• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

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

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,651
Likes
8,992
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. :)
Haha, I think I've lost the thread of this thread... Which distortion are we talking about here?
 

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,076
Likes
1,595
Location
Stow, Ohio USA
Fundamentally, I took the OP as asking whether %THD measurement for distortion applies equally to speakers as well as amplifiers. The only speaker distortion measurement that ASR (and most others) are supplying is %THD, so we have to use that.

Given this, would you buy an amplifier that had single digit %THD measurements (just like some of the speakers)? The initial response seems to be that the speaker distortion is masked at bass frequencies (implying that higher distortion would be acceptable for the amplifier at bass frequencies). Later in the thread, I mentioned that some speakers had measured 2% THD (at midrange frequencies). Now there appears to be less confidence (@ctrl post) over how well masked the THD might be in some circumstances.

In either of the 2 cases above, the question is whether the %THD measurement applies equally for amps and speakers? From what I read in the thread, the answer appears to be that the speaker measurement is different than the amplifier one. While I can see this reasoning, unless the speaker test is not repeatable, then it should have some value. So, just trying to drive out that value proposition so it might be applied when making a major purchase decision. :cool:
 
OP
Frank Dernie

Frank Dernie

Major Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
5,128
Likes
11,252
Location
Oxfordshire
Thread Starter #123
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.
Personally I have considered DACs, preamps and many power amps to have been audibly transparent for decades with only noise an occasional problem.
Since I have known speakers to be the limiting factor for 50 years I have been choosing speakers which I think have low distortion and minimal cabinet talk for decades.
I do hear considerable differences between speakers which I have always assumed was frequency response, distortion and power handling capacity. Now I can add evenness of the off axis FR too :)
Since reading Toole's book I have considered the shape of the lateral response a bit more but I have a fairly large room, listen to dynamic music - orchestral stuff where ffff really is a lot louder than pppp so a powerful enough amp and speakers which don't distort at higher levels are what I still look for.
Pretty well none of the speakers so far tested here would be any use to me in this room.
I currently have Tune Audio Animas which have high directivity, high efficiency low distortion but a relatively busy waterfall, which seems to make less difference than I expected.
Goldmund Epilog 1&2, never seen any measurements, chosen by ear for instrumental timbre and clean bass.
Devialet Phantom Silver, chosen because I was intrigued by the technology.
KEF LS50
Harbeth P3ES and Xtenders.
Yamaha NS1000M which were my main speakers for years.
ProAc EBS which are basically a 50l BBC style cabinet with ATC drive units.
 

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,076
Likes
1,595
Location
Stow, Ohio USA
Wow, saw your @Frank Dernie Anima post before, but did not know they were yours! And I thought my CBTs were different. I can usually extrapolate a speaker sound based on experience, but no frame of reference for those. So how big is your room and what amp are you using?

Also, would love to hear how the Goldmund and Devialet speakers compare?
 
Last edited:
OP
Frank Dernie

Frank Dernie

Major Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
5,128
Likes
11,252
Location
Oxfordshire
Thread Starter #125
Wow, saw your @Frank Dernie Anima post before, but did not know they were yours! And I thought my CBTs were different. I can usually extrapolate a speaker sound based on experience, but no frame of reference for those. So how big is your room and what amp are you using?

Also, would love to hear how the Golmund and Devialet speakers compare?
The room is 16x30 ft. I am using the Animas with a Job INTegrated at the moment, the Goldmunds with a Devialet Original d'Alelier.
I could easily live with the Phantoms from a SQ pov but they are controlled by a phone or computer app which I hate.
I am also concerned by their longevity since they are all surface mount components whereas everything else I have is old and repairable.
 

