Master Thread: Are Measurements Everything or Nothing?

[Sal1950]

So yes, measurements are helpful, but a normal person has to test the speaker in his room to be able to choose the best sounding speaker.

Obvious flaws like resonances and distortion are of course something you do not want to ever have, but variations in frequency response can in my opinion be sometimes preferred and might be beneficial to your room. Should manufacturers aim for nothing else than flat anechoic FR? In my opinion they can if they see a benefit for most people (a good example is the low freq shelfing to compensate placement near rear wall).

A "normal" person doesn't know his butt from a whole in the ground about reviewing speakers in his room. Best he can do is chose a preference like exagerated bass or hearing hot treble as increased detail. In a large number of cases he'd be better much better served by just following the "recommended" speaker list on ASR, Erins, Archimago's, AudioHolics, etc.. Sometimes they find neutral speakers a bit bright and forward on first listening but with time will learn to appreciate the clearer window to the source. After you get used to listening to a transparent system, colored rigs with messy tubes or speakers with un-natural tonal presentations become disturbing.
YMMV

Thanks for having a go Jim, but it ultimately doesn't matter, because it wasn't my hypothetical.

The fact that Matt is all too willing to treat it as my hypothetical and twist that into an opportunity to judge me as a person with bin-worthy thoughts, says a lot more about him than about me.

That's just Matt being Matt as usual Newman, I can't believe his butt doesn't get sore from always riding the fence. He always tries to cover his choices for using a C-J tube amp to screw up some really nice speakers. So he takes the "it's all good" Stereophile view of audio.

So you are the keeper of the gate, as it were? Here to save the world from anyone with an opinion not matching your own? Thanks for that!

What we do here is try to educate folks on the science of audio and de-program them from the BS they learned from the audiophool world.
Do yourself a favor and open your mind to the fact that audio is a solved science and has been for a few decades, there isn't any magic that quality measurement gear and an educated handler can't reveal.

 

[solderdude]

Listening to speakers (in a demo room) is about equally important/telling as knowing what Klippel measurements have to say.

That said.... one really needs to audition speakers in the actual room they are going to be used in connected to the gear they are going to be used with.
One might want to take a few days time for this at least, move the speakers around a bit (as far as the WAF allows ) and listen to various music (recordings) in the later hours of the day.
Even turning it up loud for a short moment can be informative.

A room can make or break speakers simply because a space in a house is very likely not going to resemble a 'good listening room'.
While EQ may help for some tonal balance issues at a specific listening position it would be preferable if the speakers would already sound good without it.

For amplifiers... they just need to be able to drive the connected speakers as loud as you want/need them to play without sounding nasty (power thing).

For the rest of the chain ... everyone has their own 'imagination' (perception) and those devices really do not have to be 'listened to at home' before buying.
Of course every one is free to feel that each one has a 'sound' and buy based on their own deception. That's fine.
When one wants to discuss this (the alleged but always clearly heard) 'sound signatures' and 'house sound' of components... take it elsewhere.... where they care.

Trying to convince 'ASR folks' they can hear this or that using gear that has been tested here and is deemed 'audibly transparent' is bound to be met with disbelief and FLAK.

'ASR' is more interested in ensuring the basics are 'right' (measurements) and exclude certain gear based on that or consider it a usable device (there are so many that are totally usable).
What measurements can't predict is longevity (of controls) and operational issues nor whether or not certain pieces of equipment/cables are prone to be plagued by ground-loops etc.

 

[john2017]

What we do here is try to educate folks on the science of audio and de-program them from the BS they learned from the audiophool world.
Do yourself a favor and open your mind to the fact that audio is a solved science and has been for a few decades, there isn't any magic that quality measurement gear and an educated handler can't reveal.

Tell you what.. My gear is listed below. My room is 19x15 feet and 8.5 foot ceilings, speakers are the Polk Legend L800, driven by Carver Ravens, Mcintosh C2800 preamp using their FA2 module. Qobuz streaming is my primary source now. Specifications should not be too hard to pull up, and given that Amir has already trashed the Carvers... With that information, tell me how my system sounds...

 

[Mart68]

This began on the subject of how much can one predict the sound of ANY loudspeaker from the measurements, with people chiming in as to the usefulness of listening as a deciding factor.

What form would this prediction take? How is it communicated? Words? Translating sound into words isn't easy. Plenty of room for misinterpretation.

The prediction is the measurements.

What's the objective? To pick a speaker that will deliver good results in given room with a wide variety of musical programme? That seems like an ideal objective to me. Yes, you can absolutely do that from the measurements - it's the objective behind all the research that has been done.

