View Full Version : Level matching in audio comparisons ...
jandl100
12-10-2011, 14:57
... how critical is it to match levels precisely?
Some folks say any comparisons are invalid without precise SPL matching.
Personally, I'm sure that I can hear differences between components in transparency and imaging, for example, to a large extent irrespective of differences in level.
What say you?
If it's a serious comparison, I reckon the fewer the variables the better :)
Louder sound levels sound more "dynamic" and yield more details.
Reid Malenfant
12-10-2011, 17:48
Hi Jerry, i tend to agree with you, i reckon i can hear differences even without precise matching ;) I'd suggest that would be on my own setup & with a CD i know well.
However, i reckon differences would be even more apparent if accurate level matching was done between the differing components :)
+1 here for level matching.
A fair comparison would imo be to match input and output impedance's as well.
One should seek to remove as many variables as possible (level adjustment being one of the easier ones to remove).
Level matching is vital to ensure you are hearing the real differences not ones induced by an increase in volume.
It has been shown time and time again by people Like Prof Dianna Deutch of University of California in San Diego, that a difference of little more than 0.3dB can completely fool people into thinking one sound has better quality than another.
If it's easy enough to accurately implement matched levels then you might as well do it but maybe it's not so simple. It depends on the type of comparison you are making. If you compare the first 10 seconds of a track or maybe even a whole track and then swap kit, probably here level matching is fairly important. If you are doing longer term listening to work out which equipment is say musically more satisfying for you then level match matching may be the last thing you want. When systems don't exhibit enough detail or dynamics it's tempting to turn them up louder to give the impression they are sounding better. For longer term listening comparisons I suggest using whatever level sounds "right" for the system config you are listening to.
I would say its vital too. The louder the level the greater the difference between silent and max loudness. Effectively louder level it has greater audiable resolution so should highlight more detail. Audio comparisons are hard enough as they are without adding the extra variable of volume.
Butuz
I would say its vital too. The louder the level the greater the difference between silent and max loudness. Effectively louder level it has greater audiable resolution so should highlight more detail. Audio comparisons are hard enough as they are without adding the extra variable of volume.
Butuz
Then I suggest that when making comparisons it's also important to re-run the tests at differing but matched volume settings......for speakers and amps this could be particularly relevant.
Hi Clive,
If it's easy enough to accurately implement matched levels then you might as well do it but maybe it's not so simple. It depends on the type of comparison you are making. If you compare the first 10 seconds of a track or maybe even a whole track and then swap kit, probably here level matching is fairly important. If you are doing longer term listening to work out which equipment is say musically more satisfying for you then level match matching may be the last thing you want. When systems don't exhibit enough detail or dynamics it's tempting to turn them up louder to give the impression they are sounding better. For longer term listening comparisons I suggest using whatever level sounds "right" for the system config you are listening to.
+1 (particularly the bits in bold)!
I think the distinction that you're making between accurately measuring mere sound (which, IMO, an obsession with level-matching panders to), and assessing and appreciating the 'musicality' of a component or system, is an important one.
Like others, over the years, I've somehow managed to make valid and rewarding decisions in audio, and assemble a superb sounding system, by simply level-matching (approximately) by ear, when making comparisons between equipment, not to mention being able to assess many other aspects of music reproduction, without the need for measurements, pertaining to level-matching or otherwise.
Just because something sounds louder doesn't automatically mean that your ears will perceive the sound as more musical. More sonically 'impressive', perhaps, but not necessarily more musical - and it's what sounds most like real music that is ultimately of most importance to our enjoyment of the recorded music our systems reproduce!
Yes, if you can level match accurately, then do it, as indeed others have said, it removes another variable, but it's not the bee-all-and-end-all. Just trust your ears, gut instinct and experience with hi-fi, which is what I do when assessing equipment or ancillaries. It certainly hasn't done my system, or me, any harm.
And how many ordinary audio enthusiasts own the necessary test apparatus to do the job of equipment level-matching properly, anyway? I certainly don't want my listening room looking like a bloody science lab!
I shall comment more later :)
Marco.
I think the distinction that you're making between accurately measuring mere sound (which, IMO, an obsession with level-matching panders to), and assessing and appreciating the 'musicality' of a component or system, is an important one.
Yes, thank you, I should have put it like that in the first place! :cool:
And just what does a matched level mean? if a 0.3dB change can produce a change in the way people perceive, then just how should the matching be done. What if (for example) one or both of the things being compared has (at least in the system) a non flat frequency response? Just what does matched mean in that case, Average RMS value of a wideband noise signal? Or a spot frequency level? I can imagine situations where each type of matching requires a different gain setting.
