Speakers distortion

[andreasmaaan]

I agree. But that goes for the frequency response correction of the base tone as well: if the size of the base tone is somewhat altered by build up of resonance or nulling you can simply forget that FIR filter will be able to do anything about it. But that still doesn't imply there is no sense in frequency response correction, so the same logic can be applied to "distortion correction" as well, right? It could still make sense to do it..

I actually don't believe in in-room frequency response correction in most (not all) circumstances - except of course in the room's modal region (or if the speakers measure poorly or have a difficult polar response in the first place). So I'm consistent at least

 

[Krunok]

@Krunok I actually think the idea you’re describing doesn’t sound so different from the one Don mentioned. It would be interesting if anyone had a copy or link to those MIT papers.

I do think the effectiveness would be limited for the reasons Cosmik mentions, although - at least theoretically - i don’t see why it wouldn’t be possible to build a very sophisticated model of a transducer and compensate for its nonlinear (not only harmonic) distortion quite effectively.

As I told him, distortion that would be additionally generated with added inverse phase componenets would be negligible as those added componenets have very low amplitude so their distortion components would be so low they would be deeeeply burried into the noise floor.

 

[andreasmaaan]

As I told him, distortion that would be additionally generated with added inverse phase componenets would be negligible as those added componenets have very low amplitude so their distortion components would be so low they would be deeeeply burried into the noise floor.

Wouldn't they need to be of the same amplitude as the distortion components they were trying to correct? These therefore must be audible, otherwise there'd be no point trying to correct them.

 

[Krunok]

I actually don't believe in in-room frequency response correction in most (not all) circumstances - except of course in the room's modal region (or if the speakers measure poorly or have a difficult polar response in the first place). So I'm consistent at least

Ahaaa - ok, sure, in that case you're consistent!

As you can see I am probably the opposite case. I have measured frequency response not only at my LP but on several other places in my room (which is alnot 50m2, so quite large) and I have found that all of the position benefited from room EQ. LP of course benefited most, but that was anyhow my aim.

 

[Krunok]

Wouldn't they need to be of the same amplitude as the distortion components they were trying to correct? These therefore must be audible, otherwise there'd be no point trying to correct them.

Yes, they would. I think I didn't explain myself well - again!

Let's try like this - if base tone is of 80dB 2nd and 3rd harmonic would be at say 20dB and for them I generate inevrted phase tones. Now, when they will be played by loudspeakers as 20dB tones additionall distortion would be created but that will be based on their 20dB amplitude and thus will be very low.

 

[andreasmaaan]

Ahaaa - ok, sure, in that case you're consistent!

As you can see I am probably the opposite case. I have measured frequency response not only at my LP but on several other places in my room (which is alnot 50m2, so quite large) and I have found that all of the position benefited from room EQ. LP of course benefited most, but that was anyhow my aim.

Sure, and I do think there can be some benefit in cases where the speakers don't measure flat or don't have a smooth polar response in the first place, or where first reflections are so close in time to direct sound that our brain struggles to separate them from each other (as is effectively the case in the room's modal region).

PS I hope you don't get the impression I think this is a bad idea @Krunok. I actually think it's interesting and has potential (and something similar has seemingly been tried already). I just think I'd do it at the level of the transducer, rather than the speaker+room, since the room is going to mostly be contributing linear distortion rather than nonlinear distortion. This will just tend to confound any attempt to measure the nonlinear distortion.

Let's try like this - if base tone is of 80dB 2nd and 3rd harmonic would be at say 20dB and for them I generate inevrted phase tones. Now, when they will be played by loudspeakers as 20dB tones additionall distortion would be created but that will be based on their 20dB amplitude and thus will be very low.

Yes, sure. So best case scenario you would completely cancel out all transducer distortion, and worst case scenario you would exactly double it.

Depending on how well you did it, the reality would lie somewhere in between.

 

[Krunok]

Sure, and I do think there can be some benefit in cases where the speakers don't measure flat or don't have a smooth polar response in the first place, or where first reflections are so close in time to direct sound that our brain struggles to separate them from each other (as is effectively the case in the room's modal region).

And taht is exactly the case in my room, as you may remember from our previous conversation on that topic. That is the reason why I am so "obsessed" with in-room measruement and correction as furniture and shape of the room doesn't allow for better placement.

