Removing camera sound / better noise reduction

[deinkarrieretrainer]

Hi,

I’m recording a video course but unfortunately, the microphone and the camcorder don’t like each other pretty much. There’s always a background noise like in the first and last 5 seconds of this example (downloadable .mp3 in my dropbox with the noise and some words):

https://www.dropbox.com/s/nar77mpdef8amlt/S2310019_Audio.mp3?dl=1

I used the noise reduction (Reduction -12 to -20 dB, Sensitivity 6 and Frequency smoothing 3) but the result isn’t very satisfying, sounds like underwater. Audacity version 2.3.2.

Do you have any other ideas how I could repair/enhance the audio without having to take new videos?

Greetings, Benny

[DVDdoug]

without having to take new videos?

In Hollywood, they just re-record the audio while lip-syncing to the video. All on-location movie dialog is re-recorded in the studio.

There’s only so-much you can do with processing… That’s why pros still record in soundproof studios with good mic position, etc.

Noise reduction works best when you have a constant low-level background noise (like a little background hum or tape hiss). It works best when you don’t really need it…

[deinkarrieretrainer]

In Hollywood, they just re-record the audio while lip-syncing to the video. All on-location movie dialog is re-recorded in the studio.
There’s only so-much you can do with processing… That’s why pros still record in soundproof studios with good mic position, etc.

Thanks for your answer. They are round-about 18 videos… re-recording with lip-syncing would be more time-consuming than accepting an underwater sound or just reducing the cam sound slightly with noise reduction
But I guess I’ll try for fun whether lip-syncing is possible. First step to Hollywood

[kozikowski]

First step to Hollywood

Bring strong coffee. It’s called “Looping” where a short segment of video (film) is presented over and over and the actor presents the lines repeatedly getting closer and closer to picture sync each pass.

Next loop.

It’s not the most fun you can have.

Koz

[kozikowski]

first and last 5 seconds of this example

Your sound recorder has automatic gain control—AGC. The good news is it will always record something and it doesn’t matter what your voice volume is. The bad news is it can’t tell the difference between voice and background noise.

It produces Noise Pumping.

I don’t understand how you got a video camera to make that much noise, unless the noise is really coming from your room and not the camera. Is that computer fan noise?

There is one possibility. Because of the way AGC works, the noise is never there at the same time as the voice. You might be able to use the Punch tools to “cover up” the beginning and ending noises without affecting the rest of the voice.

Punch Copy/Paste

A pair of plug-ins for simplifying “punch in” editing. Unlike the standard paste function in Audacity, the “punch” paste replaces the selected audio with just the right amount of pasted audio to match the duration of audio being replaced, thus the speed and timing of the track is not changed by pasting. These plug-ins were developed from a feature request on the forum (https://forum.audacityteam.org/t/alternate-equal-length-paste-command-resolved/28836/1) to streamline the work flow of replaci…

Koz

[deinkarrieretrainer]

There is one possibility. Because of the way AGC works, the noise is never there at the same time as the voice. You might be able to use the Punch tools to “cover up” the beginning and ending noises without affecting the rest of the voice.

Thanks, I’ll try this and get back to the forum afterwards. That was my impression, too. One can only hear the camera when I’m not talking.

I don’t understand how you got a video camera to make that much noise, unless the noise is really coming from your room and not the camera. Is that computer fan noise?

It’s definitely the combination camcorder - lapel microphone. I tried it with cable-bound microphones and the noise is a lot less.
In the future, I’ll use a different microphone or record the audio with the computer to lay it over the video afterwards.

[jademan]

If you cannot get Punch to work for you, here is something to try:

1. Duplicate the track (Ctrl+D)
2. Create a label at 4 seconds and at 5 seconds. This will allow close alignment via yellow guide-lines.
3. On the original track, Silence the audio from 0 to 4 seconds. (Select the audio, then Edit > Remove Special > Silence Audio.
4. On the original track, Fade in the audio from 4 to 5 seconds. (Effects > Fade In), then repeat the fade in a number of times.
5. On the duplicate track, Effect > Amplify by, say -20dB, on the entire track.
6. Select the audio from 4-5 seconds and Fade Out.
7. Select the audio from 5-end and Silence Audio.
8. Select All tracks, then Tracks > Mix > Mix and Render

I hope this helps.

[deinkarrieretrainer]

If you cannot get Punch to work for you, here is something to try:
(…)
I hope this helps.

I tried, but I don’t know if I did it correctly, but didn’t worked for my purpose. The first seconds were silenced, but there still were the noisy parts later in the track.

You might be able to use the Punch tools to “cover up” the beginning and ending noises without affecting the rest of the voice.

I tried the Punch Copy/Paste, but since English isn’t my mother tongue, it was a little bit hard to understand how it works. I guess I understood now, but it didn’t work the way I intended.


Thanks to all of you for your help!
My “solution”: I re-tried the noise reduction. From other videos, I copied ~15 seconds of the camera noise to get the reduction a better sample. After trying different settings, I finally found one that isn’t too disturbing and you still hear some camera sound, but I’ll just leave it like that now.

Greetings, Benny