-
Notifications
You must be signed in to change notification settings - Fork 0
Stitching Film Scans with Different Tracks
Ideally, all film that we scan is uniform, in that it is one film stock with none or one type of audio track. For example, educational classroom films are often a color print with optical audio. However, sometimes film has been compiled in a way in which there are different types of film stock or soundtracks on a single reel. Most of the time, we want to maintain the integrity of the object as one object, so we must do more than one scan or pass over the film to accurately capture the different tracks present. For example, our film scanner will not capture both optical and magnetic audio simultaneously, so these must be captured in two scans. Before you try to resolve this situation, it is important to first assess the situation based on the chart below:
- Uniform Picture (all positive or all negative), but Different Soundtracks
- Two scans are necessary, one all the way through with primary soundtrack, and one with secondary soundtrack beginning at first frame of secondary soundtrack
- Different Picture, Uniform Soundtrack
- Black and white with color can be on one reel together, but negative and positive together is best to separate. Coordinate with the film archivist to separate the negative section from the original film and assign a new unique ID and metadata to the negative section. Continue with a scan for each new film object.
As an example, let's consider one of the most common situations you will run into: you are well into the scan of a newsfilm with a magnetic soundtrack, when you come across a clip that has optical sound. Let's say it's around 40 minutes into the film. This is how you will proceed:
- Continue with the magnetic scan as normal, stopping when you get to the end of the film and complete the scan.
- Locate the first frame of the clip with optical sound. Switch the sound option in the scanning software to the new soundtrack type.
- Complete a new scan beginning with the first frame of this clip and ending when the clip with optical sound ends. Name it something simple, like wsbn_0111_optical.
- Scrub through the timeline again before dismounting the film to ensure that there are no other abnormal tracks.
Stitching the files is fairly straightforward, but it's best to first consider what outputs we need for a film scan and what outputs will need to be corrected. By default, every film scan (with sound) has these output files:
- Preservation Matroska (mkv), created when the RAWcooked utility stitched together the individual DPX frames from the scan with the audio (wav) file from the scan.
- The DPX files are fine, as they should be the same in the primary scan and the secondary scan. However, the audio (wav) file needs to be replaced.
- Mezzanine Quicktime (mov), created within the scanning software at the end of the scan.
- The picture in the mov file will be fine, but the audio will not. It will need to be replaced, so a new mov will need to be created.
- Access H.264 (mp4), created down the road by derivative scripts. This will not be made until later stages of file processing, so as long as we correct the other files now, we will not need to replace the mp4. In summary, we need a new Preservation Audio (wav) file, and a new Mezzanine Quicktime (mov) file.
Premiere is one of the most straightforward programs to use to create the two new files that we need. Follow these steps:
- Import the Mezzanine Quicktime (mov) file that the scanning software created, the Preservation Audio (wav) from the primary scan, and the Preservation Audio (wav) file from the secondary scan into a new Premiere project. You will also need to import the mov from the secondary scan to help line up the secondary audio.
- Drag the primary mov file onto the timeline first to establish the appropriate file specifications. Drag the Preservation Audio (wav) from the primary scan onto an additional track. Keep it in sync with the primary scan mov file.
- Import the secondary mov onto an additional track.
- Find something distinct in the first frame of the secondary mov, like a spot of dirt or dust that is not on surrounding frames. Locate that exact frame in your primary scan mov, cut the tracks (audio and video of primary mov) with the razor tool at this location, and then line up your secondary mov to this spot. Toggle the secondary video file off and on at different points in the timeline to make sure it matches.
- Place the Preservation Audio (wav) from the secondary scan onto the timeline in an additional track slot at the cut point. It should line up with the point where the secondary clip begins.
- Leave all of the primary mov video track, but unlink it from its audio. Delete its audio entirely. Leave the primary preservation audio in all spots except where the secondary audio will be. Delete it there. Delete the entire secondary mov video track. Delete all of the secondary preservation audio except where its needed. This should result in three populated tracks: the video from the primary mov, the preservation audio from the primary in most places, and the preservation audio from the secondary scan where the secondary clip is in the timeline.
- Export a new preservation wav, followed by a new mezzanine mov. The specifications for each are maintained on this page.
Finally, you need to replace the files from the original scans. Name the replacement files to match the structure of the primary scan files. Replace the Preservation Audio (wav) file in the file directory that's going to get stitched by RAWcooked with the new wav file you just created. Replace the mov with the new mov you just created. Delete all other files from the secondary scan.