Bright High End Speakers = Bad Room?

Most speakers aim for a ruler flat reponse which is too bright for most folk. This is why speakers must be custom tuned to the customers ears. You wouldn't walk into a store and buy a pair of shoes in any old size would you? If you did it would probably be too big or small. Speakers that are not custom tuned are likewise too bright more often than not. You can either buy whats available in the market or have a pair custom tuned or modified. 

A ruler flat response on axis only matters in an anechoic chamber and/or listening very close to the speakers. The room response is likely to be much different which is what my post speaks to.
I do not think overall that most people consider a ruler flat response too bright, especially since it is the room response they respond to, not the anechoic measured on-axis.
It would be much easier to "tune" a high end too bright for a user outside of the speaker, whether electronically, or with room treatment and positioning. Cross-over design is tightly tied to phase integration and physical speaker parameters. It would be difficult to tune any given speaker to your tastes without impacting the fine balance of other parameters the speaker vendor was targeting.

especially since it is the room response they respond to, not the anechoic measured on-axis.

The room response is obviously related to the anechoic response. 

This is why I specifically stated on-axis anechoic response as opposed to integrated energy response or something similar.

Kenjit wrote: "a ruler flat response... is too bright for most folk." 

Floyd Toole, Bruel & Kjaer, and I all agree with Kenjit on that point.

Atdavid wrote: "I do not think overall that most people consider a ruler flat response too bright."

Toole conducted extensive double-blind listening tests and found that most listeners prefer a gently downward-sloping response trend, both for the first-arrival sound and for the early reflections. Most people perceive a "flat" response to be "bright", and a gently-downward-sloping response to be "flat".

But there was one group who consistently preferred a flat response: Recording engineers. To them, speakers are a tool, and in general the more revealing the better. 

Toole did find that, when listening strictly for pleasure, many recording engineers prefer the gently-downward-sloping trend. 

Microphone manufacturer Bruel and Kjaer identified this gently downward-sloping response as being desirable long before Toole confirmed it.

Duke

But there was one group who consistently preferred a flat response: Recording engineers. To them, speakers are a tool, and in general the more revealing the better.  

Do you have a link to the study that mentions this? Thanks

Toole conducted extensive double-blind listening tests and found that most listeners prefer a gently downward-sloping response trend,

what was the explanation?

" Do you have a link to the study that mentions this? Thanks "

He describes it in his book, don’t remember whether or not he included a reference. I’ll try to remember look for it this evening.

" what was the explanation? "

I don’t recall offhand what Toole said about it.

I have read that it may have to do with the absorption of high frequencies as sound travels significant distances through the air. At a live performance the distances involved would reduce the amount of high frequency energy reaching the audience, whereas the relatively short path lengths from instrument to microphone and later from speaker to listener do not attenuate the high frequencies significantly.

Duke

Chances are very good Toole’s test system was overly bright. Oh, well, back to the drawing boards. Shut the cave door and back to pigmy country! 

audiokinesis, we may be in agreement and we may not be.

A speaker that is ruler-flat on-axis, anechoic chamber, will have a typical room response that slopes off at high frequencies. It is this response, the room response that slopes off that people prefer predominantly. Generally first reflections will always be sloped as they are off axis. I copied pieces of my post below as I wanted to be quite clear that ruler-flat on axis anechoic response is much different from room response and it is room response that people respond to and determine as too bright.

In Floyd Toole’s latest presentation he heaps praise on a speaker with ruler flat on axis anechoic response. He also states in his excellent 00 paper on Maximizing Loudspeaker Performance the need for a "flattish, smooth, axial frequency response". His personal preference at that time was room friendly loudspeakers (i.e. most rooms).  Do you know when that first study was done?  If it was old, it could be an artifact of speakers at the time?

The real solution, for professionals as well as consumers, is loudspeakers that deliver similarly good timbral accuracy in the direct, early reflected and reverberant sound fields. This can be described as a loudspeaker with a flattish, smooth, axial frequency response, with constant directivity (which together result in flattish, smooth, sound power).

audiokinesis2,102 posts10-30-2019 4:49pmKenjit wrote: "a ruler flat response... is too bright for most folk."

Floyd Toole, Bruel & Kjaer, and I all agree with Kenjit on that point.

Atdavid wrote: "I do not think overall that most people consider a ruler flat response too bright."

