THD has a much bigger effect on sound than you think

[pma]

It's a lot more complicated. Frequency/time dependency in the transfer function requires much more extensive modeling/simulation. The main issue I have with it is how to make it easy to specify such a function that would resemble some realistic device, without asking the user to enter complex numbers into the app If you have some ideas on how to make this easier to specify, let me know.

There's already a simple, bandwidth-limited feedback simulation built-in, but this has very small frequency dependency. Here's an example of a device with no feedback:

View attachment 161004

Same device, but now with 20dB of negative feedback:
View attachment 161006

Polynomial transfer function inside frequency dependent feedback. More later.

 

[Holmz]

Let's tone down the emotion folks ...

 

[pkane]

Polynomial transfer function inside frequency dependent feedback. More later.

That part is already there Currently it's a bandwidth-limited feedback loop. I can let the user change the bandwidth, which is easy. Or possibly define some sort of frequency response curve for the feedback loop. But somehow I don't think this will be used a lot

 

[solderdude]

You claimed "everyone" can hear it.

What if I am one in a 100000 that could hear it and 99999 out of 100000 can't hear it?

Then your statement is false, right?

It's at least 2 out of 100000

 

[pma]

That part is already there Currently it's a bandwidth-limited feedback loop. I can let the user change the bandwidth, which is easy. Or possibly define some sort of frequency response curve for the feedback loop. But somehow I don't think this will be used a lot

Good. To make absolutely sure we mean the same, please let me post a circuit example with frequency dependent loopgain (feedback factor) and corresponding distortion vs. level at 100Hz and 5kHz. (R2/R1 release feedback 100x to emphasise nonlinearity for educational purposes. Overall gain remains 1x at low frequencies.). Frequency response V(out)/V3 is shown as well.

P.S.: I think it is important if the SW allows for a sort of frequency dependent nonlinearity, as such is real world. How many time it will be used by the users is not so important, IMO

 

[DualTriode]

The sound energy in music in the 1-4khz range is 30-40db lower than bass, 15-30db than midrange.

Harmonics from bass make their way into midrange, harmonics from mid into the 1-4khz range.

Even if these harmonics are at a low level compared to the original tone, in the octave they intrude upon, they are loud (the ear is sensitive there and sound from music is at much lower levels.

A simple example is the 5th harmonic of 200hz is 1khz. In music, there is significantly more energy at 200hz than 1khz. Harmonics at relatively high levels make their way into lower treble and higher midrange sound energy in music.

0.1% harmonic distortion is totally unacceptable. Treble from 2-3khz up is just 0.1-1% the level of bass. 0.1% harmonics from bass are a HUGE problem to 0.1-1% treble in the signal. Even 0.01% THD is still 1% distorted treble. 0.1% THD is 10% distorted treble. Discuss.

(yes, these aren't exact numbers, but they're close enough to reality - a bigger problem than most think)
Mostly talking about odd order from cheap poorly designed solid state amplifiers with higher amounts of it (this sentence addressing a post using THD from a speaker as an example)

Hello OP,

You have no clue, you have no facts you have no evidence.

Just for grins I am posting measurements of the speakers that I listen to.

Your homework is to convert the calibrated SPL measurements to percent distortion.

Thanks DT

 

[pma]

I hope that everyone here understands that with decay of frequency response at lower frequencies the distortion automatically gets higher and that the harmonic distortion component may become more audible than the fundamental in case of low frequencies and small speakers.

 

[voodooless]

I hope that everyone here understands that with decay of frequency response at lower frequencies the distortion automatically gets higher and that the harmonic distortion component may become more audible than the fundamental in case of low frequencies and small speakers.

Bose Accoustimas anyone?

 

[mike7877]

Hello OP, You have no clue, you have no facts you have no evidence.

Hello, what's all the irrelevant charts for?

Billy Idol - Daytime Drama - there are many periods (about a third of the time for most of the time, with a few periods up to several seconds long) where bass (which I consider to be up to about 200hz) is over 30db higher than everything above 1khz.

And that's just the song I was listening to when I saw your post.

You hear highs through bass, they're not masked by them. And the highs are 30db lower than the signal level
So if you have harmonics on the lows at -60db, they are just 30db lower than the signal above 1khz.

Now, above 1khz is 30db below the signal 1/3 of the time. Do you not hear anything above 1khz for 1/3 the time? Of course you do. Is it plausible that clarity is affected by harmonic distortion? Quite.

