THD has a much bigger effect on sound than you think

[mike7877]

I think I get what you're trying to say. But the thing is, you would only get an increase in amplitude in the frequencies where the harmonics exactly overlap? And even then it's not: Distortion1 + Distortion2 = DistortionTotal, it's: SQRT(Distortion1^2 + Distortion2^2) = DistortionTotal. Meaning that the biggest contributor will by far be the dominating one. You would need a crapload of distortion over a wide frequency range to get anywhere remotely close to a 2dB increase?

You don't need to increase the entire signal by 2db for a small part of it to change significantly (ie. a voice vs the rest of the band)

 

[solderdude]

So, subtract 15db for the crest factor and 20-30db for treble being that much quieter. You're at -35 to -45db now. If harmonic distortion is at -60, it's 15-25db down. Audible, easily audible

Odd and even harmonics are present in every musical instrument as they are part of the tone. When you very, very slightly increase the level of each harmonic (due to distortion) that tone does not change one bit and thus you cannot hear the distortion.
You can probably hear it with a pure constant tone but NOT with music.

 

[Inner Space]

Third harmonic is NOT (!) related to the music. It is like a wrong note. It is a clash, more intrusive than white noise.

Is that right, though? Isn't the third harmonic a perfect fifth, above the octave? Which is a foundational interval in music, surely, and perfectly consonant and stable, even if compared to unison and octave. To say a perfect fifth is "not related to the music" would come as a surprise to most folk.

 

[mike7877]

Is that right, though? Isn't the third harmonic a perfect fifth, above the octave? Which is a foundational interval in music, surely, and perfectly consonant and stable, even if compared to unison and octave. To say a perfect fifth is "not related to the music" would come as a surprise to most folk.

If it's not a part of the scale, it is not related

 

[Inner Space]

If it's not a part of the scale, it is not related

Isn't a perfect fifth part of the scale? Try it on a piano.

 

[Pdxwayne]

I think I get what you're trying to say. But the thing is, you would only get an increase in amplitude in the frequencies where the harmonics exactly overlap? And even then it's not: Distortion1 + Distortion2 = DistortionTotal, it's: SQRT(Distortion1^2 + Distortion2^2) = DistortionTotal. Meaning that the biggest contributor will by far be the dominating one. You would need a crapload of distortion over a wide frequency range to get anywhere remotely close to a 2dB increase?

Yeah, also hard to separate other issues like IMD, noise, etc.

Ignore IMD for now, assume the example below is just from harmonics:

This is a 5 watts chart. Assuming that the bass tones are taking most the 5 watts. Assuming 1khz is at -55 dB for a music passage. The noise alone at -50db is already louder than the fundamental 1khz tone.

 

[mike7877]

Isn't a perfect fifth part of the scale? Try it on a piano.

If it's not in the scale it's not related. If it is in the scale and another note is being played, it's not related. These two are true 99% of the time. Clash

 

[FrantzM]

Masking thresholds for a 1kHz tone masker at different levels masking tones.​

At 1 khz at high levels even the 5th harmonic is masked only 45 db down. It will also mask any lower level music in that area. You'll only hear the lower level higher frequency music when the main tone has gone down. So the fact the harmonic distortion might be closer to the level of music at the higher frequencies won't matter. It is all masked.

Was waiting for someone to point toward the poorly bderstood phenomenon called « masking ». Thanks @Blumlein 88

 

[mike7877]

Was waiting for someone to point toward the poorly bderstood phenomenon called « masking ». Thanks @Blumlein 88

People seem to be forgetting that the treble is 20-30db lower in level than the bass/mid bass most of the time, sometimes with periods of silence in between. It's that level which you use to subtract from, not the bass (you hear through the bass). If the harmonic is at -50db and the treble is at -30db? 20db difference. If it's during a break? It's all there is.

