Hi to Everyone !
I seem to see a continuous and enthusiastic search to achieve the lowest possible levels of harmonic distortion in circuits
Every now and then I read that below a certain level (that I have not understood well) any improvement in terms of lowering harmonic distortion practically becomes inaudible in normal listening conditions and can be appreciated only with the best testing lab gear
Question Is there any agreement on what this level is?
Thank you very much and best regards
gino
P.S. I well understand that The lounge should be a place to discuss topics that are not audio related,
But i have sometimes doubts so generic that do not fit in any specific discussion and other members could easily get annoyed by maybe trivial topics
Thank you very much
I seem to see a continuous and enthusiastic search to achieve the lowest possible levels of harmonic distortion in circuits
Every now and then I read that below a certain level (that I have not understood well) any improvement in terms of lowering harmonic distortion practically becomes inaudible in normal listening conditions and can be appreciated only with the best testing lab gear
Question Is there any agreement on what this level is?
Thank you very much and best regards
gino
P.S. I well understand that The lounge should be a place to discuss topics that are not audio related,
But i have sometimes doubts so generic that do not fit in any specific discussion and other members could easily get annoyed by maybe trivial topics
Thank you very much
Last edited:
It seems somewhat well established that 0.01 % (-80 dBc) is about the threshold of audibility for harmonic distortion for those with good ears. Belcher did some work on that in the 1970s.
Some may argue that if some can hear 0.01% let's just use 0.001% (-100 dBc) as a target to have some margin.
Some may further argue that pushing beyond -100 dBc is "a numbers game" and I would tend to agree with them ... at least to a point. I happen to enjoy that "numbers game" and know that it takes real understanding of physics, materials and such to push beyond, say, -120 or -130 dBc, so I tend to value those designs. Also, if you improve distortion in an audio circuit, you tend to improve a wide range of other parameters as well. And that's worth taking.
I would also argue that if someone was only able to achieve -100 dBc from, say, a DAC rated at -130 dBc I would be rather disappointed.
Hope this addresses some of your question.
Tom
Some may argue that if some can hear 0.01% let's just use 0.001% (-100 dBc) as a target to have some margin.
Some may further argue that pushing beyond -100 dBc is "a numbers game" and I would tend to agree with them ... at least to a point. I happen to enjoy that "numbers game" and know that it takes real understanding of physics, materials and such to push beyond, say, -120 or -130 dBc, so I tend to value those designs. Also, if you improve distortion in an audio circuit, you tend to improve a wide range of other parameters as well. And that's worth taking.
I would also argue that if someone was only able to achieve -100 dBc from, say, a DAC rated at -130 dBc I would be rather disappointed.
Hope this addresses some of your question.
Tom
Hi, about this : Great numbers do look great on paper and one more 0 after a few of them in THD might sound even better just because the " placebo " effect it might have. ( considering we can't hear difference between that low THD , and it's only in our mind that it sounds better )
I just wondered why go 0.00005% THD if the speaker drivers have distortion in the Percentage range.
How can we tell a 0.1% from a 0.01% THD amp on speakers that have 1% THD themselfes ?.
How is Speaker distortion different from Amplifier distortion ?.
In amplifiers it's the signal with harmonics , clipped signal . In speakers there is " mechanical distortion ? " , the sound waves created, cone flex etc ? What does " make " distortion in speakers, and I'm not talking about over driving them.
I can hear a superb sounding full range 5" inch driver sound like crap on vocals if the cone excursion on bass is big, the bass sound good but the vocals become " weird ".
I just wondered why go 0.00005% THD if the speaker drivers have distortion in the Percentage range.
How can we tell a 0.1% from a 0.01% THD amp on speakers that have 1% THD themselfes ?.
How is Speaker distortion different from Amplifier distortion ?.
In amplifiers it's the signal with harmonics , clipped signal . In speakers there is " mechanical distortion ? " , the sound waves created, cone flex etc ? What does " make " distortion in speakers, and I'm not talking about over driving them.
I can hear a superb sounding full range 5" inch driver sound like crap on vocals if the cone excursion on bass is big, the bass sound good but the vocals become " weird ".
I've tested ( blind test ) different amps, ic's , discrete, diy. at the same power levels , a few watts, where THD was close to each other , and I would prefer one over the other, I think there's much more to the "Pleasantness" of sound than THD. from one design to another.
That's because it's not a HD, that's IMD, a unlinear distortion. It happens when the Bl isn't linear and/or symmetrical anymore. Additionally, the driver goes surely into the progression of the suspension.
E: You can get significantly more max spl if you use a hpf/subsonic in the range below the fb. That saves a huge amount of excursion and the distortion goes down immensely.
Last edited:
The harmonic specta is important to consider, too. Low levels of higher harmonics can quickly become offensive, but even 1% of H2 on its own will likely not offend.
Maybe good to remember that HD and IMD are caused by the same nonlinearity.
It's just two different ways to measure the one nonlinearity.
Jan
I doubt that. H2 is a harmonic distortion and as such not linear.
Yes, you are clearly right. I changed a few times what I wrote but I can't track down how that brain fart came together. I'm embarrassed and apologize for that stupid mistake. 😕
Hi thanks a lot for the very kind and valuable advice It helps for sure
i listen only to digital sources these days So my guess is that -100dB can be quite achievable by dacs
Some power amps instead are more challenged in getting low distortion figures
I read that someone speak highly of amps with a distortion spectrum far from unreprochable
https://www.stereophile.com/images/222Dar1082fig08.jpg
Perhaps we have an idealized idea of hi-fi and its purpose. But hi-fi devices and loudspeakers are also musical instruments that can be in tune or out of tune, which cannot be captured by the usual "distortions". It is, for example, about material resonances and their composition.
