Distortion and Other Measurements ?????

I've been reading a lot lately and had the following questions
based on information I came across here:

"Experimental evidence shows the ear to be most sensitive at around 3kHz, and to have a second, smaller sensitivity peak around 12kHz. If we arrange for the noise gain (1 - H ( ω ')) to have corresponding dips at these frequencies then there will appear to be less background hiss despite a small overall noise power gain due to the noise-shaper."

Here is a link to this from a DAC Theory link. The quote is
near the end of the link past the Noise Gain table and right
past the filter caculation b(k) = -1....

Introduction to digital audio


QUESTION

Wouldn't it be better to use 3 KHz as the foundation for distortion
measurment and lowering distortion than using traditional 1 KHz signal?

If our ears are more attuned to 3 KHz and 12KHz, does it make sense to
measure what our ears are mose sensitive too than the typical 1 KHz 10KHz?

After all, we are DIY AUDIO yes?

AND

The goal is to make our own stuff or modify it to sound the best, yes?

YES

I know it would be a task to refit older gear with the "right" componants for these frequencies, but would the effort be worth the task?

Any one here try it? yes, no, maybe?

What was the result? good, bad, no different, or observations?

I'd do it myself but I am lacking the resources in equipment,
methodology, and background.

I am interested because I think it might be a worth while pursuit
for all the reasons this is the DIYAUDIO site.

Thank you in advance for your thoughtful commentary.
 
I would like to point out that 2nd and 3rd harmonic distortion at 1 kHz produce 2 kHz and 3kHz respectively. These distortions would sound louder than the same amount of distortion of a 3 kHz test tone, which would be at 6 & 9 kHz. Some designs have distortion so low that it is below the noise floor at 1 kHz so 10 kHz is the most useful test frequency. Any single frequency is a trade off.

Often on old gear the best single improvement in distortion can be had for free. That is by adjusting the bias on the output stage. In class AB amps, crossover distortion is the biggest offender and it does produce high frequencies. It is dangerous, though, as you can easily kill an amp by over-biasing it. The other problem is that you almost certainly need an oscilloscope or similar to get it right. So there's my 2 cents.
 

1audio

Member
Paid Member
2004-03-24 5:16 am
SF Bay Area
He is right. A 1 KHz fundamental will highlight 2 and 3 KHz (and higher). Adding even a simple spectrum analyzer via a sound card to the output of any 70's or later distortion analyzer will tell you a lot more than any number and is pretty easy if not free once you have the analyzer.
 
Measuring the Audio Spectrum & Beyond....

Ok,

Barefoot, Damian, Samuel points very well noted.

Samuel, and others, if we wanted then to test the audio Spectrum,
Which values frequencies to test and which process(s) would we want
to use?

I'm just putting it out here so please correct/direct...

one pass or multiple passes/sweeps?

THD/THD + N

f @ Hz:
10, 20, 30, 60, 120, 240, 480, 960, 1920, 3840, 7680, 15360, 30720, 61440.

And/Or

Other f ?
 
There's no single procedure that catches it all--in the end you'll have to understand the behaviour of the DUT and the circumstances of its use, and select the appropriate measurements.

For a basic analog DUT, e.g. a preamplifier, a comprehensive distortion measurement series would comprise e.g. a sweep in a 1-2-5 sequence from 10 Hz to 50 kHz, for the 2nd and 3rd harmonic as well as THD (up to the 10th harmonic). This is done at several output levels, e.g. again a 1-2-5 sequence from 100 mVrms to 5 Vrms. Ideally this all is repeated with different load resistances (e.g. no load, 2kOhm, 600r), and perhaps even with different source resistances (e.g. 20r and 600r).

As you can see this easily generats a lot of data which may often have substantial redundancy and asks for appropriate interpretation. Frequent application of such extended sweeps is only feasible with automated measurement procedures.

Samuel