Monte McGuire wrote:

Hey someone explain this to me. It's an old story & I'm sure it's a basic
digital audio concept that still confuses somewhat. I'm sure it's probably
operator error. I'm not putting the blame on PT - insert a DAW of your
choice,
I'm just conducting my own "tests" with PT.

I open up a session, import about a 30 second 48/24k BCW Guitar Audio File

This is a mono file... right?


Correct.

pan it to the left about half way, listen to it.

...which has been panned through a panpot to a new stereo bus into a new
file...


I dunno, okay. I'm just listening to it in a PT Session with no master fader or
any of that stuff. I import to Track 1 & leave it there just move it to the
left about half.


I bounce to disk Interleaved
Stereo & also Split Mono. The files come back with a different peak gain
level.

Which part is different: the original files and the two new ones, or are
the two new ones different from each other?


The 2 new ones are the same but different from the original.


If the latter, then what
app are you using to determine 'peak level'.

The Audiosuite Gain Plug which has Check Peak on it but...I've done this before
using Nuendo & Sound Forge to Check. This time I just kept it in PT.


If that app treats
interleaved stereo files differently than a collection of mono files,
then you'll have a discrepancy even if the audio is the same for both
stereo files.

How about importing both files into a session, inverting one pair of
channels and summing the result. If it's not zero, then you have a real
problem. If it's the same, then the app that reports 'peak levels'
treats interleaved files differently than multiple mono files.

I understand the summing law thing sorta. Somebody explain to me why the
bounced file is NOT going to sound different than listening to the original
file in PT played in real time. Seems to me the difference in gain is going

to
affect what you're hearing, including your perception of WIDTH. I'm not
asking
about the 3db law & all that. I'm saying the file has physically been
altered,
no?

Sure, the original file is mono and both new files are stereo, created
via the panpot. The only issue would be if the split stereo and
interleaved file had different bits (that are not attributable to
dither). This is the part that I claim should not happen. If it does,
then we need to document and report this ASAP, since it's highly broken
behavior!!

However, having created lots of almost full scale peak limited files
both using stereo interleaved and split stereo output destinations and
having never seen any overload problems, I know the possible 'gain
error' is less than .2dB. In fact, I'd suspect zero gain error, but I
know for a fact that red lights never went off, so there could have been
no more than .2dB of gain applied to a mix or else I'd see the peak
lights go off on downstream gear.

Come to think of it, I've never seen a negative gain either - peak
levels from files usually end up staying at -.2dBFS in my world (because
I set L1/L2 to use that output headroom) and none of the resulting files
I produce have anything but that output headroom. So, the possible gain
error is much less than .1dB, if it exists at all.

All the evidence I've seen points to zero problems that are so gross as
to manifest themselves as a channel gain error. Most of what I have
found to be legitimate complaints are resolution problems that affect
signal quality in a way that would not be easily detectable with a
standard level meter, issues such as how dither is handled internal to a
mixer or plugin, something that has very little error power.


Regards,

Monte McGuire


I guess. I'm just curious why this happens.
To my feeble mind it seems when you do sum in there ....something happens.