Harmonic distortion : Kef R3 speaker versus Revel Performa 3 F206

[Lemanss]

I'd like to compare the harmonic distortion of the Kef R3 speaker with the Revel PERFORMA 3 F 206 speaker. I tried to combine the graph published by Erin's Audio Corner for the Kef and the one published by Amirm for the Revel. It's a bit of a workaround, and I'd be reassured if someone could tell me it's correct. Thank you for your understanding if I'm wrong. Thank you for this wonderful website, which has shaken things up considerably in France.

 

[Lemanss]

 

[stoneeh]

You cut off the labels on the y axis. For at least he 86 dB graph though, it is already still obvious that you aligned the graphs of both speakers wrong.

 

[Doenerkunde]

The distortion graphs for the Kef R3 are relative (= compared to the main signal). The distortion graphs for the Revels you picked are absolute. You cant compare the two like this.

Relative graphs show the difference between the main signal and distortion. At the lower frequency bound the Revel and the Kefs both have rising distortion AND a falling level of the main signal. The relative graphs for the KEFs incorporate BOTH into their distortion curve, while the absolute graphs of the Revels do not. This is the reason distortion shoots up way more in the KEFs graphs.

You could have instead chosen to use the relative distortion graphs from the Revel review, but they are converted to percentages, while the relative distortion graphs for the Kefs are given as difference in decibel, so you'd have to convert one of the two (each -6dB equals -50%, so -40dB is 1%).

 

[Lemanss]

Thank you, Doenerkunde.
I see you understood my question.
I understand your answer.
But my brain understands better with examples.
You told me -40 dB corresponds to 1%.
I'll make a modification to the Kef ratio to see if I've understood correctly.
Could you give me the figures for:
-50 dB
-60 dB
-70 dB

I'm 59 years old and I remember logarithms, but I haven't used them for over 30 years.

 

[Doenerkunde]

Thank you, Doenerkunde.
I see you understood my question.
I understand your answer.
But my brain understands better with examples.
You told me -40 dB corresponds to 1%.
I'll make a modification to the Kef ratio to see if I've understood correctly.
Could you give me the figures for:
-50 dB
-60 dB
-70 dB

I'm 59 years old and I remember logarithms, but I haven't used them for over 30 years.

Here is a useful calculator tool for conversion between decibel attenuation and percentages:

THD to dB - convert percent % to decibels dB percentage voltage % vs per cent converter THD+N total harmonic distortions calculation signal distortion factor attenuation in dB to distortion factor k in percent decibel damping - sengpielaudio Sengpiel

Calculate THD and distortions decibels dB to percentage % per cent percent voltage % vs converter and dB signal distortion THD+N total harmonic distortions factor attenuation damping - Eberhard Sengpiel sengpielaudio

sengpielaudio.com

-50dB -> 0,32%
-60dB -> 0,1%
-70dB -> 0,03%

 

[Curvature]

If the only data you have access to is through an image, then use of this tool is necessary: https://automeris.io/ I have made enough errors eyeballing graphs that I don't let myself come to any conclusions without having all data for comparison loaded into the same program, converted and scaled appropriately.

For the F206 data, we have absolute SPL and THD or relative % THD. I used the latter. The former image is too low resolution and requires extracting the FR trace and the THD trace, then normalizing the THD to the FR. It's more involved than extracting the % THD trace and converting the results to dB.

Erin's distortion measurements are anechoic, while Amir's are not, so room noise corrupts the results (edit: they are still accurate enough to give the general trend).

 

[Lemanss]

Well done, Curvature I was painstakingly trying to create this curve with the AI, following the path given by Doenerkunde, when you posted it already done. Would you be so kind as to create it for 96 dB as well? I know I'm exaggerating, but the work was so well done for 86 dB.

 

[Curvature]

Well done, Curvature I was painstakingly trying to create this curve with the AI, following the path given by Doenerkunde, when you posted it already done. Would you be so kind as to create it for 96 dB as well? I know I'm exaggerating, but the work was so well done for 86 dB.

Sure. Although I encourage you to learn to use the tool. I've used it to check poorly labeled charts in academic books and journal articles, not just in audio.

In this case I'm using Amir's absolute SPL measurement instead of the relative % THD. To compare with Erin's chart, I need to normalize the THD SPL to the on-axis magnitude response, and to do that I need to interpolate the datapoints to get the relative figures for the same frequencies. Few extra steps and a touch of math.

https://www.audiosciencereview.com/forum/index.php?threads/revel-f206-tower-speaker-review.53691/
https://www.erinsaudiocorner.com/loudspeakers/kef_r3/

Posting both the 86 and 96 for ease of comparison.

 

[terryforsythe]

Erin's distortion measurements are anechoic, while Amir's are not, so room noise corrupts the results (edit: they are still accurate enough to give the general trend).

Both Erin and Amir use a Klippel Near Field Scanner System (NFS). Erin's Klippel NFS is in his garage, not in an anechoic chamber. Indeed, the reason for using the Klippel NFS is to avoid having to use an anechoic chamber. Did Erin measure the R3 before he got his Klippel? That may be the case, I don't know. EDIT: Nope, he used the Klippel. Indeed, in the review he states "All data collected using Klippel’s Near-Field Scanner."

 

[Curvature]

Both Erin and Amir use a Klippel Near Field Scanner System (NFS). Erin's Klippel NFS is in his garage, not in an anechoic chamber. Indeed, the reason for using the Klippel NFS is to avoid having to use an anechoic chamber. Did Erin measure the R3 before he got his Klippel? That may be the case, I don't know. EDIT: Nope, he used the Klippel. Indeed, in the review he states "All data collected using Klippel’s Near-Field Scanner."

There are settings in the Klippel NFS that determine the treatment of measured data. Erin, from the beginning, has presented anechoic distortion derived from NFS measurments. Amir has, IIRC, presented nonanechoic distortion measurements in almost all his reviews. Same goes for his waterfall measurements, which are clearly nonanechoic in all instances.