OnStar Trek Into Darkness, we wanted to try something new for our sound workflow, and that was to keep a running 5.1 mix in the editors' reels throughout the entire editorial process. During Post on Super 8, J.J. Abrams had mentioned to his Post Supervisor, Ben Rosenblatt, that he wished he could have better sound while editing, and a "temp dub as you go." He wanted to be able to screen the rough cut with no notice, allowing him to edit right up until it was time to show it, instead of having to lock a version of the cut a week in advance in order to do a temp sound edit and mix in ProTools.
When Media Composer v6 came out and we saw that we could do a surround sound mix inside MC for the first time, Ben decided to go ahead with this experiment. We started digging in to exactly how MC handles 5.1 sequences, what pitfalls there might be between systems that had a 5.1 speaker setup and ones that had only stereo, and after that was all figured out we started building a 5.1 edit suite at Bad Robot. We mounted a projector, installed speakers, rigged up a screen, tuned the room to match the theater upstairs, bought two Artist Mix consoles and an Argosy console to hold everything, and started making the most complex temp soundtrack ever contained within 16 mono tracks.
Our first task was to figure out how Avid had designed their new 5.1 functionality, and what effects that might have on the editors. We wanted this process to be as seamless as possible for them, so if something about having a 5.1 sequence got in the way of their ability to cut, that would be a problem. Thankfully, we found that a 5.1 sequence will automatically and gracefully fold down on stereo-equipped systems, and you might not even notice that the sequence is set for 5.1. The biggest change we had to make across the entire Editorial team was that everyone had to work in stereo or 5.1. Many editors still like to work in Direct Out, but unfortunately if you are planning to mix in 5.1, Direct Out is not an option. All panning is done with clip settings and keyframes, and you need every available inch of space on the timeline, not to mention many other reasons why Direct Out is not ideal.
Once that change was made and we'd double-checked that everyone's mixers were set properly, we started receiving reels from the editors as they finished a first pass with J.J. They were working chronologically, so we did as well, and the first sequence we designed and mixed was the prologue on the Red Planet.
When Avid released Media Composer 6.5, there were a host of new audio controls that didn't exist previously. We wanted to take advantage of them, but we didn't want to upgrade the whole show all at once on the day the new version came out. This left us with a mix of versions, which is also not desirable, and meant that we had to be conscious of their differences, such as the maximum number of simultaneously playable tracks. In order to keep the editors' sequences as easy to cut with as possible, and since they wouldn't be able to play more than 16 tracks at one time anyway, we worked within the confines of 16 mono tracks for the whole movie.
The next time we do this, I think we'd stick to 16 tracks so that the sequences aren't any more of a hassle to cut with, but we will likely use a different track layout that would be established from the start. It would be something like:
Eventually we did have to upgrade almost everyone to 6.5.2, when we started constantly running into the maximum number of clip references that a bin could contain. Our timelines were too full to fit into the old bin constraints, though that max level had been set a long time ago and I just think it was forgotten about and never updated. When we brought it to Avid's attention, they sent us a patch to hold us over, and then released the new bin reference limit as a feature of 6.5.2.
5.1 panning in Media Composer is done on a clip-by-clip basis. Which track a clip is cut in on doesn't matter as long as your audio output is set to 5.1, 7.1, Stereo or Mono. Direct Out is the only setting to avoid for a surround sound sequence. To set a sequence to 5.1, look for the sequence setting in the upper left hand corner of the Mix Tool.
Once you put a sequence into 5.1, you can just leave it there even if you then move to a stereo system. All the surround sound panning is retained even if someone switches the sequence setting back to stereo or cuts the clips into a new sequence.
A clip can be panned to a specific channel or to a mix of channels, with the exception of the subwoofer. Each clip has an LFE level you can set in the big panner window that determines how much of the low end of the clip is sent to the LFE channel, but you can't send a clip exclusively to LFE.
There are several ways to pan a clip, and which one I use often depends on what I need to do. The first and fastest way is to use the small 5.1 panner above each channel in the mix tool. You can drag the little yellow dot anywhere in the grid and that will set the clip pan.
