Cancelling background sounds for a movie, For a student film taking place outdoors |
![]() ![]() |
Cancelling background sounds for a movie, For a student film taking place outdoors |
Mar 9 2010, 06:49
Post
#1
|
|
![]() Group: Members Posts: 494 Joined: 16-April 03 From: Pittsburgh, PA Member No.: 5997 |
So I was helping a friend get set up to shoot his film. I've helped him before in outdoor situations, and one thing that brings film to a screeching halt is when you have background noise that you didn't want, like a siren, tires squealing, etc. Typically this is pretty far off.
My friend told me that his audio recording unit has two microphone units and can record in stereo. He generally records his actors in mono with one person on boom, since he only has one boom mic that he can borrow. I was thinking of lending one of my mics that is non-boom, recording the background as a separate stereo, then eliminating the background sound in post-processing by inverting my mic signal and combining with the boom signal. Problems that I see: My mic and his mic are nowhere near the same and it'll be impossible to level match the background in the field (it may be possible in post-processing, but I've never tried it). Worst case, I drop my mic signal post processing, and it'll be just the same as if it never happened. I personally think it's worth giving it a try, but I want to get opinions of people who've done this sort of thing before. We're hoping to finish filming this weekend WITH sound, so getting feedback on this soon is critical. Thanks all. |
|
|
|
Mar 9 2010, 08:58
Post
#2
|
|
|
Group: Members Posts: 126 Joined: 16-August 03 Member No.: 8386 |
This is near impossible. To get complete cancellation from two signals they must be 180° out of phase. When two microphones are spaced apart you introduce all kinds of crazy phase relationships between the two signals. This will prevent your idea from working.
I think many films re-record the dialog in a studio after the actual filming, and sync up in post production. The actor watches thier performance while they re-record the speech. |
|
|
|
Mar 9 2010, 10:46
Post
#3
|
|
![]() Group: Members Posts: 39 Joined: 25-February 07 Member No.: 40962 |
yep, that's next to impossible. best thing you can do is operate the boom as well and as near to the actors as you can, and, very important, but often forgotten: ask them to speak quite loud!
then, in the post, you can try out spectral editing (i think adobe audition has it, not sure about soundbooth). algorithmix renovator could do the job, but this is a veeeery expensive pice of software. This post has been edited by Polouess: Mar 9 2010, 10:46 |
|
|
|
Mar 9 2010, 14:30
Post
#4
|
|
|
Group: Members Posts: 128 Joined: 12-December 07 From: Cleveland, OH Member No.: 49500 |
Also record some background noise (from the same mic position) during quite moments. Then you can splice bits of this noise in to replace some of the problem noise.
-------------------- Kevin
|
|
|
|
Mar 9 2010, 14:34
Post
#5
|
|
![]() Group: Members Posts: 494 Joined: 16-April 03 From: Pittsburgh, PA Member No.: 5997 |
Thanks for the responses, Iain and Polouess. Knowing this in advance will save the pain of carrying a mic and a stand to the location, which given it's relative remoteness, is a minor pain to haul equipment to.
Oh well. |
|
|
|
Mar 9 2010, 15:47
Post
#6
|
|
|
Group: Members Posts: 581 Joined: 17-August 09 Member No.: 72373 |
I'll point out that what you're proposing is in fact how a directional microphone works. There are two or more microphone elements or entry ports for sound in a directional mic. Off-axis "Ambient" sound coming in from the back is removed by combining it with an inverse of what's coming in the front. Boom mics are usually quite directional.
|
|
|
|
Mar 9 2010, 21:03
Post
#7
|
|
|
Group: Members Posts: 2116 Joined: 24-August 07 From: Silicon Valley Member No.: 46454 |
I think many films re-record the dialog in a studio after the actual filming, and sync up in post production. The actor watches their performance while they re-record the speech. Right! Almost all dialog is re-recoded as "voice over" in the studio, and sound effects are re-created. This generally isn't practical with amateur productions. In the "old days" many movies were shot in a sound stage. Many TV shows are still shot in a studio, but most modern movies are now shot on location.Even with access to the best-available software, professional recording is still done in soundproof studios on good, low-noise, equipment with professional talent. The exception is news gathering where you often do get noise! Noise reduction software can work great when you have a constant low-level background noise (like tape hiss). If the noise is bad, sometimes you can make some improvement, but often the cure is worse than the disease... If you're going to record live on-location, the best you can do is get the microphone as close to the speaker/actor as possible. Use a directional microphone, and if outdoors use a wind sock. And if you get a bad noise interruption... "Take 2". |
|
|
|
Mar 9 2010, 21:41
Post
#8
|
|
|
Group: Members Posts: 675 Joined: 23-February 05 Member No.: 20097 |
If you're going to record live on-location, the best you can do is get the microphone as close to the speaker/actor as possible. Use a directional microphone, and if outdoors use a wind sock. ...or carefully-placed - and taped - lavs (lavalier microphones). If good-quality wireless systems are prohibitively expensive, I've heard of (though never tried myself) hiding a small recording device of some kind - for instance, a flash recorder - on the talent and recording the lav straight into that (most lavs are designed to run on +5VDC, and many small recorders can provide that...you may have to make up your own cable, though). As long as you provide some kind of sync sound at the start of each take (hand clap, actual slate clapper, etc.), you should be able to line things up in post, and I've heard of even consumer digital recording devices (which are freerun - not able to lock to an external sync signal) being stable enough to stay in sync for up to three minutes or so...although I suspect that's pushing it. However, since most takes don't go nearly that long (unless you're trying to recreate Hitchcock's "Rope" Obviously, there's the potential for this to become quite laborious in post, depending on the number and lengths of the takes, but if they're really strapped for cash, you/they probably have far more time than money, anyway. -------------------- "Not sure what the question is, but the answer is probably no."
|
|
|
|
Mar 10 2010, 02:25
Post
#9
|
|
|
Group: Members Posts: 19 Joined: 22-August 09 Member No.: 72540 |
The way the pros do it is ADR, additional dialogue recording. You bring in your actors to the studio and they watch the tape and redo their lines over it. Cuts out the background noise. It's also used when the director likes the take but feel the voice needs a different tone to it or something.
Sometimes they don't and you can often hear it. In the movie Serenity there is a scene between Nathan Fillion and Ron Glass at night, outdoors. Fillion's lines were ADR'd but Glass's were not. As such the noise level rises audibly when Glass is speaking. It's not extreme, but you can hear it. ADR may be a pain, but it is how you do it when the sound on the day has noise. If you want some of the noise from the scene you also do some recording of the general background noise out there (with nobody talking) and the sound editor can then lay that down at a low level under the dialogue as they see fit. In terms of sync issues with audio and video recorded on the scene you can avoid problems so long as all your devices do timecode. You either sync the timecode (hard to do on consumer devices) or simply have something to allow you to sync it later. A slate can do a good job of that as it produces a visual and audio cue. Now as for audio devices if you can't get an audio device that does timecode, just feed the audio signal to a camera and record it there instead. All DV cams do timecode on the tape. Import all your data to your NLE and then sync it up. There will be no noticeable drift as it'll stay locked to the timecode to a fair degree of accuracy. With ADR the sync issues are all with the actor. They just need to be able to sync up their voice with the earlier take. You can try having them watch it, or watch and listen (with headphones). Just record the ADR take straight to the NLE and it'll all be good. |
|
|
|
![]() ![]() |
|
Lo-Fi Version | Time is now: 23rd May 2013 - 12:28 |