I cannot trust the Harman speaker preference score

[tuga]

It seems to be a common misunderstanding to think that removing all reflections from the listening room results in 'more fidelity' in reproduction. It is a very understandable and seemingly logical conclusion - of course, but it doesn't seem to hold true with human listeners.

You are conflating fact with listener preference.
Some listeners prefer some harmonic distortion too... Recording/mixing engineers use it all the time (see SoS piece Analogue Warmth).

Why is room-induced acoustic distortion good and electronic signal distortion bad? Or loudspeaker cabinet diffraction?
Besides, neither music nor audio equipment producers have any control over how the end listener's room will affect or if you prefer "enhance" the end result.

Wishful thinking...

 

[tuga]

Really? Have you actually read the whole book? Have you heard properly recorded multi-channel music on a good setup? Why is it you always take the contrarian viewpoint towards multiple researchers who literally defined the way the whole industry regards speaker development? Is there any point to me writing this post when you have thousands of posts that basically just repeat the same old boring positions with 0 cited studies and 0 knowledge backing them up, over and over, in hundreds of threads on this forum?

I think we both know there is no point in me writing these words to you, so I'll leave further responses up to the others in this thread. I wonder if you understand how privileged you are that people continue to seriously engage at all.

Multi-channel may be better than 2-channel but it is still flawed.

For a realistic reproduction of the original soundfield you need this inconvenient setup (n.b. speakers are in an anechoic chamber):

 

[dominikz]

You are conflating fact with listener preference.
Some listeners prefer some harmonic distortion too... Recording/mixing engineers use it all the time (see SoS piece Analogue Warmth).

Why is room-induced acoustic distortion good and electronic signal distortion bad? Or loudspeaker cabinet diffraction?
Besides, neither music nor audio equipment producers have any control over how the end listener's room will affect or if you prefer "enhance" the end result.

Wishful thinking...

The casually condescending tone of you post makes me not want to engage further - except to say that it is my humble opinion both your reasoning and conclusions are wrong.

 

[tuga]

The casually condescending tone of you post makes me not want to engage further - except to say that it is my humble opinion both your reasoning and conclusions are wrong.

We disagree. There's nothing wrong with that.
I would add that room-generated envelopment compensates for lack of it in stereo reproduction. But it is a trade-off. You can either get sharp imaging and a more accurate "recreation" (reproduction) of the capture ambience or a wider soundstage and images and increased envelopment.

A couple of snippets from "Acoustics 
Of
 Small Rooms" by Kleiner & Tichy (if you don't mind me quoting a source other than Toole):

Spaciousness and diffusivity

Localization of externalized single sound field components was shown to be fairly straightforward but dependent on many factors. Localization of sound field components that have identical sound levels at the ears will depend on further factors such as phase difference.
When sounds are correlated, such as a monophonic signal that is presented binaurally, the auditory event occurs inside the head, inside head localization (IHL). If the sounds at the ears are fully uncorrelated, such as two separate noise signals that are presented binaurally, there will be two auditory events, one at each ear.
An interesting effect can be heard when presenting a monophonic wide bandwidth noise signal in stereo (over loudspeakers or headphones) if the stereo signals are out of phase. The noise frequency components below 2 kHz are then perceived as spatially diffuse—having spaciousness— whereas those for higher frequencies are perceived as located between the loudspeakers (or for headphones, IHL occurs). The time difference in the low-frequency components provides phase cues that are ambiguous thus providing apparent sound field diffuseness, whereas the high-frequency sounds are analyzed by their envelopes and those will be identical at the two ears causing a located auditory event.
Similarly, when a wideband noise signal is provided over headphones to a listener and one of the headphones is fed with the signal delayed by a millisecond or more, the sound is perceived as diffuse.
What constitutes a diffuse sound field is thus different in the physical and psychoacoustic domains. In the latter, a diffuse sound field is that that provides non-locatedness of sounds or, alternatively phrased, that provides a sound that is located over all spatial angles (or rather upper hemisphere in a concert hall that has sound-absorptive seating).
In physics on the other hand, a diffuse sound field is defined as a sound field where all angles of sound incidence have equal probability, where the sound from each spatial angle is out of phase, and where the energy density is the same everywhere.
Obviously, the two ideas of what constitutes diffuseness are different in the two sciences. A physically diffuse sound field will also be psychologically diffuse but not necessarily the reverse. From the viewpoint of listening, it is of course the psychoacoustic properties that are of importance, not the sound field properties.


