Speakers distortion

[andreasmaaan]

It seems to me one (rough) solution would be to fix the mic position, do a sweep to capture the response, then use that to adjust relative levels in the distortion sweep. The in-room response should not be affected by volume (until you knock something around, or down) so check at low level then do the higher-power distortion measurements. This could be automated; for each frequency, check the levels of the fundamental and each harmonic using low-level test tones, then incorporate that into the distortion reading.

This would not correct problems in the recording (sensing) chain, however, be it mic overload, input buffer or ADC distortion, etc. And moving the mic or anything else that could change the response will of course impact the results.

In the distant past I used a Matlab program that sequentially generated and recorded the fundamental and ten harmonic tones (or to 20 kHz, highest I chose to test) of the desired test tone at 60 or 70 dB SPL'ish in short bursts, then did a burst power sweep at the fundamental. A simple loop would step through multiple test frequencies. That used GPIB control of a generator and spectrum analyzer; these days you would just use a sound card or something like REW.

Can someone provide a link to the GedLee measurements for the thread-weary?

In the first post of this thread there's a download link for an Excel sheet that can calculate entered distortion values and then provide a script to be pasted into matlab that then spits out a GedLee Metric number. Works more intuitively than it sounds

 

[DonH56]

In the first post of this thread there's a download link for an Excel sheet that can calculate entered distortion values and then provide a script to be pasted into matlab that then spits out a GedLee Metric number. Works more intuitively than it sounds

Thanks! Weird, link did not work until I tried to reply...

 

[SIY]

The latest version of Virtins Multi-Instrument has GedLee built in.

 

[andreasmaaan]

Btw, it should be mentioned (not for you SIY or Don, ofc) that the Gedlee metric is somewhat limited in that it doesn't directly account for IMD, which may not follow mathematically from THD, nor does it account for the level dependence of masking or the equal loudness curves, meaning that by definition it can't give a correct result except for a specific (unspecified) level - to the extent that it can give a "correct" result at all.

 

[Thomas savage]

You guys just need to get a mate round and cart the speakers outside.

Few beers bit of measuring, what could go wrong..

 

[SIY]

Btw, it should be mentioned (not for you SIY or Don, ofc) that the Gedlee metric is somewhat limited in that it doesn't directly account for IMD, which may not follow mathematically from THD, nor does it account for the level dependence of masking or the equal loudness curves, meaning that by definition it can't give a correct result except for a specific (unspecified) level - to the extent that it can give a "correct" result at all.

What it does do is give a much better correlation to perceived "cleanness" than just THD. It's not as universal as Earl makes it out to be, but it's (IMO) far more useful than THD.

I'm hoping to finish up my ABC-MRT article by spring- I have high hopes that this gives even better correlation to subjective perception.

 

[DonH56]

You guys just need to get a mate round and cart the speakers outside.

Few beers bit of measuring, what could go wrong..

For starters, it is currently about 12 degF (-11 degC) and snowing hard at my house, hard on speakers (and testers). Maybe just go with mates and beers, skip the carting and testing?

 

[andreasmaaan]

What it does do is give a much better correlation to perceived "cleanness" than just THD. It's not as universal as Earl makes it out to be, but it's (IMO) far more useful than THD.

100% agree.

I'm hoping to finish up my ABC-MRT article by spring- I have high hopes that this gives even better correlation to subjective perception.

Please post it when it's up

 

[andreasmaaan]

You guys just need to get a mate round and cart the speakers outside.

Few beers bit of measuring, what could go wrong..

The only speakers I have here at the moment are these. Worth it you reckon?

 

[SIY]

Please post it when it's up

Depending on my editor, it will either be in AudioXpress or (if we're really lucky and get useful results) an AES paper. If it totally sucks and shows no promise, I'll post it here as a caution.

I'm working with one of the AP guys on this, which has made the whole process hugely better than I could have managed on my own.

Back to sauteeing mushrooms.

edit: link to some background

 

[Krunok]

So no-one is doing GedLee measurements? Hmmm, I may need to correct this. But not today, I'm chained to the kitchen preparing New Years' dinner.

Thank God I'm done with the cooking for today!

Sure thing, I'll do it if you provide "how-to cook book".

 

[Krunok]

Even when you think you are holding your head still there are small movements. More so when listening casually without trying to hold your head still. These movements are used by your hearing to effect what you hear. JJ has mentioned this now and again.

So maybe something like the mmm method your brain is doing this statistical calculation all the the time. Which will alter the effect of standing waves on what you hear vs what you measure.

Ok, I see what you mean. Once I made correction averaging the response in a slightly wider "window" than one point sweep it turned out that SQ is much less sensitive to head movement than it was when I made correction based on a single point sweep. For that very reason I prefer MMM method.

 

[Krunok]

The only speakers I have here at the moment are these. Worth it you reckon?

Sure they are - just give it a shot!

 

[Krunok]

Guys, wife ordered to log off now, I wish you happy (and sound!) New Year!

 

[Floyd Toole]

Hi. I just noticed that my name was mentioned. I have not been following this thread but my simple overview of the current topic follows.

Non-linear distortions originate in a non-linear input/output relationship which changes the waveform by adding spectral components to the audio signal. Rather than quantifying the non-linearity itself, we probe the system with simple signals; single or multiple pure tones being the most common. Then we measure what comes out and, knowing what went in, a percentage distortion can be calculated. Problem is that because of simultaneous masking all of the distortion components are not heard. So, the measured percentage distortion, which includes all components, is wrong. The correlation with objectionable sounds when listening to music is poor. We need evaluations that incorporate some properties of the human hearing system's inherent masking in order even to make a serious start.

