and listening to something having the artificial stuffs or harshness in mixes in a revealing speaker can make it sound bad at times, I do think it is the reason why some prefer particular house curve say the BBC dip or so, coz their favourite tracks comes from the era mixing with those speakers, but then I do hope some time in the future we can have a real transparent speaker in all studios and at home so one can really know is the speaker honest or not.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Do We Want All Speakers To Sound The Same ?
- Thread starter MattHooper
- Start date
OP
- Joined
- Jan 27, 2019
- Messages
- 7,336
- Likes
- 12,302
- Thread Starter
- #262
For this reason I audition speakers with a wide range of different recording qualities, from excellent to "problematic" or thin sounding or whatever. Some have distortion I'm familiar with, so it's interesting to see how each speaker handles the distortion. Ideally, I can hear it, but it's not emphasized in a way that puts me off the music.
usually when something is unexplainable from any measurements, where distortion, transient response, IMD and frequency response are the same, are more or less sighted, brand or confirmation bias, this sort of coincident have come in way too often in all sort of subjective reviews, like for expample, Audio-GD dacs, when it's ok to good sounding, and you got told that this brand has X attribute, the subjective opinion would go to that side, not to say that for paid reviewers usually more bias comes.
Don't get me wrong, from the measurements the JA speakers are good, but then when it just magically sound better in unmeasurable way, usually it's our brain fooling ourselves, when you can't hear the wiggleness in the FR, why do you think that minor, beyond measurement capability THD/ IMD can be percieved?
thing is, how you know what the distortion really sound like? without a standard for transparency, you just don't know, much like saying colour, without a well calibrated monitor, with a standard colour chart, how do you know the blue in the screen is really blue, or off? By experience 99.9% just don't do it right, in photography, standard calibration is done for serious jobs, so when in audio, why it just becomes different, why not the in record FR profile of music should be just reproduced the closer the better
Most of the songs I test with don't sound BAD on a good speaker, some of them even sound good, but they do sound noticeably worse on bad ones. Here they are if you are curious. https://open.spotify.com/playlist/0NAPaCXMPgEAkASn4FKpko?si=36b61456d7da4a6b
While listening to this list just now, I realized my APOEQ was shut off! It works!
Last edited:
similar experience for me, and my thought is simpler: if I can't hear distortion in pure tone generator and in the sine sweeps during my measurement runs, the speaker should be fine, and that in such case, when some pieces of music sound broken, distorted, noisy etc. it should be the track mix itself is broken, not the speaker, much like in old systems the vinyl don't have the abundance of noise in background, whereas the E version of it in a transparent dac usually just make it sound noisy, that is the recording itself is broken, not the high resolution dac.Most of the songs I test with don't sound BAD on a good speaker, some of them even sound good, but they do sound noticeably worse on bad ones. Here they are if you are curious. https://open.spotify.com/playlist/0NAPaCXMPgEAkASn4FKpko?si=36b61456d7da4a6b
While listening to this list just now, I realized my APOEQ was shut off! It works!
Of course in the end we all listen to the music in the recordings, not listen to the speaker, so nothing wrong when prefer the speaker itself with the not transparent aspect which happen to mask the defects in the mix.
