Distortion in loudspeakers

[RayDunzl]

Will give it a go next measurement round and follow-up.

Try this once:

Sweep at intervals of 5dB, say from 50dB on up.

Measured distortion will be high at low SPL as noise dominates the measurement, and will fall with each higher SPL measurement, until the real distortion rises out of the noise floor, then the readings will rise.

50dB sweep - distortion is room noise, not the speakers.

65dB - higher signal to noise, so lower distortion

80dB - this is the about the turning point - SNR is highest before real distortion comes in

85dB - rising Distortion to Noise ratio

95dB -rising more

 

[RayDunzl]

I have tried different sweep speeds?

Yes, but I don't remember it making a whole lot of difference, but I might not have understood the ambient noise contribution when I did it.

 

[Blumlein 88]

View attachment 5486 View attachment 5486

Yes, but I don't remember it making a whole lot of difference, but I might not have understood the ambient noise contribution when I did it.

Well I just tried it and some fluctuation from one run to another, but nothing consistently different about speed of sweep in the measuring. First the fastest sweep. Then I did a sweep only to 2khz with a 1 meg sweep which is much, much slower.

This was for a single LSR305. I was using the Umik usb so the displayed levels on the graph are accurate. Normally without the fridge or HVAC on my room shows a total sound level of something in the upper 30 db range. With the fridge on add about 10 db at least. It is barely noticeable when the fridge is on to your ear, but it is all in the lower frequencies.

 

[Blumlein 88]

Okay so to test with low end noise I made some pink noise played over a stereo pair of speakers in the same room. I steeply filtered out everything above 100 hz and except for a band between 2khz and 3 khz. The noise looked like this.



Without the noise from the speakers this is the noise in my room with a nearby fridge running.


Finally here is a regular 256k sweep showing distortion. Just showing 2nd and 3rd harmonics for clarity. The noise was between 2 and 3 khz plus below 100 hz and was playing continuously during the sweep. Notice the neat effect of how that noise was mistaken for 2nd harmonic and 3rd harmonic distortion at different frequencies.


I also tried several low end spot tones like Ray showed earlier. Though not shown, I had distortion below 1% on those.

 

[March Audio]

Some amateur observations from my recent research with "distortion":

It doesn't surprise me that they look similar. Looks to me like you might be measuring the noise floor in the room vs the speaker output and not so much the speaker distortion itself.

Distortion measured in a sweep is (in my experience) higher than distortion measured for a steady tone. Take a few of those and compare.

Room noise - especially that inaudible rumble that permeates modern living - contributes greatly to the lowest frequency sweep distortion measurement.

Taking sweeps recently - 5dB increments - 85dB was the level where distortion contributed by the speakers just began to rise above the "distortion" contribution of the noise floor in the sweep.

You might go a little louder - 95dB- if you aren't in a silent chamber and want to compare distortion from the speaker and not so much from the background noise of the enviroment. A little more stress will separate the better and worse speakers, too.



REW allows you to specify the addition of distortion up to the ninth harmonic for your listening pleasure.

My experience also. Just building a dsp sub and distortion close micd is much much lower than with mic at listening position

 

[RayDunzl]

Notice the neat effect of how that noise was mistaken for 2nd harmonic and 3rd harmonic distortion at different frequencies.

Thank you for the confirmation of my theory of ambient noise being seen as a "distortion floor" in a sweep measurement which I hadn't really confirmed so carefully...

 

[oivavoi]

Thank you for the confirmation of my theory of ambient noise being seen as a "distortion floor" in a sweep measurement which I hadn't really confirmed so carefully...

Doesn't this means that distortion measurements really should be taken in an anechoic chamber?

 

[RayDunzl]

Doesn't this means that distortion measurements really should be taken in an anechoic chamber?

Got one? I don't.

If you are publishing product specification, that's one thing. I'm just a hobbyist here.

Once the harmonic frequencies generated rise out of the noise floor, the reading for THD becomes usable. Even if the numbers aren't 100% accurate you can see what's going on with single-tone RTA. If the harmonics are still buried in the noise floor, you either have a lot of noise or levels of distortion bordering on inaudible.

THD+N would still be skewed, but I have to wonder how much N a speaker generates on its own, when the signal is a sine.

Just take the noise floor into account as part of the landscape.

I just go ahead and measure everything at the listening position under "normal" conditions because that's where and how I listen, just to get an idea of what I am or am not hearing.

 

[Blumlein 88]

Doesn't this means that distortion measurements really should be taken in an anechoic chamber?

Would be nice. It mainly means take into account noise will limit what you can measure domestically in the lower frequencies. Looking at room noise, the level at 2khz and above is 30 or more decibels lower than levels at 100 hz and less. At the higher frequencies you could measure distortion below .1% if a speaker is able to do it.

You might be interested in this earlier thread about noise in the listening room.
http://audiosciencereview.com/forum...hat-level-is-noise-heard-in-your-system.1013/

It appears most of us have low enough noise to hear somewhere between 10 and zero db SPL in our rooms. The microphones could hear it for measurements too. So if you were measuring distortion with the sweep at an 85 db level you probably could get measurements in the upper frequencies of distortion -70 or -75 db pretty cleanly. That would be distortion below .03% if a speaker is clean enough to show that.

 

[mitchco]

Try this once:

Sweep at intervals of 5dB, say from 50dB on up.

