In article ,
Justin Ulysses Morse wrote:
agent86 wrote:

If the argument were that the mathematical calculations within the digital
summing bus were themselves inherently flawed, I could buy that. (Or I
could at least consider it a point worthy of discussion.)

This is the argument for analog summing. We know that digital summing
should theoretically be perfect but so far all of the claims that
real-world software is in fact perfect seem to grossly underestimate
some software companies' ability to screw it up. That's my current
view of the situation, anyway. Unfortunately this is one of the most
difficult areas to perform conclusive, objective tests because there
are just so many variables.

My take on it is that most folks have crappy DACs that don't sound
good with complex program material. Basically, they don't sound good
playing back entire mixes. If you split your program out so that each
channel only handles a single instrument, bad cross modulation stuff
happens less.

The problem remains that your mix is now analog and must be converted
again, and people don't seem to want to monitor with a DAC on the
other side of that conversion. So, it's not a fair comparison at all.

IMHO, you can avoid all of these problems by buying a good DAC.

But that would also raise the question of why one would expect the
mathematical calculations within the DA & AD converters to be any
less flawed.

The DACs themselves simply aren't asked to perform this particular kind
of calculation.

Yeah... they get to do much more complicated ones... upsampling a wide
word at a low sample rate to a narrow word at 64x to 256x the original
sample rate. Lotsa multiply-accumulates, and we haven't accounted for
the calculations inside of the A/D's decimator, needed to get back
into the digital realm.


Regards,

Monte McGuire