stevenswall

Addicted to Fun and Learning
Forum Donor
Joined
Dec 10, 2019
Messages
575
Likes
370
Location
Orem, UT
The room is 16x30 ft. I am using the Animas with a Job INTegrated at the moment, the Goldmunds with a Devialet Original d'Alelier.
I could easily live with the Phantoms from a SQ pov but they are controlled by a phone or computer app which I hate.
I am also concerned by their longevity since they are all surface mount components whereas everything else I have is old and repairable.
No kidding about the Phantoms. The SQ is good enough for me, great I'd say, as is the even dispersion, but I returned them for the same reasons... Software and repairability.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,064
Likes
10,935
Location
Riverview FL

scott wurcer

Major Contributor
Audio Luminary
Technical Expert
Joined
Apr 24, 2019
Messages
1,398
Likes
2,408
I wondered if IMD was something that could be produced in the air... A valid question, I suppose...
The SPL required is huge, air has to have distortion because the rarefaction half of a wave is limited by being a vacuum but this is at ~194dB SPL. Shockwaves caused by heated gas can exceed this, Krakatoa was estimated at 300dB.
 

tuga

Major Contributor
Forum Donor
Joined
Feb 5, 2020
Messages
2,396
Likes
2,269
Location
Oxford, England

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,651
Likes
8,992
Fundamentally, I took the OP as asking whether %THD measurement for distortion applies equally to speakers as well as amplifiers. The only speaker distortion measurement that ASR (and most others) are supplying is %THD, so we have to use that...

...the question is whether the %THD measurement applies equally for amps and speakers? From what I read in the thread, the answer appears to be that the speaker measurement is different than the amplifier one. While I can see this reasoning, unless the speaker test is not repeatable, then it should have some value. So, just trying to drive out that value proposition so it might be applied when making a major purchase decision. :cool:
Ok thanks, I see.

This is a very confused thread IMHO. Speaker distortion is not fundamentally less audible than amplifier distortion. I hope my posts haven't contributed to the impression it might be.

However, what's important in terms of audibility (of any harmonic distortion) is not particularly the % distortion, but also (1) its frequency, (2) its harmonic order, and (3) its envelope (by which I mean its relationship to SPL).

(Also important is IMD, which is not measured here as you mention.)

There are various reasons why two different kinds of harmonic distortion, although similar in terms of %, might be radically different in terms of audibility. Arguably, there are reasons why most loudspeaker harmonic distortion might tend to fall into the "less audible" category for a given % (loudspeaker distortion tends to be lower-frequency, lower-order, and highly level-dependent). But there is no fundamental difference between the audibility of loudspeaker distortion vs amplifier distortion.

(Sorry btw if this isn't news to you @Winkleswizard. I'm posting this again mostly to try to keep the thread on track.)

Given this, would you buy an amplifier that had single digit %THD measurements (just like some of the speakers)? The initial response seems to be that the speaker distortion is masked at bass frequencies (implying that higher distortion would be acceptable for the amplifier at bass frequencies).
I wouldn't buy an amp with 1% THD, since this level of THD will definitely be audible with at least some signals, and since there are plenty of cheap amps on the market that perform way better.

Having said that, I agree with the part of your post in bold. If an amp produced 1% distortion below 100Hz and 0.01% distortion above it, that would be very strange, but quite acceptable IMO. Looking at it in practice though, and based on the measurements on this site, if amps tend to distort more anywhere, it's actually in the high frequencies.

Later in the thread, I mentioned that some speakers had measured 2% THD (at midrange frequencies). Now there appears to be less confidence (@ctrl post) over how well masked the THD might be in some circumstances.
2% at what SPL though? The speakers on this site have, until a few days ago, been measured for harmonic distortion at SPLs ranging from 82dB to 100+ dB.

If a speaker measures 2% HD in the midrange at loud levels, but in its intended application is to be listened to at only quiet to moderate levels, it may never reach anywhere near 2% HD.

Also keep in mind that music tends to have a spectrum similar to pink noise, which means that for music playback, the SPLs in the midrange are likely to be significantly lower than in the bass. So, just because a speaker measures 2% midrange HD at say 90 or 100dB, does not mean it will play at such levels in the midrange under normal circumstances.

Another way to put this final point would be to say that the intended application (SPL) is also important in interpreting distortion graphs for speakers.
 