Toole conducted extensive double-blind listening tests and found that most listeners prefer a gently downward-sloping response trend, both for the first-arrival sound and for the early reflections. Most people perceive a "flat" response to be "bright", and a gently-downward-sloping response to be "flat".

atdavid OP12 posts10-30-2019 3:14pmA ruler flat response on axis only matters in an anechoic chamber and/or listening very close to the speakers. The room response is likely to be much different which is what my post speaks to.I do not think overall that most people consider a ruler flat response too bright, especially since it is the room response they respond to, not the anechoic measured on-axis.

A speaker that is ruler-flat on-axis, anechoic chamber, will have a typical room response that slopes off at high frequencies

why would it slope off in a room that is less absorbent than an anechoic chamber?

In an anechoic chamber, typically what we are showing is the on axis response typically tailored to be perfectly flat.

The room response is a combination of the direct and reflected. Higher frequencies are typically absorbed more by things in the room, even just painted drywall, but they travel farther distances, so also attenuated more by the air. Add in the highest frequencies tend to be the most directional, so it may be flat on-axis at 20KHz, but 10db down at some angle off axis, and room response is sum off on axis and reflected off axis.

The room might be the problem. Every room sounds different.  Interesting discussion. I have 4 two channel systems in 3 different rooms.  Moving the same speakers ...my  B & W 703s and Monitor Audio Silver 10s and a couple of others to different  rooms yields very different tonal balances.  Moving the speakers to a different wall in the basement STUDIO room yields a very different balance as well as radically different imaging.  The room could be the issue.  I haven't found my B & W 703s to be too bright and I don't think all "high end" speakers are too bright. I have heard some that were ...."too cool" for my taste but most sounded articulate and wonderful.  Some cheaper units are very brittle and hard.  "HEARD THEM AT THE SHOW" is not a reliable judgment of any product. Until you have the units in your room with your components you really won't  know if you will like the sound or not. Years ago, my dealer would allow me to audition a product in my room before a final decision to purchase was made. I didn't like a very expensive abd beautiful set of Thiels I auditioned. They sounded a bit bright and, to me, too analytical. Others loved them.  I didn't have enough guts from my power amp ....so perhaps that's why I didn't like them.  Who knows? If you don't like B & Ws or a "bright" speaker try something else. (I like the Monitor Audio Silver 10s better!) Enjoy the music!

Kenjit quoted me: "But there was one group who consistently preferred a flat response: Recording engineers. To them, speakers are a tool, and in general the more revealing the better. "

And then he asked: "Do you have a link to the study that mentions this? Thanks" 

I thought it was in the third edition of his book but I just skimmed all of the chapters that seemed promising and couldn’t find it. Nor is the index helpful. If I can find my copy of the first edition I’ll look there. My recollection is that Toole was the source, if/when I come across it I’ll let you know. 

When I have been asked to design custom studio monitors by the acoustician who designed the studio, he has specified "flat on-axis response" as the target. 

Duke

I often get the impression that a good deal of the members on this forum appreciate to be fooled by a speaker being non-bright on bright recordings, that is quite rolled-off in the highs. They seem to call it musical. And name other speakers more true to the recording bright. Then there are speakers that are more bright than they actually should be, which may be fine as a working tool identifying problems in a recording. The end result of course a lot depending on the room and the listening distance. 

Speakers that are very 'revealing' are often considered to be bright. 'Bright' speakers have become something to fear and avoid. I refer to many of them as 'clear'. Padding a speaker down in an effort to flat line on a chart can/will soften the sound by removing detail. High end audio is high resolution. If the speaker is capable of sounding natural at a high resolution then I find it to be wonderful clarity. Why remove detail that you've worked so hard to attain? However, it will be critical of poorly made recordings. Over processed music out of a high resolution speaker can and will sound edgy. Good recordings are amazingly clear and sound incredible when done right. 

Good quality headphones are extremely clear. I began use them as a baseline for the clarity of my speakers. This put me on the 'bright side' of neutral and I much prefer it.

Agreed. Don't want the speakers to do the eq-ing. There are other tools for that where you actually know what is done and only when it's needed.

It all come down to enjoyment.  Even if you sit there thinking to yourself, Wow, these speakers are really accurate, how wonderful!, if you cringe and shudder even time the massed violins come in forte, then there's not much point to it, is there?  In music, emotional enjoyment > intellectual enjoyment.

Then there are speakers that are more bright than they actually should be, which may be fine as a working tool identifying problems in a recording.

Pro speakers for recording rigs in studios with consoles (64 channels, etc) are designed to be bright, or chosen for their unnatural brightness...as a form of revealing what is put in a mix. So that it is spotlit and thus manipulable, workable, mutable, changeable. Where slight changes in the recording can be heard. It is purposely exaggerated and unnatural.

