On Sun, 10 Dec 2006 10:12:22 +0000 (GMT), Jim Lesurf
wrote:

In article ,
Glenn Richards wrote:
Don Pearce wrote:

It actually looks like the original CD was mastered with more level
compression than the vinyl - which is here. Very odd - may be just
the way the different DAW programmes present the horizontal scale
that causes this, though.

Looks like it's just displayed differently. The spikes at the beginning
of the track hit almost full scale, just like on the CD.

FWIW When I was doing some analysis of CD tracks I tried various ways to
produce statistics for the dynamics. I came to the conclusion that the most
revealing method was to avoid 'rms' or 'mean' powers. The problem being
that these could easily fail to indicate brief peaks which might reach
0dBFS.

I therefore used the method of breaking the track into 100 millsecond
chunks, finding the peak sample magnitudes (left and right) in each chunk,
then doing the statistics on the resulting series of peak values. This
makes it much easier to spot excessive compression leading to clipping.

The programs I used only run under RO, not Mac or Windows. But Keith Howard
of HFN has done a Windows version which should appear on his
audiosignal.co.uk website in due course.

Article on this in the current (cover dated December) issue of HFN. Note
that the comment at the end of the article regarding the programs is
incorrect!

Slainte,

Jim

Jim, Adobe Audition does its stats in a way that appears to be
somewhat similar, although the default window is 50mS rather than 100.
The output data looks like this:

http://81.174.169.10/odds/stats.gif

And a power histogram like this:

http://81.174.169.10/odds/hist.gif

I don't know what the internal maths is like, but I'm sure I could
make a pretty good guess. Would that be the kind of thing you get from
your programme?

d

--
Pearce Consulting
http://www.pearce.uk.com