Help understanding graph from a sound measurement

[seb123audio]

Hi, I have a graph that has been created from a "Distortion Product Ratio H2" test.
The parameters for the test are output volume and the frequency sweep range (20Hz-1kHz) and a sweeptime of 5 seconds. 3 channels have been recorded but just ignore all but one. I am trying to understand how to interpret a single channel's result.

Here is the graph:

Link to image: Imgur: The magic of the Internet

If I am not mistaken, the test does the following:


- A)It sweeps through the frequency range for a duration of the sweeptime.

- B)For each and every frequency in that range: the dBSPL of the 2nd harmonic frequency of the currently played frequency, is recorded as a value.

- C)The result is a graph were each value in the graph shows the difference between the 2nd harmonic and its fundamental in dBSPL. The x-value is the
fundamental frequency. The y-value is the difference between the dBSPL of the fundamental frequency and its 2nd harmonic expressed in dB.

- D)Formula: <Fundamental frequency in dBSPL> - <2nd harmonic in dBSPL> = <distortion product ratio for that fundamental frequency in dB>.



Am I correct in my interpretation?
If I am correct, were can I find a source for this? I have looked and looked but can't find any site explaining it.

 

[adason]

i see no image

 

[seb123audio]

i see no image

Fixed it.

 

[DF96]

On the Audio Precision website it says

In the Distortion Product Level view, the graph displays rms level. In the Distortion Product Ratio view, the graph displays the ratio of the rms level of the distortion product to the rms level of the fundamental tone.

 

[Elvee]

I do not see why the graph seems to be centered about ~-5dB.
In fact, for H2 (or any H), the values should all be largely negative, unless there is something really wrong with the test setup or the DUT (can happen, but should be fixed before trying to interpret the results)

 

[cbdb]

According to the graph the distortion varies 80db! Isn't +40db 10,000% or more distortion? It might be normalized +60db or something but that's confusing. Any way very bad distortion if it's from an audio device.

 

[DF96]

I am going to guess that this is a plot from a speaker test, done by someone who is still learning about dB and reference levels.

 

[PRR]

It might help if we knew what the driver is.

If the cone is smaller than your palm, it is very reasonable that above 150Hz the THD at high level is near 10% (-20dB); and below 150Hz the speaker is not making much fundamental and the distortion and stray wheeze reads higher than the fundamental (positive dB).