The existing metrics, like harmonic or intermodulation distortion are useful to design engineers, but the only truly meaningful number is zero.

In order to meaningfully interpret distortion measurements above zero, the metrics must involve elements of psychoacoustics: masking being the dominant one. A few have been tried but I know of none that have been tested sufficiently to achieve widespread acceptance. But it is a topic that I have not closely followed in recent years - I'm now retired. There is a discussion, with references in my book. The good news is that over my nearly 50 years of evaluating loudspeakers it has been extremely rare for non-linear distortion to be an audible factor in sound quality evaluations of conventional "hifi" products. Recently popular small wireless and "smart" loudspeakers face real challenges in that respect.

Measuring non-linear distortion in a room makes a difficult situation even worse, because amplitudes of distortion products are affected by reflected sounds which are highly dependent on the venue, the loudspeaker location and the mic location. Background noise is another challenge; such measurements need to be done in a very quiet anechoic chamber or equivalent. Being able to make a measurement does not make it a useful measurement.

My standard example of human tolerance for non-linear distortion is the LP. In the early years LPs were the source of program material for loudspeaker listening tests, so I made it my business to examine the performance of the medium. I published several articles in Canadian audio magazines on the topic, and it got to the point of making a test recording that contained music and test signals. There were easily audible differences between the master tape and what came off the LP - those inherent problems are still with us. The levels of measured non-linear distortions of all conceivable kinds was shocking - whole percentages being common, and, during incipient or real mistracking, off the charts. Masking is a powerful factor in what we hear. Needless to say, digital audio was a great improvement. But for some, obviously, the placebo effect is profound .

 

[solderdude]

A very quiet anechoic chamber and a measurement microphone with extremely low distortion and mic amp is not something that we all have in our sheds.
Perhaps keep to measuring FR and room compensation and worry less about the speakers used (when looking at HD below 0.1%) ?

I run in similar problems in measuring distortion on headphones. My cheap mic has high 2nd harm. distortion so can't measure below 0.1% to 0.2% for 2nd harmonic. Also room noises are always present interfering with the measurements. Low freq. pass almost un-attenuated through headphones and test rigs.

Maybe only look for distortion levels above 0.1% and about 1% below 50Hz. Most distortion below it is moot and most likely very difficult to hear while playing music.

 

[JohnPM]

Using a tone and looking at the RTA to determine distortion levels from loudspeakers is likely to give fairly misleading results, since one doesn't see what the response at the harmonic frequencies looks like relative to the response at the fundamental. Stepped sine is a better option to make use of the RTA, but is very time-consuming. Sweep measurements are most straightforward and let you see directly what the fundamental response looks like to help guide interpretation of the harmonic levels, but much interpretation remains. There is some discussion of this in the REW help.

 

[Theo]

Masking is a powerful factor in what we hear.

If I may say so, can your overview be summarized by "trust your ears" and "distortion figures are not really meaningful"?

it has been extremely rare for non-linear distortion to be an audible factor in sound quality evaluations

Does this means that our own attempt at measuring speaker distortion is a waste of time?
So, are we back at choosing our speakers by DBT in our living room?
Or perhaps should we just enjoy the music that our ears have been calibrated to. I bought my speakers in 1981... and I still like them as long as I don't play the music too loud that is ~<80 dB SPL at my chair location - Anyway, I live in an apartment...

 

[Krunok]

Using a tone and looking at the RTA to determine distortion levels from loudspeakers is likely to give fairly misleading results, since one doesn't see what the response at the harmonic frequencies looks like relative to the response at the fundamental.

This argument sounds a little odd to me. I don't need to see anything, I just need REW to calculate correct distortion figures. Am I supposed to read from the graph I'm seeing what is distortion and what is not?

Sweep measurements are most straightforward and let you see directly what the fundamental response looks like to help guide interpretation of the harmonic levels, but much interpretation remains. There is some discussion of this in the REW help.

I found that distortion figures I'm getting with sweep differ significantly from the ones I'm getting with RTA. Can you please explain which ones should I trust and how comes they differ so much. With frequency response measurement I have learnt that averaged RTA differs quite a lot from single sweep but when I make 7-9 sweep measurements and averaged them then both measurement tend to converge to the same values. With the distortion it is not possible to average sweeps yet you seem to claim that singel sweep is more relevant than RTA which averages a number of measurements. Or I got you wrong?

 

[JohnPM]

This argument sounds a little odd to me. I don't need to see anything, I just need REW to calculate correct distortion figures. Am I supposed to read from the graph I'm seeing what is distortion and what is not?

The RTA shows you the level of your test tone and the levels of harmonics at the measurement position. What it doesn't show is what the level of a test tone at the harmonic frequencies would be. The levels of the harmonics are effected by both the amount of distortion and the response at the measurement position - for example, suppose your mic happened to be sat at a perfect null for 2 kHz. A test tone at 1 kHz would show no 2nd harmonic distortion, regardless of the actual level of 2nd harmonic distortion produced. In-room responses are not flat, so all harmonic levels are effected. For more on that read Temme's paper, you may find a copy online. For more general information on distortion measurement this old B&K application note is good, though it predates sweep measurements.