How much does the speaker matter?

[Affwhire]

Sorry for the clickbait-y title. It was done tongue in cheek. Obviously, the answer is generally "a lot".

But, I am wondering about a specific situation...

Say that you have a well-treated listening room, an existing "2.2" system, and either already use some software solution, minidsp+dirac live, or trinnov for sub integration and correction....

And, you're generally happy but considering a speaker upgrade (mains/satellites only) partially to solve a specific problem and partially because "it feels like time".

Now consider that 2 specific speakers are kind of bubbling to the top of your research, at different price points (say, one is double the price of the other), both well-regarded...but only one of them has been thoroughly measured and you can't find a spinorama for the other. You're confident they're both good...but not directly comparable via measurements and not both available to demo at the same place, and subjective impressions of both are all over the map.

Short of ordering both and going back and forth (which is always a headache)....

1. How would you go about comparing?
2. Is it reasonable to rely on the correction to even-out probably minor "defects" in one?
3. What else might make the decision for you?

I am obviously in this exact situation. I'm new to the forum, but I've been reading for years. I can post more details if necessary, but I'm honestly more interested in answers to those big-picture questions. And, yes, I'm seriously considering having one drop shipped to amirm just so I can see the measurements of the other speaker....I mostly want to figure out how that's going to work with the shop/vendor's return policy before I contact him about it.

Any thoughts?

 

[Head_Unit]

Buy a bluetooth speaker and be happy; send me the rest of the cash. Sigh, I suppose that solution won't appeal...some thoughts:
- I worked with an automotive correction system similar to Audyssey, where actually a coworker of mine went. Each system differs, however there still seem real-world limitations to how much correction can do. Better speakers and a better room still improve things.
- More money generally gets better overall speaker quality. My very rough rule of thumb from years of loudspeaker engineering is that stepping up to the next quality plateau takes about 3X the money.
- I realize your question is a general one but can't really comment without specific model numbers. There are so many variables: 2-way vs 3-way, ported (how low?) versus sealed, cabinet construction, and very importantly the aesthetic LOOK of the product DOES matter a lot, unless you enjoy looking at unattractive stuff on a daily basis. I don't hence I spent like triple to get a rube red washer/dryer set with pedestal instead of boring white without.
- I'd probably tend to go for the most expensive one presuming there was a reasonable return solution (noting that shipping costs have gotten very expensive). But that would be really modified by the other factors. I might well punt and postpone and save until I could get something even more expensive. That's OK because speakers last for literally decades, hence it is worth expending more funds if at all possible.
- I would never ever buy speakers that weren't returnable without listening(unless something pretty inexpensive).

 

[amirm]

I wouldn't buy anything unmeasured unless the company provides proper ones (e.g. Genelec, Neumann, Harman, etc.). You are taking a chance otherwise. As you say, if you drop ship it to me, I can measure it and if it is within return window, you can choose to return it.

 

[tuga]

Any thoughts?

Unless you are able to correlate your preference with measurements then measurements will only tell you how accurately the speaker reproduces the signal but not how good it sounds to you. The chance that you will like a speaker which performs well in the test bench is reasonable but there's guarantee that it will happen.

Nothing beats listening for yourself, in your home, with your system and favourite music.

 

[amirm]

Unless you are able to correlate your preference with measurements then measurements will only tell you that how accurately the speaker is but not how good it sounds to you. The chance that you will like a speaker which performs well in the test bench is reasonable but there's guarantee that it will happen.

Nothing beats listening for yourself, in your home, with your system and favourite music.

His ad-hoc listening is no guarantee of anything. The only reliable information comes from measurements. Everything else is a distant second from that point of view.

 

[Killingbeans]

Is it reasonable to rely on the correction to even-out probably minor "defects" in one?

I wouldn't count on it. If the speaker has discrepancies between the on-axis and the off-axis response, an attempt at correcting the direct sound might just exaggerate problems in the early reflections and get you back to square one.

 

[Doodski]

- More money generally gets better overall speaker quality. My very rough rule of thumb from years of loudspeaker engineering is that stepping up to the next quality plateau takes about 3X the money.

