Three Primary Workflows
There are three paths for how shot footage gets to the editing room, and they are dependent on how it originates on set and what the final delivery format should be. Projects may be shot on film or tape, or using the new tapeless, file-based cameras in which footage is captured on a hard drive or a computer card. Projects may be released on film or video, using celluloid, videotape, or digital formats. The resulting paths can be film to film, film to video, or video to video (which is basically the same as video to film). Once you get the material into your NLE, the editor's editing process is virtually the same, no matter which path the project needs to take. It is up to you, as the assistant, to make sure the digital footage is created properly to make the transition from the initial format to the final release format.
I discuss finishing formats in more detail in Chapter 14, "To the Finish—Ho!!," but let's take a look at some of the possibilities. (In the following workflow overviews, I may use a few unfamiliar terms that are described in more detail in later chapters.) Our imaginary film, Silent Night, Silent Cowboy, is taking a typical path for many features and television shows. It is shooting on film—35mm in our case—that will be transferred to videotape for editing and then released in movie theaters on film. Let's start by discussing that path first.
Film to Film
The 35mm film from the first shooting day (Monday) leaves the set and arrives at a lab for processing, where the film negative is developed. After the film is developed, it is prepped and sent to a telecine house, where the film is converted into a video format.
On set, each roll of film that goes through the camera is called a camera roll and it is given a unique camera roll number. A typical system for numbering is to start on the first day with camera roll number one and continue sequentially throughout the entire shoot. Many productions shoot with more than one camera. The cameras are indicated by letters such as A, B, and C. On the first day, if production shot three rolls from the A-camera, two from the B-camera, and one from the C-camera, six camera rolls would go to the lab, labeled A1, A2, A3, B1, B2, and C1.
Prepping for telecine involves marking the film negatives in such a way that the telecine operator can find a particular frame in the film and link it up in a specific way with the video frame. This is done with key numbers. Figure 4.1 shows the number KK 23 1234 5677. The bar code next to it is readable by the telecine machine.
Figure 4.1 A piece of film with imprinted markings. The dot to the right of the key number identifies the frame associated with that exact key number. (Courtesy Kodak Corporation)
Every piece of negative, no matter which manufacturer it is from, has key numbers imprinted when it is created. As the negative is processed, the numbers become visible. Kodak's Keycode added a tiny bar code on the same edge that can be read by the telecine machine.
To make your job easier, the telecine operator looks for a frame that falls exactly on a key number as indicated by a little dot after the actual key number (see Figure 4.1). At the lab, the person prepping the negative for telecine will punch a small hole in the first frame that falls at a key number for each new camera roll (or whenever the key code numbers don't run sequentially). You can see this hole punch on the third frame in the figure.
That frame is then lined up in the telecine machine and aligned with a particular video frame. The film is locked together with the video and they are run in sync, copying each frame of the film onto the video.
Of course, it isn't that easy. The frame rates of film are not the same as in video (especially in the United States) so an adjustment has to be made. Somehow, six new frames of video need to be created for every second of film. Every 24 frames of film needs to be converted to 30 frames of video. This is accomplished by a process called 2:3 pulldown, which takes clever advantage of the fact that each video frame is actually made up of two video fields made up of alternating video scan lines that essentially draw one-half of the image on screen. When this video is projected, the fields are interlaced together, resulting in one complete video frame.
Let me explain interlacing a little better. Video images are created by shooting electronic signals onto a surface that displays them as a series of lines drawn on that surface from left to right, one line at a time. Because of the quality of the phosphorous in most televisions at the time these standards were created, engineers decided to shoot them as alternating lines. At first, all the odd lines are shot onto the screen (see Figure 4.2). Then, the even lines are drawn (see Figure 4.3). But because the phosphorous would hold the image for a slight amount of time, the human eye would perceive the image as made up of one complete image (see Figure 4.4). Each time one set of scan lines is drawn, it is called a field.
Figure 4.2 To create one full video frame, the odd-numbered scan lines are drawn onto the television screen at first. This creates the video frame's first field.
Figure 4.3 After the first field is drawn, the even-numbered lines are drawn to create the frame's second field.
Figure 4.4 When we view the two fields together, we perceive both as being drawn at the same time and we see the full video frame.
So, what is the trick to creating these 30 frames of video in each second (made up of 60 fields of video) from the 24 frames of film? The easiest way to think about it is to consider a group of four film frames, which we'll call the A, B, C, and D frames (and which take up 1/6 of a second). We now need to convert them into five video frames for that same 1/6 of a second.
In Figure 4.5, you can see how this is accomplished. The first film frame out of every set of four (called the A-frame) is converted into two video fields, as normal, and this becomes the first video frame. The second film frame, the B-frame, is converted into three fields—the full second video frame and the first field of the third video frame. The third film frame, the C-frame, is converted into two fields—the second field of the third video frame and the first field of the fourth video frame. The fourth film frame, the D-frame, is converted into three fields—the second field of the fourth video frame and the entire fifth video frame.
