Is distortion really a problem for music reproduction ?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
There have been a number of well documented studies on the subject over the past 100 years or so. Google may turn up a few. Earl Geddes - a member here (Gedlee) has done some himself.

The audibility of distortion may be different in electronics and in loudspeakers. I seem to recall that some studies pointed to that, but you'd need to verify it.
i have to wonder how you alter one without the other.
dont we use amplifiers with speakers?
 
I tend to focus on noise rather than distortion. That's because distortion as shown in your first plot is an artifact of the measurement stimulus. Single sinewaves are highly improbable in music so in my view are an idiotic way to test an amplifier.

Single sine-waves, square-waves and sawtooth waves exist all over the place in synthetic / electronic music, such as trance and chiptunes.

Are you suggesting square-waves do not exist anywhere at all in normal pop music?

I think a square-wave at different frequencies is difficult for a DAC or amplifier to produce, thus it is a measurement pertaining to skill in difficulcy.
 
I'll have to re-read the thread in depth, but it's a valid question and worthy of discussion. My first concern is the OP cites Stereo-pile as an authoritative source. That's about as unbiased and objective an opinion as you'd get by asking a prostitute in a brothel how you should spend your money :)

The banner says we can't discuss religion or politics and indeed I won't. I will, however note that our hobby is like those two taboo subjects (and money too), is a breeding ground of all manner of ugly half-truths and outright lies. A little bit of science and objectivity go a long way, but unfortunately often leave the fragile human ego as a gibbering idiot when cold harsh facts put him in his place!

Here's a web site that is an excellent dose of reality in our fantasy-plagued hobby:

Why We Believe
 
Single sine-waves, square-waves and sawtooth waves exist all over the place in synthetic / electronic music, such as trance and chiptunes.

Curious - do you have a specific example where a single sinewave (at full amplitude) is found? I'm interested to learn more.

Are you suggesting square-waves do not exist anywhere at all in normal pop music?

I mentioned nothing in respect of squarewaves, they've been introduced by you.

I think a square-wave at different frequencies is difficult for a DAC or amplifier to produce, thus it is a measurement pertaining to skill in difficulcy.

If its truly a squarewave it'll be impossible for any physical thing to reproduce it. If its bandlimited then why would it be any kind of problem?
 
Member
Joined 2004
Paid Member
Curious - do you have a specific example where a single sinewave (at full amplitude) is found? I'm interested to learn more.



I mentioned nothing in respect of squarewaves, they've been introduced by you.



If its truly a squarewave it'll be impossible for any physical thing to reproduce it. If its bandlimited then why would it be any kind of problem?

Sine wave at FS in a track: Wanna Love You- Akon https://www.youtube.com/watch?v=GJzF7H2e3Tw The bass line actually goes very close (less than 1 dB) to full scale. I have been using it to test and adjust limiters in speakers. The bass is inaudible on computer speakers. Its pretty much a sine wave at about 70 Hz I think and really is loud and low when the system is capable. Its also a festival of misogyny so I don't play it in mixed company, except when they are young. Needless to say the Youtube track has been expurgated.

Kind of by definition any reproduced square wave will be bandlimited. Analog systems to 40 KHz at best, digital systems can get to about 80 KHz. Very few recording microphones get to 20 KHz lots roll off closer to 15 KHz or less. Acoustic risetimes faster than that would most likely be unhealthy for humans either directly or the source would be.
 
Member
Joined 2009
Paid Member
Hi and thank you a lot for the very interesting advice.
I want just to say that i am not convinced that this is the best amp in the world.
I am convinced instead that even with this level of distortion the sound can be very good.
Like it would be good for any other amp without serious design flaws ... like to little power supply for instance.
I think that max available clean current at the speakers terminal is key.
Just to explain. Some very famous speakers like the Apogee Scintillas with their 1 ohm impedance were called the amps killers with some amps even catch on fire.
Other amps with higher current delivery just drove them more or less properly.
So to end, provided an electrical compatibility between amp and load ... life will be beautiful also with a little distortion.
Of course the amp i mentioned has a huge power supply.
At 150 kUSD i think it can be done.
Thanks a lot again and kind regards, gino
 
Last edited:
Curious - do you have a specific example where a single sinewave (at full amplitude) is found? I'm interested to learn more.

At full amplitude? I don't think so, it's just normal music. Is the amplitude important in the sinewave?

I don't think a normal instrument like a violin would create a perfect perfect sinewave, but perhaps it can create waves which are fairly close?


I mentioned nothing in respect of squarewaves, they've been introduced by you.

If its truly a squarewave it'll be impossible for any physical thing to reproduce it. If its bandlimited then why would it be any kind of problem?

Sorry, you're right, I'm the one that started the square-wave tangent. I apologize.

In theory, a perfect square-wave can not be perfectly reproduced, at least that would be very difficult, but that's not what I'm quasi-addressing here, it's the skill in which the DAC, amplifier or transducer executes the square-wave, which we could measure for instance on a scale of 1 to 1000.
 
