Frequency Normalized Distortion

[dlr]

John,

The March 2026 edition of audioXpress has an interesting article on transducer distortion and measurements. It's by Steve Temme of Listen, Inc. who proposed a distortion measurement algorithm (1993) he called Frequency Normalized Distortion. He explains and demonstrates in the article that the low frequency distortions may not necessarily be an issue when examined from the frequency normalized aspect for a given transducer. His arguments are solid. I think it would be beneficial to have this option in REW. Given what I see in his presentation, it may be a fairly complex addition, though. Of course it would also add overhead to the processing required. I suspect that the data exists in REW for that.

Is this something you could consider adding to REW? It would likely be a new graphing tab in the main window and the Overlays. Or the overlays would have additional options for "Measure to plot". This would make for some confusion in some users, but I think that the addition would be very beneficial as another aspect of distortion analysis of transducers.

 

[John Mulcahy]

That's the "Use harmonic frequency as ref" option.

 

[dlr]

That's the "Use harmonic frequency as ref" option.

Ah, I had no clue. Thanks, I'll check that out.

 

[dcibel]

https://www.listeninc.com/wp/media/2026/02/Article_Frequency_normalized_distortion_AX_March-2026.pdf

From a quick scan, these features are already in REW via the "use harmonic frequency as ref" and "plot harmonics at the harmonic frequency" options. Use dbr Y-axis scale to view a normalized result.

FWIW Low frequency distortions cause all sorts of modulation issues higher up in frequency, which can be measured and listened to via FSAF measurement. Don't ignore low frequency distortion.

 

[dcibel]

John beat me to the punch.

 

[dcibel]

For what it's worth @John Mulcahy , the THD figure on the distortion chart is still shown in the "traditional" way even when "plot harmonics at the harmonic frequency" and "use harmonic freq as ref" are checked. It's not bothered me in the past as I usually disregard THD as I'm more interested in the individual harmonic traces.

 

[John Mulcahy]

Not really. From the help:

Plot harmonics at the harmonic frequency changes the way the harmonic orders for sweep measurement distortion are plotted - rather than plotting the distortion at the frequency of the fundamental it is plotted at the frequency of the harmonic, so (for example) the 2nd order harmonic distortion level for 1 kHz would be plotted at 2 kHz, where the distortion occurs. This makes it easier to correlate harmonic levels with the level of the fundamental at the harmonic frequency and also helps distinguish the effects of external noise, since noise affects all harmonics at the frequency at which it occurs. The THD trace is not affected by this control, it will continue to show the THD value corresponding to the harmonic levels at their fundamental frequency. However, when plotting harmonics at the harmonic frequency the level of the fundamental at the harmonic frequency (which in this graph mode is the cursor frequency) is used as the reference, so Use harmonic frequency as ref will be selected and the control for it disabled. If that was not previously selected the THD trace will update.

 

[dcibel]

Not really. From the help:

Plot harmonics at the harmonic frequency changes the way the harmonic orders for sweep measurement distortion are plotted - rather than plotting the distortion at the frequency of the fundamental it is plotted at the frequency of the harmonic, so (for example) the 2nd order harmonic distortion level for 1 kHz would be plotted at 2 kHz, where the distortion occurs. This makes it easier to correlate harmonic levels with the level of the fundamental at the harmonic frequency and also helps distinguish the effects of external noise, since noise affects all harmonics at the frequency at which it occurs. The THD trace is not affected by this control, it will continue to show the THD value corresponding to the harmonic levels at their fundamental frequency. However, when plotting harmonics at the harmonic frequency the level of the fundamental at the harmonic frequency (which in this graph mode is the cursor frequency) is used as the reference, so Use harmonic frequency as ref will be selected and the control for it disabled. If that was not previously selected the THD trace will update.

I've highlighted in red the part I was referring to, which is the THD trace.Reading through the AudioXpress article, the focus is on a modified THD calculation, which I don't believe REW is doing, although all other aspects are covered.

 

[John Mulcahy]

It is doing it. Choosing the plot harmonics option forces use of harmonic as ref, so whether the THD plot changes depends on whether that was already selected or not. Plot harmonics does what it says on the tin, changes how harmonics are plotted.

 

[dcibel]

So it is, never mind then