Distortion Surface: old measurement, new approach

There will be some inherent roughness because this is the nature of distortion. It's not always stable and while the AP has various settings that determine what value to "catch", the noisy aspect of distortion cannot be totally eliminated.

The above were arrived by taking measurements at 31 points from 20 to 20kHz. Only up to 10kHz is plotted because 2nd harmonic goes past 20kHz after a 10kHz fundamental. But given the lessons learned on the MIL/MIB DAC, perhaps would be wise to still plot distortion to 20kHz.

That last thought aside, we won't see any narrow peaks/valleys in the measurements above. When it comes to FR where we do see narrow peaks and valleys, we get around this by applying smoothing 1/9, 1/24, etc.

As far as roughness of the distortion surface, who knows? We don't have enough data yet. I can say that with headphone measurements, narrow peaks in the distortion surface probably isn't a good thing and may likely corroborate irregularities in the linear distortion (frequency response) measurements. "Distortion" distortion often refers to non-linear distortion. It's non-linear because we can't correct it like FR, which is linear.

 

Here are the same SW51 results, but distortion is in ratio (not absolute). Same settings per above, but the other channel. For amps, I prefer the ratio.

H2


H3


H4-10

 

@Lyander, this is your request for Piety

Nitsch Piety
High Gain
300-ohm load
90kHz Bandwidth
AP "Cheat" Mode On
Channel 1

Second Harmonic Distortion Product Ratio


Third Harmonic Distortion Product Ratio


H4 to H10 Distortion Product Ratio

 

Here is THD+N (or -SINAD) for the above. Like @atomicbob's measurement of the MJ3 above, noise is really the limiting factor. And even then, taking things into context, no one is going to hear this noise.

 

Oh, I figured if anyone wants to play with this / update it to superbestaudiofriends, SBAF, whatever, put flames, nuclear mushroom cloud, etc.



The Jolly Roger represents independent thought. The banner the skeleton is breaking is a tessellated old Head-Fi logo. It represented hype train back then, but today means blinding following "authorities" as "truth" (e.g., Harmon consumer preference research target, which by the way, has changed ever few years, etc.), simplistic thought (black and white approach, numbers for the sake of numbers, litmus tests based on singular SINAD at output levels that would either permanently destroy hearing or kill a headphone after a period of time).

The pyrate theme was chosen because Currawong (Head-Fi moderator back then) admonished me as to why I insisted on being a "pariah". I wasn't insulted. I was honored!

 

Probably not unless there is something noteworthy. I feel the base of H2, H3, H4-10 is already on the cusp of too much information. However, if there is something interesting, it may be worth to show it. The effort to produce the graphs is much higher than a quick look in instrumentation.

 

You are my hero! May singular "SINAD" be left to what it's supposed to be.