PS I hope you don't get the impression I think this is a bad idea @Krunok. I actually think it's interesting and has potential (and something similar has seemingly been tried already). I just think I'd do it at the level of the transducer, rather than the speaker+room, since the room is going to mostly be contributing linear distortion rather than nonlinear distortion. This will just tend to confound any attempt to measure the nonlinear distortion.

Oh, not at all- I'm sure that tranducer that would be compensated that way by the factory would be a much better idea but there are so many good passive speakers around us that deserve to be enhanced by DSP correction as effectively as possible.


[QUOTE="andreasmaaan, post: 135381, member: 2367"
Yes, sure. So best case scenario you would completely cancel out all transducer distortion, and worst case scenario you would exactly double it.
Depending on how well you did it, the reality would lie somewhere in between.[/QUOTE]

IMHO every correction depends very (and here I mean VERY) much on how well measurements have been done. In reality I would expect certain imporevements meaning some of the distortion copmpnenets would end up lowered, but certainyl not by 100% as that would require a perfect match and that is simply not possible, nor I think it is necessary to be able t hear improvement.

 

[andreasmaaan]

And taht is exactly the case in my room, as you may remember from our previous conversation on that topic. That is the reason why I am so "obsessed" with in-room measruement and correction as furniture and shape of the room doesn't allow for better placement.

Yeh I do remember. That's why I think it worked particularly well in your case, plus the fact that your speakers have that extra upward (?) facing tweeter IIRC

IMHO every correction depends very (and here I mean VERY) much on how well measurements have been done.

Amen.

 

[Krunok]

I just think I'd do it at the level of the transducer, rather than the speaker+room, since the room is going to mostly be contributing linear distortion rather than nonlinear distortion.

What do you think of the following example: let's assume my room will resonate at 120Hz. In that case wouldn't the 2nd harmonic disortion component of the 60Hz base tone be "amplified" in a same manner as would be the 120Hz base tone?

 

[Krunok]

Yeh I do remember. That's why I think it worked particularly well in your case, plus the fact that your speakers have that extra upward (?) facing tweeter IIRC

Uh, I really don't know why I love them. Yet another example that love is blind!

P.S. for the things to be worse it's an upward facing bass driver firing up to 1800Hz together with front facing unit. You can't imagine what a mess fo reflections that creates..

 

[andreasmaaan]

What do you think of the following example: let's assume my room will resonate at 120Hz. In that case wouldn't the 2nd harmonic disortion component of the 60Hz base tone be "amplified" in a same manner as would be the 120Hz base tone?

Sure, but then how do you remove the 120Hz distortion harmonic from a recording with both a 60Hz and a 120Hz tone present (assuming all are in phase with each other) without lowering the level of the 120Hz tone present in the recording?

This seems to me to be a linear problem when looked at in-room. Correcting the in-room response will best be done linearly, i.e. by reducing the level at the 120Hz mode, regardless whether or not it is distortion. Correcting the nonlinear (distortion) component would best be done at the level of the transducer.

I'm thinking on my feet here though - perhaps I've missed something...

 

[Krunok]

@Krunok I actually think the idea you’re describing doesn’t sound so different from the one Don mentioned. It would be interesting if anyone had a copy or link to those MIT papers.

That would be great, because if you guys think that idea is ok I may try to put some effort into making it work.

 

[Krunok]

Sure, but then how do you remove the 120Hz distortion harmonic from a recording with both a 60Hz and a 120Hz tone present (assuming all are in phase with each other) without lowering the level of the 120Hz tone present in the recording?

Ok, so let's assume 2 tones scenario - 60Hz at 80dB and 120Hz at 60dB being played. If we don't correct it 60Hz tone would generate 2nd harmonic at 120Hz at say 15dB which would superpone with our second base tone of 120Hz/60dB, correct? So, if DSP now generates a 120Hz/15dB inverted phase tone it will hopefully lower that superponed 120Hz combo - and it will do the right thing lowering it as it was ebing raised by distortion component of the 60Hz tone!

 

[andreasmaaan]

Ok, so let's assume 2 tones scenario - 60Hz at 80dB and 120Hz at 60dB being played. If we don't correct it 60Hz tone would generate 2nd harmonic at 120Hz at say 15dB which would superpone with our second base tone of 120Hz/60dB, correct? So, if DSP now generates a 120Hz/15dB inverted phase tone it will hopefully lower that superponed 120Hz combo - and it will do the right thing lowering it as it was ebing raised by distortion component of the 60Hz tone!