OP
Frank Dernie

Frank Dernie

Major Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
5,128
Likes
11,252
Location
Oxfordshire
Thread Starter #131
Looking at it in practice though, and based on the measurements on this site, if amps tend to distort more anywhere, it's actually in the high frequencies.
HiFi News quote distortion at 100Hz, 1kHz and 10kHz. I have always considered the 10kHz reading to be completely pointless since even the first harmonic is inaudible, at least to most people.
Am I wrong?
 

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
2,081
Likes
4,877
Location
Suffolk UK
HiFi News quote distortion at 100Hz, 1kHz and 10kHz. I have always considered the 10kHz reading to be completely pointless since even the first harmonic is inaudible, at least to most people.
Am I wrong?
You're not wrong that the 10k harmonic(s) is inaudible, but there's merit in measuring distortion at those three frequencies, as one can then reasonably assume that they will be fine in between. There may be some further merit in measuring at 20Hz, 1k and 10k as that then covers the whole audible spectrum. One issue with high frequency distortion is that whilst the harmonic of 10k will be inaudible, the non-linearity mechanism that gives rise to harmonic distortion also brings about intermodulation distortion, so intermodulation products between, say 10k and 11k will fall at 1k, 2k etc so well audible if of sufficient amplitude. If the harmonic distortion at 10k is sufficiently low, then it's fairly safe to assume that the IMD will be similarly low. I don't know of a mechanism in nan audio product by which harmonic distortion is low, but intermodulation distortion is high.

S.
 

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,651
Likes
8,992
I'd add that, even if the harmonic distortion sweep is quite clean, this isn't a guarantee that there isn't significant IM distortion.

I did my best to explain the issue here.

EDIT: fixed link.
 
Last edited:

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,076
Likes
1,595
Location
Stow, Ohio USA
Ok thanks, I see.

This is a very confused thread IMHO. Speaker distortion is not fundamentally less audible than amplifier distortion. I hope my posts haven't contributed to the impression it might be.

However, what's important in terms of audibility (of any harmonic distortion) is not particularly the % distortion, but also (1) its frequency, (2) its harmonic order, and (3) its envelope (by which I mean its relationship to SPL).

(Also important is IMD, which is not measured here as you mention.)

There are various reasons why two different kinds of harmonic distortion, although similar in terms of %, might be radically different in terms of audibility. Arguably, there are reasons why most loudspeaker harmonic distortion might tend to fall into the "less audible" category for a given % (loudspeaker distortion tends to be lower-frequency, lower-order, and highly level-dependent). But there is no fundamental difference between the audibility of loudspeaker distortion vs amplifier distortion.

(Sorry btw if this isn't news to you @Winkleswizard. I'm posting this again mostly to try to keep the thread on track.)



I wouldn't buy an amp with 1% THD, since this level of THD will definitely be audible with at least some signals, and since there are plenty of cheap amps on the market that perform way better.

Having said that, I agree with the part of your post in bold. If an amp produced 1% distortion below 100Hz and 0.01% distortion above it, that would be very strange, but quite acceptable IMO. Looking at it in practice though, and based on the measurements on this site, if amps tend to distort more anywhere, it's actually in the high frequencies.



2% at what SPL though? The speakers on this site have, until a few days ago, been measured for harmonic distortion at SPLs ranging from 82dB to 100+ dB.

If a speaker measures 2% HD in the midrange at loud levels, but in its intended application is to be listened to at only quiet to moderate levels, it may never reach anywhere near 2% HD.

Also keep in mind that music tends to have a spectrum similar to pink noise, which means that for music playback, the SPLs in the midrange are likely to be significantly lower than in the bass. So, just because a speaker measures 2% midrange HD at say 90 or 100dB, does not mean it will play at such levels in the midrange under normal circumstances.

