• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Need for a general statement?

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Do your homework and increase your confidence(supportable) one way or the other. Opinion is not worth much in real terms re hearing in the broader arena. :rolleyes:

See my comments above citing the two studies I mention, in which thresholds lower than 0.1% were found to be reliably detectable.

And I don't appreciate the the tone of your remark. The scientific basis for my statement is clearly laid out in my comments on this thread - perhaps you didn't read them?

Anyway, I'm just posting these studies in another thread right now.
 

Wombat

Master Contributor
Joined
Nov 5, 2017
Messages
6,722
Likes
6,459
Location
Australia
See my comments above citing the two studies I mention, in which thresholds lower than 0.1% were found to be reliably detectable.

And I don't appreciate the the tone of your remark. The scientific basis for my statement is clearly laid out in my comments on this thread - perhaps you didn't read them?

Anyway, I'm just posting these studies in another thread right now.

" I wouldn't be confident .................. Probably ......................... probably". Your words. That is not confidence.

I don't appreciate your indignance. :(
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
" I wouldn't be confident .................. Probably ......................... probably". Your words. That is not confidence.

I don't appreciate your indignance. :(

Let me rephrase my original reply to your comment then:

I'm 100% certain based on scientific studies that 0.1% distortion is readily detectable in many circumstances.

What I'm not certain of is exactly what the distortion profile would have to be for this level of distortion to tip above the threshold, since the research I'm aware of addresses this question only indirectly. I believe based on the much evidence concerning distortion audibility that upper order harmonics would be much more likely to be audible than lower lower order harmonics at this level.

I'm posting these studies in another thread to allow for further discussion and analysis by other members.

EDIT: here is the new thread on these two studies.
 
Last edited:

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
3,440
Likes
9,100
Location
Suffolk UK
Let me rephrase my original reply to your comment then:

I'm 100% certain based on scientific studies that 0.1% distortion is readily detectable in many circumstances.

What I'm not certain of is exactly what the distortion profile would have to be for this level of distortion to tip above the threshold, since the research I'm aware of addresses this question only indirectly. I believe based on the much evidence concerning distortion audibility that upper order harmonics would be much more likely to be audible than lower lower order harmonics at this level.

I'm posting these studies in another thread to allow for further discussion and analysis by other members.

EDIT: here is the new thread on these two studies.

Perhaps so, but I don't know of any mechanism in amplifiers that could generate high levels of upper harmonics without also generating much higher levels of lower harmonics. In other words, amplifier distortions have their highest levels of second or third harmonic, then reducing for higher harmonics. If anyone knows of a mechanism that doesn't do this, I'd very much like to know about it, as every spectrum analysis I've seen of amplifier distortion shows higher harmonics reducing in level.

S.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Perhaps so, but I don't know of any mechanism in amplifiers that could generate high levels of upper harmonics without also generating much higher levels of lower harmonics. In other words, amplifier distortions have their highest levels of second or third harmonic, then reducing for higher harmonics. If anyone knows of a mechanism that doesn't do this, I'd very much like to know about it, as every spectrum analysis I've seen of amplifier distortion shows higher harmonics reducing in level.

S.

I agree with you. I didn't mean to suggest that some devices had only upper order harmonics or higher levels of upper order harmonics than lower order harmonics. Rather, I was noting that the threshold of audibility of upper order harmonics seems to be lower than the threshold for lower order harmonics. Thus, two devices might have the same absolute level of THD, but if one device has its harmonic distortion concentrated in the lower orders while the other device has it spread across both lower and upper orders, the distortion created by the latter device is more likely to be audible.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Perhaps so, but I don't know of any mechanism in amplifiers that could generate high levels of upper harmonics without also generating much higher levels of lower harmonics. In other words, amplifier distortions have their highest levels of second or third harmonic, then reducing for higher harmonics. If anyone knows of a mechanism that doesn't do this, I'd very much like to know about it, as every spectrum analysis I've seen of amplifier distortion shows higher harmonics reducing in level.

S.