Yeh, but the better solution would be to lower the level at 120Hz exactly as much as the room "amplifies" it (i.e. classic room mode correction), meanwhile using an inverted phase 120Hz tone to reduce the distortion component of the 60Hz tone exactly as much as the transducer produces it.

I was also going to ask, how would you tackle IM distortion?

 

[DonH56]

I found a link to ADC distortion compensation, but not speakers.

It includes the following statement:


This, I believe, basically says: "You might be able to measure an ADC's performance using steady state waveforms in the frequency domain (for example), apply compensation, and measure it again using the same steady state waveforms. But don't imagine that this will automatically translate into instantaneously-correct compensation in an application that isn't based on steady state waveforms".

A very simple example: if I need my ADC to digitise a 1kHz sine wave reference, and notice that it produces some strong 2kHz spurious component in the signal, I can perfectly compensate for this by simply notch filtering the digitised signal at 1kHz. Or steeply low pass filter at some frequency below 2kHz. Or I can appear to correct the error by simply adding a suitable continuous level of 2kHz, in anti-phase to the spurious component, to the digitised signal - the Fourier transform will give the correct result. But it doesn't mean that the ADC is giving the correct output dynamically. If the 2kHz error was not, in fact, a continuous tone, but came in some form of enveloped pulsing, the dynamic output of the system will now be wrong even though an FFT accumulation appears correct. Clearly a system that appears perfect for a certain kind of steady state application may be giving the wrong output dynamically e.g. for music.

And an ADC is going to be much simpler to compensate than a speaker cone that is breaking up!

Yes, it was ADCs, that is what I do (not speakers). Sorry, should have made that clear. I was thinking of the concept. That paper is good albeit much later than the original work. Aside: I spoke with Fred a few times, exchanged a bunch of emails for a while when I was trying to implement the approach in a design, and used to work with Dan Asta (forgot Dan had taken over). Both great (and very smart) guys. Small world.

Your understanding is reasonable. It was not necessarily specific to steady-state signals; the problem is the compensation is based upon the characteristics of the signal you are trying to compensate. My efforts focused primarily around three areas: pulses (radar/lidar), general RF signals (sinusoids-ish), and spread-spectrum signals (noise-like). If you implement the table for (using) one signal the compensation did not work as well for other types. It did not fail, just provided less improvement. You build a table using signals that are as close to what you expect to receive to create the compensation. A general solution is one thing I was striving for, and is IIRC where Dan had stepped in to extend the original research (I probably have hard copies someplace in the black hole of a basement among the 50~100 boxes of old work papers). Ultimately a "simple" practical solution was not obvious and the processing far too slow to implement in real-time, at least back then (1980'2/1990's). I did win a few research contracts and had a lot of fun with it, and some of the concepts flowed into later products (not mine for the most part).

I would dispute it is harder for ADCs than speakers but have never tried to apply it to speakers. Cone breakup and nonlinearity can be characterized for various signals and amplitudes. What I am not at all sure about is if you can really, practically, apply an electrical correction to solve a mechanical problem like that, and what it would do to the sound we hear. Given a cone entering modal operation, if you apply compensating signal to counter the mode, can you do that and still create the proper output? Not clear to me either way, but again not something I have thought about.

HTH - Don

 

[DonH56]

I think I didn't explain my idea properly. I'm not sure the idea itself is sound, but from your post, as a start I believe I didn't explain it well, so let me try again:

Measuring phase:

Let's assume a single 500Hz tone is played by the speakers. I see 2 things happening there and both of them involve both, room and speaker. In fact, I don't think they need to be separated but rather treated as one, as both of them are affecting what I hear sitting at my LP.

So, first: the tone that should make 80dB of SPL at my LP actually is making 77dB as a result of speaker response not being linear and the room modes. Second: along with 500Hz tone few additional tones have been unintentionaly created and played by the speaker (summary effect of which we are calling THD). For the sake of this analisys let's assume there were only 2 of them, 2nd and 3rd harmonic. Their amplitude depends on the SPL of baseline 500Hz tone and their frequency is always 2*freq of base tone for 2nd harmonic distortion element and 3*freq of basse tone for 3rd harmonic distortion element.

