We knew on set that ADR would likely be necessary, due to exterior disturbances and the positioning of the microphone in some scenes. I was however surprised to find how quiet the majority of the dialogue levels had been recorded at, and how loud some of the room tones were. It sounded more than adequate in the headphones and the metres appeared to be floating in the appropriate -18 to -12db range, however in hindsight I think this may have been where the lines maxed out, and not the average. Having to boom op for many of the scenes as well meant that I couldn't focus solely on my meters and was often concentrating on the angling and position of the boom. I may have also dialed the volume a little too low in anticipation of shouted lines. I boosted the volume of each clip in pro tools, however this also brought up the hiss of the room massively. To combat this I used high and low pass filters in the EQ which also removed unwanted rumbles, and used a second EQ to boost the high middle filter's frequency around 3.56kHz to brighten the vocals out from the tinny acoustics of the room. I also tried to reduce the gain of the room using the X-Noise plug in, however it did not remove it entirely. This was applied as standard to all set dialogue.
Despite this however some lines were simply too quiet still, with many of Caine's (David Chafer) lines being softly spoken and Myra (Jacqueline Collins) not having a sharp voice.
My first ADR session was with Jacqueline. It took some time to configure the studio before we could record as the input/output settings are generally changed around by each user working in there. Once it had finally been set up I enabled the playlist recording function so each looped take would receive its own track. It was an enjoyable experience and it's hard not to feel very professional when sat at the head of the controls. Though familiar with mixing desks from my background in music tech, I had never had to use dubbing before, and as my first time doing this and with Jordan absent my direction was a little poor, and I wasn't too assertive about her performance if it was wrong, however in subsequent sessions I gave firmer instruction if they weren't capturing the right tone in the takes. I added a metronome track to give her a cue which i would then manually silence so as not to distract her from her line. I much prefer the intuitive nature of using a live control surface over a mouse, the mixing feels much more organic and efficient.
I did keep forgetting to hold down the talkback button after pressing it when communicating with her in the booth, however it is poorly soundproofed anyway so she could still understand the gist of what I was saying. The other main learning curve from this experience was to pay attention to every line in the scene, as the tone of Jacqueline's deliverance varied with each line and sounded inconsistent when trying to mix in with the original audio.
In order to integrate the ADR to match the acoustical properties of the set dialogue I used EQ to dampen many of the sharper frequencies and reduce the bassiness from the voice. I also used D-Verb to match the natural echo of the room, using a wet mix of around 24%.
When Philip came in to dub some of Gideon's lines I found his performance would sync better if I removed the original set take entirely from his headphones and left him to ad lib it as he found his own voice distracting.
Having Jordan here to coach his desired performance helped, although I was more confident in my own direction by now as well.
Philip also found it easier if he had someone to direct his performance at, so we sent in Paul the producer to receive his tirade.
David Chafer proved to be the most efficient dubber by far. He picked up the rhythm and was able to emulate his performances very quickly. Unfortunately after wrapping up and moving out of the room to sort through the takes I discovered that none of them were usable due to a static cable interference that went unnoticed in the studio monitors. My assumption is that with the physical mic input on channel 1 being routed to record into the digital channel 7, and with the digital track 1 being the original sound from the video export being on mute as it wasn't used I missed the fact that the fader was likelypushed up to full. This was very frustrating as David's busy schedule left the prospect of getting him in again looking unlikely.
Upon tutor recommendation and our own feelings on the performance the decision to re-voice our lead had been bandied about for some time, but it wasn't until two days before the films required completion that we actually got someone in. It felt very cruel in a way, since the original actor, Jack Simmons, had endured extremely taxing shooting conditions given the film's requirements, and put a lot of effort into his performance. It was still unconvincing for the most part, largely due to his having quite a posh accent which sounded very out of place amidst the rest of the family, and his screams were also a little irritating and made it difficult to sympathise with his plight. A large section of the torture scene in the basement remained unfinished due to my being unwilling to sift through takes of his screams. Ben, an animation student, came along having worked with paul previously, however he had no experience in acting, just a relatively neutral clear voice. He sounded ok when delivering the spoken lines, if not a little flat, however his screams left a lot to be desired, and I would probably say Jack's were more convincing. We did try to coach this by sending Jordan in to physically interact with him but it still did not produce the desired effect.
By this stage we were willing to try anyone just for the screams as we hoped the change in voice wouldn't be too noticeable. With firsthand knowledge of my housemates propensity for shouting I asked him to give it a go whilst we still had the studio, and he was that good we ended up recorded the entire film with his performance as he was far more convincing, and delivered lines with greater consistency in the scene as a whole than Jack did originally. After being so accustomed to hearing Jack's voice the dub did seem odd at first, especially as we were privy to the comic expressions on Joe's face as he recreated an entire performance, however I found the takes improved the scenes I had been struggling to create dramatic tension out of tenfold as the nuance to his delivery made them feel more authentic on the whole despite the artifice of the situation. There was a certain Irony in the fact that after a month of becoming increasingly disillusioned with the project and leaving everyday feeling like no matter what I did the film was not improving it ended up being my housemate who did the re-dub so late into the project. Joe's voice is considerably deeper than Jack looks, so to reconcile this I removed a lot of the low end frequencies and similarly dampened the higher ones and again applied a little reverb.
The session was also enjoyable because of the informal relationship, and Joe's shaking up of lines to break the monotony of hearing the same sentence over and over yielding hilarious results.
We managed to get David back in the morning before the film needed completing. I went in especially early to run tests to ensure nothing was awry this time around, and the process went very smoothly once he came in, finishing in under an hour.
Having worked with a low quality video export for the majority of my edit, once I finally got a better resolution cut I noticed some of the shoot audio we'd used as a guide to the dub did not match the visual take, which is likely attributable to shots not being marked properly. In one instance I'm sure the shoot audio used was the correct take but you could only see Jacqueline's lips move for 'you get some rest', not 'you get some rest now'. I managed to splice the end of 'rest' from another take which I don't think is too noticeable if you're not looking out for it, and we extended a cut to mask another discrepancy. One or two of the original lines do not appear to sync despite definitely being the correct take.
All in all I gained a great deal of experience in looping for this film, with the end result being a bit of a frankenstein job, but I feel the mix of deliveries adds to the quaintness of the family in the initial scenes. Though James' re-voicing is particularly noticeable to me because I recorded the replacement who is also my friend, I showed it to another sound designer who hadn't seen the film at any stage of production and he was unaware it was not the original actors voice, though he was aware of the use of ADR as a whole in the film. One inconvenience stemmed from my using the same track to record every session into as I didn't want to disturb the I/O settings in an already inconsistent multi-user space, which resulted in a huge expansive playlist containing hundreds of tracks. I also came to appreciate that whittling down a loop of 30 takes of the same sentence into a shortlist of 6 or so is enough to drive one insane.
I also assisted in the set up of the studio for the narration of A World Within Three Acres (Bushnell, 2014) , the other film Paul was producing. As our actors were already coming in he would have them read lines for a voiceover once we'd wrapped on HAFH, a much simpler process that just required one track to record into, with no looping and no syncing necessary.
No comments:
Post a Comment