Level matching sounds simple if you say it quickly.
sq225917
12-10-2011, 23:05
It depends what i'm listening for, no amount of volume matching is ever going to make one turntable run at a more stable speed than another, but for pretty much most other thing it needs doing, at least if you interested in the differences between kit rather than the differences between volume level
And just what does a matched level mean? if a 0.3dB change can produce a change in the way people perceive, then just how should the matching be done. What if (for example) one or both of the things being compared has (at least in the system) a non flat frequency response? Just what does matched mean in that case, Average RMS value of a wideband noise signal? Or a spot frequency level? I can imagine situations where each type of matching requires a different gain setting.
Level matching sounds simple if you say it quickly.
Indeed, Nick :)
It's simply a fad for those in need of absolutes. Some folk can't seem to cope unless everything is neatly pigeonholed into little objective 'black & white boxes', which serve to verify their belief system... :rolleyes:
I do despair sometimes at the lack of ability of people to function as intelligent, but fallible, human beings, and the insistance instead of behaving like logic-obsessed robots, trying never to be wrong.
Embrace your fallibility and humanness, and enjoy music and hi-fi, I say, accepting along the way that sometimes things in life aren't always 'black & white'!
Marco.
It depends what i'm listening for, no amount of volume matching is ever going to make one turntable run at a more stable speed than another, but for pretty much most other thing it needs doing, at least if you interested in the differences between kit rather than the differences between volume level
I think you'll find, Simon, that most people simply rely on their human judgement and experience in audio, and worry not if this isn't always 100% accurate. In my experience, it usually does the job plenty good enough! ;)
Marco.
Being insistent on level matching suggests to me relatively quick comparisons. OK, with experience it's likely that many people will key into what is a good characteristic and what it not so these quick comparisons can be valid. If I'm really serious about understanding the SQ of a piece of kit I will listen to it over several days, often for about a month. The real test is to listen to whole albums and decide which equipment leaves you with the biggest smile on your face. You can then analyse why this has occurred if you wish. Being analytical in the first instance runs the risk of you simply looking for pre-conceived characteristics and thereby overlooking something which it turns out is more important than you'd previously realised.
What you've just eloquently described, Clive, is the significant benefits of subjective assessment in audio. Like I've always said, measurements are valid, but they only tell part of the story.
For the other part, one has to listen and place faith in one's ears (and senses). Oh, and worry not about occasionally making mistakes... Only by making mistakes and by embracing uncertainty do we truly learn. That's the mentality of *proper* scientists with genuinely inquisitive minds!
"The greatest mistake you can make in life is to be continually fearing you will make one." ~ Elbert Hubbard, The Note Book, 1927.
Indeed.
Marco.
What you've just eloquently described, Clive, is the significant benefits of subjective assessment in audio. Like I've always said, measurements are valid, but they only tell part of the story.
For the other part, one has to listen and place faith in one's ears (and senses). Oh, and worry not about occasionally making mistakes... Only by making mistakes and by embracing uncertainty do we truly learn. That's the mentality of *proper* scientists with genuinely inquisitive minds!
"The greatest mistake you can make in life is to be continually fearing you will make one." ~ Elbert Hubbard, The Note Book, 1927.
Marco.
It seems like we're in violent agreement Marco! Nick's point was good too, how on earth do you level match to 0.3db when no kit will be matched to 0.3db across the entire frequency range?
It's simply a fad for those in need of absolutes.
No, it isn't; the argument for ignoring level matching is simply an excuse for people who can't be bothered. I state this as an apparent absolute (pun intended) in the same spirit as your statement is made Marco ;).
(It can hardly be called a 'fad', practically everything else in hi-fi is a fad compared to the idea of level matching when comparing changes)
___________________
Making a few adjustments to the volume control is hardly an onerous job. It's bonkers to introduce added variables when they don't have to exist; you wouldn't change two components at once and then try to extrapolate the contribution of each. If you are specifically attempting to compare changes, then that is the task - if you are listening for enjoyment, then that is the task.
If you are doing some kind of long term comparison then obviously, it is impractical to do level matching (I'm not much of a fan of such comparisons though).
Hi Mark,
Making a few adjustments to the volume control is hardly an onerous job.
I know, and that's precisely what I do when making equipment comparisons, but I don't need any test apparatus to measure the results! Once I adjust the volume control on my preamp, and satisfy my ears that things are level-matched, that's good enough for me.
When I said that level-matching was "a fad for those in need of absolutes", I meant those who insist on the necessity of a bloody machine to verify it! ;)
Marco.
If you are happy to buy that way then no one would think of stopping you. But if you truely want the test to be as accurate of possible you level match.