Consider these things when figuring how "masked" harmonic distortion from the lower registers actually is.

No, the above is not a perfect example, so to make it a little better and paint the picture clearer I'll add that the same third of the time, above 500hz is 22db down.

The sounds for the aforementioned third of the time above 1khz (or 500hz), are components heard the other 2/3 of the time (also important to consider)

I can't address everyone who says I don't know what I'm talking about or am just wrong, but this post is to all of you. Organization and syntax aside, I put this post together over a few minutes and am not rewriting it to be pretty. I've made my points and they're obvious

Edit: I feel compelled to add, speaker THD is measured at a level, say 98db, over the audible spectrum. The tweeter's peaks are usually at least 10db lower than bass, 30db otherwise - so the harmonic distortion added by them should be considered at a measurement 10db -30db lower, 88 and 68db in this example. THD is lower in tweeters at quieter levels.
Point: the amount of THD added by tweeters in music with 98db peaks is much lower than some might assume (the speaker's HD doesn't mask all amplifier HD, not by a long shot). Don't ignore this last point

 

[mike7877]

You are assuming that instantaneous treble tones 35 db down from a bass town are instantly evident (distortion is temperally colocated) and making the further assumption that something harmonically related another 15-20dd below that is detectable in music.

I see your point if there was a random reduction of 35db it wouldn't be instantly evident, but in a repeating pattern (music), it is. Think about the fact that you can use your ears to isolate one sound from many others, one instrument from dozens. Maybe it takes you a couple bars before it becomes evident. Maybe a couple of bars the first time, and the second time you know its there so you pick up on it right away. Maybe you never heard it properly and you don't know how it should sound so you think nothing's wrong. I think this last point is the reason for a lot of people thinking THD isn't a problem - they've never heard legitimately high fidelity music. Expensive doesn't mean good, the reviews on this site go a long way to proving that.

Reading the posts in this thread, I think I should say you don't usually hear THD as a tone, unless it's a second harmonic of bass in a crap or overdriven speaker. Above a minimum level (the subject discussed) THD causes music to have a lack of clarity - the low level harmonics of instruments are interfered with, and their character is changed. You may no longer be able to tell if something was recorded or synthesized. This conversation isn't about if THD is "tolerable" or "immediately evident"

And third harmonic is more often NOT going to sound good with the music and disappear into it (aka be related).
If a perfect fifth is in the scale, it could be related.
But if a chord is being played which doesn't include it, it'll probably clash.
Or if another note is being played which is next to it it'll probably clash.
*If the perfect fifth of every note in a chord is played, they definitely clash.* (and that's music)
Put another way if you still don't understand - just like harmonizing doesn't always sound good even if you're in the right scale, a third harmonic does NOT always fit. Just because you can sing a scale it doesn't mean you can harmonize

 

[DualTriode]

Hello, what's all the irrelevant charts for?

>>>>

Hello,

Again without a clue.

The Plots are calibrated Frequency Response plots for a 15 inch woofer, a 10 inch midrange and a Compression Driver / Wave guide tweeter.

The lower Plots for each speaker are the Harmonic Distortion HD2. HD3 and Total Harmonic Distortion for each driver.

The very thing you claim to to speaking about.

Again, your homework is to convert the the calibrated SPL and HD measurements on the plots to percent distortion.

Thanks DT

 

[mike7877]

Hello,

Again without a clue.

The Plots are calibrated Frequency Response plots for a 15 inch woofer, a 10 inch midrange and a Compression Driver / Wave guide tweeter.

The lower Plots for each speaker are the Harmonic Distortion HD2. HD3 and Total Harmonic Distortion for each driver.

The very thing you claim to to speaking about.

Again, your homework is to convert the the calibrated SPL and HD measurements on the plots to percent distortion.

Thanks DT

No, what I'm talking about is in my posts above yours, #309 and #310. If you're being vague about my original post (#1) being a little off, I realised a long time ago, but not soon enough (I can't edit it). The discussion has progressed since then (page 16 now)

Regardless, it doesn't affect anything I said in posts 309 and 310, both of which I'm happy to discuss.

You asked for info (by stating I didn't have any) so I gave you info and applied reasoning. If you think what I said in 309 is wrong, explain how and why it's wrong (simply stating I don't know what I'm talking about is inadequate)
Don't forget to consider what I put in post 310 while you're doing so.
Looking forward to a comprehensive and civil discussion of relevant information