a reason.

 

[kimmosto]

So the question begs what a good (but also practical) solution would be to smooth out bass response over a wider area in a small room (which multisubs can do) but without having to resort to large drivers in a resistance enclosure. FIR filtering? Or is pre-ringing also a potential issue. Or just take the loss and accept a compromise in dynamic reproduction of transients.

This is off-topic, but who cares so:
FIR is not probable option at LF - at least for normal PEQs. Adequate frequency resolution requires huge amount of taps which causes long latency. Low fs helps. So economical electrical approach is FIR for XO to keep excess GD as short as possible, and all EQs with IIR PEQs and shelfs. Woofer integrated close to the mid+tweeter enables higher XO frequencies such as 300-400 Hz. That is not a problem for FIR engine. Additional latency could be 5-6 ms with 4th order linear phase slopes.

About a week ago I tuned one this kind of XO. It was full active 3-way with lower XO at 400 Hz, 4th order linear phase. Additional latency setting was 5 ms. fs 48k and taps 512+512 for woofer and mid (128 for the tweeter of 2" coaxial driver). These are not my speakers and not my preference, but occasionally I can measure frds, zmas and enter DSP settings for some other(s).

 

[Jiraya369]

This discussion reminded me about one interesting test. I scanned maybe 100 piano recordings and listened with some midsize or larger speakers with minimum-phase response. It was ridiculously easy to distinguish differences in transient force i.e. differences how much recordings had HP filtering or some other dynamic processing by sensing haptic feedback from the hammer and string.
Haptic feedback has been very low or missing with every multi-way speakers with very long excess GD at LF. Correct tonality is there, but otherwise just soft and powerless melody.

when im in a bullshitting competition and my opponent is this guy