There are quite a few but most do not fit "classical" amplifier distortion curves. I think, with only a few seconds' thought, that most of the mechanisms relate to instability and/or coupling. Insufficient loop margin can cause higher harmonics to be emphasized and add high-frequency tones to the signal. The latter may not be harmonically related to the signal; ringing is one example. LF instability is also possible, e.g. "motor-boating" caused by an oscillation in the power supply bleeding through to the output. As for coupling, well, anything coupling enough to influence the output can cause all sorts of bad things to happen and strange sounds to result. Clock coupling in digital circuits, SMPS coupling, etc. can cause high-frequency (and low-frequency) tones in the output. Overdriving the amplifier at any point (any stage within the amplifier from input to output) can cause distortion "spray" weighted toward higher frequencies (and again may not be harmonically related to the signal) depending upon the behavior of the amp when overdriven.

Etc.
 

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
3,440
Likes
9,100
Location
Suffolk UK
There are quite a few but most do not fit "classical" amplifier distortion curves. I think, with only a few seconds' thought, that most of the mechanisms relate to instability and/or coupling. Insufficient loop margin can cause higher harmonics to be emphasized and add high-frequency tones to the signal. The latter may not be harmonically related to the signal; ringing is one example. LF instability is also possible, e.g. "motor-boating" caused by an oscillation in the power supply bleeding through to the output. As for coupling, well, anything coupling enough to influence the output can cause all sorts of bad things to happen and strange sounds to result. Clock coupling in digital circuits, SMPS coupling, etc. can cause high-frequency (and low-frequency) tones in the output. Overdriving the amplifier at any point (any stage within the amplifier from input to output) can cause distortion "spray" weighted toward higher frequencies (and again may not be harmonically related to the signal) depending upon the behavior of the amp when overdriven.

Etc.
Yes,but all your examples are those of a broken amplifier, or one being used outside its design envelope.

An amplifier that isn't unconditionally stable isn't fit to be used in polite company, an amplifier used into clipping is being abused, and and as for clock or SMPS coupling, that's not harmonic distortion, that's noise at a specific frequency, unrelated harmonically to the fundamental although showing up in a conventional THD measure. I don't consider amplifiers either abused or broken as being valid, although I accept that there such amplifiers being sold commercially, some even for high prices. Some even have a following amongst the terminally stupid.

A normal, decent amplifier, anything I would recognise as an amplifier, doesn't do any of these things.

S.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Sorry, you did not specify the operating conditions, and I took distortion more broadly, natch... Seen too many amplifiers through the years that you would find impolite. I suppose load sensitivity is also out?
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Yes,but all your examples are those of a broken amplifier, or one being used outside its design envelope.

An amplifier that isn't unconditionally stable isn't fit to be used in polite company, an amplifier used into clipping is being abused, and and as for clock or SMPS coupling, that's not harmonic distortion, that's noise at a specific frequency, unrelated harmonically to the fundamental although showing up in a conventional THD measure. I don't consider amplifiers either abused or broken as being valid, although I accept that there such amplifiers being sold commercially, some even for high prices. Some even have a following amongst the terminally stupid.

A normal, decent amplifier, anything I would recognise as an amplifier, doesn't do any of these things.

S.

I'm not sure you need to turn to extreme/defective cases to find examples of amps that have what would generally be considered to have quite acceptable levels of THD, but with enough of it higher order to question whether it really would be inaudible.

Here's a quick side-by-side of Stereophile's measurements of two amps that have similar THD. The amp on the left has its strongest harmonic (the 3rd) at around -76dB, a bit of 5th and a touch of 2nd and 7th, but nothing much above the noise floor beyond that. The amp on the right has its strongest harmonic (the 2nd) at about -79dB (so, lower in level than the strongest harmonic of the other amp), but then has a string of higher order harmonics out to (apparently) infinity, all between around -90dB and -105dB.

I'm just eyeballing these graphs obviously, but both amps look to have about the same THD (the determination of which will always be dominated by the strongest one or two harmonics). Still, I'd say the amp on the right is more likely to produce audible distortion than the amp on the left, due to the higher order harmonics being so many and being 25-ish dB higher in level.

Untitled.png
 

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
3,440
Likes
9,100
Location
Suffolk UK
I'm not sure you need to turn to extreme/defective cases to find examples of amps that have what would generally be considered to have quite acceptable levels of THD, but with enough of it higher order to question whether it really would be inaudible.