We know what aspects of the speaker to look at that will determine if it can achieve that. Smooth on and off axis FR, low distortion, and if you're not intending to deploy subs, an extended bass response (which will then be tailored with EQ). Perhaps some passive treatment to augment the furnishings.

Since speakers that conform to these parameters will present the sound in much the same way we can already move away from thinking that it is essential to listen to them before making the choice.

The alternative is trying lots of speakers at random in the hope that one, by happy accident, will suit the room characteristics. That's just fumbling about in the dark. If you're not even demoing them in your own room, then forget it. Why would anyone think there's some value there?

 

[solderdude]

Hard floor ?, carpet(s) ?, room correction ?, furniture ?, curtains? bias towards gear ? personal taste ?

With that information, tell me how my system sounds..

All that matters is that you enjoy it and if you are not bothered by higher levels of IM distortion at high SPL (your speakers might distort even more at that level) then that is fine.
All that can be said with certainty is that signal fidelity is not as high as most ASR folks would like it to be.
That does not mean it cannot sound great to you or people visiting you and 'sounds better' than other systems you heard. I figure it will.

The Polk Legend measures fine. The C2800 too. The choice for the Carver simply is what rocks your boat. It does have enough power for sure.

 

[Dumu]

A "normal" person doesn't know his butt from a whole in the ground about reviewing speakers in his room. Best he can do is chose a preference like exagerated bass or hearing hot treble as increased detail. In a large number of cases he'd be better much better served by just following the "recommended" speaker list on ASR, Erins, Archimago's, AudioHolics, etc..

Our definition of "normal" person is clearly quite different. I would say a normal person is able to do a simple FR measurement and/or do a basic A/B comparison with his old speakers. People who can interpret measurements so well that they can reliably tell what a pair of speakers would sound in their room is clearly above normal and wouldn't need to test anything.

IMO the more your room deviate from the standard listening test room, the more important it becomes to try the speakers in your own room.

 

[Sal1950]

Tell you what.. My gear is listed below. My room is 19x15 feet and 8.5 foot ceilings, speakers are the Polk Legend L800, driven by Carver Ravens, Mcintosh C2800 preamp using their FA2 module. Qobuz streaming is my primary source now. Specifications should not be too hard to pull up, and given that Amir has already trashed the Carvers... With that information, tell me you my system sounds...

No thanks, Download the REW measurement software (free) and get a umik ($100) to go with it. Learn to use the software, measure your system at the listening position and post the results. That would be a very educational experience for both of us and open a path to really improving the sound of your system.
What a shame really, I owned a bunch of Carver designed gear over the years, to include a Phase Linear 700B power amp and a 1000 Dynamic Range Autocorrelator Noise Reduction System. Great gear in it's day and I'd bet a months Social Security that the 700B would kill your Ravens (second worst power amp ever measured here) both in quality measurements and sound quality. Bob used to be interested in designing great gear that moved the SOTA, today the bean counters are making the design decisions.
My rig, approx 1980.

 

[Newman]

I actually do own Mr. Toole's book (3rd and 4th edition).

Oh, the 4th edition has hit the streets? Good to see.

Also good to see that you have invested in quality information. And there is a lot in them, so it is not my expectation that you have memorised it all. I know I haven't.

I'm am in no way science denialist, and I really appreciate the information measurements provide to us. That means that I also recognize our limitations, such as that it is incredibly hard for humans to translate measurement data to actual sound we hear. Maybe someone can learn to do that after thousands of measurements in one single room, but that requires so much time and effort that it is beyond any regular person and thus very limited use in comparison to listening&measuring the speakers in their own room.

Then let the good reviewers do it. But more importantly, they don't have to or need to be smart enough to precisely predict a speaker's sound in 'your exact room'. The clues to this are indeed in the research. So when you say to me, "I think you are vastly underrating the room's effect to the sound" - I'm just saying what Toole has found. This is the area of possible denial that I meant.

Obvious flaws like resonances and distortion are of course something you do not want to ever have, but variations in frequency response can in my opinion be sometimes preferred and might be beneficial to your room.

I've just got a funny feeling that your primary evidence in support of that statement is your own sighted listening experiences at home. But maybe I am wrong, so, what is your primary evidence?

Should manufacturers aim for nothing else than flat anechoic FR? In my opinion they can if they see a benefit for most people (a good example is the low freq shelfing to compensate placement near rear wall).

Again with the low frequencies. A bit like your desktop monitor example.