The issue comes with too many people assuming they are being called names because they dont do it.
At the risk of putting words into peoples mouths, I think the OP was saying that to get a true idea in an academic sense if product X is different to product Y you need to level match.
This is different compared to how anyone goes about deciding on whatever they buy or install in there home.
From a personal viewpoint If someone writes something on a forum and they dont like people discussing it and potentially disagree-ing with it, they shouldnt write it in the first place.
Hi Jonty,
If you are happy to buy that way then no one would think of stopping you. But if you truely want the test to be as accurate of possible you level match.
Absolutely - I don't dispute that in the slightest. And if people need to use whatever 'gubbins' they consider necessary in order to verify the accuracy of their equipment level-matching, then good luck to them!
My point is that I'm perfectly happy level-matching, when making equipment comparisons, by using the volume control on my preamp, and satisfying my ears that the sound is indeed so. I don't need to measure the results in order to obtain the accuracy I seek, or use that information to confirm that what I'm hearing is correct.
In nearly 30 years of owing high-end hi-fi equipment, and achieving truly superb results with countless different systems, never once, short of a stylus pressure gauge and alignment protractor, have I needed to use any form of measurement apparatus to optimise the sonic performance of my system, or aid my enjoyment of recorded music through it :)
Marco.
jandl100
13-10-2011, 18:11
As usual, I'm pretty much in 100% agreement with Marco. (We only differ in a major way in our opinion of Tannoy speakers :))
Another point re: level matching is that some CD players (and amps etc) have significantly different tonal presentations.
As an example, Jake (Lord Mortlock) brought his massive Marantz KI Pearl Sig CD player round. The bass on that was magisterial to say the least. In comparison, the Cyrus CD8X that I had recently had a light & breezy presentation - the bass weight difference is remarkable indeed. What is the point of level matching in those circumstances? If the SPLs are equal, then the Cyrus would be playing significantly louder over most of the audible spectrum!
Reid Malenfant
13-10-2011, 18:16
Hi Jerry, i guess you need to level match in the area that the ear is most sensitive, IE between 1 & 5KHz.. In that case your Cyrus would sound like it had weak bass compared to your description of the Marantz - sorry :(
jandl100
13-10-2011, 18:25
No need to apologise, Mark. :)
A (fairly ;)) sensible answer ... but we are then venturing into the somewhat subjective area of over what frequency range to equalise the SPLs.
Whereas Marco's approach of just listening at the level at which each component sounds its best (if that is a fair summary of Marco's approach!) and then doing comparisons seems def superior to me based on my own experience.
I'm happy if your opinion differs though. :)
Reid Malenfant
13-10-2011, 18:30
Actually it doesn't, i'm of the very same opinion that a decent appraisal can only be done over an extended period of time & not back to back so to speak :)
If i was to level match though, it'd be in the frequency range i mentioned.
Whereas Marco's approach of just listening at the level at which each component sounds its best (if that is a fair summary of Marco's approach!)
It's nothing more complicated than that, Jerry, simply because to my ears it doesn't have to be. As they say, Keep It Simple Stupid (K.I.S.S) :cool:
It's a principle that I have applied, with great sonic success, to every area of my system building and equipment analysis methodology since day 1 of my interest in audio!
Marco.
I’ve never understood how it gets to “I only listen” or, “I only measure” in such debates.
Surly the idea is to use all the tools at ones disposal to evaluate an audio system :scratch:
Seems a bit like being given both hands and feet then saying I only use my toes to manipulate stuff coz it makes it more challenging :mental:
However, whatever puts a smile on your face and of course a smile on the faces of those who use fingers and toes ;)
Sure, John, but why use tools you don't need?
Like I said, the only time I need to measure anything is when I have to align a new cartridge or set the tracking force. Other than that, my ears are plenty good enough for analysing anything else I need to in audio!
I'll leave the measuring of this and that to those who build equipment. Ultimately, I'm a listener and music lover, not a builder :)
Marco.
I’ve never understood how it gets to “I only listen” or, “I only measure” in such debates.
Surly the idea is to use all the tools at ones disposal to evaluate an audio system :scratch:
Seems a bit like being given both hands and feet then saying I only use my toes to manipulate stuff coz it makes it more challenging :mental:
However, whatever puts a smile on your face and of course a smile on the faces of those who use fingers and toes ;)
Yes, I agree, I use as many tools as I can to see what is going on and that includes ears. But those arguing for level matching have not defined just what it means, just that "it" needs to be done. Re, my point about is it single frequency, wideband, narrowband. And does the choice of matching strategy potentially preselect the outcome in some cases?
Yes, I agree, I use as many tools as I can to see what is going on and that includes ears.