This means you should never think that a studio monitor speaker is even remotely suitable for neutral and balanced home use. They are emphatically not.

Neutral speakers for mastering, is another thing altogether.

This is true most of the time, so much so that it is a logical maxim. There will be exceptions or people taking exception with such statement. Sure. Whatever... But it is very much a norm in recording studios. Horses for courses.

NS-10 speakers from Yamaha, for example... are trash. Utter trash, and not suitable for home use or balanced listening to actual recordings in your home. Unless one is a masochist.

The anechoic and/or quasi-anechoic measurements people are used to seeing are supposed to be flat.
In a room you should have a descending response from around 100-150 Hz. This is what all of the automated room-correction software tries to do, more or less. They don’t attempt a ruler flat response for exactly the reason stated by the OP.


There are some other things going on in tweeters I'd like to mention though.


B&W and some other "high end" speakers use a ragged tweeter response designed to elicit a sense of detail. This is a trait I blame Stereophile for hyping/promoting as "neutral." It isn't.


Brands which shy away from this kind of tweeter trickery are Magico, Vandersteen and YG Acoustics.  However, that doesn't mean the rest of their response is ruler flat/accurate either. Just that they at least don't try this hyper-detail trickery.  Of course, I call it trickery, you might call it a wonderful feature. :)


Having said all this, I also find that the floor area behind/between the speakers is an unexpected place for harshness in a speaker sound. If you have a solid floor or rack between the speakers, it’s worth experimenting with blankets/pillows in the area.

I find that the most natural and comfortable to listen to speakers are the ones that employ some subtle version of "The BBC Dip".

Highly revealing, spacious, well done lower slope crossovers, fast, tuneful, all of it, with a slight BBC Dip. Clean and warm, as the mix demands. A properly done BBC Dip, in my design experience....has the lowest part of the dip centered around 3.5khz and then goes back up again, and is about 2db deep, overall, as a best potential compromise.

And then it’s good all around and it can play all day, at any volume. From soft to break the windows levels...

Everything else that might be off.....is generally just bad equipment choices. (It is a multistage learning curve and individually committed to as an act --same goal, and very probably a different path for each person)

This graph has it as being slightly exaggerated but warming up to being ideal.

This would come down to personal preference and likely room response as well. For me, I am not a big fan of this design technique, though the reason to implement it, to invert the peak in the Fletcher-Munson equal loudness countours and "soften" the sound makes sense. From a methodology standpoint, it is similar to the "loudness" setting that boosts bass response at low volumes where we have poor sensitivity to bass. It is sneared at in audiophile land, but there is a solid reason to do it when you can’t turn the volume up.  I would not be surprised if there was a variation in the level of dip in the loudness contours from person to person.

teo_audio1,177 posts10-31-2019 11:52amI find that the most natural and comfortable to listen to speakers are the ones that employ some subtle version of "The BBC Dip".

Highly revealing, spacious, well done lower slope crossovers, fast, tuneful, all of it, with a slight BBC Dip. Clean and warm, as the mix demands. A properly done BBC Dip, in my design experience....has the lowest part of the dip centered around 3.5khz and then goes back up again, and is about 2db deep, overall, as a best potential compromise.

@teo_audio

This means you should never think that a studio monitor speaker is even remotely suitable for neutral and balanced home use. They are emphatically not.

B&w, pmc, atc, kii audio, etc many are used for both. 

@teo_audio 

That curve is about what I normally do with digital eq to "tame" my loved ATC 150asl somewhat. Maybe up to 4db around 3,5 khz where they cross to the tweeter. Except for a flat response in the bass.....

Can't really agree that most pro monitors are especially exaggerated in the highs (above 8 khz). Maybe they are just not rolled-off and a little more direct in their dispersion. Most often they have controls to take down/roll-off the treble. NS-10 of course very special. Made only to work with the mids.

Teo_audio wrote: " I find that the most natural and comfortable to listen to speakers are the ones that employ some subtle version of "The BBC Dip"."

Agreed - a bit of dippage centered somewhere between 3 and 4 kHz does seem to correlate with long-term fatigue-free listening.

One way of seeing the BBC Dip is "inverting the Fletcher-Munson peak", as atdavid so eloquently described it.

But here’s another way of seeing it: The BBC dip is right smack where the bottom end of most tweeters is, where most tweeters have an off-axis pattern flare. The result is a net excess of energy in this region (the off-axis energy perceptually adds to the on-axis energy), and the BBC dip is one way of addressing that.

