Ann: Performance Specifications, page-5

To me this doesn't look like anything wonderful.

First of all, they have only tested a single chip and provided theoretical results for arrays.

They also haven't specified if the test results relate to maximum output, or as speakers are usually tested, at 1 watt input power. I realise the 1 watt input would not be transferable to this tech, but they should at least specify if this is max or not and what the all important total harmonic distortion is at these levels and compare this with the 'typical' mobile phone speaker distortion.

Distortion is key, there is no mention of this. If it was good news about that, it probably woudl have been in there.

They have not mentioned dimensions for the single chip they tested. How big is it? If it's the 1cm by 1cm unit?

It shows in comparison to the 'typical' mobile phone speaker, a better performance between 200 and 700hz, but this better performance is below 65dB, already so low that it's probably only relevant if the phone is close to your ear.

Also, the 'theoretical' 64 chip speaker, (i'll assume these charts mean max) looks like it can produce 93dB at 100hz. To compare to a home or car speaker, it's about the same, maybe 3dB above average, but look at it tail off after that. The all important low frequency range below 100hz down to 20 hz, conveniently not shown on the graph (leaving out the most important part) looks like if you project the straight line, it's heading for sub 70dB at 20hz. This is awful sensitivity and useless for home hifi or subwoofers, car speakers etc. You will need larger arrays to achieve something like what we are used to from traditional speakers. Looking at the chart, maybe 1024 chips to achieve 100db at 20hz, and 4096 chips to achive 110db at 20hz.

A good set of home hifi speakers can easily exceed this.

How big and how expensive is a 4096 chip array? if each chip is 10by10mm, that would be 640mm by 640mm.

That's not too bad, but I'd like to know the cost, the distortion figures and see actual results for these larger arrays, instead of theoretical data on a chart.

The actual measured results for the single chip, even if the distortion figures were comparable, I don't think would excite a mobile phone manufacturer. It shows much better performance above 16,000hz, but that's for the dogs, most humans cant hear much above this frequency.

It shows better efficiency below 700hz and between 900 and 16,000, but the actual power savings I expect, will be negligible in a phone power budget.

If they can make the 4096 array for $20, I'd be interested in this company. That means they have to bring the price of each chip to 0.49 cents (half a cent.)