Ann: Performance Specifications, page-97

Thanks for going to the trouble there MattyE.

I stand behind my calculations however I have made an incorrect assumption as far as I can conclude. This is clearly a complicated argument. In defense of my calculations,

First of all, what you have written here: 'Nice One has the fundamentals wrong. SPL is ALWAYS 20log, not 10log. You only use 10log when calculating dBm and we are talking about dBspl - honest mistake? Nice One should be aware that dBm is commonly used in RF power calculations and sometimes to describe audio amplifier gain, but not in this context."

I clearly specified that there was 20dB difference between 0.1m and 1m in my methodology. Adding back 10dB was to convert from 0.1w to 1W. So the 10log you say I have used includes a tenfold increase in audio power. I think you or your source has misunderstood this. But I now understand that in their calculations they have not used 1W as they assume the micro speaker cannot handle this amount of power. My assumption that it could handle 1w was where I was incorrect.

If they would just include these assumptions and calculation/conversion details in the chart notes, like most technical documents would, then there would not be this confusion.

So I can go back to my initial conclusion, that a single chip AP speaker outperforms the typical phone micro speaker between 100hz and 900hz, albeit at pretty low levels that as I mentioned before may not be that relevant in the real world. Other benefits of the AP speaker for a mobile phone application might be a flatter freq response once equalised. Maybe distortion might be better but it's not compared on these charts.

From this exercise we have potentially gathered from your unknown source, that these charts on the AKP announcement show MAX SPL. This is very important information.  

If you compare the low frequency max spl response, which as I have mentioned is all important and will set the limit of efficiency for a AP hifi speaker, as the better responding higher frequencies will need to be equalised flat, then a theoretical 64 chips speaker, the best performing one on that graph, reads 86db MAX SPL at 100Hz.

Going beyond the micro speakers now, I suggest you research what a bookshelf speaker's sensitivity at 100HZ is at 1 Watt at 1 metre (not max spl), then compare it with this 86dB MAX SPL figure for the 64 chip ap speaker then consider the bookshelf speaker can be driven with much more than one watt. This comparison should say something to you about the size of the array needed to make enough SPL to compete with conventional speakers. Maybe that could be another question for your source Matty.

Here is one that is pretty small http://www.mordaunt-short.com/products/alumni-2

	Column 1	Column 2
0	Dimensions (h x w x d)	176 x 108 x 145mm

It has a rated sensitivity of 86dB at 1 watt/metre. Power  handling is 15 - 80W recommended.

Looking at the AP chart, the 64 chip theoretical array at maximum power could only achieve what this litte bookshelf speaker can do at 1 watt. So what size array would you need to match the performance of this bookshelf speaker? Much more than 64 chips that's for sure. 256 chips for another 10dB above 86 = 96dB, 1024 for another 10 to 106dB, 2048 for 116dB I'm guessing at least 1024 to 2048 chip array.