-
Matching re-recorded VO to original audio
I’ve been working on a project that keeps getting revised for about 4 months now. The issue I’ve run into is that the script slightly changed recently, and I have to record some replacement VO for one or two lines (total of about :10 seconds) in a 90 second clip. I recorded the replacement VO with the same mic I used originally, but in a different location (original was outside in a field during summer, while it’s currently winter).
The re-recorded VO doesn’t quite match the sound of the original on-camera audio, and I can’t re-record the on-camera stuff due to it originally being shot outside in the summer. I’ve tried playing with the 3 Band Equalizer to get it close enough to the original that you wouldn’t notice the change, but can’t get it to where I like it. I don’t need the voices to be identical, but pretty close would be great.
Is there another filter I should be using? I could try using STP, but have very little experience with it and would prefer to use FCP7 (but I am not opposed to using STP). Any thoughts?
-
5 Replies
-
It may help to grab a sample of “room tone” from the original footage and blend that with the new audio.
There IS software out there that can sample audio and create a “model” of the sound space that can be applied to new audio and makes your problem go away. Unfortunately, I can’t recall a name at the moment, it may be Izotope, or Adobe’s sound editing program.
-
I believe Soundtrack Pro, which you already have, has this feature, as well.
-
Failing that, it is a LOT more work, but you could try building the track up from short phonemes copied out of other words in the original track. These won’t always have the inflections correct, so it is a hit-and-miss process and time-consuming. But OH SO GRATIFYING if you can pull it off: then you feel like Gene Hackman in “The Conversation” or something. And the client thinks you are a magician.
-
there a few parts to this.
first yoo need to “FILL” out the area you have removed the original words from, using the original bg audio. find some parts of the clip where there is a stretch of clear audio.
obviously it has to match , so looking closer to the area you have cut may be best, and you can expect to join a few bits together, as well.
the result should be a perfectly smooth version of the original where the words just disappear, but the bg audio remains.next your recording does have to mach pretty well.
same mic helps a lot.
there are so many other things involve that sometimes EQ alone wont help.
sometimes a spot of reverb is requires, but maybe not if your original was outside.if your re-recorded sound has a room feel to it, maybe you need to do it again in a more dead environment.
the final thing to resort to is MASKING,
which is adding some extra sound that covers that troublesome area, like a plane going overhead, or a telephone ringing if the scene is inside.or as someone else suggested, try re-buiding the vocals from other words.
nick
-
Thanks everyone – not knowing how to make an audio ‘model,’ I decided to take parts of everyone’s feedback. The first thing I did was to take teh audio in to STP and removed the background noise using Set Noise Print/Remove Noise giving me some pretty clean/clear audio. Then I found the room tone/fill from the original shoot and put that underneather the new, noise printed (can I call it that?) audio so there were no noticeable cuts/changes in quality. I played with the EQ a little – added a tiny amount of bass – just to see if it would help (and it did).
The resulting audio didn’t exactly match – I can tell they were recorded at different times, but no one else who QCd could really tell unless I pointed it out, which is all I was looking for in the first place.
Thanks to everyone for their help!
Reply to this Discussion! Login or Sign Up