Auditory source width and image precision

As we listen to sounds, the apparent width of the auditory event, often called the auditory source width (ASW), will depend on many issues. To those listening to stereo or multichannel recordings of sound, it is quite clear that the width of the array of phantom sources treated by the recording or playback is determined by not only the layout of the loudspeaker setup in the listening room and the directional properties of the loudspeakers but also on the listening room itself. The more reflections arriving from the sides of the listening room, the wider will the ASW be. However, the ASW will be frequency dependent above 0.5 kHz and a 2 kHz sound arriving at ±45° relative the frontal direction will produce maximum ASW [38,39]. This is to be expected since the masking by direct sound is the smallest for this angle of incidence of early arriving reflections [16]. The ASW also depends on the low-frequency content of the signal, more low-frequency energy increases ASW [38,40,41]. Psychoacoustic testing shows that the spatial aspects of the early reflections are primarily determined by the reflection spectrum above 2 kHz [33].
Reliable data for sound reproduction in small rooms are difficult to find. A single omnidirectional loudspeaker judiciously placed close to the corner of a room may well create as large an auditory image as a conventional stereo loudspeaker setup placed out in the room as discussed in Chapters 9 and 11.
Using digital signal processing, the ASW can be made to extend far outside the bounds set by the stereo baseline. Sound field cancelation techniques


Symmetry

Early reflected sound will confuse hearing and make the stereo stage and its phantom sources appear incorrectly located or even blurred. As explained in Chapter 8 the listener’s placement of the phantom sources is dependent particularly on the transient nature of the sound that comes from the loudspeakers so it will be affected by the early reflected sound from the room surfaces. The early reflected sound will also affect the global auditory source width for an orchestra for example and may make it extend considerably beyond the baseline between the loudspeakers.
In asymmetric rooms where the walls on the left and right of the listener have different acoustic properties, the stereo stage may become biased towards the wall that reflects the most. The curve in Figure 8.23 shows the dependency more clearly for different levels of unbalance as applied to the center phantom source in a stereo loudspeaker system. The intensity will then be higher at that ear and the sound stage distorted. This distortion is usually compensated by changing the balance in amplification between the stereo channels.
At low frequencies in the modal region, symmetry may not be desirable since someone sitting in the middle of the room may be on or close to modal node lines. One way of avoiding such node lines is to make the room asymmetric in the low-frequency region.
This can be achieved by having an asymmetric rigid shell surrounding the inner room which is symmetric for mid- and high frequencies by suitably reflective side walls, ceiling, and floor. The inner room must be open acoustically to the outer shell at low frequencies, for example through ventilation vents, and similar large openings, for example at corners. In this way, one can have the desired listening position sound field symmetry for mid- and high frequencies while at the same time have asymmetric conditions in the modal frequency range. Bass traps to control the damping—and thus the reverberation times—of these modes can be placed between the outer and inner shell. It is important to remember though that noise transmission to the surrounding spaces will then be dependent on the sound isolation of the outer shell that must be physically substantial.

 

[Inner Space]

Also, can you clarify why you believe these spectral errors would not be audible in mono but would become audible in stereo? Research seems to suggest that such spectral aberrations are much more easily audible in mono vs stereo.