I think the bigger issue than "oh they (and we) can't predict the exact sound in my room" argument, is the "my room is a fixed invariable constant" issue. That is just begging for second rate results, because small room acoustics are so badly behaved. The argument that goes, "oh I need to get them home so I can determine which versions of sh*te it sounds like in my room because of the room", does not sound like precision modelling is the issue.

Think of it this way (using the findings from research to benefit at home):-

1. To our enormous relief, it turns out that above a few hundred Hz the effect of the room on the sound quality is much less important than we may have previously thought, and we largely hear through the reverberation into the direct/first arrival sound, and once that sound is flat and smooth and extended, it is going to be highly satisfying. What's more, this can be seen in measurements better than by listening at home.
2. For the above frequencies, 'total room sound' is relatively uncritical and never an overriding consideration to the point where one would sacrifice the direct sound in order to gain something in the indirect sound. It mainly affects our choice of beamwidth and smoothness of off-axis response. If we are determined to stick with stereo and its shortcomings, a wider beamwidth and smooth off-axis response can be utilised to advantage. As we move into multichannel audio it becomes less and less important. Don't fixate on it (says the research). Given this, one can still pick a suitable speaker for its off-axis measurements being generally more suitable to your 'room personality' while understanding that it isn't game-changing stuff. Letting go of the "fixed room paradigm" is going to be far more effective than picking between two speakers with the right overall off-axis pattern but with small variations.
3. Below a few hundred Hz, you are gonna need some help. This shouldn't be left up to the speaker, and unlike #2 it is a very important part of our satisfaction. The total summed sound of speaker plus room is dominant to our perception in this band, and small room acoustics are going to dominate to such a degree that the idea of the speaker's innate bass response being a big deal is naive. The room sound is bad enough that it almost doesn't matter whether the speaker's bass was tuned for wall proximity or not. Something needs to be done to correct for the room and to create a smooth frequency response. The speaker's main job is to have sufficient low frequency SPL output to handle the correction that will be applied, and to extend as deep as you need it to (with or without subwoofer), and do so without signs of strain. Again, you can get all the necessary data from the right measurements.
4. Give regard to the sighted listening effect. Don't spoil your enjoyment (and even your opinion of the sonics) with a speaker that looks ugly to you, is from a country you have issues with, looks puny or elephantine to you, is so cheap you don't trust it or so expensive you resent it, is a brand you don't respect, etc. The research indicates that this doesn't just change one's happiness about non-sonics: it also impacts one's perception of the sound quality.

cheers

 

[Mart68]

Tell you what.. My gear is listed below. My room is 19x15 feet and 8.5 foot ceilings, speakers are the Polk Legend L800, driven by Carver Ravens, Mcintosh C2800 preamp using their FA2 module. Qobuz streaming is my primary source now. Specifications should not be too hard to pull up, and given that Amir has already trashed the Carvers... With that information, tell me how my system sounds...

The speakers are a good choice, but I can't find an impedance trace for them. I suspect they are not the best match for amplifiers that have output transformers, so that's a wild card. Otherwise, assuming a benign acoustic, I predict you'll get consistently good sound with most programme.

 

[Dumu]

Oh, the 4th edition has hit the streets? Good to see.

It's available at Google Play Books atleast. Physical copy might not be available yet.

I do not agree with your points 1&2. They might hold true for most rooms, but in my experience (actual measured response), the whole shape of the frequency graph can be drastically affected by the room and furniture, not just the low end. Sometimes a flat measuring speaker can sound very bright and sibilant in one room, and be perfectly fine in other.

Also there can be huge difference in the end result if you go from one coaxial driver per speaker to a tower speaker with multiple drivers. Directivity plays one part, but so does the physical placement of the drivers. It is very hard to predict the end result for such a change without hearing them both.

 

[Mart68]

I do not agree with your points 1&2. They might hold true for most rooms, but in my experience (actual measured response), the whole shape of the frequency graph can be drastically affected by the room and furniture, not just the low end. Sometimes a flat measuring speaker can sound very bright and sibilant in one room, and be perfectly fine in other.

Whilst this is true why would you want to complicate things further by choosing a speaker that does not have a smooth anechoic response?

If the room is reflective that can be dealt with quite easily assuming some leeway with furnishing choice.

 

[Newman]

I do not agree with your points 1&2.

The numbering is mine, but the information is not just an opinion nor mine.

They might hold true for most rooms, but in my experience (actual measured response), the whole shape of the frequency graph can be drastically affected by the room and furniture, not just the low end.