Processing phase:

Let's assume we have created a lookup table which contains 2 independent variables (frequency and SPL/amplitude of the base tone) and 2 dependent value (SPL of 2*freq and SPL of 3*freq of base tone) which we measured.
Lets assume our table has resolution of 5dB for the SPL base tone variable.
Let's also assume our table has some appropriate freq resolution, probably in log scale for practical purposes.

We feed that table into convoluton engine in the same manner we feed it with FIR filter. At one slice of time convolution engines sees that it has to process 2 tones: 522Hz at 73dB and 1875Hz at 57 dB. Engine calculates amplitude correction for 522Hz and 1875 Hz base tones based on FIR filter as it normally does and modifies signal slice acccordingly. In addition, engine looks ap the "distortion response" table at the closest points (say it is 500Hz/75dB for the first tone and 2000Hz/60dB for the second tone). Engine reads from the table what are expected distortion harmonics amplitude of the tones to be generated by the speaker/room for those 2 base tones and inserts into signal slice the same tones as harmonic distortion components of those 2 base tones would be, but with the opposite phase (or applies some other more clever cancelling mechanism). Engine moves to next slice of time. And so it goes..

So, do you think it would work?

I wouldn't begin to guess at this point. Try it and let us know. As I said, it sounds similar to the compensation method to which I referenced and Cosmik linked. Note phase is significant and not just amplitude, and by inference time relationships matter. The table can get very complex very quickly when you try to handle multiple tones, but at audio rates you can do a lot of processing. I suspect there are AES papers addressing the problem of compensating speaker distortion in that manner but I have not followed the AES in many years (decades; my career took a turn far from audio).

Note this is not at all what I thought the goal was, i.e. simply to measure the distortion (magnitude) of a speaker whilst in a room.

 

[andreasmaaan]

That would be great, because if you guys think that idea is ok I may try to put some effort into making it work.

I would definitely love to see you make some progress I'm a current AES member so let me know if there are any papers you'd like me to go through that might be relevant.

I've had a quick look and found these so far:

• http://www.aes.org/e-lib/browse.cfm?elib=12323
• http://www.aes.org/e-lib/browse.cfm?elib=17483
• http://www.aes.org/e-lib/browse.cfm?elib=2032
• http://www.aes.org/e-lib/browse.cfm?elib=7878
• http://www.aes.org/e-lib/browse.cfm?elib=7935
• http://www.aes.org/e-lib/browse.cfm?elib=13001
• http://www.aes.org/e-lib/browse.cfm?elib=12134

I also found this interesting master's thesis on this topic, and a very interesting book here which goes into quite some detail.

And there's also this slide show from a presentation by Dr Klippel, which actually covers a lot of what has been discussed in this thread. It's from 2003 so hopefully/presumably he's progressed further with these ideas since then.

It's very interesting to see the conclusions he reaches:

PS: I haven't read any of these yet! Just started searching.

 

[AnalogSteph]

It's an interesting problem for sure. The crux is that a speaker is basically a series of linear frequency responses and nonlinearities (you will often see the 3rd harmonic following the frequency response at 3x the fundamental, for example). Even a simple model that just takes L(I) nonlinearity into account would look something like this:
[H1(f)] --> [A(x) = ax + cx³ + ...] --> [1/H1(f)] --> [Hcone+room(f)]
(H1(f) is there to account for variation of nonlinearity over frequency.)

I'm guessing you could probably invert something like this if you had a DSP active speaker with known constant levels, but good luck finding out all the coefficients by measurement. Probably adapting a linear speaker model would be worth a shot. And all of that is assuming time invariance, which may be a bit of a stretch once a hot voice coil comes into play. This could get almost arbitrarily messy, depending on how comprehensive a distortion cancellation you want. And that's just for one driver individually. Doppler distortion in a coax would go uncompensated.

 

[DonH56]

@andreasmaaan - Interesting. I can find analogs to the DFEs (decision feedback equalizers) and other circuits used for active, dynamic compensation in my current world of high-speed (multi-Gb/s) SerDes testing.

@AnalogSteph makes good points.

I have not done significant research into audio transducers for a long, long time if ever.

 

[Blumlein 88]

snip

Note this is not at all what I thought the goal was, i.e. simply to measure the distortion (magnitude) of a speaker whilst in a room.

Yeah that was my thought too. I'd like to know how effectively distortion is being measured before moving on to something else.