I cannot trust the Harman speaker preference score

[anotherhobby]

Thank you again Dr. Toole for chiming in. You mention that using EQ below 500Hz and multiple subwoofers is very important to listener experience. If a person uses that advice, gets a high quality AVR, multiple subs, and good speakers based on their spinorama is there any advantage, assuming an average sized room and listening distance, or reason to get tower speakers over standmount speakers since subwoofers will be in play?

Some bookshelf speakers get midrange distortion from the bass freqencies hitting the single midrange driver at high volumes, even with with an 80 Hz crossover. Depends on if your speakers exhibit this, or if you even notice. Mine did, and I noticed, and I hated it. I don't know how high of a crossover it would have taken to mitigate this with the Focal Aria 906's I had, but I didn't like how higher crossover points sounded anyway, so I fixed the issue by moving to towers. The perceived sound quality difference was huge to me, but if you don't listen at loud volumes, and/or your particular bookshelf speakers do not distort at the volumes you listen to with your setup, then it's likely of far less importance. Disclaimer: All of the above is just my personal opinion, experince, and preference.

 

[Philbo King]

57 pages so far!
Here's my take as a recording studio owner/operator:
The absolute flatness accuracy of a speaker/room pair is paramount to getting a mix that works on a variety of playback systems. If I add weighting to the FR curve, such as Harmons HF rolloff, I no longer can make mix decisions based on what I hear.

A flat response is not preferred by the public; usually it's said to be 'too bright' and lacking in bass. But if I use the Harmon curve on my monitor system the mix will end up overly bright and slightly weakened in bass, compared to a mix done on a flat system. Because the mix engineer will *always* compensate for perceived excess and lack across a non-flat FR.

Look at an alternate reality - let's say test partipicipants were given a graphic EQ to be adjusted how they prefer, and everyone ended up with some variant of a 'smiley' curve (mids cut, highs and lows boosted). Would that mean that it would be prudent to put an 'average of the smiley curves' onto every playback system? I find it mildly ridiculous, personally.

As far as using it for home stereo use, that is another issue entirely. Add whatever weird EQ you want and no one is any worse off (except possibly your neighbors).
But to do it in a recording studio mixing environment would be irresponsible and foolish and doing your clients a disservice by creating inferior product. YMMV...

 

[fineMen]

A flat response is not preferred by the public; usually it's said to be 'too bright' and lacking in bass. But if I use the Harmon curve on my monitor system the mix will end up overly bright and slightly weakened in bass, compared to a mix done on a flat system. Because the mix engineer will *always* compensate for perceived excess and lack across a non-flat FR.
... to be adjusted how they prefer, and everyone ended up with some variant of a 'smiley' curve (mids cut, highs and lows boosted).

What is 'flat' referring too, the direct sound 'on-axis' or the sound power output, or the resultant sound field in the room?

Another anecdotal observation: eventually I changed my personal focus when evaluating the subjective qualities of a speaker. Formerly I was quite obsessed with bass, now I ignore bass nearly alltogether. The reason is simply that I learned how to adjust it anyway to my liking using an equalizer. Exceptions apply: e/g KEF LS50 is not managable in this regard. When adjusted to my liking I always find that the bass measures quite linear, especially not emphasized. That is because since lately I got used to really clean midrange playback. Clean in regard to harmonic and intermodulation distortion, and of course smooth directivity. Most consumer grade speakers disregard the mids.

 

[Philbo King]

What is 'flat' referring too, the direct sound 'on-axis' or the sound power output, or the resultant sound field in the room?

My system is measured at listening position with a measurement mic so 'resultant sound field' would apply here.

 

[krabapple]

57 pages so far!
Here's my take as a recording studio owner/operator:
The absolute flatness accuracy of a speaker/room pair is paramount to getting a mix that works on a variety of playback systems. If I add weighting to the FR curve, such as Harmons HF rolloff, I no longer can make mix decisions based on what I hear.

A flat response is not preferred by the public; usually it's said to be 'too bright' and lacking in bass. But if I use the Harmon curve on my monitor system the mix will end up overly bright and slightly weakened in bass, compared to a mix done on a flat system. Because the mix engineer will *always* compensate for perceived excess and lack across a non-flat FR.

