On Fri, 20 Jan 2012 16:26:54 +1100, Trevor wrote:
"Scott Dorsey" wrote in message
Hint: there is no one definition for an "over" because you can in fact
not measure how far over the signal has gone once it has clipped.
Don't need to, more than 3 or so consecutive samples at dFS is usually
considered over
Steady on - your qualification of "usually" corresponds with Scott's
"there is no one definition". Some meters use 3 samples, some use more,
some use less.
If you want to nitpick about whether clipping occurred or not, it isn't
always easy to define. For instance it's quite possible for two
consecutive samples at full scale accurately to represent a peak between
those samples which was over full scale value. Whether that results in
actual distortion on output depends on whether the reconstruction filter
can cope with levels over FS. That's one reason why people aim for
anything between -0.1dBFS and -1dBFS maximum for peak levels.
Of course hundreds of samples in a row at FS is clipped - nobody's
disputing that, but how often is it that bad with commercial CDs?
--
Anahata
-+- http://www.treewind.co.uk
Home: 01638 720444 Mob: 07976 263827