A Broad Discussion of Speakers with Major Audio Luminaries

[goat76]

But those limits of audibility do not account for simultaneous and forward temporal masking - the sounds that generate the distortion also prevent us from hearing all of it - that is why the simple technical measurements of distortion are misleading. Newer methods take masking into account, but they are not standardized or widely used.

I just did the Klippel distortion test for the first time. Thanks for reminding me it existed. I came out close to average on most metrics, slightly worse on one, and significantly better with one test signal. I'm amazed at 88 - my hearing is definitely degraded and I have tinnitus. I suspect that knowing what to listen for is part of it.

I ended up with the result of -33, but I must say that I felt a bit distracted by the low-quality audio of those sound clips. To me, they sounded like 128kbps audio files or something, and it felt like I would be able to detect the distortion at even lower levels if the audio files were of higher quality to begin with.

Does anyone know what the audio quality is for the Klippel distortion test?

 

[Cote Dazur]

Question about the test, are you doing it using headphones?
Should it be done on our best speaker system?

Never mind, it is self explanatory when starting the test.

-9

 

[Newman]

I'm happy to hear that Andrew has a bright future ahead of him.

...at doing distortion tests. Not at evaluating loudspeaker sound quality. Doubly not at evaluating loudspeaker sound quality in sighted listening.

 

[Newman]

What are the audibility thresholds for distortion at different frequencies?

Check out these two papers by Geddes and Lee, you will see it is not so simple.
paper 1
paper 2

cheers

 

[IPunchCholla]

I ended up with the result of -33, but I must say that I felt a bit distracted by the low-quality audio of those sound clips. To me, they sounded like 128kbps audio files or something, and it felt like I would be able to detect the distortion at even lower levels if the audio files were of higher quality to begin with.

Does anyone know what the audio quality is for the Klippel distortion test?

-33 with music? You have my condolences! Luckily the absolute best I could manage was -15.

 

[goat76]

-33 with music? You have my condolences! Luckily the absolute best I could manage was -15.

I thought the result I got would be average around here. I did the test using the T. Chapman track.

 

[IPunchCholla]

I thought the result I got would be average around here. I did the test using the T. Chapman track.

View attachment 557101

It might be. I might just not be that sensitive to distortion. Given that I seek out interesting distortion techniques when playing music, it might just be part of my sound processing? Either way, I test well in terms of my hearing frequency ability for my age ( I have relatively less loss than most people my age (54) and can still hear above 17k). I was just amused by my disappointment when I took the test that I couldn’t do better, even though being able to hear distortion isn’t necessarily a good thing. I mean it’s good if you are a sound engineer but just seems like it might get in the way of enjoying music. It’s like being sensitive to a single pea under 20 mattresses might be good if you’re a professional mattress tester but not so great if you just want to sleep.

When I went to the Vegas Sphere, we had nosebleed seats. One of the subs above my head had a resonant rattle. It was terribly distracting for the first part of the show, but after a few minutes my brain just filtered it out and the sound was awesome. I could intentionally choose to hear it after that, but as soon as I stopped concentrating it disappeared. If it works like that, training myself to hear it better might make sense. But having to hear it would be horrible.

When it comes to music I prefer to focus on why that note/chord, accent, articulation, rhythm, etc than worrying about recording flaws.

 

[goat76]

It might be. I might just not be that sensitive to distortion. Given that I seek out interesting distortion techniques when playing music, it might just be part of my sound processing? Either way, I test well in terms of my hearing frequency ability for my age ( I have relatively less loss than most people my age (54) and can still hear above 17k). I was just amused by my disappointment when I took the test that I couldn’t do better, even though being able to hear distortion isn’t necessarily a good thing. I mean it’s good if you are a sound engineer but just seems like it might get in the way of enjoying music. It’s like being sensitive to a single pea under 20 mattresses might be good if you’re a professional mattress tester but not so great if you just want to sleep.

