Speaker Distortion - how is it specified and how is it measured?

[alanf]

New member here. Bear with me if this is a dumb question or if it's covered elsewhere...
In the good Polk R700 speaker review on this site, there are many helpful charts of measurements. But nothing of distortion that I can see. I would think that a setup that sweeps a very low distortion 30Hz-20kHz sine wave into the speaker and measures output power at the input frequency divided by all power at frequencies other than input frequency would be very useful. Just plug the test microphone into a spectrum analyzer. One would make these measurements/calculations while sweeping both input frequency and input power. SNDR, SNR, THD, THD+N, etc. Measurements like these are common for amplifiers. Why not for speakers too?

 

[shadyJ]

New member here. Bear with me if this is a dumb question or if it's covered elsewhere...
In the good Polk R700 speaker review on this site, there are many helpful charts of measurements. But nothing of distortion that I can see. I would think that a setup that sweeps a very low distortion 30Hz-20kHz sine wave into the speaker and measures output power at the input frequency divided by all power at frequencies other than input frequency would be very useful. Just plug the test microphone into a spectrum analyzer. One would make these measurements/calculations while sweeping both input frequency and input power. SNDR, SNR, THD, THD+N, etc. Measurements like these are common for amplifiers. Why not for speakers too?

Electronics don't have to deal with acoustics. Acoustics makes measuring distortion fairly difficult in a standardized manner.

 

[TLS Guy]

New member here. Bear with me if this is a dumb question or if it's covered elsewhere...
In the good Polk R700 speaker review on this site, there are many helpful charts of measurements. But nothing of distortion that I can see. I would think that a setup that sweeps a very low distortion 30Hz-20kHz sine wave into the speaker and measures output power at the input frequency divided by all power at frequencies other than input frequency would be very useful. Just plug the test microphone into a spectrum analyzer. One would make these measurements/calculations while sweeping both input frequency and input power. SNDR, SNR, THD, THD+N, etc. Measurements like these are common for amplifiers. Why not for speakers too?

You can't really do it. For one thing the maximum power varies by pass band. For instance a tweeter only handles a fraction of the power of a woofer.

You can't really do steady state measurements on a speaker as the VC gap will heat quickly. If it does not blow the speaker there will certainly be thermal compression.

Lastly, I'm pretty sure you would not want to know, and ignorance is bliss.

However, good designers know intuitively where to place their resources. This comes from knowing the energy/frequency sources of program.
This guides you how to apportion resources. I am certain my designs are influenced by my musical preferences. I make conscious decisions on the power band resources. This is actually very much akin to what organ builders do when specifying an organ for a customer.

So, if an organ is moved, a builder will always revoice it.

I am one of the few lucky guys who can design, build and voice to space as a total design integrity, just like a custom organ builder would.

 

[ski2xblack]

In the good Polk R700 speaker review on this site, there are many helpful charts of measurements. But nothing of distortion that I can see.

Linear distortion amounts to deviations in frequency response, so in that review the freq response data is *all about* distortion.

You may be thinking more in terms of non-linear distortion. Speakers do produce non-linear distortion, typically in amounts orders of magnitude greater than that produced by electronics. Alas, it's just inherent to a solid object (speaker cone) pushing on air (very squishy medium) when trying to convert electric signals into sound, typically low order.

I'm sure that James would have relayed any distortion as far as resonances or obnoxious audible tells in the review. Impedance sweeps and waterfall plots expose those.