Loudspeakers are a mechanical example. If I place LS on wood or steel or aluminum or stone or rubber, the audibly tuning will be different every time. However, it will not be possible to detect this with the usual current peep measurement methods. And if it is, then it will not be easy to translate into hearing measurement methods.
Loudspeakers are a mechanical example. If I place LS on wood or steel or aluminum or stone or rubber, the audibly tuning will be different every time. However, it will not be possible to detect this with the usual current peep measurement methods. And if it is, then it will not be easy to translate into hearing measurement methods.
hi thanks a lot and yes i get the point Speaking of distortion i understand speakers are the weakest link in the chain
i do not knoe why but the only think that fascinates me about speakers are the drivers design and construction
After all drivers set the max performance that i can get from a speaker
I am not in LPs listening but also MC preamps can be very challenged by distortion
The problem with power amps could be the high gain they have to provide I really do not understand why they have decided for 1V power amp sensitivity while line stages can be designed to provide up to 10V with very low THD + noise
In the old days the line stage gain was commonly around 6-7 times
in some idiot tests i have tried to reduce the gain of an amp causing its destruction ... i knew and know nothing about stability A very mysterious issue For instance i do not know if it is the same to reduce the feedback resistor or increase the resistor to the ground from a stability point of view
Another question for the The Lounge
What i like of high power amp is the immense power reserve they usually have because the caps are charged at higher voltages
What i do not like is their very high gain
hi and thank you very much indeed for your advice I thought about odd order harmonics being more offensive
I have seen in some 1kHz sine simulations a raise of a peak at 3kHz higher than the 2kHz peak Quite strange
I understand that could be much more offensive of a higher 2kHz peak
I have the FiiO K3 (2021) using The ES9038Q2M DAC . Nothing super fancy here and it claims : -120dB THD .
There are way lower THD DAC's out there.
@ginetto61 Amp I'm using is Pioneer a702R 0,007% THD. There are better THD amps here on this forum.
I wouldn't buy new amps, Something I love about old tech that say Made in Japan, not China.
Old amps , swapping all the bad caps or DIY is the way IMO.
I wouldn't buy new amps, Something I love about old tech that say Made in Japan, not China.
Old amps , swapping all the bad caps or DIY is the way IMO.
Both can be done, usually a combination of the two. feedback resistor should be same value as the resistor to ground in the input path ( most of the time , I don't think it applies to all designs ).
Let's say input resistor is 50K and feedback is 50K with 1K to ground , 51 gain, you could increase the resistor to ground to 2.2K for 23 gain .
but most amps have 20 - 35 gain in the main amp , not the pre amp.
! ! ! And going lower gain might cause instability depending on design. ! ! ! .
I don't find the high gain in amps a problem. I usually keep my amp at 60% - 75% and change the volume from my DAC ( digital volume control ) .
Some might say it isn't good because signal to noise ratio is bad this way , keeping the amp " open " at high volume , feeding it lower signal.
But I can't hear hiss , noise even with my ear next to the speaker. and this way I don't accidently feed to much signal to the amp and cause distortion.
@ginetto61 What I dont really like is new music being so damn LOUD. I mean they don't have dynamic range anymore , the difference between the quite parts and the loud parts are almost non existent, they crank the gain so much and use tons of compression. It sounds lifeless.
Incredibly good question, and unfortunately you won't find a good answer to this.
Even at organizations like AES and the like, it's hard to find any proper research and literature on this.
The correlation between any number and actual audibility (incl psycho-acoustics) is extremely poor.
(fact not an opinion)
The people who tried to make a bit more sense of this all (like Geddes and some others), were never taken serious.
So be prepared for (yet another) thread full of personal subjective anecdotal stories and discussions that just fight over nit-picky details.
Even at organizations like AES and the like, it's hard to find any proper research and literature on this.
The correlation between any number and actual audibility (incl psycho-acoustics) is extremely poor.
(fact not an opinion)
The people who tried to make a bit more sense of this all (like Geddes and some others), were never taken serious.
So be prepared for (yet another) thread full of personal subjective anecdotal stories and discussions that just fight over nit-picky details.
That is factually untrue;
https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
page 36
Although somebody with some basic knowledge would already guessed that from the name "intermodulation distortion".
Meaning we have at least a couple of things intermodulating with each other = "interact" with each other. (to put it simply)
You can't intermodulate with only one component.
HD distortion has just a simple linear relationship, the best example of this is compliance in a loudspeaker.
An example of intermodulation would be BL vs cone excursion.
Where the higher frequencies intermodulate with the change in BL caused by the cone excursion.
Those are mechanical examples, but there are equivalent electrical examples to this as well.
On the scale of audibility intermodulation distortion (IMD) is far more audible than just HD distortion.
https://purifi-audio.com/blog/tech-...o-whats-the-point-of-low-distortion-drivers-4
Which is also completely logical, since IMD creates additional byproducts.
Some of which don't correlate well anymore with the original signal = always more audible.
But there is absolutely a plethora of articles and papers on IEEE and AES, as well as good other literature and books on this subject, so I am not gonna discuss that here.