First of all, a live concert is not "sound reproduction in a room". What concert is performed in a "room"? Rather it will be in a venue such as a concert hall where many people can gather. This will have very little in common with the acoustics of and sound reproduction in a "room" like is commonly used to mean "in a domestic listening space" e.g. in your home. Acoustic coverage of, and acoustic treatment (e.g. diffusion, adsoprtion) in, these different spaces is not at all the same.
The only thing that listening to an actual performance will tell you is what that exact performance sounded like in that acoustic space. But can you then go back home and replay that exact performance (a recording of it?) in your home on your home hifi system? Probably not. So any comparison is apples to oranges, especially since there is no guarantee that even IF you captured that exact performance in a recording that recording would be accurate to what you heard where you were seated! I only know of Siegfried Linkwitz, who attempted to make a recording of e.g. the SF Symphony and then went home to play it back on his speakers to judge their accuracy. He was 1 in a million in that way.
You can go on and on down this rabbit hole. Let's face it, music playback is just "for fun". There is no absolute standard or goal to reach, and no perfect reproduction chain exists including the recording end of that chain. So you might as well just find a setup that give you pleasure when you listen to whatever recordings you happen to like. For YOU that will be about as good as it gets. If there are measurements of loudspeakers out there that are able to point you in that general direction, and techniques retarding how to best deploy the loudspeakers and make them sound the best in your own personal listening space, then use these to your advantage.
Hi thanks for the valuable advice and i agree with you Many years have passed since my last pop or rock concert
I stopped to attend classical concerts because i am lazy and i do not like people around me coughing or making noise
But my point was different When i had the opportunity to listen to a real full range speaker i could not go back to small bookshelves
I lost any interest in them They are severed
When i play a Toccata e Fuga i like speakers that make the walls trembling Not speakers that tremble
I cannot understand how people can be satisfied by small speakers Maybe they do not have a clue about what they are missing
and they are missing a lot
I start to think that a very good 10" is a minimum requirement for a realistic experience
To hearing the bass from a pair of JBL L166 was a revelation An audio life changing experience
Another one was when i first listened a 3d soundstage from dq10s driven by Audio Research amps
Right ! ok i have been carried away by my love for lab instruments
In reviews i like to read the listening impression but i get excited only by the lab reports
To correlate lab measurements and listening impressions it is as difficult as it is fascinating
i do not understand why this goal seems to fascinate only me
I remember an article by Lynn Olson about the Ariel project
in the end he supported the choice of selecting a driver based on listening tests
it is something that with all the good will I cannot accept
What you experience is a speaker with so much headroom that there is no compression on the drivers at all. You can clearly see hat on thest like done by Erin (google him) who always does compresson tests on speakers, and big woofers, compresson drivers and so are in a hifi situation never stressed at all.
A 8" woofer or smaller is vey limited in how much air it can move and so how loud it can go before distortion kicks in, idem with dome tweeters. Compression drivers and big woofers have way higher limits than dome tweeters and small woofers. They don't necesairly go lower or so, but do it with much more ease. With compression drivers it also includes a controlled directivity (when done right) untill you go to the woofer. I love fullrange drivers (the right kind at least), but that is their weak point, they have little to no headroom. That's why i did build some bigger systems for others and are now busy with my own variation (now that i have some experience with those). JBL is one of the best in that with their bigger systems (Cinema series or the 43xx series) today, but not the only one who does such things.
it's not at all "only you".
it is just very difficult (impossible) to find a general consensus.
There have been scientific studies to find a "general" preference for how speakers should sound.
check out Floyd Toole, here is an interview with Erin.
but often we are just discussing what we like and we try to find "scientific" explanations why we like it - or in the worst case we try to justify or generalise our opinion by pointing at a graph that seems to confirm our very subjective feeling.
it's nice to realize we like truth (low distorsion?) over everything else. but "truth" is a very vague concept in psychoacoustics and may quickly become very moralising.
"I like red better than blue" is a personal, subjective preference.
"I like red better than blue because of the longer wavelength of red light" is a pseudo-scientific attempt to "rationalise" a subjective preference.
A good conclusion would be:
"red light has a longer wavelength than blue. subjectively I like red colour better than blue. Thus it seems I prefer colour with longer wavelength."
Thank you very much for the very helpful explanation This behaviour could be predicted on the basis of distortion measurements
I have seen a case of a small speaker that was showing very low distortion at 96dB/1m above 200 Hz
but huge levels of distortion below that point
My point is that there is no need to play a pipe organ track to predict that the result will be disastrous
I am pretty sure that at JBL they perform distortion measurements to assess their speakers performance
Why i cannot find them in the hifi speakers datasheets ? actually in some there are
https://www.madisoundspeakerstore.c...6-6.5-ceramic-cone-woofer/midrange-neodymium/
Last edited:
They are not in the datasheet because many don't care enough, but there are tons of external tests done by Erin en Amir from ASR and others that show the differences. The JBL 4367 was tested by Erin: https://www.erinsaudiocorner.com/loudspeakers/jbl_4367/
And there were more tested, but you need to search those tests on independent sites, not on sponsored audio mags (who are bias to their sponsors) or on the site of the brand itself. And many reputed speakers don't measure that good actually. Especially small bookshelfs have a hard time to give bass and are very often high distortion below 100 or 50Hz. Some do it right like the Neumann KH150, but even then they run into physical limits that no small speakerset can avoid wthout a relative big subwoofer.