One day I was curious about whether the BBC dip would be subjectively desirable if there were no off-axis energy flare to compensate for. My designs are constant-directivity over much of the spectrum including that region, so I dialed in some gentle dippage centered between 3 and 4 kHz. What I heard was educational to me.

Yes the speaker became more forgiving, but the harmonic richness was degraded. There was something obviously missing relative to having smooth response through that region. The loss of something quite desirable - timbre and texture and a natural-sounding richness - caught me by surprise.

So while the BBC dip does invert the Fletcher-Munson peak, I have come to believe that its primary raisin d’etre is to compensate for that off-axis energy flare at the bottom end of the tweeter’s range. And if that flare isn’t there, ime the BBC dip would be a step in the wrong direction.

One thing that peak in the Fletcher-Munson curve DOES tell us, in my opinion, is where the stakes are the highest.

Duke

Duke,
That is an fantastic post!
I had a feeling the general "view" may have also been tied to typical tweeter emission, but did not feel confident enough in what the "typical" speaker was in this community to suggest it.

I always put effort into room treatment and response trying to achieve flatness which could be why I am not a fan of the dip since I have addressed excess energy already.

Professional studio techs check their mixes on several different speakers.  The goal here is to provide a pr9duct that "translates" the artist's music to there listeners.  A "good mix" or "master" will "translate" well on different systems.  i have found that the recordings I really am moved by manage to "translate" well on a number of different systems from by car audio, a modest compact system as well as my attempt at putting together a "high end" system  There are no perfectly "flat" or "neutral" speakers in existence and if there were, that "perfect" speaker would sound different in different rooms and with different components. The pros put a lot of effort into perfecting the acoustics of their rooms and still, no two rooms sound exactly alike.  Put together a system that "translates" the music not only to your ears but to your heart and soul.  Trust your ears and your heart and ENJOY THE MUSIC. 

Duke,
That is an fantastic post!
I had a feeling the general "view" may have also been tied to typical tweeter emission, but did not feel confident enough in what the "typical" speaker was in this community to suggest it.

I always put effort into room treatment and response trying to achieve flatness which could be why I am not a fan of the dip since I have addressed excess energy already.

The second point is that this area of the frequency spectrum is the peak of the ’intelligence’ read for the human ear.

men use it to tell where the threat is coming from (directional, closeness) in the jungle scenarios, etc. Women use that frequency spectrum a bit differently, they use it to tell the difference between the cries of children, re health, etc. They use it to understand the sounds of the very very young.

so men can be excited by peaks in that area of the spectrum, and women will be powerfully turned off and agitated by it, in all the wrong ways. men think it is exciting and invigorating, in some critical ways, and women think of that sound as someone killing or torturing children.

So if you want spousal harmony...as an audiophile in a shared space....it is easy.

Put in the effort.....and .....Don’t go there.

~3-4khz, no peaks, no resonance, no honking, nothing but a clean slight/subtle dip.

When a speaker does the screech thing, it is usually screeching in the mentioned frequency area. I personally won’t even go into the given room at an audio show, I can hear it from outside the door. This is also the top range of the ear canal resonance range areas.

Things are a bit conflated in this (canal resonance vs hearing peaks vs intelligence reads), but it is notably real. Whatever we build, or make, we make absolutely sure this peak area of importance does not exist as a problem.

And that is how you help get the other 50% of the human race into the audio world.

Note that the most favored ’high slope required’ mid-bass drivers start to go nutty and slurred in this area of the frequency range. I would not touch any of those drivers with a barge pole. It leaves you down to about maybe a dozen or less drivers that actually function correctly - in the entire world of available drivers.

the market aim is flat FR.

this is too bright for my taste. even in a untreated room, i dont find the highs gets tamed enough.
and a flat FR in a well treated room is too bright for my taste as well.

the solution is either eq, or try to tame the HF response with resistors if you have a 4 binding post.

if you look at the AN-E measurements, its no wonder so many love that speaker. its flat from 100hz to 1000hz, then from 1khz up to 8khz its down 4db in comparison. this gives a much smoother sound.