One might agree or disagree, but after all the subjective is king debates I do still believe the speaker should add/alter the sound as little as possible, while personal preference or even "fixes" on certain tracks, should be done either run it through some SW to remove the noise, or just EQ the tone curve in a click
Dumdum
Senior Member
From my work in cars I can tell you freq response on and off axis is only one aspect of how a speaker sounds and whether it pleases your ears or not
You can eq two cars the exact same in room eq wizard and they will sound different because of the environment
Is the midbass snappy, does it have impact, is the midrange dry and revealing or warm and non fatiguing, does the treble have shimmer and air or is it very neutral but accurate
The above are all characteristics of the drivers and there relative alignments in the enclosure
For example in my car I have to dial in some extra around 60-120hz to get a realistic sense of drums and impact, yet I did a van the other day that didn’t need that and it had very good snap and impact without being overblown, if I tuned my car the same it would sound flat and a bit lifeless
My point is that the sound of two speakers can measure the exact same response and be identical off axis also, but they can sound very different, the design, drivers and finally the environment and listening position relative to the drivers can make a huge difference to the sound and all of that is why you should ideally spend a fair bit of time and effort picking your speakers for your room
So do we want all speakers to sound the same… for me measurements can only tell half the story (and definitely can’t take the environment into account
You can eq two cars the exact same in room eq wizard and they will sound different because of the environment
Is the midbass snappy, does it have impact, is the midrange dry and revealing or warm and non fatiguing, does the treble have shimmer and air or is it very neutral but accurate
The above are all characteristics of the drivers and there relative alignments in the enclosure
For example in my car I have to dial in some extra around 60-120hz to get a realistic sense of drums and impact, yet I did a van the other day that didn’t need that and it had very good snap and impact without being overblown, if I tuned my car the same it would sound flat and a bit lifeless
My point is that the sound of two speakers can measure the exact same response and be identical off axis also, but they can sound very different, the design, drivers and finally the environment and listening position relative to the drivers can make a huge difference to the sound and all of that is why you should ideally spend a fair bit of time and effort picking your speakers for your room
So do we want all speakers to sound the same… for me measurements can only tell half the story (and definitely can’t take the environment into account
If I didn't misunderstood the anechoic off axis i.e. directivity plus the environment (room, car), is the thing you are actually referring to, in room measurement on/off axis can't possibly be EQed to be the same throughtout the space with different speaker precisely due to anechoic on and off axis differences plus the speaker placement difference, above schoreder frequency our ears are smart enough to distinguish mostly from direct sound vs reflected sound, so in the car example, if you measure in driver seat for FR 1, and another seats for FR2, 3, 4 etc. if you EQed FR1 in car A to FR1 in car B, the other seats would have wildly different FR, that's why anechoic response in determine relative speaker difference is the thing to do, not in room response.
But true the measurements along won't be able to be converted fully into actual percieved sound in a room, just a broad estimation,
Dumdum
Senior Member
I tune for one seat for car audio competition, time alignment, eq, crossovers etc are all timed for the drivers seat
It doesn’t sound bad from the passenger seat, but the centre is in the left corner of the screen and the left stage is very compressed due to it being timed for the right front seat (uk rhd)
Two seat is doable now with some car dsps having upmixers but you have to conpromise Sq and staging and imaging for both seats and it will never sound as good as it does from one seat
Effectively I do to cars what you do in a house with a single seat for critical listening and an equilateral triangle… shift out of the seat and the stage is not correct with timing and response mismatches
that's for sure, and since the driver's directivity in cars are a lot more constrained compared to any home speakers, it will vary a lot more from car to car, in cars usually I just don't bother EQ too much, as it's really difficult to get the schroeder frequency point right and the reflective surfaces are way more complicated to home audio, plus the main point: when driving I prefer listen to engines and road noises so I focus on driving, not relax in music.
Please show me that you had EQ-ed them to the exact same frequency response. I assume you had used more that 100 filters with filter Qs of as high as 15 and linearized phase afterwards?
My point here is: you probably have not even succesfully EQ-ed linear distortion to be the same. I can hear differences of low decimal counts depending on the position in the band and obviously filter width.
Dumdum
Senior Member
If you measure with a spatial average it’s fairly easy to see where the measurements begin to vary so I tune with a single mic location upto that point (I set myself a limit of 350hz and use smaart to sort the impulse response and also mid to midbass, midbass to front sub and front sub to rear sub crossovers and phase
It’s quite impressive when you have sub bass with subs in the boot but it comes from the stage, speaker height is irrelevant as long as it’s all in phase and well setup
It’s why I don’t understand the parroted get tweeters at ear level in houses (well I do when the tweeter is firing parallel to the ground…) but why not raise drivers up to get better responses in a house if it gives better results, we don’t hear height very well so eye/ear level is where the stage ends up if drivers are all tuned to be as one effectively… you don’t notice the centre channel is under the tv for example… it sounds like it’s where you look… or at least mine does
I am soon to make a set of floor standers for my place which may end up as being hung from the ceiling to get a better reflection path for the left driver that matches the right driver far better (I have a bannister rail in my front room around the stairs in the rear left corner… it’s better with a hand rail and spindle vs when I first moved in and had a solid plasterboard wall around the stairs…) the room feels bigger and the acoustics got better also… but I had hoped the spindles would diffuse the reflections a little less than they do, it still sounds better than any houses I’ve heard but I’ve not really heard good two channel in a house
My next experiment is to see if I can diffuse some rear reflections with absorbers near the floor standers as the window behind the left and the double door Georgian balcony behind the right contribute differently and also skew the stage a touch
Dumdum
Senior Member
Talk about complicating things… no I have not looked at distortion in a car because it has far too many reflections and room modes to even contemplate doing it, but that is one of the things that will effect the sound… we sit asymmetrical in cars
As for eq, you seem to underestimate the power of modern car audio dsps, my ten channel dsp has 60 bands available per driver/channel if I want them (I don’t ever use anywhere near 30 generally), and they are eq’d to roughly +/- 0.5db with a spatial average and variable smoothing in room eq wizard if required, but having gone to those lengths before it is not required in a vehicle due to reflections
There is zero point in very fine (1/48) smoothing above 2k give or take
We use house curves and aim to tune to those, but a single house curve can sound different in one car vs another
I can guarantee you if I tuned two different cars with unique installs to the same house curve they will likely sound different, drivers response to stimulus can vary, distortion etc all contribute
My point is to the original post… just the same as the car we can take two different branded floorstanders with identical response, distortion etc but they will sound different potentially due to drivers, enclosure etc or are you saying you’d expect them to sound identical based on measurements alone? And that’s before we introduce them to an environment like a front room, so let’s assume they are in an anechoic chamber… should you be able to predict the sound with currently available objective measurement techniques?