I myself tend to love 10" or 12" woofers mostly. They fit my listening position (fairly small room) and some still go from very low to clean relative high to meet a good compression driver in horn somewhere reasonable on relative high volume for that space. Big woofers mostly demand a 3 or 4 way, that cost a lot more, and is way more complex to design and build (be it active or passive filtered).
And there were more tested, but you need to search those tests on independent sites, not on sponsored audio mags (who are bias to their sponsors) or on the site of the brand itself. And many reputed speakers don't measure that good actually. Especially small bookshelfs have a hard time to give bass and are very often high distortion below 100 or 50Hz. Some do it right like the Neumann KH150, but even then they run into physical limits that no small speakerset can avoid wthout a relative big subwoofer.
I myself tend to love 10" or 12" woofers mostly. They fit my listening position (fairly small room) and some still go from very low to clean relative high to meet a good compression driver in horn somewhere reasonable on relative high volume for that space. Big woofers mostly demand a 3 or 4 way, that cost a lot more, and is way more complex to design and build (be it active or passive filtered).
As Adason mentions, directivity is definitely in the mix. In fact, it makes the idea of "two speakers with the same frequency response" kind of silly -- what frequency response is it that is he same? The on-axis, the average in the listening window, the overall average? The anechoic response or the full in-room with all reflections response with some amount of smoothing -- or without any smoothing? It's hard to imagine they'd get two speakers of even the identical model to measure the same for that last case.
Because you certainly can't just go by the on-axis anechoic response unless the listener's head is in a vise and he's in an anechoic chamber. Because off-axis sound (with unique response shaping for every pair of angles) will reflect in a room and also finds its way back to the listener.
Because you certainly can't just go by the on-axis anechoic response unless the listener's head is in a vise and he's in an anechoic chamber. Because off-axis sound (with unique response shaping for every pair of angles) will reflect in a room and also finds its way back to the listener.
I think the following parameters correlate well between measurements and listening impression:
Cumulative Decay Spectrum (waterfall plot) - should be as abrupt and even over the audio band as possible
Minimum Group Delay
Flat frequency plot - it is best when the frequency plot is 1 to 2 dB/decade falling towards the top
Directivity - I am not sure what is best: constant directivity or gradual beaming towards the high freq.
Distortion, IMD - of course
Cumulative Decay Spectrum (waterfall plot) - should be as abrupt and even over the audio band as possible
Minimum Group Delay
Flat frequency plot - it is best when the frequency plot is 1 to 2 dB/decade falling towards the top
Directivity - I am not sure what is best: constant directivity or gradual beaming towards the high freq.
Distortion, IMD - of course
cumulative decay spectrum and flat frequency plot both point to the same issues, which are resonances in the passband of the driver. If the spectrum is even and without resonances the frequency band will be flat an distortion will be low.
Directivity also plays a roll and constant directivity is the best but very hard to do over the whole passband of a speaker, You need cardioid woofers with the same dispertion of the tweeter to have a full constant directivity and almost nobody archives that.
But a gradual beaming is much better than an irregualar dispertion, and if a big part (all but the bass) is constant directive and you have good acoustics and/or room correction you get +95% of the way there.
But the reason many like big woofers, even for hifi on normal levels, is the lack of distortion in the bass, the other factors play less a role in that judgement, and many big drivers beam faster than smaller ones and have resonances way lower in the freuency range, so that is there disadvantage. But the lack of distortion makes up for that for many.
Directivity also plays a roll and constant directivity is the best but very hard to do over the whole passband of a speaker, You need cardioid woofers with the same dispertion of the tweeter to have a full constant directivity and almost nobody archives that.
But a gradual beaming is much better than an irregualar dispertion, and if a big part (all but the bass) is constant directive and you have good acoustics and/or room correction you get +95% of the way there.
