So we now have a speaker with >5% distortion which isn't audible subjectively. Is a SINAD of 96dB poor?

[andreasmaaan]

Might be why tweeters still seem to sound fine even with relatively high levels of 2nd-order distortion, f1+f2 is high enough that we don't care too much and f1-f2 is more or less outside their reproduction range.

Agree wholeheartedly with your post, bar this point

E.g:

2kHz + 3kHz = 5kHz

7kHz - 4kHz = 3kHz

etc..

 

[andreasmaaan]

I would say - it IS possible. See my measurements on tweeters under voltage and current drive and quite significant difference in odd harmonics IMD components.

http://pmacura.cz/speaker_dist3.htm

Very interesting, thanks

Do you have a theory as to why the levels of distortion below 2kHz were so much higher with current drive, though? (I'm looking at the final two graphs on the last page.)

 

[AnalogSteph]

Response is probably up due to resonance, so larger excursion.

If this is not an argument for using tweeter dropper resistors even in an active speaker, I don't know what is. Mind you, Esa Meriläinen's tweeter results weren't nearly as impressive-looking (unlike several of the mid-woofers he tested, in some of which 3rd dropped by 30+ dB).

Anyone got a clue what the distortion mechanism is that produces all of the 2nd in domes, even beyond the range where suspension dominates? Would seem to be electrical in nature but not related to inductance nonlinearity, which is eliminated by current driving. Offset BL(x) curve?

Agree wholeheartedly with your post, bar this point

E.g:

2kHz + 3kHz = 5kHz

7kHz - 4kHz = 3kHz

Fair enough. That said, there usually isn't that much treble energy to begin with so I wouldn't expect something like 7-4 to be very big, and for 2+3 I would expect some auditory masking. On the whole, it could definitely be a great deal worse. The lower average sum signal levels plus generally good tweeter level handling alone would make for a relatively relaxed environment.

Partitioning the frequency range is commonly used in radio receivers to good effect. There 2nd-order IMD is commonly regarded as the one that's easy to get rid of. 3rd is much harder, and is what is mainly being looked at when evaluating dynamic range. This premise doesn't hold with a super wideband frontend, of course. I had little difficulty provoking some 2nd-order intermod in an AR7030+ back in the day with a rather laughable antenna (though shortwave levels no doubt still were substantially higher then), and that's got about the highest dynamic range 1st mixer of its day.

 

[Kvalsvoll]

Probably already answered now, but bear with me.

5% is definitely audible, even 0.1% can be audible, it all depends on the harmonic distribution. Lower h is less audible due to masking, around 2% 2h is borderline, while 5h will be audible at much lower level. Distortion in speakers tend to be dominated by low order harmonics - but not always.

Then what is audible may not necessarily be destructive for sound quality.

And good speakers do not have 5% distortion. The system playing Panda Dub here now will never reach close to 5% for normal or even quite loud listenting, because then you need to reach spl levels well above 110dB - and that is valid to below 20Hz. At lower levels, the best speakers have total distortion below 0.1% at higher frequencies, higher harmonics even lower.

Distortion in speakers also tend to increase with increased spl, so that distortion for continuous tones are quite low, while distortion in transients is not very audible, so it does not matter much if distortion reaches high levels at very high spl.

 

[pkane]

Definitely.

I just thought of an idea.

Throw in the 32-tone IMD that has been done for DAC/AMP/etc.

It shows THD, IMD, even frequency response in one shot.

Here's my iMac's speakers measured by the built-in microphone... I wouldn't say this is awesome performance, considering all 100 tones are supposed to be the same amplitude

(white is just the distortion, blue is signal plus distortion)

 

[typericey]

Sorry to thread jack. But I have a sort of related question:

Is distortion compounded?

i.e.
DAC = .1%
Amp = .1%
Speakers = 1%
Total system distortion = 1.2%

Same question goes for SINAD. If, say, my DAC's SINAD is 100dB and my amp 90dB. What is my "system SINAD"?

