Tips for successful ADR? Please weigh in...

Matthew Bennett

Well-known member
Anyone have any good tips for a successful ADR session?

Heres what I've gleaned so far from my research:

1. Be organized, Break everything down to as small a bit of dialogue as possible. You should only work with a sentence or so at a time. It's simply too difficult to keep sync on a long passage of dialogue.

2. It's more helpful to have the performer look at the sound directly (perhaps just at the waveform in an audio editor?) than their visual performance. They are trying to mimic the SOUND exactly, not their physical performance.

3. Record the audio in a neutral, echo and reflection free environment with good mic technique, (close shotgun, proper lav technique, etc) and add the reverb, echo, BG noises back in later in the mix.

4. Be patient and keep working until you get a recording which is as perfect a mimic as possible of the timing and emotion of the original performance.


Anything else?
 
When doing outside scenes listen to the pitch of the voice you are trying to mimic. When we are outside there are less percieved low frequencies so we naturally raise the pitch of our voices to be heard over wind etc. this can be replecated in a vocal booth, but has to be pointed out to the actor.

over time you develop a good perception of what repetitions will sync with a bit of playing around, and what won't. this involves listening to the breaks between the words ensuring they are as before. the words themselves can be sped up/slowed down, but very often the breaks between the words cant be manipulated as easily.

record first and try and sync the most dubious ones when the actor is still around. after you develop this perception I just mentioned you can be confident that you can sync the majority of the lines another time. don't waste time by syncing everything when the actor is there. call backs are a nightmare though so you have to be on the ball.

I have only ever recorded ADR with sound only in the headphones. occasionally I get the actor to come through to look at the screen if there is a facial expression that needs to consider, that I can't explain through talkback, but mostly its just sound.

I use protools so the way I do it is to loop the line (or half line or what ever) while recording, and hover over the track mute button, so the actor can hear the line as many times as necessary, then when the track is muted, thats their cue to act. that way you control if you think they need to hear it more, or whatever.
Another way to do it (which is best when there are lots of spaced out short lines) is a run-through : to put all the lines on a big long loop and mute the track after each line, so you will hear the line, then act that line, then the next one etc etc. This run-through method is less accurate, but faster. use your judgement wheather the lines are easy enough to do this.

don't use a studio mic, they are too high quality. use something like a 416 and mic from above, not in front.

you don't need all the actors in the scene to attend at the same time. that just confuses things.
 
Thanks! That's extremely helpful.

Why would say don't use a studio mic because they're too high quality? Does a certain sort of 'lesser' quality help sell the recording as being live?
 
well yes, exactly! when you hear people speaking, they don't sound like a radio broadcaster. you need to record far less detail than that. you get a similar problem when you use a 416 straight on. i started out using a 416 not from above, but straight on and suffered for it for the same reasons, cos I've had to eq so much detail out of the dialogue to make it believable. you can do it, but why not just use the right mic to start with. When practicing, try an outdoor scene for your own interest, with a studio mic, and a 416 boomed above, and sync both, you'll see instantly what i mean.

by the way, you always hear people talk about not using a shotgun inside etc etc, and I suspect that less than half of people who know that, don't know why. The reason you can use a shotgun in a vocal booth is because (hopefully) its been deadened and treated to minimise reflefctions.

the natural sound you hear from a human voice indoors is a combination of direct sound and reflected sound. ..we naturally use indoor reflections to be heard in a room, (and can speak lower volume and pitch compared with outdoors where there are no reflections) so hopefully you can see why a hyper is a better indoor choice already..

Now, the main reason that a shotgun is the wrong coice is a fundamental aspect of shotgun design. the more directional they can make it, there is a trade-off to be considered which directly affects off-axis frequency colourisation. ie, the sound arriving at the sides is not natural.

I suppose if a shotgun could be designed with non-colourised side rejection (more like the schoeps hyper) then we could indeed use them inside.. but my understanding is that there is a clear compromise between directional properties and off-axis frequency colouring properties.

did that make any sense whatsoever? :) i feel like i've just rambled off topic a little.
 
Scooby,
Yes.. thanks for more excellent info.. I could listen to this all day.

I've just recently been exposed to the concept of a shotgun 'photographing' the effect of sound in a room, giving the listener a sense of space along with the picture... thus reinforcing what is seen visually with an audio 'picture'..

This is what I'm doing : ADR for picture shot in a mid-sized apartment, originally with only an omni lav mic... the original track has either too much bass or too much echo, so I'd like to scrap it...
I'm trying to record the ADR in a small, sound blanketed room, with an overhead shotgun, and then sync and try to recreate a proper 'sound picture' in post, based on camera distance, etc..
I figure a nice recording of voice in a small 'dead' chamber is the best thing to start off with, then reflections could be added later to taste?

Am I on the right path.. any more tips or words about the effects of shotguns indoors?
 
Hi Matthew,

Never having done ADR myself, be forewarned... I'm aware that some Indies use their "ghetto studio" (inside automobile - obviously luxury cars work best). Certainly this would require an understanding talent. On that note, never have the talent immediately do take after take - a sure way to cause emotional loss (take breaks between takes). Of course, mix the B-roll ambient room tone video with your ADR. As always, the audio engineer religiously wears headphones and takes careful notes with timecode stamps. In post production, use cuemarks to tag important syllables for subsequent syncronization. Be sure to use fade in/out to avoid noticeable clicks during editing (and don't forget to add the ambient noise from the B roll back in).

BTW, are you aware of the psychological aspects of voice and video? Have you seen the McGurk effect in action? Don't miss this clip (or Flash). Wikipedia has several links on auditory illusions for your entertainment and education.

No, the auditory illusions isn't simply theory. In Joshua Paul's "Digital Video Hacks" (pg 201), he cites the following:

[ The most impressive use of the (McGurk effect) I have personally experienced was on "The Larry Sanders Show." We had a scene where one character, Artie, had his back to the camera and another, Larry, was speaking a line. We needed to drop one of Larry's lines, but we couldn't cut away. So, the editor borrowed one of Artie's lines from a take we weren't going to use and placed it where Larry was speaking. When watching the scene, even though Larry's lips were moving, your focus turned to Artie and your mind ignored the fact Larry was supposed to be talking. ]

BTW - Did you know Mel Gibson's voice wan't used in "Mad Max" and why India has the best ADR experts? Or that folks claimed Marion Brando used to mutter his lines (making them unusable) so he could do ADR with the final cut to ensure his lines had the best emotional continuity? (don't let your talent hear about this one!)

The point is, consider ADR can be far more than meets the eye (ear, brain...)

Finally, here's a link for more ADR reading if you care to signup.

Have fun and keep us posted, Michael
 
Biggest thing I would add is to make sure to get all the breath sounds. These are often overlooked because they are only really noticed when they are missing.

As for whether to use the three beep method, video, or just playback and wing it; it really varies from talent to talent. Some work better with a video reference while it completely screws up other actors. Try one method, if it isn't working well, try a different one.

Good luck
 
Back
Top