As a general rule, technical people like a very low THD. This is a preference arising from knowledge of the technology, and a belief that an amplifier should amplifier correctly, with no distortion. However, many non-technical people, possibly a majority of audiophiles, often like a 'sounds effects' amp. Given that all recording processes, amps, and speakers ALL add distortions, the truth is that we all listen to distortion whenever we use audio gear. It is also obvious that there is a spectrum of audiophiles, those who love clean, low distortion sound and others that like an 'effects box', and clearly all audiophiles lie somewhere across this spectrum.
Tube amps give high levels of H2, typically at about 45dB down on the fundamental for zero global feedback amps. The triode does this, and the transformer adds some H3 as well. When you measure a tube amp at listening levels it generally delivers around 1% THD or even higher. This is a horrific figure compared to SS amps, but there are lots of audiophiles that love it. In uncovering this fact we identify a market where manufacturers can pitch their products....
It seems some harmonics are OK, but others are verboten. In truth, the higher and odd order harmonics are particularly objectionable and really screw with the musicality and add listener fatigue. But it is important to recognise that all amplifiers distort, some more than others, and some with different types of distortion which you analyse by measuring the harmonics of a fundamental.
So, H2, H3 and H4 are OK (within reason) but H5, H6 and beyond are NOT OK. It is important to say that were are talking here about single tome testing, not music, which introduces much more complexity which leads to intermodulation.
A simple metric here is to calculate the percentage of THD is comprised ONLY of H2+H3+H4. If that percentage is higher than 99% (use FFT on LTSpice), and THD is less than 0.2% the amp will likely 'sound' pretty good. This is a simple preference; nothing is better than another, it is simply that some listeners like one, others like another.
Just how you define 'good sound' is moot. Again, you can use statistics but only by comparative methods. Find 25 audiophiles, and play the same music on two or more amps with same speakers and source. Simply choose which audiophile selects the 'better' sound. You are not measuring anything of course, this is subjective, but you are finding the 'preferred' amp which will give you a sample, not a good or bad judgment. This might lead you to the type of amplifier sound YOU like, and this could be an experiment you tested jumt to identify a market point to pitch a future product.
I ask these questions to figure out if the audiophile sound conundrum has changed in the last ten years. In recent years there is small change; measurement, which is always improving. It is now possible to analyse the harmonic profile of an amp with simple test equipment (eg QuantAsylum QA403). From the harmonic profile and the THD, it might be possible to design for a particular amplifier 'sound' for a suitable sector of the market. And I agree that a particular 'sound' is highly individual, but if 80% of the audiophile group, say 25 people, like a particular amp (a SET over an AB SS, for example) you could direct your R&D to the amp which brings more sales.
HD
Tube amps give high levels of H2, typically at about 45dB down on the fundamental for zero global feedback amps. The triode does this, and the transformer adds some H3 as well. When you measure a tube amp at listening levels it generally delivers around 1% THD or even higher. This is a horrific figure compared to SS amps, but there are lots of audiophiles that love it. In uncovering this fact we identify a market where manufacturers can pitch their products....
It seems some harmonics are OK, but others are verboten. In truth, the higher and odd order harmonics are particularly objectionable and really screw with the musicality and add listener fatigue. But it is important to recognise that all amplifiers distort, some more than others, and some with different types of distortion which you analyse by measuring the harmonics of a fundamental.
So, H2, H3 and H4 are OK (within reason) but H5, H6 and beyond are NOT OK. It is important to say that were are talking here about single tome testing, not music, which introduces much more complexity which leads to intermodulation.
A simple metric here is to calculate the percentage of THD is comprised ONLY of H2+H3+H4. If that percentage is higher than 99% (use FFT on LTSpice), and THD is less than 0.2% the amp will likely 'sound' pretty good. This is a simple preference; nothing is better than another, it is simply that some listeners like one, others like another.
Just how you define 'good sound' is moot. Again, you can use statistics but only by comparative methods. Find 25 audiophiles, and play the same music on two or more amps with same speakers and source. Simply choose which audiophile selects the 'better' sound. You are not measuring anything of course, this is subjective, but you are finding the 'preferred' amp which will give you a sample, not a good or bad judgment. This might lead you to the type of amplifier sound YOU like, and this could be an experiment you tested jumt to identify a market point to pitch a future product.
I ask these questions to figure out if the audiophile sound conundrum has changed in the last ten years. In recent years there is small change; measurement, which is always improving. It is now possible to analyse the harmonic profile of an amp with simple test equipment (eg QuantAsylum QA403). From the harmonic profile and the THD, it might be possible to design for a particular amplifier 'sound' for a suitable sector of the market. And I agree that a particular 'sound' is highly individual, but if 80% of the audiophile group, say 25 people, like a particular amp (a SET over an AB SS, for example) you could direct your R&D to the amp which brings more sales.
HD
That's a strong word, but you are correct. It almost seems like religion with some people.
Which would you prefer (and it is simply a preference with no judgement on my part) in a speaker level amplifier?
9th harmonic dominant at 120dB down from the fundamental. No other measurable harmonic distortion above 120dB below the fundamental. Distortion characteristics remain constant throughout the rated frequency response 20-20k (not just 1kHz), power spectrum (not just 1W), and into loads varying from 2R to 16R.
2nd harmonic dominant at 45dB down from the fundamental. No other measurable harmonic distortion above 120dB below the fundamental at 1W into 8R 1kHz. However, the distortion rises linearly with power delivery to rated power, and the distortion characteristics are not consistent with load and/or across the frequency spectrum.
Neither of those would be considered 'atypical' in today's world, I don't think.