Figure 4.5 The 2:3 pulldown converts four film frames into ten video fields (or five video frames) to adjust for the film rate of 24 fps to video's 30 fps.
The sequence is then repeated for the next set of four film frames—two fields, three fields, two fields and three fields. This alternating sequence of two and three fields is what gives this conversion process its name—the 2:3 pulldown (also called 3:2 pulldown).*
The problem with this conversion process arises when an editor is working. If the editor happens to cut from the third video frame in the 2:3 sequence to the fourth frame in another 2:3 sequence, the cut actually occurs between two video frames that are made up of a combination of two film frame images. Where, pray tell, should you make the film cut when you attempt to match the film to the editor's video edit? Because the editor is working in the 30-frame-per-second mode, it is difficult to accurately match back to the 24-frame world.
The solution is to take the 30 frames per second video and convert it back into 24 frames per second inside the digital editing machine. As long as we can accurately determine which video frames are which frames of any 2:3 sequence, the computer can easily throw away the artificially created frames.
To do this, however, the computer needs to know just which frames on the video are A-frames, which are B-frames, and so on. Actually, it only needs to know the orientation of one of the five video frames for it to deduce the proper order of them all (aren't computers smart?). Traditionally, the frame used is the A-frame.
When the telecine house transfers the film to videotape, normally you request that they make cuts on the A-frames and that they put their A-frames on the "zeroes and fives." This means that every time they make an edit in the videotape transfer—which will happen whenever they stop and then restart the recording process, usually at each new take—they should make sure the A-frame falls on the time codes ending in either a zero or a five. In other words, if the video frame at time code 1:13:23:05 is an A-frame, then the next A-frame will fall at 1:13:23:10.
This is why the lab will prep the negative for telecine at every key number jump by punching that frame, as we saw in Figure 4.1. The telecine operator can easily see what frame contains an actual key number and make sure that it falls on an A-frame.
If your telecine house has not put the A-frame on a zero or five, you can still tell which frame is the A-frame by examining the time code burn-in on the videotape. As you rock the videotape forward field by field, the A-frame should have no time code change between the fields. The B-frame, which is made up of three fields, will have no time code change between the first two fields but will have a change between fields two and three. The C-frame will have two fields with a time code change between them. The D-frame, made of three fields, will have a change in between fields one and two, but not between fields two and three.
After the film has been converted to video images, the telecine house will go through each take, one by one, and synchronize the sound up to the picture image. This is done in several ways, but the most accurate way is to find the exact frame where the second assistant camera person has clapped the picture slate (see Figure 4.6) and align it with the exact moment on the matching sound track where the slate sound can be heard. This moment is called the first modulation of that sound.
Figure 4.6 A "slate." The stick at the top pivots, so when it is pushed down onto the slate it makes a distinctive sound that can be easily synchronized with the picture frame where the stick closes.
Once this common synchronization point is found, the rest of the take should match. This process is called synching the dailies. When it is all done, we say the "dailies have been sunk" or "sunk up."
When the telecine house finishes synching the dailies, they either send a tape or a hard drive with the fully sunk-up material to you in the editing room. If they send a tape, it is typically a low-cost tape such as an HDV/DVCam.
Working With Telecined Film
For a film-to-film workflow, the next thing to discuss is what sort of information you need to input into your digital editing system.
Along with the video dailies from the telecine house, you also should receive some paperwork and some computer files (either by e-mail or on a CD-ROM). The paperwork consists of a telecine log and, occasionally, a report of any problems occurring in the telecine (the sound is too loud, a take is out of focus, and so on). You will also receive information that tells the computer how to read the file. The data in that file is one kind of metadata that your NLE uses to keep track of the digitial files.
A typical telecine log contains information for every take on the tape including:
- The name of the take (the scene and take number, for example 10-5 for Scene 10 Take 5).
- The starting and ending time code numbers.
- The Keycode of the first frame of the take (this is the key number for the take that is automatically read by a machine at the telecine session).
- The time code of the first frame of sound used (if the on-set sound recordist used a time coded machine; if he or she didn't, synching at the telecine house will be an expensive nightmare).
- The camera roll.
- The sound roll unless the shot was MOS.
- The date the take was shot.
- Other assorted information you may or may not need.
Several standard formats are used for these files—FLEx, Evertz, Keyscope, ALE, and ODB. The two most commonly used formats are FLEx and ALE (a proprietary Avid format). Some telecine houses can create logs directly in FLEx and ALE format. Otherwise, you need to translate the file into a format appropriate for your system.
Since each videotape typically holds up to 60 minutes of dailies, and each tape gets its own FLEx file, you may receive several files if you have a lot of dailies. The files are quite small, so several files can fit on one disc. Every day, a new disc is delivered with your editing tapes. An example of a FLEx file is shown in Figure 4.7.
Figure 4.7 A FLEx file shows every take telecined for digitizing into the NLE along with the metadata that enables you to get back to the original film, if necessary.
You will also get a videotape, typically an HDV/DVCam, with your dailies. This tape must be created to very detailed specifications. Figure 4.8, on page 38, shows an example of an order that the postproduction supervisor submits to the telecine house.
