IM distortion

[acoustat6]

Hi, Does the program show distortion? I am sure I have seen some reference to it but cannot find it again. I thought it showed up during initial measurements but now cannot find it. If it does show distortion, would it be speaker or electronics distortion or some room artifact.
Thanks,
Bob

[brucek]

From the REW HELP FILES under SPECTRUM, scroll down to the heading DISTORTION.

In part:

distortion.jpg

brucek

[quest100]

From the title I believe the OP was asking about intermodulation distortion. My understanding is that IMD can only be measured with two (or more) input frequencies. Typically you would produce a signal that had, say, 30 and 72 Hz sinusoids and look for responses at 42 and 102 Hz. More response would occur at m*42 +- n*102 where m and n are (small) integers. See http://www.hometheatershack.com/foru....html#post6991 for how the IMD tests are done for this sites subwoofer tests.

It seems like it would be nice to add measurement of IMD to REW. Any possibility of this?

[JohnM]

It would be possible to add IM calculations but I don't think they have much value for low frequency devices like subwoofers. Ilkka no longer makes those measurements for the subwoofer tests.

Any device which produces intermod distortion will also produce harmonic distortion, IM is a more useful measurement when the harmonics are outside the range of frequencies that can be measured. Examples would be RF devices where operation is typically over a narrow frequency range within a band and out of band content is heavily suppressed by other elements of the device being measured or audio systems when operating at higher frequencies. For lower frequencies there are no difficulties measuring harmonic distortion so non-linearity is easily determined from that.

[quest100]

Does this mean that if you know the harmonic distortion at two frequencies that you can calculate the IMD? Is it a simple calculation? The reason I ask is that it is my understanding that HD, at reasonable levels, is inaudible while IMD is readily apparent. Most music contains lots of harmonics and this masks the HD. IMD creates new frequencies and does not sound musical.

[JohnM]

The approximation is pretty straightforward although precise calculation is a little more involved. The first order intermod products (i.e. the sum and difference of the 2 input frequencies) will be approximately twice the level of the 2nd harmonic. For a more rigourous explanation try this.

[acoustat6]

Hello, Thanks for the responses thusfar. I was asking about IMD, which apparently the REW does not calculate. It has been mentioned that it is not "relavent" to subs or freq below 200hz, is IMD (if REW was able to measure IMD) useful and mesasurable for the main speakers?
thanks,
Bob

[JohnM]

IM measurements can be useful to assess distortion behaviour at high frequencies, when the harmonics would be beyond the system measurement bandwidth. Whether it is useful depends on whether you could use the result to take any action. What did you have in mind?

[DrWho]

I'm not convinced that THD and IMD are perfectly correlated. For instance, motion of a diaphragm creates a doppler effect when more than one tone is being reproduced. Even with perfect cone motion (so no THD), there will always be an IMD component. It's one of the main reasons to go with horns (it limits cone excursion).

That said, there's no reason the spectrum analyzer couldn't be used with an external two tone test signal and then IMD distortion manually calculated.

Also, I think there's an IEEE article that discusses how IMD can be better calculated with some kind of swept signal (it's been a while since I've read it).

In other words, I think there is more to IMD than just the high frequency trick to check small signal non-linearity with bandwidth limited measurements.