Is it possible to eliminate IMD in coaxial speakers with DSP?

[Liopleus]

When listening to designers of coaxial speakers, a lot of them talked about limiting intermodulation distortion by limiting the excursion of the woofer surrounding the tweeter. This is done through either adding another separate low-frequency woofer in Genelec, KEF and Kali design or using a very large woofer in the MoFi SourcePoint.

Since IMD is caused by the movement of the woofer surrounding the tweeter changing the sound of the tweeter, I'm curious if it's possible to predict the exact change in waveform of the tweeter caused by the movement of the woofer with the original signal, and add another signal of inverse phase to cancel out the IMD (or at least reduce it).

I don't have much knowledge in physics or speaker design, and I know that IMD is affected by volume, driver configuration and many other factors. However, I want to know if it's possible to eliminate IMD with DSP provided enough information being supplied to a DSP with enough computing power.

If it's possible, the performance of compact 2-way coaxial speakers could be greatly improved so I'm curious if there's any research on this.

 

[ernestcarl]

When listening to designers of coaxial speakers, a lot of them talked about limiting intermodulation distortion by limiting the excursion of the woofer surrounding the tweeter. This is done through either adding another separate low-frequency woofer in Genelec, KEF and Kali design or using a very large woofer in the MoFi SourcePoint.

Since IMD is caused by the movement of the woofer surrounding the tweeter changing the sound of the tweeter, I'm curious if it's possible to predict the exact change in waveform of the tweeter caused by the movement of the woofer with the original signal, and add another signal of inverse phase to cancel out the IMD (or at least reduce it).

I don't have much knowledge in physics or speaker design, and I know that IMD is affected by volume, driver configuration and many other factors. However, I want to know if it's possible to eliminate IMD with DSP provided enough information being supplied to a DSP with enough computing power.

If it's possible, the performance of compact 2-way coaxial speakers could be greatly improved so I'm curious if there's any research on this.

I've never heard of DSP being used for that kind of specific application... besides, the woofer movement may be too "dynamic" (even on the small scale).

Alternative is to use a coaxially mounted horn waveguide. It should altogether avoid or at least significantly reduce the kind of IMD seen at very high/extreme SPLs. For example, Fulcrum Acoustic's RM28. Due to the HF horn loading and larger mid-range woofer pairing, I would expect to see significantly less compression or DSP limiting applied in the higher frequencies -- although, normally, this should not be an issue anyway in nearfield/midfield studio monitoring.

The main compromise is there will be more diffraction and loss of frequency response linearity seen at increasingly off-axis angles. From what I've observed, generally, as long as the listening position stays within +- 30 degrees axially, the frequency response should be quite linear enough for critical listening applications e.g. studio monitoring work.

FIR DSP correction here is applied (amongst other things) to treat any significantly consistent/regular occurring horn-reflection-induced resonances.

Temporal Equalization—TQ™ | Fulcrum Acoustic

Temporal Equalization TQ™Rather than choosing a compromise between two competing attributes, we physically optimize the attribute that can’t be addressed with DSP, and solve the other problem with DSP. The Fulcrum DSP Philosophy – TQ™ Means Solving the Other Problem

www.fulcrum-acoustic.com

I'm no real expert on this subject myself, but, I think the concept may be similar to Acourate's "room" virtual bass array (VBA) filtering. Unlike listening rooms though, these horn reflections are physically stable enough to effectively treat/correct for much more precisely via DSP.

 

[fpitas]

I'm guessing that DSP could compensate, if it knew the exact cone movement.