Recording - October 2013 - 70

if there is any reverberation in the original
recording, the reverb tail following the last word
of the sentence will be overlaid on top of the
breath. Preserving that reverb tail is very important to natural sound. Cut after the breath, not
before it.
Altering pronounciation
It may be necessary to alter perfectly correct pronounciation in order to maintain the
pacing of speech. The classic example
given by Mortimer Goldberg (see the sidebar) is with this original:
Joe and David were going fishing and he
decided, that is, David decided, to ask Bill
along.
And the editor, in the process of tightening
things up, wants to remove “he decided,
that is.” Now, if the editor removes that
directly, the effect is very unnatural because
of the two consecutive D sounds—the composited phrase “and David” doesn’t sound
right; if a person spoke it directly they would
tend to blend the two Ds together. Therefore
the editor has to do just that, take a little
wedge of sound out after the “an...” part in
order to preserve the pacing and avoid the
effect of what Goldberg calls a Jump Cut.
In conversation a person will often correct
an error immediately, giving you an opportunity to cut that correct word right in where it
belongs. Again, Goldberg uses the example:
“After his application was received, he
was given an applointment— appointment— for an interview.”
Here, he points out that the corrected
word is fine, but the problem is with the transition between “given an” and “appointment”, because the reverb from the “an” will
fall into the beginning of the word “appointment”; so cut inside the word if at all possible. And, cutting at the point of the P makes
it very easy to find the edit point solidly
because the plosive is easy to identify,
whether you’re working on tape or in a
DAW.
Outtakes
One thing that Goldberg talks about is the
need to keep outtakes around, because they
give you material where you can retrieve
70

RECORDING October 2013

just one word (or even one vowel!) with the correct inflection, to paste over words
that may not be inflected properly after a repair cut is made. His example is:
“And so the building was scheduled for demolition, because as I said before,
the property was sold to the developer.”
Cutting out the last clauses and ending the sentence with demolition would
tighten this up considerably, but because the word “demolition” is in the middle of the sentence, with rising inflection, it will seem abrupt and out of place
as the end of the sentence. Also, of course, the reverb tail will have to be cut
off of it. So, this is when you root around in the outtakes to find a word ending in “-tion” in order to replace the ending of the word.
It’s very difficult to use a heavily inflected word except in the original sentence
it came from, and likewise it’s hard to find a word to paste into a heavily inflected sentence to make it sound smooth and natural. One problem: if people are
interrupted, they may return to their speaking with a very different inflection, and that
can make it very hard to cut the interruption out and make the statements follow one
another smoothly. The same goes for removing sudden applause.
Sometimes the best thing to do is to leave the interruption in, or replace it with a
different and shorter interruption to tighten things up a bit. The safest solution is to
avoid cutting after a highly inflected word, unless the next word can be voiced naturally. If it cannot, inserting an interruption like a cough or even an “um” can help
smooth things along.
The wonderful thing about the digital world is that you can cut and paste repeatedly from the material you’re working on. In the tape world, once you cut the tape,
it’s cut.... but in the digital world you can make copies of one sound over and over
again, so there is no reason to root around in the outtakes bin to make these fixes.
Other things a speech editor might do
As mentioned above, the speech editor might actually insert “er” or “um” sounds
in order to mask a bad edit point, one where either the reverb tail was damaged
or there is a change in background noise. Sometimes adding in a loud background
noise can help mask an abrupt noise level or speech quality change. It’s worse
when the background noise is rhythmic, such as recordings made on a train,
because there the background noise has to remain a constant beat with any edits
or the editing will become obvious.
In a radio situation, the fellow doing speech editing might also be taking recordings made at different times and places and making them sound the same. He could
be using equalization to do this, to make the vocal tones similar. He could also be
adding a common background noise track in order to make them sound like they
were recorded in one place. Often the speech editor is the only person who works
with the material at all before it is aired, and so a lot of responsibility can fall on
him or her.
Edit vocal tracks
Editing music vocals is a lot like editing dialogue, except that you have to keep
the beat and maintain the pitch. You may have to keep it precisely, you may not
have to keep it accurately, you may have to be ahead of it or behind it at specific
times in the song. Pay attention to what is going on and look at where the singer is
with respect to the beat throughout, and try to preserve those variations.
But, keep in mind the exact same things said above about dialogue editing. If you
cut a word out of one place and paste it into the other, do the breaths seem natural? Is the reverb tail natural so that one word joins into another? Sometimes you can
get away with damaged reverb tails in vocal tracks once the instruments are
brought in. And, of course, all the comments above about inflections in speech go
a thousandfold for singing.
With vocals we mostly have multiple complete takes, and we “comp” the takes
together, switching from one to the other at various points in the song. Sometimes,
though, you can’t get more than one complete take, for example in a live recording. Sometimes you get a singer who just can’t hit that one note a second time no
matter what. You can fix those, and you can fix them in a way that sounds natural,
but you have to do it while listening to just the vocal track in isolation (or at most
the vocal track and a drum submix) so you can concentrate on the vocals. All of our
speech editing techniques, though, are possible.
Edit drums
The ability to edit drum tracks before overdubbing the rest of the band is a powerful tool, and like all powerful tools, it can be abused.
First thing, if your DAW has a “snap to grid” function, pretend it doesn’t. I’m serious, that is a thing of pure evil—and in this audio engineer’s opinion, that goes double for the new time-stretch/squeeze functions that are appearing all over the place.

Recording - October 2013

Table of Contents for the Digital Edition of Recording - October 2013