To be specific, I'm saying the errors would be equally audible in mono and stereo, but because their locations would be audibly tied to the cabinets, they would corrupt the development of free-floating phantom images. The question I was answering concerned plausible stereophonic illusions, not speaker quality per se.

 

[sarumbear]

We design and test speakers in mono but always demonstrate them in stereo.. As Floyd writes in his book, the dominant factor in perceived spatial attributes for stereo tests is the recording itself. And the data from the 1983 test supports that. I've heard the Mirage M1 at NRC and in his home. They sounded wonderful. Very spacious in his large living room, but too diffuse , unfocussed sound stage in the NRC listening room without curtains on the rear and wide wall to absorb the strong rear reflection. More proof that the desired directivity depends on the room acoustics and to some extent the recording.

80 dB is the average level that varies up and down depending on the dynamics and crest-factors of the music. That is what most people consider comfortable in a domestic room.

I am sorry, has there been a issue on the forum software? I cannot see anything you said above is related to what I asked, which I copied below and highlighted for clarity.

I have one question: since the Harman tests done almost 30 years ago, audio reproduction technology has improved quite a lot. DSP was not even a thing then. Were there recent test done using modern speakers where the score is verified to be correlating to listener preferences? Will you personally see the following to come out as an expected result?

Sonos Roam: 5.5
JBL M2: 5.1

Thank you for taking time to comment.

 

[abdo123]

@sarumbear Can you please not waste the gentleman's time with frivolous comparisions?

for the trillion time the measurements of the Sonos Roam are of lower resolution compared to the JBL M2 so the comparision between these two scores is meaningless.

 

[sarumbear]

...Mono tests are simply more reliable and sensitive methods to assess the performance of a loudspeaker.

Stereo is 1950s, and the research focus now is immersive audio.

There lies the issue, as I see it. Research shows that a single speaker should be listened in order to judge the quality of speakers and as 3D/immersive listening can alter that quality of a the listening experience, research community has left the concept of quantifying individual speaker quality behind.

But here on ASR it is still gospel.

 

[RobL]

Kudos to Dr. Olive for striding into this pile. We need to start a pool on how long this thread will circle the drain.

 

[sarumbear]

@sarumbear Can you please not waste the gentleman's time with frivolous comparisions?

On ASR everyone is equal. He is indeed an esteemed member of the industry and needs respect, which I have done that on my post. He quoted my post hence he has answered me directly. If he wanted to ignore me he wouldn't have commented my post. If you don't find it odd than so be it, but it has nothing to do with you. Ignore my post and move on.

for the trillion time the measurements of the Sonos Roam are of lower resolution compared to the JBL M2 so the comparision between these two scores is meaningless.

What is meaningless is the score!

 

[maverickronin]

Dr. Toole has said this as well both in his book and on this very forum, many times in fact: "What is needed to deliver a more credible sound field to listeners is a multichannel system. All else is compromise, especially two-channel stereo - so we play around attempting to extract from a directionally and spatially deprived system some sense of realism."

This is frequently skipped when discussing his work. I think what it boils down to is convenience. Building a multi-channel setup is *less convenient* than a stereo setup. You have to figure out where to put several more speakers, etc.

The part that bugs me is that the stereo-fixated folks rarely state that what they're looking for is primarily convenience. They still consider a stereo system to be an uncompromised audio system and pretend that you can reach the best experience possible using current technology with stereo speakers. But you can't. That's a fact that is as strongly proven as research can show, a fact that you cannot escape no matter how eloquently you wax about stereophonic soundstages and music and emotions.

The "problem" with multichannel is that essentially all music is still only released in stereo and that's the reason it's "skipped". Mch is theoretically better and, allows one to do much more creative things in a mix, but outside of movies, the content just isn't there. Maybe you're lucky and the niche recordings fit your tastes perfectly. It doesn't fit mine very well.