Excellent Teo. Duke is right on. Having a TACT room control device I can analyze the in room frequency response of any loudspeaker and devise target curves to give the speaker any response curve I want. I can load 9 different response curves and switch back and forth between the nine.
A flat in room response curve is not just bright. It is way too bright and will make anyone with normal hearing wince, even kenjit. I can not relate this to the speaker's anechoic response curve because I do not know it.
I tested Watt/Puppies at a friends house. They are a very easy to listen to loudspeaker and seem to have fine detail. They were very popular among the press. Measured, they were down 10 db at 20K and had a 2 dB dip centered on 3K. One would have to believe they were intentionally designed this way. When I made them flat they were painful. 
Then there is the problem of volume. Our hearing sensitivity changes with volume. Refer to this link   https://www.kmuw.org/post/loudness-and-fletcher-munson-curve  So, a speaker that sounds balanced at 80 db will sound bright and bass heavy at 90 dB. Conversely, a speaker that is balanced at 90 dB will sound honky like a table radio at 80 dB. What this means is that any system has only one right volume at best. If you have  loudness compensation you have two right volumes. All this pertains to each individual recording! Without loudness compensation each recording has one right volume level. With loudness compensation you have another "right" volume at a lower level. 
By bouncing back and forth between the 9 different compensation curves this can be easily demonstrated to anyone except maybe cleeds. 
The TACT demonstrates that this can be managed in the digital domain using dynamic loudness compensation. It can be programmed to hop from one loudness compensation curve to another based on volume so no mater where you set the volume control the music sounds exactly the same which is very spooky. So now I have the unit programmed with two compensation curves both down 6 dB at 20K, one flat through 3K and another with a 2 dB notch filter centered on 3K for edgy recordings. Both  have 4 different volume levels. This takes up 8 presets. The ninth I keep perfectly flat for demonstration purposes. The dynamic loudness compensation has an on and off switch so I can play all 9 presets with or without loudness compensation. I always keep it "on." The different volume levels are so I can keep the TACT up against it's maximum digital volume depending on the volume of the source material. This helps to maintain maximum digital resolution (bit depth). 
TACT of course is out of business and no other unit on the market has this capability yet. Hopefully that will change before my TACT dies. 
None of this applies to strictly analog systems. As I have mentioned before I run my ARC phono amp into a Benchmark ADC so I can input it into the TACT and take advantage of it's magic. Nobody thinks running the turntable strait analog sounds better. Once you are in numbers you can do virtually anything to the signal without any increase in distortion.
Being able to modify the frequency response of the signal any way you want is a great way to learn what happens if you do this, that or the other and I am sure more units that can do this will be forthcoming from companies like Anthem and Trinnov.

I wish I could bold this in 100 point font. THIS!!!    Volume has an enormous impact on perceived tonal balance, far more than almost anything. I think I may even make a topic about it!

mijostyn1,269 posts11-01-2019 2:11pm Without loudness compensation each recording has one right volume level.

Fully agree. Wish I could find a TACT unit. My solution at the moment is a sonos streamer to martinlogan Unison for the ARC and then Antimode 2.0 for individual settings. Antimode has four channels to play with. Finally different dacs.

Here's a bit of an academic view from a professional musician and studio owner who tracks/mixes/masters many clients.   My belief is that you cannot understand this topic completely just thinking about the speaker and room alone.   I'll make a few truisms and see if people agree.

1) Flat Response speakers with accurate phase in a decent room properly communicate what the microphone is picking up when you're tracking.

2) The ideal speaker to mix on is the speaker the end user will listen on.   And by the way, it won't be the speaker above.   If you make your mix satisfy you on speaker X, then with a similar room, it will likely satisfy (approximately) the end listener.

3) End listeners do not listen on razor flat speakers with accurate phase (by the way that also means no ports).   And I believe that historically, a typical end user speaker didn't have excellent high end and thus were a bit muddy.

OK.   What does this all mean?   My experience is that when I mix on my Dunlavy Mk-VI speakers which are razor flat or Genelec, in the control room, it sounds just like the mics and the mix sounds incredible.   Then when I play back in the car or elsewhere, it's muddy.    

My conclusion is that the accurate studio speakers help you to here exactly what you recorded but......you need to adjust you mix by listening and evaluating on muddier speakers.   Then, your mix will sound right there (and too bright on your Dunlavy's).

I think it's a historical problem.   If audio began for consumers with flat speakers, we wouldn't be having this conversation.   Historically, engineers mixed bright for the average expected consumer speaker.   Then we had to continue forever.   I mentioned this to Bob Hodas (who tuned the rooms at Abbey Road) and he replied with "You've learned a powerful lesson on room tuning".

I'm scared to hear the response to my dissertation.   Thanks,  Anthony

Interesting and really something I have thought a lot about. Could you also indicate how and how much you typically compensate for lesser gear. That could help the ones that actually have better than average gear to tune it more properly.