When David Cronenberg screened rushes (also known as dailies) on the Alliance Communications production of his virtual-reality, sci-fi thriller eXistenZ, which wrapped early in July in Toronto, the much-lauded director was treated to the highest-quality digital sound perfectly in sync with the film print, thanks to a new process using the latest in post technology.

With the support of Avid Toronto’s Jeff Krebs and the cooperation of Magnetic North and deluxe toronto, the film’s picture editor Ron Sanders was able to convince the production to use the new process, which makes syncing sound much easier and allows rushes to be screened on a projector with synced digital sound from a hard drive played on a Pro Tools sound editing system.

Sanders says the new process, which combines various digital technologies, has generated a lot of interest from other editors and tech types who have dropped by to see the setup. ‘As far as we know we’re probably the first to do all of this in this combination,’ says Sanders, adding that other editors are no doubt working with similar systems.

Though the costs to set up the system were not insignificant (two modified Panavision cameras were needed, an Avid media station was required at Mag North, and a computer and Pro Tools system at the screening room), the new system saves the substantial cost of transferring the sound to 35mm mag stock and the expense and time of paying assistants to sync up the mag stock to the picture.

‘No sound editors use mag film anymore, they all work digitally,’ Sanders points out. ‘So in order to screen rushes you had to transfer everything to mag. It would be hundreds of thousands of feet of stock and two or three assistants and all their time, and that stuff would never be used again.’

On Cronenberg’s last film, Crash, Sanders says printed takes were screened silent. Synced picture and sound rushes were available only on tape. ‘Other people have used a dat tape or a D-88 and locked up to that, but the logical extension was to go right to a workstation where you don’t have any physical medium at all,’ he says.

Sanders says the process also saves first picture assistants from having to digitize media, freeing them to do more useful things.

And because the sound stays digital throughout the new process there is essentially no generational quality loss, until the mixing stage, meaning that dialogue evaluations can be done right on the Avid.

Here’s how it worked: As the negative was exposed on set, veteran sound recordist Glen Gauthier recorded the sound separately using a state-of-the-art Nagra D (digital Nagra field recorder). A modified Panavision camera with an Aaton Code Generator burned bar code into the negative opposite the key code when the picture was exposed. The Nagra D also generated this code and it was ‘jammed’ into the Aaton clock at least twice a day, which gave the quarter-inch set audio and the film a common time code, making the task of putting film and audio in sync very easy.

At the end of each shooting day the audio was shipped to Magnetic North, where the original recordings were loaded from another Nagra D (which Gauthier must lend to the video house for each production he works on) directly onto a hard drive which is attached to an Avid Media station set up in the transfer bay.

After Toronto’s The Lab processed the negative based on the director’s selections, the lab rolls were shipped to Mag North for the Telecine video transfer.

The colorist, in this case Bill Holly, put each lab roll up on the Rank and transferred each roll starting at the ‘Picture Start Mark’ on the Academy leader of the negative lab roll. The mark served as a reference point between the negative and film print, which was made from the negative and screened with audio in a viewing theater as rushes.

The film was color timed and transferred in sequence to D-Beta video. In addition to reading the key code, the Rank was modified to include an Aaton code reader and was also able to read the Aaton time code burned on the neg.

Both the Aaton and key code were logged into the media station computer as text logs. The Aaton code log matched the code on the audio and allowed picture and sound to be put in a sync relationship.

The D-Beta was then digitized onto the drive through the Avid media station. The logs of the film-to-tape transfer containing key code, Aaton time code, camera roll, slate and take-in and -out points – which were created by the video colorist as a text file – were then imported into the Avid software.

The digitized picture, now called ‘media,’ was broken down into individual shots based on the logs.

Because the sound had been digitized and already existed as media, and the time code on the sound matched the Aaton burn time code on the film, the audio could be put into an immediate sync relationship through an Avid function called ‘autosync.’ This function takes separate audio and picture clips with common time code and creates a synced clip. These are the clips used by Sanders on his Avid for cutting.

(The autosync function makes the process applicable to other types of production such as tv series, mows and commercials, because, in all cases, the time it takes for the editor to sync picture to sound is significantly reduced.)

The assistant editor received the media files from Mag North on a media shuttle hard drive and copied them onto his Avid’s hard drive. The media shuttle was then sent back to Mag North where the synced audio was striped back onto the master D-Beta video transfer of the picture so that vhs dubs could be pulled for Cronenberg, the actors, executive producers, etc.

deluxe made a print of the negative as usual.

To do the digital rushes process, the editor assembled a sequence on the Avid which reflected all of the film selected for that day. The rushes were laid down in lab-role order, with each lab roll starting with a different hourly time code. Lab roll one was hour one, roll two was hour two, etc.

The hours corresponded to the number of lab rolls and the sound and picture would match the print that had been run off the negative at deluxe. However, they had to be given a common start mark.

This start mark was the Academy leader, as the picture start mark on the media matched the picture start mark on the print. The zero hour of each code was at the picture start mark, so lining up the film print in the projector at the picture start mark was like parking the Avid’s cursor on the picture start mark on the digitized media.

To screen rushes in one of deluxe’s theaters, the editor exported the audio media only to a Jaz drive, as well as exported a data copy of the lab roll sequence. The Jaz drive was connected to a Macintosh computer in the screening room that was installed with Pro Tools software. The data was imported into the Pro Tools system, allowing it to access the audio media which had been exported on the Jaz drive from the Avid.

The print was strung up on the projector at the picture start mark on the Academy leader, and the daily lab roll sequence imported into the Pro Tools existed as a similar sequence. The picture start mark on the sequence was represented by the zero hour of the time code for the first lab roll to be projected that day.

But since a film projector was mechanical and didn’t generate time code, and the sound sequence on the Pro Tools needed to chase time code, a box called a jsk was attached to the projector, which generated a code for the audio on the Pro Tools to chase.

With lab roll one in the projector cued to the picture start mark, the projectionist punched the hour-one time code into the jsk generator box. The Pro Tools was put into ‘online mode,’ waiting to receive code to chase.

When the projector was started, the jsk box generated hour-one time code and the audio sequence on the Pro Tools leapt to the zero point at the hour-one mark and chased the code it was getting from the jsk, which was slave to the projector. If everything started at this common point, then the audio would run from the hard drive in sync while the print was running on the screen through the projector.

At the end of each lab roll the ascending number code was punched into the jsk box so the Pro Tools chased the proper time code for each lab roll. The process was repeated for each roll.

Sanders says the evolution of the new process will be when the transfer house is able to send the media through fiber-optic cable to the editor’s system.

‘They would be able to send a half-hour worth of rushes to us in three or four minutes,’ he predicts. ‘There would be nothing physical going back and forth at all, which is the logical extension of the thing. Everybody would work on a file level rather than drives going back and forth.’