What if your pre-amplifier had a 2nd harmonic dominant characteristic, but the distortion product was in the inverted phase to the second power amplifier above. Would that alter your preference for your amplifier? Let's say it was an old tube pre-amp you adored.
I like it all, and I like to fool myself into thinking I can hear a difference with some things. Heck, I'm old, and I can barely hear out of my left ear now anyway. It's fun, and every now and again, I pick up a tidbit of useful information in these types of threads.
Cheers to all that have fun with it and share some tidbits.
For someone so passionate about their statement, there seems to be a lot of information refuting it.
As a general rule, THD can, and often does, vary with the input frequency, in my experience. Perhaps share what you mean precisely and share a reference.
These graphs are commonly distributed (THD vs. frequency). The example below, I might consider "good".
Perhaps I misinterpreted what you mean?
it doesn't say anything about phase, what if the phase is 160o off , it doesn't say nothing about rise time and fall time, can it play asymmetric , how many v/ns can it does
It is just good to give an idea of right/wrong when designing the same machine, like if the components are lower thd, if they sound is likely to improve when thd goes down.
It is just good to give an idea of right/wrong when designing the same machine, like if the components are lower thd, if they sound is likely to improve when thd goes down.
I had amps with 0.001% thd, very bad amps sound terrible,
I am not advocating amps with like 0.5 %, which I consider clipping limit.
It has been proven many times that we cannot hear 0.01% thd, and that 12bit 44khz is way enough to capture all the sound.
there is nothing above 12khz in music anyway, most of the information is around 60-500, where phase and noise floor needs to be very accurate.
I am not advocating amps with like 0.5 %, which I consider clipping limit.
It has been proven many times that we cannot hear 0.01% thd, and that 12bit 44khz is way enough to capture all the sound.
there is nothing above 12khz in music anyway, most of the information is around 60-500, where phase and noise floor needs to be very accurate.
What? Why would THD have anything to do with that?
You said ...
I have shown evidence of that being wholly incorrect and/or I misunderstood your meaning. Stay on topic if you are responding to me, please. I can't follow your train of logic re: phase or anything having to do with THD.
Otherwise carry on and have fun.
You don't understand that THD lacks this information,
phase , and rise and fall times are complementary to THD,
only the THD figures is not an indication of sound quality.
It is only practical to give an idea of the power/resolution of ONE signal continuous.
So it is practical to service a component, check biases, find a fault.
phase , and rise and fall times are complementary to THD,
only the THD figures is not an indication of sound quality.
It is only practical to give an idea of the power/resolution of ONE signal continuous.
So it is practical to service a component, check biases, find a fault.
I understand this perfectly well, thank you.
So what? Sometimes they are, and some times they are not.
I agree completely that is is not the SOLE indicator of sound quality, but it can be an indicator.
That is the counter opposite of what you previously said, if I interpreted you correctly, but let's just move along. You and I likely won't come to a common understanding, and that's just fine.
And much, much more... (to me).
I am telling you it is not an indicator of quality when you choose an audio component, meaning that an upgrade can be realized in qualitative and quantitative sound quality by retrograde to an amplifier which produces more thd at certain wattage
It’s a huge wall many run into by downgrading their system by replacing a component by one with more power less thd,
N Pass talked a lot about this first watt more important than the 60-100 watt upgrade
It’s a huge wall many run into by downgrading their system by replacing a component by one with more power less thd,
N Pass talked a lot about this first watt more important than the 60-100 watt upgrade
You don't need to tell me anything. I absolutely disagree with you, and that's perfectly OK.
I think THD is / can be an indicator of quality (when viewed with the correct lens, IMO).
You first claimed THD was the same for every frequency (by definition), then you said that it was only for one frequency (which is closer to being correct). Now you're saying that it is not an indicator of quality at all. We can disagree, and we can live peacefully together.
That's fair. I can see easily why an amplifier with higher THD (along with other things changing) could be both objectively and subjectively "better". What I'm curious about is why you seem to think I may disagree with that or may not understand that.
I don't have any evidence to the contrary or to support it... so I'll just say... OK... whatever to anecdotal thoughts. I don't agree or disagree.
Well, I'm not sure that's exactly what he meant (and it's a quote from someone else), but we can agree that First Watt and the first Watt might be very important. I support that fully.
I always enjoy the fun chats. We don't have to agree on everything, but I think I understand your POV a little better.
Cheers!
Last edited:
Does THD accurately predict good sound quality?
No
Does THD accurately predict record to sound transfer quality?
Yes
Don’t mix these two things and leave record creation for musicians. Let them add any distortions they want into the record and don’t add “salt” on your own even if they are pleasant. Enjoy what is in the record. If you don’t like it- do not listen and choose other record.
No
Does THD accurately predict record to sound transfer quality?
Yes
Don’t mix these two things and leave record creation for musicians. Let them add any distortions they want into the record and don’t add “salt” on your own even if they are pleasant. Enjoy what is in the record. If you don’t like it- do not listen and choose other record.
But what if my body is less sensitive to "salt" than yours? What if the "salt" that I need to add makes my brain perceive it more closely to what those that don't need to add salt perceive or closer to what the "engineer intended".
Why would you care?
It sounded awfull but only because i had a single signal reference and that there was other harmonics at lower level than H2/H3.
Guess that with an actual musical signal it would be impossible to spot, neverless i would advocate for 0.01% THD,
not because it is absolutely necessary but because it is feasible without much components and effort and that low THD
also mean low IMD.
It is just my take. You do what you want with “salt”. I would like to hear closer to what sound engineer heard. This approach would be closer to standardization. Otherwise there is no agreement on which side and how much “salt” is added. I don’t afraid of loosing freedom here as reflections plays a huge role on top off that. But at least some standardization i would like to have.