Look at an alternate reality - let's say test partipicipants were given a graphic EQ to be adjusted how they prefer, and everyone ended up with some variant of a 'smiley' curve (mids cut, highs and lows boosted). Would that mean that it would be prudent to put an 'average of the smiley curves' onto every playback system? I find it mildly ridiculous, personally.

As far as using it for home stereo use, that is another issue entirely. Add whatever weird EQ you want and no one is any worse off (except possibly your neighbors).
But to do it in a recording studio mixing environment would be irresponsible and foolish and doing your clients a disservice by creating inferior product. YMMV...

Dr. Toole discusses data on the speaker/sound field preferences of studio engineers versus typical consumers, in his book.

 

[Floyd Toole]

Some bookshelf speakers get midrange distortion from the bass freqencies hitting the single midrange driver at high volumes, even with with an 80 Hz crossover. Depends on if your speakers exhibit this, or if you even notice. Mine did, and I noticed, and I hated it. I don't know how high of a crossover it would have taken to mitigate this with the Focal Aria 906's I had, but I didn't like how higher crossover points sounded anyway, so I fixed the issue by moving to towers. The perceived sound quality difference was huge to me, but if you don't listen at loud volumes, and/or your particular bookshelf speakers do not distort at the volumes you listen to with your setup, then it's likely of far less importance. Disclaimer: All of the above is just my personal opinion, experince, and preference.

I said: "It is partly a matter of visual aesthetics - a large bookshelf on a stand vs. a medium sized floor stander. I say medium sized because it is not necessary to duplicate the very low frequency output supplied by the subs. So, high-pass filtering a medium sized tower will allow it to play a bit louder. Otherwise, as you conclude, there is no need for anything beyond a sturdy standmount loudspeaker." So, "large" and "sturdy" are not elaborated on, but clearly imply that the bookshelf speakers need to have woofers capable of significant output. Ideally, they would be three way designs - a rarity in the marketplace - like the Revel Gem 2 that I use in my system for surrounds. So, a medium floor stander is generally a safer choice. For home theaters there are suitable loudspeakers (e.g. JBL Synthesis), but they may or may not be cosmetically acceptable for living room use.

 

[Floyd Toole]

My system is measured at listening position with a measurement mic so 'resultant sound field' would apply here.

As has been pointed out several times, the first requirement for subjectively judged "neutral" timbre in double-blind listening tests is a flattish direct sound - i.e a flattish anechoic on-axis/listening window response. You are measuring the steady-state combination of direct and all reflected sound - a different thing. Depending on the listening distance, the dispersion characteristics of the loudspeaker and the acoustical environment this will take on a measure of downward tilt; more as the distance increases. The "Harman" (with an "a") curve you mention is what results from highly rated, well-designed forward-firing loudspeakers in typical listening rooms. It is not a target curve, and certainly not one to be applied in some of the aberrant control-room acoustics in existence. Check out the thread "what is your favorite house curve" for a fairly chewy discussion on this topic.

 

[AdamG]

fairly chewy discussion

 

[anotherhobby]

I said: "It is partly a matter of visual aesthetics - a large bookshelf on a stand vs. a medium sized floor stander. I say medium sized because it is not necessary to duplicate the very low frequency output supplied by the subs. So, high-pass filtering a medium sized tower will allow it to play a bit louder. Otherwise, as you conclude, there is no need for anything beyond a sturdy standmount loudspeaker." So, "large" and "sturdy" are not elaborated on, but clearly imply that the bookshelf need to have woofers capable of significant output. Ideally, they would be three way designs - a rarity in the marketplace - like the Revel Gem 2 that I use in my system for surrounds, where they are probably more than is needed. So, a medium floor stander is generally a safer choice. For home theaters there are suitable loudspeakers, but they may or may not be cosmetically acceptable for living room use.

Exactly, and I completely agree. Since the person asking didn't mention size, I thought I might help them by also providing an example of a medium sized bookshelf that I know of personally where bad distortion, even with dual subs, was resolved by stepping up to a tower. I apologize if I come off as being disagreeable.

 

[fineMen]

My system is measured at listening position with a measurement mic so 'resultant sound field' would apply here.

I once tried to implement a 360° omnidirectional speaker. Alas, I tuned it to flat on-axis response. It failed in being too bright in-room.