Do you mean the cost of speakers at manufacture stage and design or at retail?
Your comment made me remember how many speaker brands that I've retailed and I made a approximate list of them. I've never experienced a situation where 3x the budget needs to be expensed for a better speaker. Most of the time just going to the next model in a speaker line is enough to get a substantial change never mind choosing from another brand. Sometimes the same price, sometimes $50 more or a hundred more is all that's required. If you mean manufacturing then that's a whole different pricing structure.
energy
b&w
axiom audio
mirage
infinity
kef
rft
sony
pioneer
yamaha
celestion
cerwin vega
jbl
magnat
mb quart
jvc
kenwood
mission
panasonic
technics
pioneer
polk audio
paradigm
Definitive Technology
koss

 

[HooStat]

Get the ones that have been measured. You will probably find that you like them a lot. Then keep an eye out for measurements on the other one. After you see them, decide whether they seem notably better, whether what makes them better matters, whether you think you will actually hear the improvement, and whether you want to go through the hassle.

You will often find that "improvements" come from better bass extension which is obviated by the use of subs. This is because bass extension differences can affect the costs of the drivers, cross-over, and cabinet. So, when bass is the "difference" between two well-engineered speakers, the payoff can be minimal if you have subs.

 

[CinDyment]

Unless you are able to correlate your preference with measurements then measurements will only tell you how accurately the speaker reproduces the signal but not how good it sounds to you. The chance that you will like a speaker which performs well in the test bench is reasonable but there's guarantee that it will happen.

Nothing beats listening for yourself, in your home, with your system and favourite music.

The old throw stuff up against the wall and hope it will stick method.

Here are 3 things about speakers I will have a hard time fixing:

- Off axis frequency response and dispersion
- Distortion
- Resonances

I can correct on-axis within reason, and I can even correct on-axis phase to some degree, and that is just a portion of the overall performance. If I don't know those first 3 are good above, I will never achieve the sound I expect in a room. There is still work, but without a foundation, it is pointless.

 

[tuga]

The only reliable information comes from measurements.

I agree with this bit.

I use measurements for shortlisting, but if he cannot correlate measurements with his preference, then it's just guesswork. At least by listening to a few speakers he will have a better idea of what he prefers.

 

[dlaloum]

His ad-hoc listening is no guarantee of anything. The only reliable information comes from measurements. Everything else is a distant second from that point of view.

But measurements don't account for psycho-acoustic preferences - the Harman work resulting in a Target curve gives us an indication.... circa 60% preferred the Harman curve.... so what about the other 40% ?!

And that's only looking at FR...

Not that I am discounting measurements - but there is no correlation mapping between measurements and personal preferences that would allow pure measurement based decision making.

Working in high end retail audio for some years, "educated" (some would say biased! ) my ears...

I can listen to a set of speakers in a demo setup for 5 to 30 seconds, and know whether they are worth spending any more time on.... or not.

And this is not necessarily price related... The little Gallo Nucleus Micro speakers primarily have sins of omission... single small driver, no bass to speak of (intended to be used as satellites with one or more subs) - but there are no nasty resonances, the sound is clean.... not overly resolving, but definitely listenable - lots of speakers at that price point, you listen for 5s and you just move on.... nasty!

Some people are so used to certain resonances, that they tune them out... One woman I knew, insisted for years that pretty much all speakers and systems sounded the same... - then at one point I had to send my Quad ESL's (original 57 version) away for repair, and I temporarily replaced them with a pair of Klipsch Forte III's (which have their fans...) - on walking into my lounge room, the comment from the "all speakers sound the same" lady, was "is your system playing in mono?" (we shared a home for some years, so she was familiar with the sound of the system).... she clearly heard a difference, she clearly thought it sounded worse, but could not pin down the reason... yet she continued to claim that all speakers sound alike...

There are a bunch of things that the Klipsch does better than the Quad's - measurable things - bass extension, high extension, maximum SPL's... but, they have flaws too... distortions and resonances - they still sound like "boxes" - and the distortions and resonances sit in the critical midrange, where the ESL57's provide a purity well nigh unequalled by others....

Can we measure these things.... yes. Do we have a relatively easy way of mapping those measurements to heard/listened experience.... no.

