Audio in film is very important, as it sets the mood and of course in the case of dialogue, helps the story progress. For this reason, it is of the utmost importance that all audio is clean and sounds the way it is supposed to. The job of a sound crew is to make all sounds (background music, Foley and dialogue) look as natural and seamless as possible, like it was originally there, rather than put in later or edited. In other words, in this industry the best compliment is no compliment at all!
ADR (Automated Dialogue Replacement)
ADR or Automated Dialogue Replacement is often necessary in films for endless reasons. Firstly, sometimes the dialogue recorded on set is sometimes not usable due to unforeseeable noises such as a plane going overhead or traffic noise. There are ways to reduce this noise (for more information read on), however sometimes it is best to just re-record the dialogue in a more controlled environment. Another reason ADR is necessary is in Animations or Games, or a narration of the actions, when the sound was obviously never there to begin with. Less commonly, sometimes a director may want to “beef up (Purcell, 2013)” a performance or change the way the line was read. Non-verbal cues such as a long sigh or a quick breath in between sentences are also sometimes recorded after the fact. The list of reasons for ADR goes on…
Something to remember while recording ADR, is (when reading from a script) to always put the script on the opposite side of the mouth to where the actor speaks from (everyone speaks more from one side of the mouth to the other). This is so that if they turn their head to look at the script, the sound is still going directly into the mic. Also take into consideration the room sound/ambience. This can sometimes be a good thing, depending on where the scene is supposed to be. For example if the scene is set in a coffee shop, some ambience is needed and later in the production, more room noise will most likely be layered on top to add to the realism. Another job of an ADR supervisor is to ensure the actor has the same amount of ‘feel’ as when they are recording on camera, otherwise the lines can sound flat or out.
Before modern technology, ADR was actually called ‘Looping.’ Purcell (2013) explains that
“[to rerecord dialogue, sound editors used to have to create a physical loop of film,] which contained sound and picture of a line to be replaced. A film loop was prepared with beeps and visual clues to cue the actor, and because it was a loop, it could repeat continuously. The actor would repeatedly hear and see a line, and then, when he was ready, recording commenced. Each time the actor heard his line, he would immediately repeat the text, in a process that continued over and over until everyone was happy. Then up went the next loop, and so on.”
These days, we have technology that allows us to loop record, with the scene usually in front of the actor, so that they can see how the line should sync up with their mouth movements. After this process, the ADR supervisor will choose the best take out of all the loops, taking into consideration the “sync, delivery, pitch, and attitude (Purcell, 2013)” of the line.
ADR should ideally be recorded as late as possible without getting too close to deadline. Therefore, around just after halfway through the dialogue editing process. However, it is obviously quite difficult to know exactly when to do this, as some lines you thought needed to be rerecorded could have easily been fixed with a new take, and other problematic dialogue may have arisen since you last made the ADR list. The key within the timeline is to be reasonable and flexible.
In terms of equipment and studios, the ADR supervisor will be informed of what microphone for the other lines in scene they are rerecording, so that they can choose a similar sounding mic for the ADR. It is also better if they know where the scenes were shot in advance, so that they can set up mics in appropriate spots in the studios (eg. spots with reverb vs ‘dry’ spots), which makes the session more efficient.
As you can see there are many different components which go into the art of ADR, there is tonnes more research out there, from knowing how to deal with actors to what needs to go on an ADR call sheet. This is just a basic overview of the processes involved. Now, if you want to try to avoid doing ADR (which generally you want to do as much as possible), read ahead to find out more about noise reduction.
Noise reduction is all about trying to preserve the quality, tone and volume of the good parts of your audio as much as possible while reducing the bad sound. This sounds a lot easier than it actually is. I am going to be specifically talking about Izotope’s RX Audio Editor.
There are three main types of ‘noise’ that can be found in an location take: tonal (something that moves up and down the frequency spectrum: ie hum from a machine such as fridge or air con, or a siren in the background), broad band (scattered all over the frequency spectrum, a jet flying overhead is considered transient broad band noise) and intermittent noise problems (ie. clicks, pops, bad cable or mic connection, lavalier mic bumps, boom pole bumps, birds)(Hirsch).
The first thing I’ll talk about is the deconstruct feature, which is unfortunately only available in RX Advanced, but is super helpful nonetheless. This feature can reduce both tonal noise and broad band noise. All you have to do is click on ‘Deconstruct,’ then set the gain level of whichever noise you’d like to bring down. You can also set the balance slider at the bottom in order to either focus on tonal or noisy (broad band) balance. Next step is to select all of the tonal issues, which can be seen on the frequency spectrum of your track. Then all you do is hit ‘process’ and wallah! No more noise. See the example in the video below:
It is important to note that you can hear an obvious pitch change in the audio, which makes it sound more muffled because it takes out the high-end, as the tonal noise was in the high-end. It would be interesting to see whether the voice could then be returned to near its original state using EQing, or maybe if they hadn’t turned the gain all the way down but only partly down whether it would make a difference to how many frequencies were taken away.
There is also a tool for removing intermittent noise problems, called the spectral repair module. For the how-to, see here. Again, you’ll notice in the ‘after’ track when the beeping isn’t there, you can here the gaps in the background noise. This could be fixed or made less obvious by then reducing the background noise, or maybe layering on some ambient noise.
As you can see, the RX Audio Editor is an awesome tool to be able to reduce noise, and I especially love that you are able to see the frequency spectrum and quite literally remove particular frequencies. For more instructions on what the RX Audio Editor can do, you can look here, where they have a whole ‘cookbook’ worth of tools to use. Although we all know it’s easier, generally cheaper and better off in the long run to just get the sounds right to begin with, we’re also lucky that we have technology like this to be able to attempt to salvage some of our audio!
As always, thanks for reading!
Hirsch, Scott. Hums And Tonal Noise Repair. 2016. Web. 1 Mar. 2017.
Hirsch, Scott. What You Can And Can’t Fix. 2016. Web. 1 Mar. 2017.
Purcell, J. (2013). Dialogue Editing for Motion Pictures : A Guide to the Invisible Art (2). London, GB: Focal Press. Retrieved from http://www.ebrary.com