But the reason many like big woofers, even for hifi on normal levels, is the lack of distortion in the bass, the other factors play less a role in that judgement, and many big drivers beam faster than smaller ones and have resonances way lower in the freuency range, so that is there disadvantage. But the lack of distortion makes up for that for many.
they do not care because they are not important ? that is my question
which measurements can provide information more useful to predict the speaker sound ?
i knew of Amir only Strangely even he performs normally distortion measurements he tends to downrate their importance
imho a big mistake They are by far the most important
And he does not perform IMD tests Even more telling The test is simple
Two impulse at the same time one at 40Hz the other at 300Hz for a woofer at 100dB/1m
then look for peaks around and between the test signals
some weak woofer will even break
this explains a lot of things indeed The mags provide measurements that do no create problems to the marketing dpts
i have the same feeling Actually the distortion graphs show this very clearly even for me
Regarding the kh150 they are great even at low Hz for a 6.5" woofer I am pretty sure they are using a beast of a woofer
in the end drivers make the sound
very interesting and thank you Actually my next goal is to understand how to select great 10 or 12" woofer for a project of mine
that is a sat above bass box concept
for the sat i have some small aluminum speakers from Teac and Kef to use as cabinet for different drivers options Maybe even small fullrange
But the woofer is the critical corner stone
I used a 10", the Scanspeak 26W8534G00 for it in a 77L sealed cabinet. But i only cross at 300Hz to a Mark Audio Alpair 10.3. Others suitable are the SB34NRX75-6, the Faital 10FE330, The Faital 12FE330 and the Faital 12RS430, all depending on how high and low that woofer has to play, and how big the box has to be. What kind of crossover you use may also play a role (especially when you use lower order passive crossovers). And i'm sure there is more that i don't know well enough to know they may be suited.
You want an relative low fs, high sensivity, enough but not too much xmax (not a hardcore subwoofer) and an flat response untill about 2-300Hz at least above the crossover frequency to the top if you stay in that same frequency region as i. You don't want that the resonances still come trough the top. Heavy resonances (mostly metal or ceramic cones) needs to be higher above the crossover frequency than damped (mostly paper or other fiber cones) resonances. A low LE also helps. A flatter cone helps also for better dispertion than a deeper cone (they beam faster).
You want an relative low fs, high sensivity, enough but not too much xmax (not a hardcore subwoofer) and an flat response untill about 2-300Hz at least above the crossover frequency to the top if you stay in that same frequency region as i. You don't want that the resonances still come trough the top. Heavy resonances (mostly metal or ceramic cones) needs to be higher above the crossover frequency than damped (mostly paper or other fiber cones) resonances. A low LE also helps. A flatter cone helps also for better dispertion than a deeper cone (they beam faster).
Last edited:
Danny at GR Research makes money selling parts kits. IMO, you just need to realize that all opinions of his should be filtered through that lens of what benefits him financially. If he didn't call polyester caps and iron core inductors "cheese" maybe the market for clariity caps and 14ga perfeect lay coils would dwindle...and don't forget the tube connectors.
I am a firm believer that not just FR but cumulative spectral decay response + dispersion characteristics go a long way as to a speakers sound.
Distortion is obviously a given.
Distortion is obviously a given.
To find the correlation between measurements & listening, you must first make the 'listening' a measurement. ie it must be a repeatable experiment. You do this with DBLTs.
When you start doing this, you quickly find certain unwelcome truths. eg
- most (all?) wannable Golden Pinnae are deaf (give random results in DBLTs).
- HiFi Reviewers are mostly deaf compared to the 'man in the street'. (There are a VERY few exceptions)
- The 'woman in the street' is more perceptive than the 'man in the street' so pay attention when your girlfriend, wife, mother says, "I don't like this new one as much as your old one."
- Good speakers are preferred by ALL types of (perceptive) listeners ... this includes, recording engineers, speaker designers, teenage pop lovers, classical musicians to da guy who designs his own microphones and insists on using his own recordings done with them. From this, I conclude there is such a thing as a 'good speaker' ... one which is preferred by everyone who can actually hear the difference reliably.
- Above a quite low standard, THD has little relevance. There are speakers with high measured THD that perceptive listeners say sound low distortion.
- A nice frequency response and CDS (waterfall) is one of the hall marks of a good speaker but not the only one.
But there is a clue. The Room Interface Profile is a naive 1D signal path which somehow incorporates the total magic. You can clearly hear it ... or the lack of it. If I am lucky, I might figure out how to 'measure' this signal path and grok da magic before I drop dead.
But how do you deal with the unwashed masses of deaf Golden Pinnae? No point designing stuff to suit their ears cos they are deaf. For them (including most HiFi reviewers), you say loudly and clearly that your stuff is Hand Carved from Unobtainium and Solid BS by Virgins.
What about a wooden box filled with sand calling it a clean earth point?
Thank you very much
Very interesting
Being used to small 2 ways listening to a big woofer was mind opening
CSD is a very telling test about the behavior of a speaker
But also imd tests are important
Now the question is where to cross a 12 inches woofer
Which cut frequency I mean
Hi thank you for the very helpful advice
His test of speaker cables connected to a tuner was puzzling
A speaker is not a tuner
But my very interest is in to find a method to select raw drivers
I guess that fr CSD and distortion together can provide most of the information needed
it depends onto what you cut and what 12" woofer. There are that I would not cut above 95Hz, while others are good to about 1kHz or even above.