So how does one select a speaker based solely on measurements? - you can't! - all you can do is narrow down the choices, by using measurements to rule out the "nasties".... after that it gets difficult.

 

[dc655321]

Sorry for the clickbait-y title. It was done tongue in cheek. Obviously, the answer is generally "a lot".

But, I am wondering about a specific situation...

Say that you have a well-treated listening room, an existing "2.2" system, and either already use some software solution, minidsp+dirac live, or trinnov for sub integration and correction....

And, you're generally happy but considering a speaker upgrade (mains/satellites only) partially to solve a specific problem and partially because "it feels like time".

Now consider that 2 specific speakers are kind of bubbling to the top of your research, at different price points (say, one is double the price of the other), both well-regarded...but only one of them has been thoroughly measured and you can't find a spinorama for the other. You're confident they're both good...but not directly comparable via measurements and not both available to demo at the same place, and subjective impressions of both are all over the map.

Short of ordering both and going back and forth (which is always a headache)....

1. How would you go about comparing?
2. Is it reasonable to rely on the correction to even-out probably minor "defects" in one?
3. What else might make the decision for you?

I am obviously in this exact situation. I'm new to the forum, but I've been reading for years. I can post more details if necessary, but I'm honestly more interested in answers to those big-picture questions. And, yes, I'm seriously considering having one drop shipped to amirm just so I can see the measurements of the other speaker....I mostly want to figure out how that's going to work with the shop/vendor's return policy before I contact him about it.

Any thoughts?

It may help responses if you provide specifics on models you’re considering. Kinda odd you didn’t lead with that…

 

[MattHooper]

His ad-hoc listening is no guarantee of anything. The only reliable information comes from measurements. Everything else is a distant second from that point of view.

Agreed that measurements will produce the most reliable information about a speaker, especially in the sense that it's objective information that can be accessed and useful to anyone (vs just a subjective opinion). And that can certainly be very helpful for someone wishing to use measurements to help winnow down their choice. Especially if they can't hear a speaker before purchasing it.

However, I don't think in person listening auditions are necessarily a "distant second" in regards to an individual auditioning speakers in person.

I've heard speakers that measure very well that didn't grab me when I auditioned them in person, and visa versa.

One could assume it may be due to the vagaries of a set up, the room, one's mood, other psychological cues (e.g. how a speaker looks) and that's all fair. On the other hand, when I audition a speaker I usually do a pretty thorough listening test, moving the speakers and listening position around (usually possible), listening from distance, from the sides, standing, squatting, close, far, etc. By the time I'm done I have a pretty good subjective impression and I've never been surprised by the sound of a speaker I auditioned elsewhere once it ended up in my home.

I've made very satisfying purchases from listening tests, where measurements thus far have not been as helpful (that is helpful in making the ultimate decision on whether to purchase or not).

 

[Affwhire]

Thank you for the responses. I'm kind of impressed with the speed and kindness, even with the disagreements.

It may help responses if you provide specifics on models you’re considering. Kinda odd you didn’t lead with that…

The omission was intentional. I wanted to get big-picture ideas without biasing people based on preconceptions of the particular speakers.

They're Dynaudio Lyd48 (which have not been measured) and Neumann 310s (which have). If you look at the reviews that are available for them (or similar speakers).....either I'm reading things wrong, or you'll see my dilemma. The 310 measurements are great. Other reviews of the Lyd48 seem like they should correct some of the issues of the Lyd5 (which look worse than the 310 but have a higher preference score with a sub for some reason)...but I don't know that. I also can't figure out what makes the KH 80s come out "above" the 310s in the preference score. 310s look better to me.

I also haven't heard either of them in person, but I have heard 120s a long time ago and was very impressed. That being said, the last time I went speaker shopping, I just ended up buying several sets in my price range and returning almost all of them. To be honest, I went with "close enough after correction and in budget", partially because after 2 months of dealing with speaker placement tweaking and returns, I was freaking sick of it.

I'm leaning heavily toward the 310s. I'm just hesitating based on cost for some reason (I have about that much money in treatments, and that was a comparatively easy decision).

I use measurements for shortlisting, but if he cannot correlate measurements with his preference, then it's just guesswork. At least by listening to a few speakers he will have a better idea of what he prefers.

