1

I'm hitting a brick wall on a key bit of sound design on a show I'm working on. It's set in a magical fantasy world and the moment I'm having trouble with is when a mind control spell is cast on one of the characters. The script calls for having both the voice of the magician and the characters audible. So I've already got two dialogue tracks - one of the character and one of the magician in sync, but the tricky part is combining those voices in a believable way where each voice is recognisable, but it's clear that the magician is exerting his will here. I'm having difficulty with both coming up with creative solutions that work and technically achieving this effect.

A few of my early brainstorms and tests included thinking along the lines of David Bowie's Space Oddity, where Bowie sings and octave higher and and double tracks and octave lower - so the result is this beautiful, eerie harmony. But with two voices that idea goes nowhere. Playing around with the usual EQs, reverbs, delays just muddies the clarity of the dialogue and it still sounds like two voices stacked on top of one another. A bit of reverse pre-delay is nice, but that's the only element I've liked so far.

I should also add that the magician ADR was recorded to exactly match the delivery of the possessed character. Perhaps this was a mistake and I should have tried to have gotten some variations, say having the magician whisper his lines, like he's whispering a suggestion in their ear...hmmmm. Some second opinions would be most excellent. Thanks in advance!

1

There are no wrong answers here… Above all, CLARITY is key. If people can't understand the words you're sunk before you leave the harbor. So whatever you do - make sure you protect the dialog.

Off the top of my head…

Approach #1: Use EQ. High-pass the character's voice around 3k (or higher, maybe 8k) so you simulate a whisper (similar to Utopia's idea). Low-pass the magician's voice around 500Hz or so to capture the body. Perhaps the two EQ curves have a good amount of crossover so you don't miss the mids? In addition, try adding some synthesized low end to the magician (ie. Lowender, LoAir) - this may heighten the menace.

Approach #2: Use Pitch. Keep the characters voice as it is but duck it in level by a fair amount. Process the magician's voice through something like UltraPitch 6-voices to give it an otherworldly feel (similar to Andy Lewis's idea). Be careful, it can start sounding faked if you push it too much.

Approach #3: Use your fader. Try dropping the character's voice altogether and just use the magician's voice. And make it front and center, no mistaking that it's him. Who knows, that could be completely wrong for your project but could be worth a shot. Sometimes, it's better to just let it jump right out and bite you than to hide everything behind gauzy reverbs and delays and such.

Approach #4: Use convolution. I have no idea how this may work, but try using the magician's dialog as an IR and process the character's speech through it. You'd probably have to do it on a word-by-word basis, but it could be incredibly cool! (Or an utter failure).

Would love to hear what you come up with - please consider posting your results so we can all benefit from your experience.

  • +1 Intelligibility/clarity is key above all else. Nice idea too! I've tryed with convolution sometimes, interesting to play with, but haven't done it enough to make a decision on how effective it is for me. – Stavrosound Jan 21 '12 at 7:22
  • Certainly, the episode just went up today: escapistmagazine.com/videos/view/tales-from-the-table/… In the end I went with a layer of the clean character dialogue, a layer of the magician with reverse pre-reverb and a track of the character dialogue with the AVOX Throat plugin. I found that using that pitch shift the character closer towards the magicians voice really helped bind things together. I think I could have pushed this further and plussed it more, but I'm happy with it. Though I think I did better dialogue design work on the robot in ep 16. – SqueakyFish Feb 4 '12 at 1:31
  • 1
    I ended up doing the sound on most of the later epsidoes, but the whole series is pretty excellent if people are at all interested: escapistmagazine.com/videos/view/tales-from-the-table All pride and promotion aside though, thanks heaps for everyone's input on this. I ended up taking a little bit of advice from each post. This kind of community is really valuable and awesome. – SqueakyFish Feb 4 '12 at 1:33
0

The reversed pre-delay was the first thing that sprung to mind, but similar to @Utopia's reverb idea, you may want to try using only one character's voice for the wet and the other for the dry signal.

Depending on how the picture is cut you may try nudging the wizard's voice a little earlier (or the possessed character later) so that it's as though the wizard is leading the character and the character mimics the wizard.

But if you're set on having them both play together, I would imagine if you don't use something akin to Vocalign or judicious use of Elastic Audio you're going to end up with doubled syllables & slurred fricatives which would muddle up the intelligibility of the lines even before you start futzing with EQ.

3

Your reference here is the Two Towers (LOTR) special features where they discuss how they created the doubling effect between Sauramon and Gandalf when Gandalf reappears as Gandalf the White. This should give you some inspiration.

Also, if the wizard is supposed to have the dominant voice, I would re-record the other characters lines whispered so the wizard's voice has more cut. Whispering takes all of the low-end out of a voice and this will help to separate and not muddy up the sound, but add a nice timbre.

If this is not possible, high-pass the actor's voice at about 1K or higher and send it through a reverb unit of your choice (perhaps a room or shorter reverb, not something longer than 3 seconds) and ONLY mix in the return of the reverb with the wizard's voice, and have NO reverb on the wizard's voice, so in effect, the reverb return is from the actor's voice and the main signal is from the wizard's.

I also find AVOX Throat to be a good plugin for this type of scenario.

In the end there is no right or wrong way. Just as long as you provide what the director envisioned!

Good luck!

  • I like the reverb idea, next time I have an opportunity to use it I will definitely try that. – Steve Urban Jan 20 '12 at 18:17
  • @Steve No problem! The idea just popped into my head. Sometimes these morph mixes can be tricky and it just takes trial and error and bouncing ideas off other people. I'm glad you like the idea :) – Utopia Jan 20 '12 at 18:57
  • Love those Antares plugins, thanks for the tip! Gunna have to keep hose in my back pocket to invest in when a show requires such dialogue treatments ;) – Stavrosound Jan 21 '12 at 2:23
0

Tried using some kind of harmoniser effect on the wizard vocal? Maybe that'll be worth trying. Or double tracking it but changing the formants of the voice with something like Melodyne type manipulation? Sounds like fun!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.