Extend ASRC to support 16KHz FS_MIN

[shuchitak]

The ASRC implementation currently supports sampling rates from 192Kz to 44.1KHz. In order to use it in sln_voice we'd need to extend it to support a minimum sampling rate of 16KHz.

[mbanth]

@shuchitak, please provide a Story Point value for this issue. Thanks.

[shuchitak]

@mbanth, Given my unfamiliarity with the design I can't provide an accurate story point value for this. I've allocated 15 days but that's just a guess looking at the size of the ASRC code.

[henkmuller]

For what it is worth, creating 24/32 inputs was surprisingly straightforward, and I suspect adding 16 as an input is not that hard. Note it will need characterisation and I may have missed something obvious. Adding 16kHz as an output frequency will be simple for most low sample rates, but for high sample rates (say 192 -> 16), one may need to fundamentally alter the method of operation since it will need a 12x reduction in sample rate. The current two-stage approach may not be good enough for that.