I’m saying measurements do not equate to the same sound characteristics, you can say if speakers should sound reasonable or good with a degree of certainty, but listening you can find nuances and they will often sound subtly different even in an anechoic chamber, that was the original question was it not?
Do we want all speakers to sound the same, my answer is no, but we can gleen an idea of whether they should sound good or bad from measurements… yes, but if you get several of Amirs reviewed speakers that scored very similarly or measured similarly and have different sounding speakers
For me measurements are clues not cut and dried statements
Last edited:
actually measurement includes the distortion and dispersion pattern, i.e. directivity, most of the on axis flat speakers will sound different in room, coz the directivity pattern is different, and in room response, though looking similar, will say differ in the downward slope, and off axis response will result in different reflected sound, so they are not measuring the same even on axis is similarly dead flat.
also even both with near perfect directivity, the beam width, and are they cardiod or not, all contribute to the final sound. these are the main reason for different cars to sound so different even given the same drivers being installed, the final combined FR can be EQed to be similar, but you can't EQ them to behave the same in reflected sounds
Dumdum
Senior Member
So you keep writing… my point is simple… will all these measurements tell you with 100% certainty that the midbass will have impact or sound a bit flat? I don’t think we can say that… so back to the original question… no is my answer
Will two identical speakers in all your measurements sound identical? My answer to that is also no in my opinion
I just tried to give a bit of background and my take on it… yet you seem to want to dissect it massively without a point regarding the original question
I didn't do car EQ, but in room, usually the decay time +distortion do tell you wether it will sound a bit flat or muddy vs impactful.
Dumdum
Senior Member
well I don't actually say using the measurements alone could predict how a speaker sound in a actual room, noone can factor in the in room response plus all those reflections will sound like, my point is that those non-FR measurements all contribute to how it will sound in an actual room, but the more treated space, similar measuring speaker including on axis and directivity, do sound very close to identical, but when those variables changes? Only God could tell.
My point is: do we want all speakers to sound the same? yes, the closer to taking the speaker out of music reproduction the better, and the better it respond to EQ through good directivity, the better. If that's possible? I don't think so due to technical constrains, and for in room, it's practically impossible to sound the same anyway, just as good as it can go. I tried to EQ some poor directivity speakers, it sounded weird, but then EQing similarly well measuring speakers on & off axis in the same room, same spot do sounded identical to me.
Dumdum,you are missing the point because you make up a problem which we did not have before: your cars. A speaker's performance and differences is with good reason shown anechoic. There is no secret that every room will change the perception of the speaker that is playing in it. So we exclude it.
This thread by now has departed from its more worthwhile question, to which degree we can also trust our sighted listening, into the territories of pure subjectivism. I do appreciate Matt's inquiry: After all, I claimed to hear differences between two bit-perfect DACs and I hold my claim.
But by now, Matt's question is basically directed against the belief that the methodology behind spinorama is founded on, and we are in a rather boring culture war with little insight.
People tend to love attributing their sonic perceptions to all kind of materials. Like cables (make no difference when resistance is chosen correctly), or when one tweeter sounds metallic because it is made from metal, and the silk dome does not. As if high frequency air pressure differences would care. It is a myth debunked and there is a thread on this forum that proves it. The same goes for woofer materials and there is a by now famous case, a user who could not buy the most modern studio monitor because the enclosure was made from MDF, not from aluminium. The typical end game of this argumentation, that is driven by the passion to hear, especially to recognize sound in its differences, will lead every argument to the typical claim that the sum of differences, the spider of the woofer, the form of enclosure etc. pp. create a unique signature that makes all the difference, but this could not be proven by measurement.