I never said otherwise. It's the importance to perceptual preference that is crucial to understand.

Sometimes a flat measuring speaker can sound very bright and sibilant in one room, and be perfectly fine in other.

I'm sure. All covered in point #2 and the room-as-invariant issue. If a speaker with a flat anechoic response is sounding very bright and sibilant, then you are between the devil and the deep blue sea. Your best hope is to do something about it without adopting a speaker that lacks a flat anechoic response.

Also there can be huge difference in the end result if you go from one coaxial driver per speaker to a tower speaker with multiple drivers. Directivity plays one part, but so does the physical placement of the drivers. It is very hard to predict the end result for such a change without hearing them both.

Are you sure that the on and off axis frequency responses were the same? Let's not confuse cause and effect.

 

[Dumu]

If the room is reflective that can be dealt with quite easily assuming some leeway with furnishing choice

The furnishing is not always up to debate

Also if you know you have a reflective room, why would you start modifying your room if you can pick up a speaker that takes away the need for those changes?

And to the original topic of this recent discussion, would you be able to tell from two different flat measuring speaker with different directivity, which one would make the problem worse and which would sound better to you?

I think that if you have to go blind, pick a speaker that measures well, but if you can audit, then absolutely test the speakers out, make some measurements to confirm what you are hearing and enjoy the music.

 

[Dumu]

I'm sure. All covered in point #2 and the room-as-invariant issue. If a speaker with a flat anechoic response is sounding very bright and sibilant, then you are between the devil and the deep blue sea. Your best hope is to do something about it without adopting a speaker that lacks a flat anechoic response.

Why would you rule out picking up a speaker with rolled of highs? The end result would be pretty much the same as if you would fix the issue with EQ?

 

[Mart68]

The furnishing is not always up to debate

True but in that case you're introducing a variable that has nothing to do with choosing a speaker via measurement or audition.

Also if you know you have a reflective room, why would you start modifying your room if you can pick up a speaker that takes away the need for those changes?

It's a more effective way of achieving the objective.

That aside, if the room was very reflective and you wanted to keep it very reflective, but not have a bright sound, does this now make auditioning multiple speakers in the hope of hitting on a solution a better option than choosing a suitable speaker by using measured performance?

 

[Sal1950]

I seriously don't get this, You men need to start standing up for yourselves and stop being a punching bag in your own home.
50/50 shared decisions is fine but hey, much of what I read here is off the tracks. I've told my X wife and GF that they can do anything they wish with
the entire house, I don't give a dang, but the living room is mine to make the decisions on audio / video gear, period. I've worked too hard for 5 decades to be bossed around in "our" house.

 

[Dumu]

That aside, if the room was very reflective and you wanted to keep it very reflective, but not have a bright sound, does this now make auditioning multiple speakers in the hope of hitting on a solution a better option than choosing a suitable speaker by using measured performance?

Wasn't this the whole question? Do you need to listen to the speakers or can you solely trust that well measuring speaker sounds always just fine? And/or can you identify accurately enough the cause of the brightness, and interpret measurements so well, that you could predict if that speaker will also sound bright.

It's a more effective way of achieving the objective.

This is very situational. I could imagine someone having very hard time to convince their partner to give up some old heirloom coffee table just so you can get rid of that nasty sibilance that she doesn't even notice.

 

[Purité Audio]

Listening in your own space is the ideal and ideally at the same time as some other ‘possibles’, measurements are really useful to sort the wheat from the chaff.
Keith

 

[Dumu]

Are you sure that the on and off axis frequency responses were the same? Let's not confuse cause and effect.

Most likely not, since the configuration already is so different. All the reflections would be different. But how would I know without listening/measuring the IR response which one produces more even response and which sound signature I prefer?

 

[Mart68]

Wasn't this the whole question? Do you need to listen to the speakers or can you solely trust that well measuring speaker sounds always just fine? And/or can you identify accurately enough the cause of the brightness, and interpret measurements so well, that you could predict if that speaker will also sound bright.



This is very situational. I could imagine someone having very hard time to convince their partner to give up some old heirloom coffee table just so you can get rid of that nasty sibilance that she doesn't even notice.

Like I said you're introducing an obstacle that need not exist. If you're really stuck with a highly reflective room you can do nothing about then you're never going to get close to optimal and you'll just have to live with that. The well measuring speaker is still the best option, regardless.

A coffee table might have a small impact on sound quality but it will not alone create 'nasty sibilance' that would be sufficient to be a deal breaker. And no speaker selected by any method will solve that problem anyway.