On Thu, 22 Nov 2001 23:10:26 -0700, newvideo@amug.org (Bill Davis) wrote:
>In article <3BFDC47E.2020606@flock.org>, Bruce Feist wrote:
>> MSu1049321 wrote:
>>
>> >Select the whole track, then select "normalize".
>> >
>> If I understand correctly, that'll change the overall gain for the track
>> so that it's at an appropriate level, but everything on that track will
>> be boosted (or lowered) the same amount; it doesn't try to figure out
>> which parts of the track are too faint and boost them more than others.
>> So, it is probably too simplistic for my needs, where there's a dialog
>> between two people and one comes across as much louder than the other.
>> Thanks for the suggestion, though.

>Actually no.
>
>The way I've always understood it, the "normalize" function looks at the
>amplitude of the various sections of your audio file and attempts to move
>ALL the waveform peaks into a single maximized range.
>
>(Jay, et al, if I'm off base here, speak up! I've spent a LOT more time on
>the video side of the fence in the past 10 years than on my old love,
>Audio - and back when I was paying my dues in radio, nobody had every even
>HEARD the term normalize!)

I think BF is correct - selecting "normalize" will only
raise the whole clip level to the point where the highest
peak (could be a single high cycle...) is just at clipping.
With multiple clips, this is not a good idea since it will
result in varying average levels for the clips. One can
select an arbitrary gain increase amount and apply it to
all parts for consistency, but this will probably result
in clipping some parts (you can export test WAV files,
look at them for clipping in a good sound editor (I like
the cheap-but-excellent Cool Edit [www.syntrillium.com]),
then go back and manually reduce peak areas (and retest).
You can also use the compressor filter in Premiere to
reduce peaks while increasing overall level averages
(also wise with this to test the results for clipping...).

[...]
>Remember, if you start with crap, normalizing, compressing, auto-gain
>controlling, or applying anything short of a magical spell on it, usually
>just turns it into more consistent crap.

True - but the quality of the crap can be noticeably
increased, especially when careful noise-reduction and
mixing is added to the above...;-)

>Ask anyone in the business and they'll tell you, there is only one BEST
>AND EASIEST path to really good audio. And that is to learn and
>consistently employ good original recording techniques.

True - but in "live" work (one take only...;-), knowing
various "saves" for sound problems is, well, quite
valuable... (I'm working on a 4-camera edit of an event
now - and I forgot to turn on the wireless receivers[!!!].
With a little sound editing effort, you would never know
it - I have good, clean "close-sounding" audio where I
needed it, even without the lavs...).