Another way to put this final point would be to say that the intended application (SPL) is also important in interpreting distortion graphs for speakers.
Thanks for sharing your knowledge and I hope our dialog clears up some of the confusion. While aligning on your major points, I think one takeaway is the value of %THD as a speaker measurement is just as it applies for amplifiers. In either case, even if the distortion is well tolerated (yes, am avoiding using the term "inaudibility" as it is regarded as an absolute), a %THD measurement outlier is indicative of a quality compromise. For the amplifier case, 2% at @2kHz would be a major outlier. My contention is that this also true for the loudspeaker %THD measurement.

Given a valid %THD measurement test (stable conditions with trusted measurement equipment), my takeaway for a speaker that has 2% THD at 2kHz is that I personally would not buy that speaker. Just as with the amplifier, this would be a major outlier that indicates a quality compromise. For a speaker that could be issues with the cabinet or other hardware, a driver issue, crossover problems or some other design compromise. All else equal, am buying something without a known quality compromise. Like @Frank Dernie, my career experience was in a field that has safety ramifications, so maybe am just setting a high bar here. ;)
 
Last edited:

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,651
Likes
8,992
Great post @Winkleswizard.

I'm still concerned though about your speaking of % HD at a particular frequency (2% HD @ 2kHz) in the absence of any characterisation of:
  • spectrum of the harmonics (2nd, 3rd, 4th, etc.)
  • SPL
  • % and spectrum of distortion at other SPLs
Every speaker reaches 2% HD @ 2kHz at some SPL. If you're looking for a nearfield monitor or a speaker to play at moderate levels in a moderately-sized room, 2%HD @ 2kHz @ say 95dB is going to be quite well tolerated (to borrow your excellent term), especially if that distortion is lower-order and the speaker performs linearly at lower SPLs.

This last point brings me to one other small factor that I think has remained somewhat obscured so far in this discussion, and that is the distortion behaviour of a speaker at a range of different SPLs.

A better driver will produce very little distortion at low levels, but of course as it gets closer to its linearity limits, distortion will unavoidably rise. A poorer driver will also produce rising distortion as level rises, but may also produce certain levels of particular harmonics regardless of SPL.

An example of such a driver is shown here (in fact, this is a good bass driver, so I wouldn't call its performance "poor", unless it were used in a 2-way design):



Notice how distortion is fairly low overall in the midrange at moderately high SPLs. If all you saw were this driver's distortion performance at 95dB, you might think it reasonably good / suitable for a 2-way design: it's basically under 2% from 40Hz up, and predominantly 2nd and 3rd order.

However, now compare its performance at the 3 different SPLs. Note how levels of odd-order distortion (3rd, 5th, 7th) basically stay the same in the midrange regardless of level.

1% third harmonic distortion in the upper midrange is quite tolerable in a nearfield monitor at 95dB SPL (it will never be listened to that loud, and our sensitivity to distortion decreases as level increases anyway).

But at say 75dB or 80dB SPL, 1% 3HD in the upper midrange would be something I'd definitely want to avoid.

What we're looking at here is an inherent nonlinearity in the driver's performance, i.e. something that is present regardless how far the voice coil is displaced. Whether we're listening at 70dB or 95dB, this driver is going to be producing just over 1% third harmonic distortion in the upper midrange.

This is the kind of issue that is just not going to be captured by a single harmonic distortion sweep at a single SPL.* For a speaker that is intended to play at only moderate SPLs or only in the nearfield, it would be something I'd be more concerned about than whether the speaker's HD were under 2% overall at higher SPLs.

This is one reason why I've been asking in a couple of speaker measurement threads if Amir can take distortion measurements at at least two different SPLs: one moderate and one loud (I would suggest 86 and 96dB).

*Having said that, an educated guess can often be made based on the relationship between the harmonics at whatever SPL the single measurement is taken at.
 

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,076
Likes
1,595
Location
Stow, Ohio USA
Great post @Winkleswizard.

I'm still concerned though about your speaking of % HD at a particular frequency (2% HD @ 2kHz) in the absence of any characterisation of:
  • spectrum of the harmonics (2nd, 3rd, 4th, etc.)
  • SPL
  • % and spectrum of distortion at other SPLs
Every speaker reaches 2% HD @ 2kHz at some SPL. If you're looking for a nearfield monitor or a speaker to play at moderate levels in a moderately-sized room, 2%HD @ 2kHz @ say 95dB is going to be quite well tolerated (to borrow your excellent term), especially if that distortion is lower-order and the speaker performs linearly at lower SPLs.

