Distorting/Transforming audio to create consistent dialogue.

Hi there

I am looking for a way in which to take dozens of pieces of an actor's audio from many films, and merge it all together into one scene, with as much consistency in levels as possible so as the viewer would not notice that all the audio has come segmented from different movies?

Example: If I was doing this with Tom Hanks. I would like to create a scene where Tom says
"My name is Forest, Forest Gump *edit* Houston we have a problem *edit* there's no crying in baseball!"
(Obviously cutting between Forest Gump,Apollo 13 and A League of their own)

The video is irrelevant I just want to make the audio clips sound as much as if they came from the same character/scene/moment as possible.

I understand it will be impossible to completely synch tone/texture/pitch exactly, but any tips to how I might get close or advice on any effects that might aid me in this would be supremely appreciated.

Cheers

It is expensive but iZotope RX5 Advanced can match ambience and EQ between different clips:

www.izotope.com/en/community/blog/tips-t...d-and-pro-tools.html

www.izotope.com/en/community/blog/tips-t...t-sound-sources.html

Another audio element that often requires matching is reverb. I don't think iZotope will do this automatically but it can selectively adjust reverb on different clips which should allow you to get a closer match: ask.audio/articles/removing-reverb-using...-s-de-reverb-plug-in

Generally in keeping dialogue consistent you are looking at few things, the dynamics (loudness), the frequency spectrum (eq), background noise and ambience (room tone).

So it helps if you break it down to the 4 parts. You may not need to worry about the background too much since you seem to be doing a montage from different sources. Denoise it if you need to, I don't use the native denoising though, I think its pretty crude and I personally think they haven't updated the algorithms since Logic was with Emagic. I do youse RX actually, if it warrants it, or NS1, W43 & WNS from Waves.

If a signal is clean enough you may be able to use Match EQ on the neighbouring sources to get them to fall in line. It may need a little additional EQ after that though. Final Cut's EQ is quite fine for this, but it does take a developed ear to be able to do this well enough without screwing up.

Dynamics is really volume matching and maybe putting some soft compression and/or a limiter on it to bring up any levels that maybe too low and to prevent any sources from clipping. There is probably a setting for the Adaptive Limiter so you can use it to catch you peaks and shave off as much as maybe 2-3dB.

If you needed to compress at all You can use FCPX's Compressor, I'd suggest Ratios below 2:1 an attack between 3-15ms and a release somewhere around 100ms, give or take. The threshold should just be low enough to take a few dB off. If some parts are still a bit soft you can try lowering it further, but not too much, you could also try limiting a bit more. The trouble with compressing dynamics is it can really bring up the ambience and the noise which can be counter productive. But its also the reason that if you denoise first, you buy yourself a little room to move.

An easier way to compress the signal would be to use a tool like Vocal Rider. It's quite transparent and works in a very similar way to doing fine level automation on the volume, except it can work much more quickly and possibly more detailed than you realistically could.