Have not got there yet... But no doubt those measures I took are in the noise floor as others have shown. Noise floor in my listening environment;

I swapped some 1.5" compression drivers and took some near field measurements. Sorry it is on another site, any feedback appreciated:

http://www.avsforum.com/forum/89-sp...pair-jbl-4722n-speakers-222.html#post50700321

 

[Blumlein 88]

Have not got there yet... But no doubt those measures I took are in the noise floor as others have shown. Noise floor in my listening environment;

I swapped some 1.5" compression drivers and took some near field measurements. Sorry it is on another site, any feedback appreciated:

http://www.avsforum.com/forum/89-sp...pair-jbl-4722n-speakers-222.html#post50700321

Those measurements that looked off over at AVS would fit with having some sort of buzzing noise or resonance that was going off all the time around 20 khz. You could try playing some constant spot tones around 3 or 4 khz to see if some 20 khz ringing is being set off and show up in the recorded output.

 

[mitchco]

20 kHz is (beyond) the limit of 1.5" compression drivers. Not too concerned about that as I can't hear it either It's the relative distortion levels between the two different drivers. The 2453h-SL has much less overall distortion than the 2432h driver that were replaced. One of the 2432h drivers has considerably more distortion than the other, like a factory defect or the diaphragm not properly gapped... I have not taken it apart yet. My comment on feedback is how I can make better distortion measurements? I.e. anything wrong or missing from my procedure. Thanks.

 

[RayDunzl]

how I can make better distortion measurements?

You have this:

Under the controls:

Uncheck that if it is checked. I don't have it figured out but it makes for funky displays like above.

 

[RayDunzl]

And here:

If you didn't notice it, it looks like the 2nd harmonic dominates the result - THD tracks the 2nd except at one little spot where it peeks out. The rest of the 2nd line is obscured by the THD line.

---

Then I see this:

http://www.avsforum.com/forum/89-sp...-pair-jbl-4722n-speakers-67.html#post50711633

Which throws the UMIK-1 under suspicion for adding 2nd harmonic distortion itself. oh well...

 

[Blumlein 88]

Home Theater shack's explanation is better than mine. Quoting from 2nd paragraph here:
https://www.roomeqwizard.com/help/help_en-GB/html/graph_distortion.html


By default the harmonic and THD plots in normalised mode use the level at the fundamental for each frequency as their reference - for example, the distortion figures for each harmonic at 1 kHz will depend on the level of the fundamental at 1 kHz. If Use harmonic frequency as ref is selected the reference will be the frequency of the harmonic - for example, at 1 kHz the 2nd harmonic figure will depend on the level of the fundamental at 2 kHz, the 3rd harmonic will depend on the level of the fundamental at 3 kHz and so on. This follows a recommendation made by Steve F. Temme in "How to graph distortion measurements" presented at the 94th AES convention in March 1993. If the response of the system being measured is flat this makes no difference to the results, but when the response is not flat (as for most acoustic measurements) it can remove the influence of the loudspeaker's fundamental response from the distortion figures. As an example, suppose the loudspeaker response was flat apart from a 6 dB peak at 2 kHz. 2 kHz is the 2nd harmonic of 1 kHz, so the 2nd harmonic level shown at 1 kHz will be increased by 6 dB due to the boost in the fundamental when using the excitation frequency as the reference. Similarly the 3rd harmonic level at 667 Hz (2/3 kHz) will be boosted by 6 dB. If the harmonic frequency were used as the reference the distortion figures would not show this boost. Using the harmonic frequency as the reference also provides a more meaningful view of distortion at frequencies below the LF roll-off of the system as otherwise the distortion levels are boosted as the level of the fundamental drops. Note that this option will not affect the traces when the plot is not normalised, but will still affect the values in the legend if the distortion figures are set to read in percent or in dB relative to the fundamental.

 

[mitchco]

ok thanks for that. With the check box off:

2432H CD1 looks out of spec. 2HD?

 

[Blumlein 88]

ok thanks for that. With the check box off:

2432H CD1 looks out of spec. 2HD?

Only out of spec by a factor of 10x or so on the distortion levels.

Off topic comment: Finally got around to installing REW on one of my Linux machines today. Hallelujah it works great! Linux, REW and UMIK-1 a nice handy simple measuring solution.

 

[Brad]

Here is a document with a lot of CD measurements
http://forums.melaudia.net/attachment.php?aid=1760
I have experienced similar problems with a CD. In my case it was with a celestion cdx1-1747. The distortion increased after too much low frequency energy went thought the driver (even with a protection cap).
There was no visible damage to the diaphragm, but the shape was slightly distorted.

 

[Thomas savage]

Here is a document with a lot of CD measurements
http://forums.melaudia.net/attachment.php?aid=1760
I have experienced similar problems with a CD. In my case it was with a celestion cdx1-1747. The distortion increased after too much low frequency energy went thought the driver (even with a protection cap).
There was no visible damage to the diaphragm, but the shape was slightly distorted.

Welcome to the forum Brad, thanks for taking the time to contribute to our microscopic corner of the Internet

 

[mitchco]

Thanks Brad, I appreciate that. Some great info there. My little Class A amps have turn off thump, so I have a 82uf protection cap in series (~500Hz roll off @ 8 ohms). Never had any issues with some of the other CD's I have had using this same setup. Can't be sure though - I should have measured the CD's as soon as I installed the system.