I know a lot of people like upmixers, but it's a lot of money to gamble when all your sources are only in stereo. Also, being so niche and lacking an open standard they are all proprietary and ephemeral. A lot of people say Auro3D is currently the best, but Auro is probably going to get crushed by Dolby and their codec and upmixers are going to disappear from newer AVR/AVPs. If you're lucky there will only be a long drought until another good one is released.

I don't think mch isever going to take off for music either. It's been tried a lot, from stuff well before my time like quadraphonic records and passive upmixer circuits to audio and concert DVDs. Even before the height channel codecs came out and we had an essentially open standard of 7.1 it was never popular for straight music.

 

[Andrej]

I suppose we can agree to disagree…
If I am talking on the phone my family and friends will say, “did you go into the kitchen?”
Or “did you just go into the bathroom?”
Or “You are back in the living room again aren’t you?”

The reflections tell us about the room. You can claim reflections are distortions, but the distortions are not reflections, and our hearing has evolved to let us work this out. Or, if you want to believe that God created us this way, that is fine too.
But all the powers of heaven do make reflections the same as distortions.

If I am not not correct, then I hope someone will provide some proof to correct my understanding.

It is all about what you desire. For those of us who are all into accuracy and hearing the recording venue, reflections mask it and thus they are noise.
For those (and you appear to be one of them) who like to hear the sound of their listening room (think of its shifting the performers into theit listening space) reflections are a source of information.
In extreme cases you could think of the theoretical situations:
-- Record in a perfect acoustic space - listen in an anechoic chamber (listeners perceive themselves in the recording venue)
-- Record in an anechoic chamber - listen in a perfect acoustic space ((listeners perceive performers in the listening space)

Given that so many recordings are close miked and mixed, with no/little recording venue captured, who am I to say that your approach is not preferred more often than not. It's all about personal preference, and it appears ours are different.

Having had the opportunity to listen to music within a large anechoic chamber (Bell Labs, Murray Hill, see the archival photo of the insides) I definitely prefer it to any typical living room by a huge margin.

 

[Andrej]

It seems to be a common misunderstanding to think that removing all reflections from the listening room results in 'more fidelity' in reproduction. It is a very understandable and seemingly logical conclusion - of course, but it doesn't seem to hold true with human listeners.

It seems that some reflections from the listening room are in fact required to make the stereo illusion stable - without them in-head localization may occur and sense of envelopment may be lost. May I refer to my previous post from a different thread:

I can also speak from personal experience that spending time in an anechoic chamber is not pleasant - I would not wish to relax and listen to music there.

We should also not forget that stereo is an illusion, and not a reproduction of any original soundfield. In addition, unlike electronics, loudspeakers have the dispersion component, and obviously 3D output cannot anyway be made the same as the 1D electrical input signal, so the argument of 'keeping true to the signal' is somewhat moot - especially considering that some listening room reflections appear to be desirable for soundstage externalization (see above).
IMHO we should instead aim for characteristics that make the stereo illusion stable and pleasing, while not being detrimental to tonality of the recording - which based on research seems to be loudspeakers with flat on-axis, and even off-axis behaviour used in normally furnished living rooms; resulting in early reflections that are similar in spectrum to direct sound.

I have tried the whole range from a fully reverberant to anechoic listening, and as I gradually moved from one to the latter, it sounded better and better (more realistic and pleasant) to me. Perhaps it is a matter of taste. And this was true for a large range of recordings. With continual adjustments in the frequency response as I went along. Illusion or not, that is my preference. And as you say yourself there is a clear logic to it. And if the presence of reflections are so much better, why are we not all clamoring to have them added to headphone listening, easy enough to do, I guess? Also, without reflections, it is much easier to tell what kind of recording setup was used, and if there is a presence of the original soundfield or not.

To my understanding, any resonant peaks that may fill-in on-axis dips (to make the on-axis FR nicer-looking) would still show-up as peaks off-axis - i.e. they would still be visible from the spinorama in the end. So I'm not sure if this hypothetical case equates to a real concern.