 

[Blumlein 88]

Sorry to thread jack. But I have a sort of related question:

Is distortion compounded?

i.e.
DAC = .1%
Amp = .1%
Speakers = 1%
Total system distortion = 1.2%

Same question goes for SINAD. If, say, my DAC's SINAD is 100dB and my amp 90dB. What is my "system SINAD"?

No it doesn't work that way.

Here is a post earlier in the thread.
https://www.audiosciencereview.com/...ely-is-a-sinad-of-96db-poor.12519/post-371017

And here is where you can find the calculator. Near the bottom of the page.
http://www.sengpielaudio.com/calculator-thd.htm

Basically if you have a source of distortion 20 db or greater than other sources that one device basically sets distortion level for the entire system.

In your example total distortion would be 1.00995% for all of it.

 

[NTK]

Sorry to thread jack. But I have a sort of related question:

Is distortion compounded?

i.e.
DAC = .1%
Amp = .1%
Speakers = 1%
Total system distortion = 1.2%

Same question goes for SINAD. If, say, my DAC's SINAD is 100dB and my amp 90dB. What is my "system SINAD"?

You can compound, but because distortion and SNR are power based (instead of voltage based), you need to compound them by taking the square root of the sum of squares. Using your numbers, the calculations are:

DAC = 0.1% = 0.001
Amp = 0.1% = 0.001
Speakers = 1% = 0.01
Total = sqrt(0.001^2 + 0.001^2 + 0.01^2) = 0.0100995 = 1.01%

When the numbers are in dB, you'll need to convert them back to ratios first. And don't forget the negative sign!

DAC = -100 dB = 10^(-100/20) = 10^-5
Amp = -90 dB = 10^(-90/20) = 10^-4.5
Total = sqrt((10^-5)^2 + (10^-4.5)^2)) = 3.32e-5

Convert back to dB = 20*log10(3.32e-5) = -89.6 dB

 

[vavan]

If that's not possible, you can probably manage ok results by getting the speaker as far from room boundaries as possible and then measuring semi-nearfield, say 20-30cm

outdoors is not an option for me but I could put each speaker in the same spot at least couple of meters away from both side walls. but I need more details please. given f208s height I suppose I could put mic pointed on axis against middle of two woofers first and also between tweeter and mid for another set of measurement. then what? use 0 degree calibarion file I guess. there are lot of options for dual and multitone generators in rew, not sure which ones should I try. oh, and probably correct options in RTA window settings also important

 

[Asylum Seeker]

There are plenty of people who deliberately choose high distorting systems like vinyl, SET valves and horns, claiming they sound more real, better than low distortion equipment. I don't suppose the frequency response is flat either!

S.

Turns out they are deluding themselves! Because, as this discussion suggests, the distortion they purposely induce and 'hear' is actually inaudible!!

 

[Blujackaal]

Turns out they are deluding themselves! Because, as this discussion suggests, the distortion they purposely induce and 'hear' is actually inaudible!!

Same could be easily said for the ones paranoid about about >0.6% @ 1KHz+ charts. Since it inaudible in music & there no hard data for audible even in sine waves, Which Etymotic said when talking about the ER4XR on head fi. Since they really skpetic people could tell the 4XR over a 2SE, Seen some mistake 0.49% as 1% in impressions?.

 

[Rick Sykora]

So, circling back to the OP, would you not buy an amp that distorts a few percent, but buy a speaker that does?

In both cases, you may assume the distortion is considered inaudible.

 

[ctrl]

So we now have a speaker with >5% distortion which isn't audible subjectively

Perhaps we should differentiate the topic a little more.

Is 5% distortion is audible at 50Hz@85dB, most likely not. At 1kHz@85dB for a music signal with little possibility of sound masking, very likely yes.

Furthermore, one should always be aware that distortions are non-linear when the sound pressure increases. This means that even with a small increase in sound pressure, the distortions can literally "shoot into the sky". Therefore it makes sense to have a buffer that is as "big as possible".

Depending on the music material, a floorstanding loudspeaker in a large room can be required to provide well over 100dB@1m for a short period of time, which can result in double-digit distortions for some speakers.