When I went to the Vegas Sphere, we had nosebleed seats. One of the subs above my head had a resonant rattle. It was terribly distracting for the first part of the show, but after a few minutes my brain just filtered it out and the sound was awesome. I could intentionally choose to hear it after that, but as soon as I stopped concentrating it disappeared. If it works like that, training myself to hear it better might make sense. But having to hear it would be horrible.

When it comes to music I prefer to focus on why that note/chord, accent, articulation, rhythm, etc than worrying about recording flaws.

It might be that I'm sensitive to distortion, but that doesn't mean I dislike the sound, and I actually use it on many tracks when I mix my music.
Adding distortion can really help bring elements closer in the mix and give them that missing three-dimensional "organic" quality that is sometimes missing when things sound a little "too clean". A great-sounding saturation plugin I use on many tracks is Waves "Magma BB Tubes"; it has one knob for even harmonic distortion and another knob for odd harmonic distortion. Klanghelm SDRR2 is also a great-sounding saturating plugin.

 

[Salt]

It might be that I'm sensitive to distortion, but that doesn't mean I dislike the sound, and I actually use it on many tracks when I mix my music.
Adding distortion can really help bring elements closer in the mix and give them that missing three-dimensional "organic" quality that is sometimes missing when things sound a little "too clean". A great-sounding saturation plugin I use on many tracks is Waves "Magma BB Tubes"; it has one knob for harmonic distortion and another knob for non-harmonic distortion. Klanghelm SDRR2 is also a great-sounding saturating plugin.

Thanks for some new aspects and clarification. Appreciated.

 

[Salt]

It might be that I'm sensitive to distortion, but that doesn't mean I dislike the sound, and I actually use it on many tracks when I mix my music.
Adding distortion can really help bring elements closer in the mix and give them that missing three-dimensional "organic" quality that is sometimes missing when things sound a little "too clean". A great-sounding saturation plugin I use on many tracks is Waves "Magma BB Tubes"; it has one knob for harmonic distortion and another knob for non-harmonic distortion. Klanghelm SDRR2 is also a great-sounding saturating plugin.

For illustration , might you give some example of pre and post mixing tracks?

 

[goat76]

For illustration , might you give some example of pre and post mixing tracks?

I don't have any before-and-after examples I've done myself, but many demos on YouTube show how saturation can help elements in a mix stand out and sound more three-dimensional. Below is one video demo of the BB Tubes plugin I mentioned.

 

[MattHooper]

I don’t see how this addresses what Mart68 wrote…

I understand you. You are saying that because stereo was how the records were made, stereo listening is as good as any other listening.

That’s not what he wrote.
Mart68 very specifically wrote:

“the fact that those recordings were intended to be listened to on just 2 channels. Aside from the quite brief flirtation with quadrophonic, which I'm sure we all agree is not the same thing at all as comtemporary multi-channel, no-one even conceived of having multiple discrete surround channels back then.”

So he was clearly talking about listening to stereo recordings in the format for which it was intended - mixed and mastered for: two channel Stereo.

It’s a funny thing (and I think Cote was getting at this) that there’s lots of talk about accuracy around here, and also the virtue of trying to close the circle of confusion so that one hears at home the art as it was intended.
Yet as soon as multichannel comes up some insist one should be adopting multichannel if one is a “serious audiophile interested in the best quality sound.”

Well, it seems to me Cote is alluding to the point that if one favours the music contained in the many recordings made in stereo, then using stereo playback is just as legitimate a choice - in fact more legitimate if one is being consistent with the “ closing the circle of confusion” approach. And the vast majority of recorded music is as we know, not multichannel nor mixed and mastered with such intent.

And once one is upmixing stereo to enjoy it more in multi channel, then to a degree you’ve abandoned the “ accuracy” thing and and closing the circle of confusion, and going essentially with “ I’m gonna alter the recording to playback in a way that I personally like better.”

So in a sense it seems like there is lip service paid to accuracy, until multichannel comes up and is advocated for, and then it’s “ it can sound better if we manipulate the original recording from stereo to multichannel so go for it if you want.”

Now I’m not for a moment arguing against anyone upmixing two channel to surround.
I do it all the time and I love it. Tone controls?
Use ‘em if it allows you to make some recordings more pleasing. It all makes sense to me.