Sine wave at FS in a track: Wanna Love You- Akon https://www.youtube.com/watch?v=GJzF7H2e3Tw The bass line actually goes very close (less than 1 dB) to full scale. I have been using it to test and adjust limiters in speakers. The bass is inaudible on computer speakers. Its pretty much a sine wave at about 70 Hz I think and really is loud and low when the system is capable.


Nice! :snowman2:

Have any examples of square-waves? I'm not of the thought school which believes they are pointless to test, quite the opposite, I just like finding cool waves in music from time to time.
 
So my question is ... when distortion is low enough for you ?

I can hear the difference with every single op-amp I have ever tried, including blind / random.

No two sound the same, unless they are the same.

Therefore, I'm in the 'no limit' school of distortion, that it needs to be extremely low to become transparent, at least I assume so.

I respect that some people can't hear the difference, because these are very subtle differences at times.

I'm not super special or anything, but it's estimated around 10% of humans are missing a neural network which processes sound called "the superior branch of the arcuate fasciculus".

These are not subtle sounds they can't hear, rather very rough sounds like pitch. Here is an article on it Neural Pathway Missing In Tone-deaf People -- ScienceDaily

Why would humans need to hear very, very fine differences at all? I have some theories on that, but the simple answer is practice and choice.

Why do we fly down ski slopes? Is it necessary for survival to fly down ski slopes? Did we used to chase bears on ski slopes and the best skiers survived? No we just ski because we decided to, choice.
 
As Jan pointed out earlier using other words, one might use the analogy from the motor industry: Hp sells cars, torque wins races.

I guess with distortion it's about the same. The .0001% amp with the wrong distortion behaviour will sell better than the .1% amp with the killer-sound.

Basically it's a matter of behaviour, not of numbers. Unfortunately it's the numbers we are thaught to obey. Experience is hardly quantifiable, so it does not sell.
 
BV said:
What "special" nonlinearity should variable frequency difference to reveal?
Non-linearities closely related to a signal-envelope-sensitive filter. For example, the bias shift in a valve output stage caused by the combination of cathode bias RC network and second-order distortion in the valve. If the difference in frequency between the two sine waves becomes small enough then the cathode voltage will vary, thus introducing extra IM which is not present for larger frequency differences.

Similar things might happen in SS circuits - anywhere something is decoupled there is a frequency below which the decoupling begins to fail. This will be well below signal frequencies, but it may be within the range of envelope frequencies. The input signal contains no components at envelope frequency but some second-order distortion can introduce them. In some cases they can then react back so a second-order nonlinearity can produce third-order IM.
 
Member
Joined 2009
Paid Member
I can hear the difference with every single op-amp I have ever tried, including blind / random.
No two sound the same, unless they are the same.
Therefore, I'm in the 'no limit' school of distortion, that it needs to be extremely low to become transparent, at least I assume so.
I respect that some people can't hear the difference, because these are very subtle differences at times.
I'm not super special or anything, but it's estimated around 10% of humans are missing a neural network which processes sound called "the superior branch of the arcuate fasciculus".
These are not subtle sounds they can't hear, rather very rough sounds like pitch. Here is an article on it Neural Pathway Missing In Tone-deaf People -- ScienceDaily
Why would humans need to hear very, very fine differences at all?
I have some theories on that, but the simple answer is practice and choice.
Why do we fly down ski slopes? Is it necessary for survival to fly down ski slopes? Did we used to chase bears on ski slopes and the best skiers survived? No we just ski because we decided to, choice.

Hi and thanks for the very interesting advice
I conclude that low distortion for you is of paramount importance
Honestly i am not sure anymore ... when an amp with relatively high distortion amazes trained listeners ... if it sounds good must be good ?
Or maybe there are other things more important than distortion
I have in mind a very simple test to see for instance the influence on perceived sound of lower damping factor
Just a resistor in series with the outout of the amp to get a damping factor of 30 on 8 ohm, like the amp mentioned
My present amp has 150 ...
Kind regards, gino
 
Distortion is definitely used at recording of rock music, it belongs to the nature of this kind of music. If the reproducing chain produces distortion of similar nature, the music becomes more involving. Same might apply to other genres, like jazz, classical, instrumental. Here I don't talk about the nature of distortion because I have little knowledge about it. But definitely not the distortion of clipping/soft limiting. Tube amplifiers usually add pleasant distortion way below limiting, even at low level.
 
oshifis said:
If the reproducing chain produces distortion of similar nature, the music becomes more involving.
If the "reproducing chain" adds significant distortion then it is not a reproducing chain but part of the musical performance. I certainly don't want this. However, those who like some distortion may prefer more distortion - perhaps the extra distortion at home compensates for the lack of volume when compared to a live gig?
 
So my question is ... when distortion is low enough for you ?
When it is truly inaudible. Pretty graphs performed on sine waves don't necessarily reflect what happens when we listen to dynamic content (music).

While speakers vary more widely in sound quality than electronics, most amplifiers do sound different from each other to these ears despite the fact that their *measured distortion* is very similar.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.