[Ardour-Users] click free editing and zero crossings?

Thomas Vecchione seablaede at gmail.com
Thu May 20 04:48:36 PDT 2010

On Thu, May 20, 2010 at 7:18 AM, John Emmas <johne53 at tiscali.co.uk> wrote:

> Most people can visualize how difficult it would be to produce such an edit
> successfully - and yet they continue to believe that editing at a zero
> crossing is somehow easier (which it isn't, for all the reasons stated
> already).

What reasons exactly, I just listed exactly why it IS easier.  If you are
referring to matching of slopes, that is the only reason I could see for it
being listed as not being 'easier' but to be honest, that isn't a matter of
ease as much as accuracy, or rather making sure there is not a drastic
change of slope, but it is much harder to in fact match the sample reference
level at anything beyond zero in most cases due to UI issues in almost every

> As others have stated previously cross-fading, whilst not bullet proof, is
> the best way to produce a click free edit in most cases.
I do agree with this, but I find it interesting people are calling the
zero-crossing a myth, when it is obviously not a myth in itself.  The belief
that by editing at a zero crossing by itself will mean you don't have a
click, is a myth, but the belief that editing at a zero crossing is the
first step to a good clean edit is far from it.  You at least have taken out
one variable, making sure the sample reference values are close to each
other so as to help prevent clicks, by making your edit at a level that has
a clean and easy visual reference.

> The problem is that the true local DC zero at the cut will seldom be the
> same as that after the splice, as Jörn's worst-case graphic illustration
> showed. The level of any harmonic artifacts that emerge will be equal to
> that local DC shift. So, ideally IMO, the method used to splice would choose
> the best local DC zero-crossing points to cut each piece such that the
> resulting local DC shift that results from the splice is as close to zero as
> possible. In Jörn's example that would simply be to choose the next (180
> degree later) zero crossing in the trailing piece.

Ideally you should not have a DC offset in your recordings, if you do that
really needs to be fixed.  As others explained it is the change in slope,
the greater the shift the more chance of an audible click.  That change in
slope can be increased by having a DC offset in one or both recordings that
does not match yes, but the end result is that sudden change in slope in the
reconstructed waveform is the big thing.

Inverting the phase is not a good thing to do as an alternative for this
> purpose, except in some limited situations. You would, for example, want to
> maintain the phase when splicing drum clips, and most percussive sounds.

Once again people are assuming that musical source is the only source that
edits happen to.  But it is far from it.  And even in musical source,
whether or not inverting phase is an option depends on the source material,
inverting phase of percussive hits in complete isolation is not an issue at
all.  It is when you have bleed from other mics that is also to be played
back at identical time, that it can become an issue.

Once again, I do believe a crossfade is the best answer, and this entire
discussion is a bit moot in Ardour as a result.  But people wanted to know
why the zero-crossing belief still existed, and it is because in other
programs that do not do automatic fades at region starts and stops as Ardour
does, doing the edit clean can be at least as viable an option.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ardour.org/pipermail/ardour-users-ardour.org/attachments/20100520/c5423658/attachment-0002.htm>

More information about the Ardour-Users mailing list