If music material provides little masking and is more like an artificial two-tone or multi-ton test signal, the threshold for distortion drops dramatically. With an artificial two-tone test, even 0.3% distortion can be perceived by listeners.

At a piano concert, a piano keystroke can be more than 30dB above average - e.g. Heather Pierson - When You Wish Upon A Star

In this case you would have to look at the masking effects of a single tone.
Source: https://de.wikipedia.org/wiki/Maskierungseffekt
(Drawing is only available in the German version)

At 80dB of the noise signal (in our case, this would be the piano keystroke at 1kHz), HD2 (second order harmonic distortion at 2kHz) would be masked up to about 40dB (1%). HD3 would only be masked up to about 30dB (0.3%). The orange masking curve with the number 80 applies.


Or think of a jazz trio with piano, for example by the ingenious Ray Brown Trio. Here too, the music material provides only a few masking possibilities for resulting distortions.


My recommendation (as always ):
Try the Klippel hearing test or just install a sine wave generator and do different tests yourself.
http://www.klippel.de/listeningtest/?v=3

 

[Rick Sykora]

Perhaps we should differentiate the topic a little more.

Is 5% distortion is audible at 50Hz@85dB, most likely not. At 1kHz@85dB for a music signal with little possibility of sound masking, very likely yes.

Furthermore, one should always be aware that distortions are non-linear when the sound pressure increases. This means that even with a small increase in sound pressure, the distortions can literally "shoot into the sky". Therefore it makes sense to have a buffer that is as "big as possible".

Depending on the music material, a floorstanding loudspeaker in a large room can be required to provide well over 100dB@1m for a short period of time, which can result in double-digit distortions for some speakers.

If music material provides little masking and is more like an artificial two-tone or multi-ton test signal, the threshold for distortion drops dramatically. With an artificial two-tone test, even 0.3% distortion can be perceived by listeners.

At a piano concert, a piano keystroke can be more than 30dB above average - e.g. Heather Pierson - When You Wish Upon A Star
View attachment 58715
In this case you would have to look at the masking effects of a single tone.
Source: https://de.wikipedia.org/wiki/Maskierungseffekt
(Drawing is only available in the German version)
View attachment 58717
At 80dB of the noise signal (in our case, this would be the piano keystroke at 1kHz), HD2 (second order harmonic distortion at 2kHz) would be masked up to about 40dB (1%). HD3 would only be masked up to about 30dB (0.3%). The orange masking curve with the number 80 applies.


Or think of a jazz trio with piano, for example by the ingenious Ray Brown Trio. Here too, the music material provides only a few masking possibilities for resulting distortions.
View attachment 58716

My recommendation (as always ):
Try the Klippel hearing test or just install a sine wave generator and do different tests yourself.
http://www.klippel.de/listeningtest/?v=3

So I think you are saying you think masking may not always be effective, so your answer is it depends?

Or is it "If you choose not to decide, you still have made a choice"

 

[andreasmaaan]

outdoors is not an option for me but I could put each speaker in the same spot at least couple of meters away from both side walls. but I need more details please. given f208s height I suppose I could put mic pointed on axis against middle of two woofers first and also between tweeter and mid for another set of measurement. then what? use 0 degree calibarion file I guess. there are lot of options for dual and multitone generators in rew, not sure which ones should I try. oh, and probably correct options in RTA window settings also important

For IMD measurements, you only measure one driver at a time (since one driver won't intermodulate as a result of signal reproduced by another driver).

With the F208, your crossover frequencies are 270Hz and 2200Hz. Measure each driver using two tones that are within that driver's operating range.

For example, in this case the midrange covers 270Hz-2200Hz. I would choose a tone towards the bottom of that range (but not within the roll-off of the crossover filter) and another towards the top. Maybe 400Hz and 1500Hz, say? You'll need to select the "custom" option in REW for this. Place the mic near the driver you're testing. The RTA window settings won't be important, as you won't be windowing this measurement (instead you're going to minimise the effect of reflections by keeping the mic close to the source and far from reflective surfaces). So you can just open the generator, generate the tones, and hit "record" in the RTA window.

