Sign in with
Sign up | Sign in
Your question

Voices too soft in background, other noises too loud

Last response: in Home Audio
Share
Anonymous
February 20, 2005 12:25:14 AM

Archived from groups: rec.audio.pro (More info?)

I've been using CoolEdit (and now Audition) to improve home tapes for a few
years, but almost always got by on noise reduction and normalize. Now I have
a challenge that's beyond my little knowledge, and would really appreciate
any tips.

I have a wedding reception video, shot with amateur equipment from the back
of the room. So the people who are making toasts or speeches are far from
the camera, their voices muffled, while the audience members near the camera
are heard loudly laughing, applauding, etc.

With this mix, Normalize doesn't reduce the large disparity between the low
voices that we want to hear better, and the loud nearby sounds that aren't
as important.

I've tried increasing the gain on sections with stretches of talking, but
it's so mixed-in with the loud, intermittent audience sounds, that this is
too crude a tool.

Visually, it's a narrow band for the talking, with frequent high bursts for
the noises near the camera. What I want is more like a wide band with lesser
variations, making the voices similar to the audience sounds in volume.

What would be a good technique to make these sounds closer to each other in
loudness? Thanks for any advice.
Anonymous
February 20, 2005 12:25:15 AM

Archived from groups: rec.audio.pro (More info?)

Ted Kerin <tfkerin@noearthlinkspam.net> wrote:
>
>I have a wedding reception video, shot with amateur equipment from the back
>of the room. So the people who are making toasts or speeches are far from
>the camera, their voices muffled, while the audience members near the camera
>are heard loudly laughing, applauding, etc.

Yup. Some idiot used the on-camera mike. Sorry about that. You can't
send it back for a reshoot either. Sounds like it's time to just treat it
as MOS footage and edit it to a nice musical track.

>With this mix, Normalize doesn't reduce the large disparity between the low
>voices that we want to hear better, and the loud nearby sounds that aren't
>as important.

All Normalize does is to increase the overall level so that the loudest thing
is at 0 dbFS. It does not do anything that turning the volume control up
by hand will do.

A limiter or compression tool will bring the low sections up and drop the
loud sections down, but you'll notice that the noise floor also comes up
when the gain is cranked up, and that all the nasty room echos that you
get from a shitty camcorder microphone pointed in the wrong direction get
exaggerated. Also, if it's a typical camcorder, you'll also have lots of
motor noise, and that will pump up and down with tha gain as well.

>I've tried increasing the gain on sections with stretches of talking, but
>it's so mixed-in with the loud, intermittent audience sounds, that this is
>too crude a tool.

A compressor or limiter will basically do this automatically, and it can
adjust the gain faster than you can do it by hand. But I suspect you will
not be pleased with the overall results because it cannot do anything about
the audience noise itself or the nasty room reverb.

>What would be a good technique to make these sounds closer to each other in
>loudness? Thanks for any advice.

I'd use a fast limiter, and crank the threshold up so that the intermittent
audience sounds are just slightly distorted. That will bring the low stuff
up as high as it'll get. Then I'd throw most of the soundtrack away and
replace it with stock music. Note that adding music at a low level can also
help hide some of the room nastiness and camera noise.

Tell the man to hire a sound guy next time.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
February 20, 2005 7:37:35 PM

Archived from groups: rec.audio.pro (More info?)

Thanks, Scott! Aggressive use of the limiter in Dynamic Processing got me
exactly the result you desribed -- for better, and for not-so-good. It's
still a nice improvement, to be able to hear the speeches without having the
other sounds at ear-splitting levels.
Related resources
!