Also, can you clarify why you believe these spectral errors would not be audible in mono but would become audible in stereo? Research seems to suggest that such spectral aberrations are much more easily audible in mono vs stereo.
Note that listening in mono does not mean listening anechoically - i.e. the listener still hears the coloured off-axis sound that such a loudspeaker would produce from reflections - even if on-axis is flat - so these kinds of issues should be audible in mono, or not at all.

Otherwise any poorly designed loudspeaker could be EQ-ed to a flat on-axis and sound great - this is IME most definitely not the case. Good directivity behaviour is needed for this to be feasible.

How much of a factor is that when listening to a mono speaker the reflections are different than when you have two speakers, unless the mono speaker is in the same location as one of the stereo speakers? Was that the case when these comparisons were made?

I believe that mono is better then stereo for these kind of experiments, as that is where the experimental evidence points.

 

[More Dynamics Please]

When weighing various opinions on technical issues those who have earned respect for their decades of scientific research also earn greater credibility for their opinions related to that research. Doesn't mean they are always right, but just more likely than the less experienced and less qualified.

 

[Pdxwayne]

May I know if Harman's blind prefernce tests ever used eq to adjust freq response of speakers? Or all speakers used as-is?

I wonder if we eq one speaker to sound exactly like the headphones preference curve when it reaches testers' ears, would it score high in blind speakers preference tests, especially when the test subjects are younger generation who primarily using headphones for music?

Headphones preference curve:

Vs speakers preference curve:

They don't match!

 

[Andrej]

When weighing various opinions on technical issues those who have earned respect for their decades of scientific research also earn greater credibility for their opinions related to that research. Doesn't mean they are always right, but just more likely than the less experienced and less qualified.

I absolutely agree! But the probability is not 1.0, and it is well known in many fields that creativity stops in one's twenties, and not having so much experience is what moves those fields forward with out of the box, unorthodox thinking. So it pays to keep an open mind. Which is not a trivial task for any of us.

 

[sarumbear]

Multi-channel records had been around for more than two decades. The Super Audio CD format was introduced in 1999. A year late DVD-A followed. They both offered uncompressed, hi-res, multichannel (7.1) audio. One look at the chart below will tell you the story that market didn't care about either format, i.e. multi-channel audio. Are we going to say that just because we now have speakers in the ceiling, the future of music playback will change? (Reference RIAA)

 

[ROOSKIE]

Think of the effect of early-reflections as watching a film in a room full of mirrors and you will understand what I mean.
I find that part confusing in Toole's book, it seems to conflate reflections in live sound (used for source location and characterisation of spatial qualities) with reflections (which overlay onto the recorded acoustics or spatial qualities and create confusion).

This is not correct.
The early reflections are just that early, they arrive so nearly in sync in time with the original they add a bit of spaciousness, give a sense of the space (ie an event taking place in your room) and affect the perceived tonality but NOT a distinct and separate image AT ALL.

They do not create the house of mirror's effect in any sort of typical room.

The early reflections and the direct on axis are one event to the human mind and constitute "the sound".

Removing them and determining if you like that sound without any reflection is just a matter of doing some listening in an anechoic chamber. You do that and report back. I am sure that would be extremely interesting and I 'd be very interested hearing if you enjoyed that musically.

This house of mirror's analogy is a very mistaken one.

 

[dfuller]

Multi-channel records had been around for more than two decades. The Super Audio CD format was introduced in 1999. A year late DVD-A followed. They both offered uncompressed, hi-res, multichannel (7.1) audio. One look at the chart below will tell you the story that market didn't care about either format, i.e. multi-channel audio. Are we going to say that just because we now have speakers in the ceiling, the future of music playback will change? (Reference RIAA)

I was going to make this point, but I seem to have been beaten to it. Yeah, multichannel is superior in almost every respect - but the entry price is absurd.

 

[Purité Audio]

Many particularly partners balk at the thought of two loudspeakers.
Keith