With the earlier mentioned quite radical minimization of mid frequency distortion (HD and IM) the actual 'tilt' of my house curve doesn't bother me too much anymore. It changes anyway with listening distance, which again depends on the room size and so on.

I personally cannot stand the 'subjectivist' approach to speaker evaluation, but that is due to the self-entitled ignorance of its proponents. The modern standard set by the Harman ranking has a lot to it. On the other hand one has to accept that due to the variety of listening conditions it necessarily stands on clay feet as we say here. I'm afraid there will be no iron rule to replace it. Let alone all the recorded musical heritage that was produced using non-standardized studios as reference.

I always said that the referring Harman quality score is a tool decidedly set-up for manufacturers. The quality oriented consumer remains still obliged to evaluate by (trained) listening test at home, whether or not the speakers fit the needs. If all participants in the business lean towards an acceptence of the standard, the chance for chosing the right speakers get better.

 

[Dj7675]

It is partly a matter of visual aesthetics - a large bookshelf on a stand vs. a medium sized floor stander. I say medium sized because it is not necessary to duplicate the very low frequency output supplied by the subs. So, high-pass filtering a medium sized tower will allow it to play a bit louder. Otherwise, as you conclude, there is no need for anything beyond a sturdy standmount loudspeaker.

Large, Sturdy, Standmount (thank you for your work bringing about the LSR32/LSR6332). Using 7 of them ...

 

[levimax]

Otherwise, as you conclude, there is no need for anything beyond a sturdy standmount loudspeaker.

Yea but that's not what is in your room https://www.thescreeningroomav.com/...te-real-world-home-theater-and-listening-room

 

[fineMen]

The "Harman" (with an "a") curve you mention is what results from highly rated, well-designed forward-firing loudspeakers in typical listening rooms. It is not a target curve, and certainly not one to be applied in some of the aberrant control-room acoustics in existence.

Consumers' listening rooms may be even more aberrant. On shelf positioning of speakers is quite common. Even floor standers get positioned close to the walls; at least in Europe for practical reasons. European houses are made of sturdy bricks, but are smaller, people sit more close to a wall.

The emphasis on a flat direct sound is one thing, but the Harman quality score takes an average listening room (listening position included?) into mathematical account.

A standard is a good thing by its own right. It is justified if it doesn't pose to much trouble to comply to it.

Would it be possible to (virtually) equalize a smaller speaker to flat on-axis sound when siitting on a bookshelf, and then re-evaluate the Harman quality score? For reference of a re-evaluation after e/q see https://www.audiosciencereview.com/...core-center-speaker-review.40156/post-1415633

 

[Floyd Toole]

Yea but that's not what is in your room https://www.thescreeningroomav.com/...te-real-world-home-theater-and-listening-room

There are six Revel Gem 2s in that room, used as main surrounds. The LCRs are bigger - why? Because they looked "cool", especially upside down, sounded spectacular, and I could afford them. They do most of the work and are loafing along - no threat of distortion even at theatrical crescendo level.

 

[Mr. Widget]

There are six Revel Gem 2s in that room, used as main surrounds. The LCRs are bigger - why? Because they looked "cool", especially upside down, sounded spectacular, and I could afford them. They do most of the work and are loafing along - no threat of distortion even at theatrical crescendo level.

Really appreciate your taking the time to share so, so much!

 

[Newman]

Large, Sturdy, Standmount (thank you for your work bringing about the LSR32/LSR6332). Using 7 of them ...
View attachment 263231

He said Medium...

 

[Benedium]

Pretty sure he said...
Large standmount or Medium Floorstander

 

[Newman]

Good point! Even though my smilies don't seem to penetrate your cornei.

And note that those standmounts are larger than some medium floorstanders,

 

[thewas]

The emphasis on a flat direct sound is one thing, but the Harman quality score takes an average listening room (listening position included?) into mathematical account.

The Predicted In-Room Response (PIR) is a statistical approximation including except the direct sound also the early reflections and sound power (12% Listening Window, 44% Early Reflections, 44% Sound Power) and thus giving smooth directivity a higher score.

 

[Floyd Toole]

Good point! Even though my smilies don't seem to penetrate your cornei.

And note that those standmounts are larger than some medium floorstanders,

And they come from JBL Pro division, not JBL consumer, so they were designed for control room applications where loud is normal and dead air is to be avoided. They were used in the Harman "reference" multichannel room for a few years.