I've been trying to wrap my head around something for a few weeks and it's
just not coming together. It all comes down to one basic question: What is
so wrong about truncation?
Some thoughts that have been in my head...If you're hitting full scale
levels using a 24 bit ad, are the most significant 16 bits coming out any
different than they would on a 16 bit ad. While wrestling with this I put
together a little spreadsheet that takes a 24 bit word, truncates the 8 LSBs
and then compares the truncated output to the 24 bit output in terms of % of
full scale. As expected, the deviations only appear on the last couple of
bits. Then I consider my recording situations (live concerts in bars/clubs)
and the fact that the levels at which the quantization error occur are far
below the ambient noise. Does that mean it doesn't matter?

This is not meant to question the merits of 24 bit - that much I understand.
This is about running a 16 bit capture device and a 24 bit source e.g.
backup on a JB3 when I do 24 bit laptop recording.

thanks

david