Yup, you need to, 'cos you build stuff! :)
I've never understood though why some end users of kit literally obsess over measurements.... For me, such things are for equipment designers to worry about.
You measure and build 'em, and we buy 'em and listen to 'em - seems a simple enough process to me! Why make it more complicated?
Marco.
jandl100
14-10-2011, 06:36
I’ve never understood how it gets to “I only listen” or, “I only measure” in such debates.
Surly the idea is to use all the tools at ones disposal to evaluate an audio system :scratch:
Hmm. I've been giving some thought to this.
I think you need to use all appropriate tools.
I'll try and explain ....
One of the reasons that I am so fascinated by hifi - and why I like to try such a lot of it! - is that to me it all sounds different. Not just tonal presentation, but more importantly dynamic expression and transparency and presented resolution and image focus and suchlike things.
I'm mainly a classical music listener, and so the natural analogy for me is listening in a concert hall. That is what recordings are attempting to emulate, if they are any good! To a significant degree different hifi kit places you in different seats in the hall. Dynamic and transparent is like sitting front centre stalls. Laid back and diffuse (pipe & slippers if you like) is sitting in the upper circle. ... and of course all gradations inbetween - rear stalls, front circle etc.
OK - now if you accept that analogy, then you would expect different kit to be played at different loudness levels in order to sound "right", as perceived loudness in a concert hall is substantially affected by distance from the performers. It really is less loud at the back! - as well as being less dynamic and more diffuse, and having a softer tonal balance. So setting equal SPL for different kit is not only irrelevant it is actually wrong!
You need to match the SPLs to the characteristics of the kit you are listening to. :)
StanleyB
14-10-2011, 07:18
I'm mainly a classical music listener, and so the natural analogy for me is listening in a concert hall. That is what recordings are attempting to emulate, if they are any good! To a significant degree different hifi kit places you in different seats in the hall. Dynamic and transparent is like sitting front centre stalls. Laid back and diffuse (pipe & slippers if you like) is sitting in the upper circle. ... and of course all gradations inbetween - rear stalls, front circle etc.
OK - now if you accept that analogy, then you would expect different kit to be played at different loudness levels in order to sound "right", as perceived loudness in a concert hall is substantially affected by distance from the performers. It really is less loud at the back! - as well as being less dynamic and more diffuse, and having a softer tonal balance.
I don't know if you have ever seen how a concert hall is wired up for recording, but it doesn't consist of just a couple of mics at the front. Some mics are placed close to certain instruments or musicians, and the recording engineer tries to level match that sound into the overall atmosphere. But those individual mics tend to be mono, and can upset the soundstage within which some instruments in an orchestra is apparently positioned.
........ So setting equal SPL for different kit is not only irrelevant it is actually wrong!
You need to match the SPLs to the characteristics of the kit you are listening to. :)
I agree. It's rather like comparing 2 cars for fuel economy by only running them at 50mph. This would only be a single datapoint, due to engine and wind resistance their comparative results may vary across the full spectrum of speeds.
IF level matching is done (and it's tough bearing mind Nick's observations) then it needs to cover the whole spectrum of levels people listen at or else it's pretty much irrelevant for most kit, most certainly for amps and speakers.
Doesn't that then defeat the whole object of the exercise?
Doesn't that then defeat the whole object of the exercise?
Why? (maybe I didn't explain it well) You could then compare are differing but matched levels. Some speakers and amps "come alive" when they reach certain levels, other may struggle to go loud. In reality it would be a royal pain but those wanting matched levels should go the whole hog.
Hi,
I`m not sure that level matching between different speaker systems can.be done meaningfully given the relatively gross differences in frequency response and from responses. Surely two speaker can show an equivalent SPL with very different response across the audio band. I.can see that level matching would be important when making comparison after adjustments etc. within the same system to avoid biasing comparisons.
Cheers,
Alex
I'd like to chime in, as I do quick abx testing everyday, during work
as I work from my home based mastering setup
before doing any kind of abx testing, level matching is absolutely crucial,
1.0dB difference and you could describe the louder sound as fuller, better defined, more highs, deeper lows, better stereo, etc..
I use pink noise, and looped separate wav. freq. from 30 Hz up to 20 kHz
rme interface users got privilege to have free tool - digicheck - brilliant
I check cables this way too, dac, effects bypass, different mix/master versions;
quickly swapping different sources
it's even easy to set headphones test - you need spl meter and pink noise as .wav or aiff file
anyway in my opinion I wouldn't dare to consider making opinions about audio without level matching
Powered by vBulletin® Version 4.2.3 Copyright © 2026 vBulletin Solutions, Inc. All rights reserved.