DSP correction of HD and IMD

Why isn't it done? It seems a simple enough exercise to build a profile of a drivers non-linear distortion and then apply pre-distortion to reduce it. Building DSP algorithms is beyond my ability but I'm sure some of the big companies could do it.

As an example, if you play a 1KHz test tone through a midrange driver and find that it produces harmonics at 2KHz - 45dB, 3KHz -49dB 4KHz -62dB etc... then you could simply add these same harmonics out of phase to the input signal. Run a full spectrum sweep at multiple levels and you can build a profile of how the driver behaves. Even a naturally poor driver could be made to perform very well... couldn't it?

 

Sort of like extending the amplifier NFB loop to include the driver?

I think for that to work you'd need to be recording the driver output acoustically in real time since the distortion levels with vary with SPL.

If you mean modelling the driver distortion and then having that run in DSP, anticipating what the driver is doing, I suppose it could work f you measured each driver individually in order to program the DSP.

 

Recording the driver output in real-time is basically servo control, but it seems that is limited, not only to low frequencies because of the time it takes for acoustic feedback to work, but also to a narrow band where the phase of the driver can be aligned with the input signal for negative feedback to work (the driver exhibits phase shift relative to the amp input signal so NFB doesn't naturally work).

These two problems can be avoided by pre-distortion of the input signal without a need for acoustic feedback provided you know what the driver is going to do. This would allow correction through-out the range. As you say, it is level dependant, but I'd have thought the required correction could be interpolated from a handful of different level tone sweeps until you hit the excursion limits of the driver.

I agree that ideally every driver would need an individual 'map', but surely a generic map could still dramatically improve things for a set model of driver.

I wonder if it could be done for IMD as well? Perhaps a 2-tone input sweep is not enough to model multi-tone behaviour.

 

Didn't Philips do this with the MFB speakers of the late 70's?

I think a lot of the commercial problems would be similar to the "integrated active speaker" scenario, where many enthusiasts (well, us baby-boomers anyway) and servicing dealers are concerned. selling an all-in-one solution isn't what "we" are programmed for. Supplying a product where everything is done for you (including minimising driver distortions in whatever way possible) just isn't what the remaining audiophool fraternity wants. The current youngsters eventually after a better sound, once their student loans are paid and they finally have a job with money to spend, may feel very differently about it though

 

I would think that the DSP idea is problematic because the HD/IMD varies a lot with level also, and even a bit drive impedance, and if you dont get your DSP inversion correctly matched it'll make things worse not better.

Oh and will varying the DSP by level also alter the latency..? (merely a thought, I really dont know much about DSP...)

A big complicated mess for little tangible benefit, I'd have thought. Good drivers used well can be very well behaved.