If you want to keyframe your pan, you have two options. You can enable keyframing for any one of these directions: Front Left-Right, Rear Left-Right, Front-Rear, Center % (if you want a center panned clip to also be sent to L & R). If you're doing a complex pan, you will often need to add keyframes in the first three settings. Unfortunately there is no way to display all the settings at once, so this method involves a lot of clicking between settings to get a pan right.
The second way to animate a pan is to change your Mix Tool mode to Auto, and use either the mini or big panner to set your keyframes. You can do this live by recording automation, or you can position your playhead where you want your first keyframe, drag the yellow dot to where you want to start your pan, then move the playhead to where you want the next keyframe and drag the yellow dot to where you want the panning move to end. As soon as you move the yellow dot, a set of keyframes are made at that location in the clip representing that pan setting.
When using this method, it is often necessary to set the yellow dot to the exact opposite end of the grid from where you want it, and then set it back to desired pan location. This ensures that a keyframe is made in all three axes. Otherwise, MC might not set a keyframe that you actually need and your panning move won't go exactly where you intended it.
For example, if you have a three point animation, from Front Left to Front Right to Rear Right, MC won't make the Front-Rear keyframe if you're only dragging the yellow dot across the top of the grid because it hasn't perceived a change in the default Front-Rear value that would warrant making a keyframe. This means that you would end up with no Front keyframe at all and a move that goes Rear Left to Rear Right without stopping at Front Right first. The video below explains this better:
As you may have seen in my article on quick dialogue cleanup, using RTAS effects was a core component to getting our dialogue in the right space. That was far from the only modification we did to the dialogue, though. We bought iZotope RX2, which is commonly used in ProTools to clean up noisy tracks, as well as Speakerphone 2, which handled all of the different futzes that are needed for a movie where people are constantly talking on all sorts of different communication devices. Both of these are common ProTools plugins, and it turns out that they work almost as well in MC. iZotope in particular requires a bit of handholding in MC, but once you know its problems you can take care to avoid them.
MC's built-in 3-band EQ is also very handy for reducing boominess, increasing clarity, and quickly taking out a problem frequency. Moreover, Matt would often use keyframing to bump or dip individual syllables in order to make sure that the dialogue was as easy as possible to understand.
One side effect of all the dialogue cleanup and futzing we had to do turned out to be a massive increase in the size of our bins. While a sequence bare of AudioSuite effects and EQs might have been 15 MB, a sequence with all those effects easily reached 100-200 MB in size. If there were multiple versions of a reel inside the bin, the file size of the bin would multiply accordingly. A bin with a full copy of the movie in it, such as the one I used to take the screenshot at the bottom of this post, could come in over 1GB easily. Because of this, and because these sequences get duplicated frequently into other bins for turnovers and outputs, the total size of the Star Trek project directory was over 100GB, and that's with some culling of old bins into another archive folder.
I handled most of the sound effects editing, though I was often given a head start by the 1st Assistant Editors, Julian Smirke and Rita DaSilva, who did first passes on ambiences and key sound effects for the editors during the assembly. Matt, Will, Robby and I would create a master list of effects goals we wanted to accomplish for our first pass, including what existing effects were working well and what we could improve. I would then go through and start removing one channel of most of the stereo effects that were already in the timeline, and then re-pan them so they still sounded right. This included flybys, where I would usually keyframe a pan in one channel instead of keeping two channels on the timeline. Effects that needed to be in stereo, such as ambiences, would of course get to keep both channels.
After I had removed as many redundant channels as possible and stripped out effects that we would replace, I dug into Robby's library and started filling the timeline back up with new fx. I like to do a rough mix and pan as I go, even though a more thorough mix pass still needed to be done after taking delivery of Matt's cleaned up dialogue tracks, whatever fx Will was making, and our temp score from Ramiro Belgardt and Alex Levy. Robby also tackled a lot of foley, since once you start down the road of making your temp sound good, you start to notice that you can no longer skip the things you would normally ignore for a temp, such as foley.
3a8082e126