I've been trying to tell Dan that I think he might be wasting his time...
he says:
When you combine two 16/44 sources to single
16/44 source, you have made a new source. Not just a new source on
paper either but a real, living, breathing source. The SBD + AUD matrix
mix doesn't sound like the SBD, doesn't sound like the AUD but is
actually a new animal altogether. (This is 2 channel mind you, the 5.1
is a horse of a different color.)
What I was able to do is take the two 16/44 sources and do a 50/50
mixdown to a new 24bit source. There was no upsampling of either
source. The two sources combined made a single 24 bit master. Dunno if
master is the right word but seeing as it is the only one, I guess it
is.
The question remains the same. Is there enough data in two 16/44
sources that when combined can make a single 24 bit version? Wouldn't
the answer be just plain math?
FYI: I left it at 44.1, so it is a 24/44 recording. In my experience it
is the bit depth that is much more critical than the sample rate. Plus
the mixdown did not allow me to set the sampling frequency only the bit
depth. I would have to "after-the-fact" change from 44 to 88 or 96,
whatever. I did not.
Anyway. I was doing some fun A/B on the boat playing the 24/44 and
switching to the 16/44. The high rez seems to have lost a bit of the
harshness and sounded smoother, easier and rounder. It is hard to find
words to describe sound but in short, I think the 24/44 matrix mix
sounded better than the identical matrix mix at 16/44. Not a night and
day change mind you, a very subtle but also very real difference.