The sad thing about the argument is, it is true in a way: the sum of small differences will never be normalized, which required an ideal object, and hence we will hear differences from speaker to speaker; but it systemically overrates this small differences and underrates some of the most basic features of loudspeakers. This is, for example, frequency linear distortion (which is not distortion of the drive units, btw), speak frequency response. Frequency response differences get drastically unterrated. Small differences wil make the speaker sound different. But not bad. ANSI/CTA 2034-A does not go down to this granular level. The preference score is an abstraction that hovers a a level higher, defines degrees of deviation from linear as detrimental, but not how degrees of deviation make perception different in ways that create a certain, sometimes enjoyable character. The reason why many people do not look into the small differences in frequency respone is that somehow, atavistic thinking is more attractive. This speaker is incarnated sound, and object of desire we can worship.
This thread by now has departed from its more worthwhile question, to which degree we can also trust our sighted listening, into the territories of pure subjectivism. I do appreciate Matt's inquiry: After all, I claimed to hear differences between two bit-perfect DACs and I hold my claim.
But by now, Matt's question is basically directed against the belief that the methodology behind spinorama is founded on, and we are in a rather boring culture war with little insight.
People tend to love attributing their sonic perceptions to all kind of materials. Like cables (make no difference when resistance is chosen correctly), or when one tweeter sounds metallic because it is made from metal, and the silk dome does not. As if high frequency air pressure differences would care. It is a myth debunked and there is a thread on this forum that proves it. The same goes for woofer materials and there is a by now famous case, a user who could not buy the most modern studio monitor because the enclosure was made from MDF, not from aluminium. The typical end game of this argumentation, that is driven by the passion to hear, especially to recognize sound in its differences, will lead every argument to the typical claim that the sum of differences, the spider of the woofer, the form of enclosure etc. pp. create a unique signature that makes all the difference, but this could not be proven by measurement.
The sad thing about the argument is, it is true in a way: the sum of small differences will never be normalized, which required an ideal object, and hence we will hear differences from speaker to speaker; but it systemically overrates this small differences and underrates some of the most basic features of loudspeakers. This is, for example, frequency linear distortion (which is not distortion of the drive units, btw), speak frequency response. Frequency response differences get drastically unterrated. Small differences wil make the speaker sound different. But not bad. ANSI/CTA 2034-A does not go down to this granular level. The preference score is an abstraction that hovers a a level higher, defines degrees of deviation from linear as detrimental, but not how degrees of deviation make perception different in ways that create a certain, sometimes enjoyable character. The reason why many people do not look into the small differences in frequency respone is that somehow, atavistic thinking is more attractive. This speaker is incarnated sound, and object of desire we can worship.
Could it be they had read what others had written before? And chatted at their favorite audio fair, with one another and the original producer, who had minced an original ascription in informal talks?
- Joined
- Oct 10, 2020
- Messages
- 806
- Likes
- 2,638
@Dumdum and @YSC An interesting discussion! IMHO we need to consider the following:
While I enjoy audio technology itself, I much more enjoy music and definitely wouldn't mind obsessing (even) less about the equipment!
- If we manage to find two loudspeakers with identical anechoic on-axis frequency response, identical directivity, identical dispersion and identical distortion characteristics, and we use them in the exact same ways and in the same position in the same room, they should indeed sound the same. However I've personally never seen two different loudspeakers that measure the same in all of these, so I'm not sure it this point brings any pragmatic value to the discussion.
- Assuming the two loudspeakers differ in one (or likely more) of the above listed criteria, I think it is fair to assume the audible difference is related to those (measurable) differences.
- EQ-ing two loudspeakers with otherwise different measurements (and especially if in different environments) to the same steady-state target curve will of course not make them sound the same, because they will at the very least still have slightly different radiation patterns, and humans don't just hear the steady-state frequency response that the measurement microphone measures. We hear a lot of the direct sound in the higher frequencies, and the exact borders where we stop hearing the steady-state response and start hearing the direct sound are unclear and likely variable (I assume depending on room and loudspeaker directivity).
- At least to my knowledge, there is no detailed catalogue describing which kind of measured performance results in which specific audible quality. Sure - we can sometimes predict general trends (bright - scooped - dark; wide - focused; etc.) but even that is not guaranteed with e.g. irregular/atypical responses.
While I enjoy audio technology itself, I much more enjoy music and definitely wouldn't mind obsessing (even) less about the equipment!Similar threads