I wouldn’t defend some of what Cote said against multichannel. But there certainly IS among certain members a constant suggestion that in sticking with stereo one is just being a stuck in the past, obdurate audiophile unwilling to upgrade to The Scientifically Established Superior Way Of Listening To Music. To that I think Cote’s point about Stereo being a legitimate choice for even the most serious audiophile, given the preponderance of music meant for that format,and/or if one has an interest in the music that had been recorded in that format . You can still be as serious an audiophile as the multi channel user, following advancements and refinements in two channel playback suited to the format for which the majority of music has been recorded.

 

[bmc0]

And once one is upmixing stereo to enjoy it more in multi channel, then to a degree you’ve abandoned the “ accuracy” thing and and closing the circle of confusion, and going essentially with “ I’m gonna alter the recording to playback in a way that I personally like better.”

In some sense, this applies to any setup which differs significantly from the one(s) used for mixing/mastering/evaluation of the original mix. Upmixing certainly has the potential to alter the spatial quality more than typical variations between two-channel systems, but the latter can still have a considerable effect. You've written frequently about spatial differences you perceive between different setups. Which presentation is more "correct"?

It's conceivable as well that at least in some cases, the mix doesn't really represent the artist's and/or engineer's original intent. It may well just be a compromise necessitated by the target delivery format, which historically has mostly been one or two channels except for niche productions. Of course, there's no guarantee that a given upmixing algorithm will do the right thing according to the original intent—but the more I think about it, the more it seems to me that attempting to preserve the (mostly unknowable) "artist's intent" is a bit of a farce when it comes to spatial aspects, given the spatial limitations of standard two-channel stereo.

 

[Cote Dazur]

I wouldn’t defend some of what Cote said against multichannel

First, I believe you represented my opinion with great care and accuracy, thank you.
I am and never was, in any way, against multi channel, what I am against is the insistance from some that it is the “way”

The opinion that it is better is based on preference, not science.
I tried it, way back then, tried it again many time, I hear it at my friends and children’s place from time to time, I understand why it can be attractive, it is just not what I like best. So not against in any way just not favouring it.

 

[levimax]

In some sense, this applies to any setup which differs significantly from the one(s) used for mixing/mastering/evaluation of the original mix. Upmixing certainly has the potential to alter the spatial quality more than typical variations between two-channel systems, but the latter can still have a considerable effect. You've written frequently about spatial differences you perceive between different setups. Which presentation is more "correct"?

It's conceivable as well that at least in some cases, the mix doesn't really represent the artist's and/or engineer's original intent. It may well just be a compromise necessitated by the target delivery format, which historically has mostly been one or two channels except for niche productions. Of course, there's no guarantee that a given upmixing algorithm will do the right thing according to the original intent—but the more I think about it, the more it seems to me that attempting to preserve the (mostly unknowable) "artist's intent" is a bit of a farce when it comes to spatial aspects, given the spatial limitations of standard two-channel stereo.

Artist intent and all is fun to argue about but I think the issue being raised is about "Hi-Fi" to the signal on the original stereo recording. It would be one thing if most systems could accurately reproduce the original stereo signal of the recording and then say this shows the limitation of the stereo format and then use upsampling to compensate for the weaknesses but that is not what is happening. Vanishingly few systems can accurately reproduce what is on a stereo recording (which includes low distortion LF and time domain information among other things). Upsampling to MC is just basically saying it is too hard to get stereo right and easier to reconstitute it into something easier which is also likely to be preferable. There is nothing wrong with this approach which is the same thing as "SET amps", tube buffers, and colored speakers. A "hardwired" non linearity is being applied to all signals with the goal being a pleasing sound rather than fidelity to the source. The inconsistency being pointed out is that some that condemn some "hardwired" nonlinearities as not being "Hi-Fi" embrace other versions of the same thing.

 

[MattHooper]

I understand generally what you mean, but I think people will rightly take contention when it is expressed this way:

The opinion that it is better is based on preference, not science.