Here's a quick side-by-side of Stereophile's measurements of two amps that have similar THD. The amp on the left has its strongest harmonic (the 3rd) at around -76dB, a bit of 5th and a touch of 2nd and 7th, but nothing much above the noise floor beyond that. The amp on the right has its strongest harmonic (the 2nd) at about -79dB (so, lower in level than the strongest harmonic of the other amp), but then has a string of higher order harmonics out to (apparently) infinity, all between around -90dB and -105dB.

I'm just eyeballing these graphs obviously, but both amps look to have about the same THD (the determination of which will always be dominated by the strongest one or two harmonics). Still, I'd say the amp on the right is more likely to produce audible distortion than the amp on the left, due to the higher order harmonics being so many and being 25-ish dB higher in level.

View attachment 13690
Why is the distortion spectrum being measured at 50Hz? That seems an odd frequency.

With a spectrum like those at 1kHz or above, I think both amplifiers are likely to be transparent as the Pass amplifier has higher distortion, but still well below 0.1% and predominantly third harmonic whilst the Evo has 20dB lower distortion albeit with more higher harmonics. However, from the 50Hz spectrums, I can't tell anything useful.

S
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Why is the distortion spectrum being measured at 50Hz? That seems an odd frequency.

With a spectrum like those at 1kHz or above, I think both amplifiers are likely to be transparent as the Pass amplifier has higher distortion, but still well below 0.1% and predominantly third harmonic whilst the Evo has 20dB lower distortion albeit with more higher harmonics. However, from the 50Hz spectrums, I can't tell anything useful.

S

I think you’ve misread the scale on one of the graphs. THD is about the same for both amps, not 20dB lower for the Evo.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
o_O

Whoops, missed that. Sorry. However, the 50Hz part still stands.

S

Yeh fair enough - I think John Atkinson explains his testing methodology somewhere on the stereophile website but I can't find it atm.

Why do you say the distortion profiles would be significantly different at 1KHz? (I'm good with psychoacoustics but a complete idiot when it comes to amplifier design.)
 

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
3,440
Likes
9,100
Location
Suffolk UK
Yeh fair enough - I think John Atkinson explains his testing methodology somewhere on the stereophile website but I can't find it atm.

Why do you say the distortion profiles would be significantly different at 1KHz? (I'm good with psychoacoustics but a complete idiot when it comes to amplifier design.)
I'm not saying they would necessarily be different, but could well be. Firstly, harmonics of 50Hz could be mains or power supply breakthrough, not necessarily real distortion. Secondly, we are far more tolerant of distortion at LF than at MF, so even highish distortion at 50Hz could be quite acceptable, and anyway would be largely masked by loudspeaker distortion at LF, which is usually much higher than at MF or HF. It's just a bizarre frequency to use as a general distortion measure rather than for a specific purpose. I appreciate that the tests may have been done with 60Hz mains so my first comment may not apply, but the oddity still remains. If I'm characterising an amplifier, I'll measure distortion at LF, probably 50Hz, deliberately to show up any mains effects, but also at 1k and 10k, bjt wouldn't present the 50Hz spectrum as characteristic of the amplifier.

S
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
I'm not saying they would necessarily be different, but could well be. Firstly, harmonics of 50Hz could be mains or power supply breakthrough, not necessarily real distortion. Secondly, we are far more tolerant of distortion at LF than at MF, so even highish distortion at 50Hz could be quite acceptable, and anyway would be largely masked by loudspeaker distortion at LF, which is usually much higher than at MF or HF. It's just a bizarre frequency to use as a general distortion measure rather than for a specific purpose. I appreciate that the tests may have been done with 60Hz mains so my first comment may not apply, but the oddity still remains. If I'm characterising an amplifier, I'll measure distortion at LF, probably 50Hz, deliberately to show up any mains effects, but also at 1k and 10k, bjt wouldn't present the 50Hz spectrum as characteristic of the amplifier.

S

Makes perfect sense, yeh :)

This measurement was done in the US so the 60Hz mains should not interfere with the measurement.

