?signal processing to match different vocalist with pre-recorded sound clip?

number six

New member
Hello,

I wanted to make Dave, from 2001:Space Odyssey, say, "open the SEED VAULT" door" instead of "open the POD BAY door." Accordingly, (this is indirect but what I have to work with) I recorded from "TV" with handheld unit to give mp3, then got this as .wav into pro tools. Once there, added "Seed vault" and added a strip of background noise, and the results is the mp3 bounce-to-disk attached.

Neglecting the TV-to-handheld issues and some distortion on the HAL 9000; and realizing that the patched-in voice is not, in fact, Dave; thoughts on what techniques could be done to make it more of a match?

So far I've tried, for various "takes" at attempted different pitches/intonations/durations of speaking:
eq, slow down, speed up, change pitch. View attachment 2018.edit.seedvault.mp3
 
First, what are you planning to use this for? If it’s just a personal thing, I don’t see anything ethically wrong with using a YouTube to MP3 converter for that short scene so that you can have cleaner audio.

Second, there’s very little you can do as far as using Dave’s voice. It’s going to sound off and funny and it’s always gonna sound off and funny, no matter what.

The best thing, IMO, is to record yourself as Dave. Emulate his voice as much as possible and then match it up with HAL.
 
+1 to the above.
Direct capture of the source is going to help a lot.

Other things that might help; It sounds like you're quite close to the microphone so your voice is full and clear.
It would make more sense to be further away and less full sounding. Try rolling off the extreme highs and lows on your voice too, to match the fidelity of the day.

Another thing that might help, in terms of smoothing transitions a bit, is capturing some of that background ambience and noise without dialogue, then using it along side your re-record.
Try to make that ambience gradually fade in and out before and after your new dialogue rather than just snapping on and off.

You could actually do the same for dialogue taken from the show; Anything to avoid sudden changes.

For example, if I wanted to replace some dialogue in a star trek scene on the bridge, I'd be making up or getting a relevant star-ship ambience track, then using that to sew my parts in.
If I can have that ambience gradually fade in over a few seconds before I speak, in addition to the existing ambience on the original 'backing track', it will take the edge off.

People will notice if all the ambient noises suddenly disappear but they probably won't notice if they're slightly louder with certain parts of dialogue, fading in and out over 5 seconds, or whatever.
 
Back
Top