Figure 4.8 A work order that a postproduction supervisor sends to a telecine house that specifies exactly what you need.
Once the material arrives at the editing room, check the tapes to ensure that they match the work order. Then, bring the video and audio material from the videotape or hard drive into your NLE. (This process is called digitizing, ingesting, or importing.) As discussed in Chapter 5, "Moving on to Editing," you will sync the dailies and give them to the editor to cut.
The editor starts cutting, and when public screenings (previews) of Silent Night, Silent Cowboy are necessary, the tape used for projection is made from an output of the NLE or from an uprezzed (higher resolution) version of the dailies. The editing process continues with successive rounds of editing and screenings.
When the editing is complete, various people create lists that enable you to recreate the edits in high-definition video. This new edited master, called a digital intermediate, is used to create a new film negative. The negative will be matched up with a separately created sound mix to create the print audiences see in theaters. (See Figure 4.9 on page 40 for a chart of this process.)
Figure 4.9 This chart summarizes a film-to-film workflow with a digital intermediate (DI).
Voilà! Your film is done.
Film to Video
The process for film to video (such as for television or home video), shown in Figure 4.10 on page 41, looks similar to the film-to-film path outlined in Figure 4.9. In fact, the process is more or less identical until the final step.
Figure 4.10 When a project finishes on tape, the digital intermediate does not need to be created. Instead an online finish is performed. The editing process is called offline.
The material is telecined, then edited and screened using a lower-resolution video format. This is typically called the offline. When the editing process is complete, an online session is performed. In the online session, a higher-resolution version is created from the original negative and then conformed to match the edited offline. The material is then color corrected and the sound is laid back onto the tape. That becomes the master from which you create all subsequent video copies.
In many cases, high-quality (usually HD) offline tapes are created at the telecine stage from all the original dailies as well as the editing HDV/DVCam tapes. This means that it will not be necessary to go back to the negative to create the masters. We discuss all this in more detail in Chapters 5 and 14.
So, once we do the layback onto a videotape for the final master, voilà! Your film is done.
Video to Video
The third primary workflow is videotape to video or video file to video. These are the two workflows chosen by the majority of low-budget films made today, an increasing number of film and television productions, as well as virtually all corporate and event videography. These paths involve capturing in a video format and finishing on video. Let's take a look at each workflow separately: videotape to video and video file to video.
Videotape to Video
In the videotape path (see Figure 4.11 on page 42) the original picture tapes and sound files are sent directly to a tape dubbing house where they are converted for use in the editing room. (Usually, audio is captured separately to allow for better quality and control, so you have both picture and sound files.) Because the conversion process usually involves creating a lesser-quality video image, this process is called downconverting. Tape copies are made from this master for editing. On television shows for which the final format exists on tape rather than film, the video master is on a high-quality digital videotape, usually on D-1, D-2, Digital Beta (Digibeta) tape, or HD tape, with a series of SMPTE time code numbers. The viewing and editing tapes made from this master tape (usually HDV/DVCam tapes) have the same time code numbers as the master, making it easy to match back to the high-quality tape at the end of the editorial process.
Figure 4.11 In the videotape-to-video path, the production captures the action on high-quality video, downconverts it to a lower resolution for editing, and then completes the project using the original videotapes.
This editorial tape is input into the digital editing system. Screenings are usually produced from a direct output from the editing system, although occasionally an EDL (edit decision list) is made of the editor's cut. We discuss EDLs in the next chapter, but for now you should know that an EDL is used to exactly recreate a cut using the master tapes (instead of the editing tapes). After every screening—and there can be many—changes are made using the NLE. At the completion of the editing process, a picture EDL is generated to recreate the editor's work on the master tapes. This edited master is color corrected and the polished sound is laid onto it.
Voilà! Your film is done.
Video Files to Video
The final workflow we'll talk about (see Figure 4.12 on page 44) is coming into play more and more at all levels of filmmaking. In this workflow, the original shoot is not recorded onto film or tape, but directly onto some kind of computer file and stored either on a hard drive or on some kind of storage card. The format might be compressed video such as P2 or HDCam EX, and the footage is delivered to the editing room on memory cards. Or, a production might use a higher-resolution format such as those produced using the RED, Viper, Sony F35, or Phantom cameras. This footage is usually delivered on hard drives. Since no videotape is involved, this is often called tapeless acquisition.
Figure 4.12 When a production shoots in a tapeless video format, no tapes need to be made to get the video into the NLE. However, the files may need to be transcoded to work properly.
With no tape on set, the files can (with some work) be copied into your NLE for editing. Generally, you will need to transcode the footage from the capture format to an editable format that can be used with your editing system. After transcoding, the video and audio files can be copied quickly into your NLE where they can be sunk as normal.
At that point, the editor will edit as in any other workflow. At the end of the process, all the color correction can be with digital files—no tapes. When that is done, it is typical to transfer the final, color-corrected film onto a high-end tape format for duplication and distribution. (In the future, with digital delivery to theaters, this might change as well.)
Voilà! Your film is done.