Given that the Evo is an older Tripath-based class D amp, I imagine that non-linear distortion increases at higher frequencies. The Pass amp probably doesn't do this although I'm not sure.

But anyway, the example was used simply to illustrate how a perfectly normal and acceptable amp may nevertheless produce relatively a lot of higher order harmonic distortion - which I think the Evo measurements demonstrate. I suggest that distortion from an amp with this kind of distortion profile is likely to be more audible than distortion from an amp with a similar THD % but where the harmonic distortion is more concentrated in the lower orders (like the Pass).

Cheers,
Andreas
 

tomelex

Addicted to Fun and Learning
Forum Donor
Joined
Feb 29, 2016
Messages
990
Likes
572
Location
So called Midwest, USA
The thing about ability to hear distortions when listening to music, (not sinewaves) is that you have to take into account loudness levels, fletcher munson curves type stuff, and then the harmonic spray of that particular amp. Then, for your to decide you are hearing distortion means it has to be pretty damn high (positive of -70db for the best of ears i would say) and how do you know you are hearing distortion unless it is way high, how do you know its not what was recorded or not? I would say in general if you hear noise, there is an issue, if you hear or think you hear distortion, it would be above -70db. I will link to something below when I find it for your own testing on music.. This is a sticky area but generalities could be made though IMO.

click on the word LINK in the third sentence
http://ethanwiner.com/audibility.html
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
The thing about ability to hear distortions when listening to music, (not sinewaves) is that you have to take into account loudness levels, fletcher munson curves type stuff, and then the harmonic spray of that particular amp. Then, for your to decide you are hearing distortion means it has to be pretty damn high (positive of -70db for the best of ears i would say) and how do you know you are hearing distortion unless it is way high, how do you know its not what was recorded or not? I would say in general if you hear noise, there is an issue, if you hear or think you hear distortion, it would be above -70db. I will link to something below when I find it for your own testing on music.. This is a sticky area but generalities could be made though IMO.

click on the word LINK in the third sentence
http://ethanwiner.com/audibility.html

Also see the first study I posted in this thread, where sophisticated measurements for distortion taking into account these factors are tested in controlled trials.
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
Maybe a combination of a strong mission statement/rules and then inforcement by moderation would be a good approach.
Ever read the stuff like this at HeadFi
"Due to the flame wars that erupt as a result, this, and the other forums (other than the Sound Science forum), are DBT and ABX-free zones and posts about either will be moved or deleted."
Something like that here from our viewpoint might be appropriate.
I have a similar position to Sal on this. I think tighter set of rules for some areas of the forum might help, say in the measurement threads that all replies need to be directly about the measurements of this product, and any listening tests needs of the product to be clearly stated as controlled. So questions about the value of a specific measurement, audibility, or subjective opinion would not be allowed. They can be discussed in other more general threads, and not used as a way of lowering the signal to noise level in the specific thread. So opinion or thought is banned by the forum, it just has to be in the correct place.
 

Thomas savage

Grand Contributor
The Watchman
Forum Donor
Joined
Feb 24, 2016
Messages
10,260
Likes
16,298
Location
uk, taunton
I have a similar position to Sal on this. I think tighter set of rules for some areas of the forum might help, say in the measurement threads that all replies need to be directly about the measurements of this product, and any listening tests needs of the product to be clearly stated as controlled. So questions about the value of a specific measurement, audibility, or subjective opinion would not be allowed. They can be discussed in other more general threads, and not used as a way of lowering the signal to noise level in the specific thread. So opinion or thought is banned by the forum, it just has to be in the correct place.

I’m still getting to grips with the review and measurements threads , I’m fairly happy with how the last big schiit thread went but there is a tendency when the discussion has reached and past it’s zenith for ill discipline to creep in.

This comes in a few diffrent forms but is a constant phenomenon, I have as a result started to close them to reply’s when I feel we get to that point.

As to new ‘rules’ and strictly inforced posting restrictions I’m not in favour of that, you are all the subjects of my judgement and that will continue to be the case.

You start making rules for everything and all you get is a bunch of guys using those rules to bash each other and ironically it actually makes my job harder as taken to the extreme so many will be in breach it gets ridiculous.

We are here to inform , encourage and have fun..
 
Top Bottom