I do have an idea. It's just hard to describe (other than "accurate") without using hi-fi weasel words. Frankly, trying to talk in those terms is difficult. It seems like everybody has a different definition of them. I do have an idea of my "target curve", but it's heavily based on things I don't like. I don't like the mid-forward sound of the ATCs I've heard, certainly not as bright as the B&Ws (though the 685s in my untreated living room sound "fine" for TV), definitely not as scooped as the Klipsch. It's really close to the published B&K curve. I want enough "oomph" for lack of a better word, and don't want them to be forgiving of bad masters/recordings....though I still haven't figured out how a speaker is supposed to actually do that....it's almost just a thing hi-fi people say as far as I'm concerned, but when I hear people describe a speaker as "forgiving", I tend not to like them. I haven't liked any Genelecs I've heard, but I've never set them up. Knowing what I know now about rooms, I wouldn't expect any of those setups to sound right, so I don't really want to write them off. But, I'm not excited about them either. It's hard to un-hear the harsh, fatiguing sound that came out of them.....it wasn't bright, it was just wrong.

I wouldn't count on it. If the speaker has discrepancies between the on-axis and the off-axis response, an attempt at correcting the direct sound might just exaggerate problems in the early reflections and get you back to square one.

That's a great point that I hadn't thought of.

I wouldn't buy anything unmeasured unless the company provides proper ones (e.g. Genelec, Neumann, Harman, etc.). You are taking a chance otherwise. As you say, if you drop ship it to me, I can measure it and if it is within return window, you can choose to return it.

Any idea what your turn-around time would be if I decide to go that way?

 

[sweetchaos]

I would like to see Dynaudio Lyd 48 measured, officially.

Here's the on-axis and directivity graphs, that we have:

https://dynaudiodata.blob.core.windows.net/media/6539/spec_for_lyd48_2017pdf.pdf

 

[Head_Unit]

I've never experienced a situation where 3x the budget needs to be expensed for a better speaker. Most of the time just going to the next model in a speaker line is enough to get a substantial change never mind choosing from another brand. Sometimes the same price, sometimes $50 more or a hundred more is all that's required.

Well as I said it's a vague guideline. But let me redefine that I don't mean just "better"-yeah if you move up from a Focal 906 to a 936 it is better (although way more expensive come to think of it). I'm talking about moving up to a whole different level of clarity, not just better but WAY better to a point of transformation. Not a sideways move just better in some ways, but head and shoulders above the other. For instance, in one setup I have some Wharfedale 11.4. At another location, Focal 936. The Wharfedale are quite nice, however they just don't have the je ne sais quoi that the Focals do. 936 are just more transparent and clear. A more extreme example is some $2k-ish Magico we listened to at Brooks Berdan. WOW no idea that "Walk On The Wild Side" had so much sound inside, it was like a totally different song.

 

[dlaloum]

Well as I said it's a vague guideline. But let me redefine that I don't mean just "better"-yeah if you move up from a Focal 906 to a 936 it is better (although way more expensive come to think of it). I'm talking about moving up to a whole different level of clarity, not just better but WAY better to a point of transformation. Not a sideways move just better in some ways, but head and shoulders above the other. For instance, in one setup I have some Wharfedale 11.4. At another location, Focal 936. The Wharfedale are quite nice, however they just don't have the je ne sais quoi that the Focals do. 936 are just more transparent and clear. A more extreme example is some $2k-ish Magico we listened to at Brooks Berdan. WOW no idea that "Walk On The Wild Side" had so much sound inside, it was like a totally different song.

I recall, when I worked in "HiFi", my rule of thumb became.... it costs 2x the price for an audible but slight improvement... and it takes 3 audible but slight improvements to make for an "Ahah!" paradigm shift, step up.... so more like 8x the price - but that was based on official published MSRP's

The base level of quality at the bottom of the market, has, I believe, gone up - especially in electronics.... but to some degree in speakers too - and my old rule of thumb of 8x the price may be too conservative... (it may be 4x nowadays ?).

But I certainly agree with the principle!

Once you get to a certain level of performance.... the next step up gets seriously cost prohibitive.