I think that will do the job (I'm no expert on REW though).

 

[andreasmaaan]

At a piano concert, a piano keystroke can be more than 30dB above average - e.g. Heather Pierson - When You Wish Upon A Star

In this case you would have to look at the masking effects of a single tone.

Good post.

Just a small quibble with this statement above: a piano keystroke contains not just a fundamental tone, but also many harmonics (plus tones that are not harmonics). So it's not quite that simple.

But I agree that looking at masking thresholds is the best way to go about determining distortion audibility thresholds.

 

[Rick Sykora]

Good post.

Just a small quibble with this statement above: a piano keystroke contains not just a fundamental tone, but also many harmonics (plus tones that are not harmonics). So it's not quite that simple.

But I agree that looking at masking thresholds is the best way to go about determining distortion audibility thresholds.

Great points on both counts. but my question already conceded the audibility issue.

 

[andreasmaaan]

Great points on both counts. but my question already conceded the audibility issue.

Sorry, didn't quite catch what you mean by the "audibility issue"?

 

[tuga]

Perhaps we should differentiate the topic a little more.

Is 5% distortion is audible at 50Hz@85dB, most likely not. At 1kHz@85dB for a music signal with little possibility of sound masking, very likely yes.

Furthermore, one should always be aware that distortions are non-linear when the sound pressure increases. This means that even with a small increase in sound pressure, the distortions can literally "shoot into the sky". Therefore it makes sense to have a buffer that is as "big as possible".

Depending on the music material, a floorstanding loudspeaker in a large room can be required to provide well over 100dB@1m for a short period of time, which can result in double-digit distortions for some speakers.

If music material provides little masking and is more like an artificial two-tone or multi-ton test signal, the threshold for distortion drops dramatically. With an artificial two-tone test, even 0.3% distortion can be perceived by listeners.

At a piano concert, a piano keystroke can be more than 30dB above average - e.g. Heather Pierson - When You Wish Upon A Star
View attachment 58715
In this case you would have to look at the masking effects of a single tone.
Source: https://de.wikipedia.org/wiki/Maskierungseffekt
(Drawing is only available in the German version)
View attachment 58717
At 80dB of the noise signal (in our case, this would be the piano keystroke at 1kHz), HD2 (second order harmonic distortion at 2kHz) would be masked up to about 40dB (1%). HD3 would only be masked up to about 30dB (0.3%). The orange masking curve with the number 80 applies.


Or think of a jazz trio with piano, for example by the ingenious Ray Brown Trio. Here too, the music material provides only a few masking possibilities for resulting distortions.
View attachment 58716

My recommendation (as always ):
Try the Klippel hearing test or just install a sine wave generator and do different tests yourself.
http://www.klippel.de/listeningtest/?v=3

One of my test tracks is the first movement of Rachmaninoff's Symphonic Dances Op. 45.

According to the Wikipedia, the ~100-strong instrumentation is as follows:

1 piccolo, 2 flutes, 2 oboes, 1 cor anglais, 2 clarinets, 1 bass clarinet, 1 alto saxophone, 2 bassoons, 1 contrabassoon, 4 horns, 3 trumpets, 3 trombones, 1 tuba, 1 timpan, 1 triangle, 1 tambourine, 1 side drum, 1 cymbals, 1 bass drum, 1 tamtam, 1 xylophone, 1 glockenspiel, 1 tubular bells, 1 harp, 1 piano, and strings (possibly 16 violins, 14 violins, 12 violas, 12 cellos, 10 double basses)


I can smell the IMD as I type.

 

[Rick Sykora]

Sorry, didn't quite catch what you mean by the "audibility issue"?

Essentially that the distortion was inaudible...

Was just seeing if the OP could be driven to a decision (or not). The question has been discussed, but no clear cut decision as best I can tell. Maybe has to be restated or done differently, but was just interested in whether there might be consensus an actionable result.

Going to mull it over a bit. May just need another approach.