[Trevor]

On 13/10/2015 12:25 AM, Scott Dorsey wrote:
geoff wrote:
No CD that I've purchased in the last decade exhibit clipping that I've
noticed, though a few are over-compressed to hell (not many
fortunately). And some I've been prompted to actually check ! Not much
'current' pop though I concede.

The problem is that in the digital world, clipping is whatever you define
it as. I tend to set metering so three consecutive FS samples light the
over light, and so that is clipping.

No, you first have to normalise the gain back to 0dBFS for that to work
since many CD's are first *severely clipped* then normalised to about
-0.3dBFS, so your clip lights will never come on. BUT the flat tops
remain regardless!


Aggressive limiting that flat-tops the signal isn't necessarily clipping,
it's just aggressive limiting.

Rubbish, peak limiting that causes flat tops IS clipping. YOU simply
don't understand what they have done, or the difference between
compression and limiting it seems.


But at what point does limiting turning
into clipping? The point at which three consecutive FS samples appear.

I'd agree with that, BUT it doesn't have to be FS, simply whatever point
they have renormalised to after clipping. That is the bit you seem not
to grasp.



Much of the k man's confusion has to do with the fact that he can't get
the difference between reference levels and loudness..

It seems you don't get the difference between 0dBFS and the chosen
maximum normalised level yourself.


and once you start
adding limiting, it doesn't matter _what_ your reference level is because
you can go infinitely high over it and still not light that red light.

Or if you create square waves from sine waves and renormalise to
anything slightly less than 0dBFS. No red light, and NO "limiter" is
required to do that!


Now... the truth is that I have seen some pop CDs that have as many as
eight consecutive FS samples...

Hell, you haven't looked much, I have seen HUNDREDS of consecutive
samples at maximum NORMALISED level, ie FLAT TOPS!


and I would call that clipping.

Me too.

But,
someone else who decides to calibrate their over light differently
might not, and that is the problem when you start using the word 'clipping'
in the digital world.

NOPE, clipping is clipping, whatever you choose as YOUR final peak level.


Loudness is achieved by extreme compression and/or limiting. If digital
clipping occurs, that is a *technical error* - not inherently part of
the hyper-compression process. And the same 'loudness' could be achieved
without any clipping.

The question is where limiting ends and clipping begins, and where that
exact line is actually is a philosophical question and not a technical one.

NOPE, clipping is the same regardless of the chosen peak level. When a
sine wave looks like a square wave, (ie completely flat tops) it is
clipped, even if the max level is normalised to -60dBFS !!!!!!!!!!!
I can EASILY do that, as can you if you want, and all the BS in the
world about whether it is clipped or peak limited does NOT change the sound.

Trevor.