2 Way Speaker Bass Distortion: Not So Important?

[watchnerd]

3 different THD @ 90dB (50 hz - 10khz, 2 m) graphs of 3 different 2-way loudspeakers, all of which have good "street cred", all from Soundstage Network testing:

KEF LS50

Revel Performa M126Be

Dynaudio Special Forty

There is about a 10dB spread here at 50 Hz.

But most subjective reviews seem to rate all 3 speakers as very good, with differences being mostly a matter of tastes and preferences.

Does this mean that bass distortion in small 2 ways just doesn't matter that much, subjectively?

 

[Blumlein 88]

I'm not even sure what distortion you are seeing in those graphs below 250 hz is distortion. It may be mostly noise. I seem to recall the NRC anechoic chamber isn't anechoic at those lower frequencies. And Soundstage says what they show in those plots is THD+N.

BTW, while my REW in room measures are somewhat close to Soundstage.net results for FR, the distortion curves are very close to what they get.

 

[watchnerd]

I'm not even sure what distortion you are seeing in those graphs below 250 hz is distortion. It may be mostly noise. I seem to recall the NRC anechoic chamber isn't anechoic at those lower frequencies. And Soundstage says what they show in those plots is THD+N.

BTW, while my REW in room measures are somewhat close to Soundstage.net results for FR, the distortion curves are very close to what they get.

So it's just useless data of no consequence?

Because above 500 Hz it's all practically the same.

 

[BillG]

Does this mean that bass distortion in small 2 ways just doesn't matter that much, subjectively?

It just means that we're relatively immune to the perception of it in that region...

 

[Blumlein 88]

I couldn't get the scaling completely identical. You see the similarity I think. Upper is Soundstage.net for Revel F12. Lower is the F12 via REW and Umik 1. I was measuring further away and the sound level at 2 meters would have been just about 90 dbSPL.

 

[watchnerd]

It just means that we're relatively immune to the perception of it in that region...

même sens, mots différents, no?

 

[Blumlein 88]

It just means that we're relatively immune to the perception of it in that region...

Yeah, at 100 hz our perception of a 40 dbSPL harmonic is very near our threshold. Yet if the woofer had 1 % distortion playing at 80 db that is what the harmonic level would be.

 

[Blumlein 88]

même sens, mots différents, no?

I do wish Soundstage net would at least show enough to display -60 or -70 db distortion curves. They stop at about .5% I suppose assuming anything below that is unimportant with music. If so then distortion above 500 hz is almost never an issue with any speaker. And measurements below 500 hz are corrupted by noise.

 

[watchnerd]

I do wish Soundstage net would at least show enough to display -60 or -70 db distortion curves. They stop at about .5% I suppose assuming anything below that is unimportant with music. If so then distortion above 500 hz is almost never an issue with any speaker. And measurements below 500 hz are corrupted by noise.

Do we know if this is incorrect? Or, is it in fact, reasonable?

Horn speaker lovers seem to make a big deal about ultra low distortion being part of the magic.

 

[Blumlein 88]

Do we know if this is incorrect? Or, is it in fact, reasonable?

Horn speaker lovers seem to make a big deal about ultra low distortion being part of the magic.

Well with headphones, and I think speakers in a quiet room, people can hear .1% using sine wave test tones.

With music it depends upon the music. 1% is usually considered good enough for nearly all kinds of music. So it isn't unreasonable. And those numbers are at fairly loud levels. With music only the peaks would have this much distortion. Most of it would be cleaner.

 

[watchnerd]

Well with headphones, and I think speakers in a quiet room, people can hear .1% using sine wave test tones.

With music it depends upon the music. 1% is usually considered good enough for nearly all kinds of music. So it isn't unreasonable. And those numbers are at fairly loud levels. With music only the peaks would have this much distortion. Most of it would be cleaner.

This sounds like:

High pass any $2K-$4k 2 way speaker (like those above) at 80 Hz to a sub and never worry about distortion at any 'normal' volume.

 

[Blumlein 88]

This sounds like:

High pass any $2K-$4k 2 way speaker (like those above) at 80 Hz to a sub and never worry about distortion at any 'normal' volume.

That sort of fits with Toole's brief comments upon speaker distortion in his book.

 

[Blumlein 88]

https://www.axiomaudio.com/blog/distortion

This is interesting. They added tones (not related to music) and found at each frequency where it was loud enough to be noticed by listeners. Generally more noticeable at lower levels (versus the average music level) as frequency goes up. I suppose some of that is less masking at high frequencies. In any case, everything had to be -45 db or higher than the average music level. Much higher

Also here is an oldie, but a goody from JGH about a Bob Carver experiment.
https://www.stereophile.com/asweseeit/673awsi/index.html

Also a thread about Carver here with link to the original article.
https://www.audiosciencereview.com/forum/index.php?threads/how-audible-is-distortion.2047/

 

[watchnerd]

https://www.axiomaudio.com/blog/distortion

This is interesting. They added tones (not related to music) and found at each frequency where it was loud enough to be noticed by listeners. Generally more noticeable at lower levels (versus the average music level) as frequency goes up. I suppose some of that is less masking at high frequencies. In any case, everything had to be -45 db or higher than the average music level. Much higher

Also here is an oldie, but a goody from JGH about a Bob Carver experiment.
https://www.stereophile.com/asweseeit/673awsi/index.html

Also a thread about Carver here with link to the original article.
https://www.audiosciencereview.com/forum/index.php?threads/how-audible-is-distortion.2047/

This should be printed on the bottom of the SINAD chart:

"Consequently it's worthwhile viewing distortion numbers with caution when making a purchasing decision. Ever-lower distortion numbers may not contribute to improved sound quality."