This last point brings me to one other small factor that I think has remained somewhat obscured so far in this discussion, and that is the distortion behaviour of a speaker at a range of different SPLs.

A better driver will produce very little distortion at low levels, but of course as it gets closer to its linearity limits, distortion will unavoidably rise. A poorer driver will also produce rising distortion as level rises, but may also produce certain levels of particular harmonics regardless of SPL.

An example of such a driver is shown here (in fact, this is a good bass driver, so I wouldn't call its performance "poor", unless it were used in a 2-way design):



Notice how distortion is fairly low overall in the midrange at moderately high SPLs. If all you saw were this driver's distortion performance at 95dB, you might think it reasonably good / suitable for a 2-way design: it's basically under 2% from 40Hz up, and predominantly 2nd and 3rd order.

However, now compare its performance at the 3 different SPLs. Note how levels of odd-order distortion (3rd, 5th, 7th) basically stay the same in the midrange regardless of level.

1% third harmonic distortion in the upper midrange is quite tolerable in a nearfield monitor at 95dB SPL (it will never be listened to that loud, and our sensitivity to distortion decreases as level increases anyway).

But at say 75dB or 80dB SPL, 1% 3HD in the upper midrange would be something I'd definitely want to avoid.

What we're looking at here is an inherent nonlinearity in the driver's performance, i.e. something that is present regardless how far the voice coil is displaced. Whether we're listening at 70dB or 95dB, this driver is going to be producing just over 1% third harmonic distortion in the upper midrange.

This is the kind of issue that is just not going to be captured by a single harmonic distortion sweep at a single SPL.* For a speaker that is intended to play at only moderate SPLs or only in the nearfield, it would be something I'd be more concerned about than whether the speaker's HD were under 2% overall at higher SPLs.

This is one reason why I've been asking in a couple of speaker measurement threads if Amir can take distortion measurements at at least two different SPLs: one moderate and one loud (I would suggest 86 and 96dB).

*Having said that, an educated guess can often be made based on the relationship between the harmonics at whatever SPL the single measurement is taken at.

Good point and great animation to show your point. Also agree about different SPLs. I had written a longer post and the SPL got dropped in the edits. The NRC already has stated %THD testing levels and would follow their lead (with one exception). The NRC starts at 96 dB @ 1 meter and backs to a lower SPL if needed. Amir appears to have done so in his last couple of reviews. Since he has to consider that he not damage loaners, my suggestion would be to limit the start of the sweep to 80 Hz for small speakers.

Also agree spectrum is important when evaluating distortion audibility. As I stated earlier, I am willing to forgo the audibility argument if the %THD measurement (at 96 dB) is higher than most other speakers at 2kHz (or anything above 1% above 300 Hz). This is based on working with batches of drivers to find matched pairs and finding some that do not match the others well (whether in FR, distortion or SPL). As I mentioned though, this is my bar. This is not workable for ASR as Amir only has a sample size of one for a given speaker....

So, when I see Amir test %THD on other speakers (under comparable SPL conditions) and they do not have a high Q distortion spike at 2kHz, I am left guessing. Since it is my money and am risk averse, I would not buy that speaker. Others might argue the inaudibility and dismiss the issue. If they are happy, then there is more than enough wiggle room in this hobby to let them enjoy their choice(s). :)

I empathize with Amir in that it is not a productive use of his time to investigate questionable test results. In this case, my advice is to watch whether the vendor is willing to comment on the test results. I would not expect them to deal with the membership, but if they do not respond to Amir at all, that is a telltale sign.
 
Last edited:

Soundstage

Active Member
Joined
Jan 10, 2020
Messages
202
Likes
122
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
Have you had the chance to conduct the experiment?
 
Top Bottom