It occurred to me we need to take into account the average to peak level (crest factor) of music - pop/rock etc. Which is 10-20db in modern music

So, subtract 15db for the crest factor and 20-30db for treble being that much quieter. You're at -35 to -45db now. If harmonic distortion is at -60, it's 15-25db down. Audible, easily audible

 

[solderdude]

Ignore IMD for now, assume the example below is just from harmonics:

You can't ignore it as a large part is IMD. This is a continuous test signal it is NOT the same as music.
Even when you would listen to it you can't hear the grass. The reason for that is each tone will mask the noise that is 45dB lower.
Compared to a clean amp you would not be able to distinguish the sound.
The noise is drowned by the continuous tones and even if you could hear it it would sound like noise added to all those tones.

 

[Killingbeans]

It occurred to me we need to take into account the average to peak level (crest factor) of music - pop/rock etc. Which is 10-20db in modern music

People seem to be forgetting that the treble is 20-30db lower in level than the bass/mid bass most of the time, sometimes with periods of silence in between. It's that level which you use to subtract from, not the bass (you hear through the bass). If the harmonic is at -50db and the treble is at -30db? 20db difference. If it's during a break? It's all there is.

So, subtract 15db for the crest factor and 20-30db for treble being that much quieter. You're at -35 to -45db now. If harmonic distortion is at -60, it's 15-25db down. Audible, easily audible

Why are you duplicating your own post #117?

 

[Inner Space]

If it's not in the scale it's not related. If it is in the scale and another note is being played, it's not related. These two are true 99% of the time. Clash

Well, I tried to be Socratic, and failed, so now I'll be blunt - you're wrong. A perfect fifth is completely and totally related to a tonic. Hearing a faint P5 due to distortion doesn't upset anyone at all, musically.

 

[mike7877]

Well, I tried to be Socratic, and failed, so now I'll be blunt - you're wrong. A perfect fifth is completely and totally related to a tonic. Hearing a faint P5 due to distortion doesn't upset anyone at all, musically.

No, you're wrong. A fifth doesn't fit all the time, nowhere near! It's not even plausible...

 

[mike7877]

Why are you duplicating your own post #117?

Because it applies and not everyone reads every post. He posted his post which made me think he didn't read mine

 

[solderdude]

This is one reason why you can't tell if something is real or synthesized

There are other reasons for that.

 

[mike7877]

There are other reasons for that.

It's one of many

"...this is one reason"

 

[Inner Space]

No, you're wrong. A fifth doesn't fit all the time, nowhere near! It's not even plausible...

You said: "Third harmonic is NOT (!) related to the music. It is like a wrong note. It is a clash, more intrusive than white noise." NB "related." Your ignorance of music is so profound I can't imagine why you endeavor to reproduce it.

 

[mike7877]

You said: "Third harmonic is NOT (!) related to the music. It is like a wrong note. It is a clash, more intrusive than white noise." NB "related." Your ignorance of music is so profound I can't imagine why you endeavor to reproduce it.

So a fifth always fits? Ok, you can always play a fifth. I give in

 

[dfuller]

By definition, they are not.

I don't think that he meant it in the terms we usually use - more that a rock kick drum is a short thump rather than a sustained note.

 

[Pdxwayne]

You can't ignore it as a large part is IMD. This is a continuous test signal it is NOT the same as music.
Even when you would listen to it you can't hear the grass. The reason for that is each tone will mask the noise that is 45dB lower.
Compared to a clean amp you would not be able to distinguish the sound.
The noise is drowned by the continuous tones and even if you could hear it it would sound like noise added to all those tones.

I know I can't ignore IMD. What else can I use to illustrate an approximation?

Also, Amir used that to simulate real music playback. What else can I use to illustrate that when bass is loud, high distortion amp likely to contribute to extra db in the highs?

And why do you keep saying masking issue? I am not saying all freq are at the same volume. I am saying typically bass are louder than treble in songs. Our hearing sensitivity in the highs are much better than the bass, right? Even bass is at 90db, we can hear 1200hz at 65db just fine. If the noise at 1200hz is at 70db, then I am sure I can sense 5+dB increase in 1200hz even when bass are in 90db.