 

[oivavoi]

https://www.axiomaudio.com/blog/distortion

This is interesting. They added tones (not related to music) and found at each frequency where it was loud enough to be noticed by listeners. Generally more noticeable at lower levels (versus the average music level) as frequency goes up. I suppose some of that is less masking at high frequencies. In any case, everything had to be -45 db or higher than the average music level. Much higher

There was a thread about this recently: https://www.audiosciencereview.com/...ass-undetectable-when-it-stays-under-20.9353/

Like I said in that thread, I think this is fairly reasonable and intuitive. A violin not played exactly in tune sounds horrible. Nobody notices a double bass that doesn't play exactly in tune. As a choir singer, I know that it sounds really bad if the sopranos don't hit the tune, whereas bass singers have some more leeway. Now this isn't the same thing as THD, but I believe the general mechanisms are the same. I've also seen studies which show more or less the same pattern: Distortion in the deep bass basically doesn't seem that important.

When I read speaker measurements which include distortion (like Soundstage and the German magazines do), I don't pay attention to distortion below 500 hz, unless it's very very high. Significant distortion above 1 khz - now that's a red flag for me.

 

[Blumlein 88]

This should be printed on the bottom of the SINAD chart:

"Consequently it's worthwhile viewing distortion numbers with caution when making a purchasing decision. Ever-lower distortion numbers may not contribute to improved sound quality."

Yes, I agree.

I typed a long post a few days ago about how some of the most revered amplifiers from the 1980s and 1990's had high power, load intolerance, and flat frequency response. Almost none of them even claimed THD less than -70 db (most were just at -60 db). I dug up detailed info on the performance of each with testing from various sources. In the end I didn't leave it posted. Something to think about though. We are much more sensitive to FR errors even small tilts over the spectrum. For even what we consider rather poor gear, FR may be the only thing to worry about.

Should Amir change his focus to pass/fail sinad (passing being 70 db just for safety) and post nothing else except detailed FR charts? The only other thing is making sure quiescent noise is low enough. You can do that for yourself. Find the max volume you'll feed your headphones or speakers. Do you hear any noise without music? No, then it is quiet enough.

 

[Soniclife]

I do wish Soundstage net would at least show enough to display -60 or -70 db distortion curves. They stop at about .5% I suppose assuming anything below that is unimportant with music. If so then distortion above 500 hz is almost never an issue with any speaker. And measurements below 500 hz are corrupted by noise.

I've assumed it's because they cannot trust the results below 45db, but I agree as they are they show very little. My take away from reading them is that even small 2 ways do surprisingly well when pushed harder than they would likely be.

 

[Blumlein 88]

http://mjaudiolab.pl/images/stories/Bryston/swedish14bsstreview.pdf

Read about this Bryston test. The one that eventually passed the Swedish series bypass testing. It failed at first because it was a half db down at 20 khz. Fixing that allowed it to pass the test.

 

[Cosmik]

There was a thread about this recently: https://www.audiosciencereview.com/...ass-undetectable-when-it-stays-under-20.9353/

Like I said in that thread, I think this is fairly reasonable and intuitive. A violin not played exactly in tune sounds horrible. Nobody notices a double bass that doesn't play exactly in tune. As a choir singer, I now that it sounds really bad if the sopranos don't hit the tune, whereas bass singers have some more leeway. Now this isn't the same thing as THD, but I believe the general mechanisms are the same. I've also seen studies which show more or less the same pattern: Distortion in the deep bass basically doesn't seem that important.

When I read speaker measurements which include distortion (like Soundstage and the German magazines do), I don't pay attention to distortion below 500 hz, unless it's very very high. Significant distortion above 1 khz - now that's a red flag for me.

I wonder if the nature of music is helping us here. Chords played higher up on the piano keyboard sound fine, but at the bottom end sound terrible. Any keyboard improviser or composer knows to start making their chords 'sparse' as they go lower. The theory seems to be that our hearing is less sensitive to the fundamentals of the bass notes and more sensitive to the overtones (harmonics), thus we begin to hear 'chords' of the harmonics, not the fundamentals, and this sounds terrible. In effect, the harmonics of the piano notes are combining in a form of 'intermodulation' (technically maybe not the right word) 'distortion' that we are sensitive to, so composers avoid creating it. Bass is often simply monophonic.

Maybe when we say "We are not sensitive to bass distortion", it's because the creator of the music has already pre-filtered the content to avoid situations where we would notice it.

 

[watchnerd]

Nobody notices a double bass that doesn't play exactly in tune. As a choir singer, I know that it sounds really bad if the sopranos don't hit the tune, whereas bass singers have some more leeway.

As a bass player, both acoustic and electric, I'm deeply offended.

Other bass players, at least, notice.

I don't care if the violists can tell; they're all deaf in one ear, anyway.