You seem to be suggesting that if preferences are involved, then science cannot be involved.

But that is not correct: science can study preferences. And come to scientific conclusions about preferences. They can come to theories that predict preferences (statistical probabilities).

And that’s what we have in much of the work cited in this forum. The research cited by Toole et al shows that under controlled conditions people’s preferences can be predicted to a high degree of accuracy.

Science can often be deflating to our sense of intuition, and one intuition is that we are experts in our own preferences: that nobody can tell us more about our own preferences, accurately, than we can.

But that doesn’t turn out to be quite true.
It’s possible in principle and often enough in practice, for a scientist to be able to tell you something more accurate about your preferences than you had presumed.

This has been shown many times in blind tests, where under sighted conditions people were sure they knew their own preferences in terms of the sound they thought they were hearing, but controlling for biases, it turns out they had a different preference. A preference that the scientist could actually have predicted better than the individual.

Personally, I have felt I really enjoyed some loudspeakers that depart from something like the sound of a Revel or some other similar loudspeaker that tends to score high in blind test, preference scores.

It may be the case that I actually preferred the sound of those errant speaker designs.

But if somebody was going to blind test me with those speakers against a top Revel speaker, as a betting man, I should probably place my bet that I’ll prefer the sound of the Revel speakers under blind testing.

 

[Cote Dazur]

You seem to be suggesting that if preferences are involved, then science cannot be involved.

Sorry, it appears that way, I am not suggesting that science cannot be involved in preference, just that preferring multi over stereo is not based on science.

But that is not correct: science can study preferences

Absolutely, it can, no argument from me, but where is the science showing multi is preferred over stereo?

 

[Thomas_A]

I thought the result I got would be average around here. I did the test using the T. Chapman track.

View attachment 557101

There is a dedicated thread also to this Klippel test, but I found the various tests being quite different in level and quality.

Distortion Listening Test

This is an interesting kind of test but, if we put aside the audible noise, we know little about what we can learn out of it absolutely speaking. Only that Mr. X has a better score than Mr. Y. Klippel knows how to interpret it for their own purpose for sure, knowing what is "real speaker", what...

www.audiosciencereview.com

 

[hex168]

Artist intent and all is fun to argue about but I think the issue being raised is about "Hi-Fi" to the signal on the original stereo recording. It would be one thing if most systems could accurately reproduce the original stereo signal of the recording and then say this shows the limitation of the stereo format and then use upsampling to compensate for the weaknesses but that is not what is happening. Vanishingly few systems can accurately reproduce what is on a stereo recording (which includes low distortion LF and time domain information among other things). Upsampling to MC is just basically saying it is too hard to get stereo right and easier to reconstitute it into something easier which is also likely to be preferable. There is nothing wrong with this approach which is the same thing as "SET amps", tube buffers, and colored speakers. A "hardwired" non linearity is being applied to all signals with the goal being a pleasing sound rather than fidelity to the source. The inconsistency being pointed out is that some that condemn some "hardwired" nonlinearities as not being "Hi-Fi" embrace other versions of the same thing.

I could be wrong about this, but I don't think upmixers to multichannel make objective measurements worse, unlike SET amps, tube buffers, and colored speakers. Along similar lines, I generally prefer the presentation from floor-to-ceiling line arrays. Certainly not what the mastering engineers were expecting as a playback system, but I would not say good line arrays measure worse than the expected stereo pair of two- or three-ways.

 

[bmc0]

Vanishingly few systems can accurately reproduce what is on a stereo recording (which includes low distortion LF and time domain information among other things). Upsampling to MC is just basically saying it is too hard to get stereo right and easier to reconstitute it into something easier which is also likely to be preferable.

My setup is capable of high SPL with low distortion and is time-aligned with linear phase crossovers. LFs are not summed to mono and spatial effects are audible down to circa 40Hz. There are certainly even better systems out there, but mine is no slouch. Upmixing is not a kludge to mitigate limitations of the loudspeakers; it is a kludge to mitigate limitations of standard 2-channel stereo.