My benchmark for several decades, has been Electrostatic speakers - particularly Quad & Martin Logan CLS... although I no longer own a set.

Taking a step up from the level of performance provided by such speakers (room and WAF permitting) becomes a seriously expensive proposition.... although there are many sideways steps/options with differing compromises.

Hearing the speakers you are considering is absolutely critical - I would absolutely not consider a set of speakers "unheard". Exception might be when moving within closely related speaker families.... where you are already familiar with the family sound.

 

[Holmz]

…
Short of ordering both and going back and forth (which is always a headache)....

1. How would you go about comparing?
2. Is it reasonable to rely on the correction to even-out probably minor "defects" in one?
3. What else might make the decision for you?

…

Any thoughts?

I would choose the one with:

1. The best (lowest) distortion
2. The least compression
3. The best step function response
4. and/or best impulse response.

So basically I would not consider many speakers without a measured set of specs.

The question of, “ Is it reasonable to rely on the correction to even-out probably minor "defects" in one?” relies on the directivity index being flat.

I wouldn't buy anything unmeasured unless the company provides proper ones (e.g. Genelec, Neumann, Harman, etc.). You are taking a chance otherwise. As you say, if you drop ship it to me, I can measure it and if it is within return window, you can choose to return it.

^this^

 

[HooStat]

Just from a philosophical point of view, I think it is better to get well-designed speakers and from those, use other factors to make a final decision. By well-designed, I am referring to the content that @Amir and @hardisj provide. That doesn't mean it has to have the best "preference score". Just good enough for your purposes.

It is very similar to buying a DAC with the lowest distortion -- it isn't required for sound quality but it shows the level of attention paid to the design.

 

[dlaloum]

Just from a philosophical point of view, I think it is better to get well-designed speakers and from those, use other factors to make a final decision. By well-designed, I am referring to the content that @Amir and @hardisj provide. That doesn't mean it has to have the best "preference score". Just good enough for your purposes.

It is very similar to buying a DAC with the lowest distortion -- it isn't required for sound quality but it shows the level of attention paid to the design.

So should more people be looking at Electrostatics?

Have "dynamics" caught up to "electrostatics" capabilities in terms of distortion?

Quad ESL-63 loudspeaker Larry Greenhill

Larry Greenhill reviewed the Quad ESL-63 US Monitor in February 1989 (Vol.12 No.2): Electrostatic loudspeakers (ESLs) have always held a fascination for audiophiles. The Jantszen, Beveridge, KLH, Acoustat, Stax ESL-F81s, and Servo-Static models in the late 1960s and '70s, as well as the...

www.stereophile.com

"This design meant near-perfect phase coherency, as shown by Quad's show-stopper demos in which two squarewaves, out of phase with each other, are fed to two Quad speakers. A microphone placed between the speakers shows that the two signals cancel out completely, suggesting very low distortion in the speakers."

Yes I had to give up my all electrostatic surround setup.... ... WAF

But we now have many more exotic drivers in the marketplace... have electrostatics met their match among more affordable and easy to live with speakers?

The Gallo's I now have are very "electrostatic like" .... but they don't achieve the transparency that my old Quad's do/did.... although they have much more bass (!). Distortion is around 1.5% in @50Hz (when BiAmped), 0.6% in the mids and 0.8% in the highs .(so averaging circa 0.7% probably...)

Quad ESL63 has distortion at around 0.03% average (reasonable for an amp, exceptional for a speaker!) - with max at around 1% at 40Hz

The Genelec 8351b - very well regarded - has circa 0.04% THD @1kHz (96db) - but circa 2% at 50Hz, and circa 0.3% @ 10kHz - seems to average around 0.15% ... (all read from Amirm's measurements @ 96db) - the Gallo above was measured at 100db, Quad ESL I am not sure at what SPL...

Martin Logan Motion 4i measured here as well, come out at around 0.15% average as well

So from a cursory glance at measurements - one would suggest that ESL's still provide a closer approximation to a perfect speaker? - My Gallo's seem to be among the lowest distortion speakers... and still they have about 20x the distortion of the ESL's.

Hmmm - this topic might be worthy of its own thread...