Current Usages and Practices

[ LiB ]

Typically, most AV projects use 29.97 drop-frame time code as a standard. Some commercial post houses will use 29.97 non-drop frame because most of the work done there is not much longer than a minute. Music-only sessions tend to use 30fps, as it has a convenient relationship to real time. Film uses 24fps but is often transferred to video, resulting in 29.97fps. The newer HD cameras may use 29.97fps, 24, or even a special 23.976 rate for direct transfer to regular videotape on location. The choices are many and the confusion runs high.

Knowing what time code you're using and why you're using it is imperative when using Pro Tools. Pro Tools has the ability to change time code formats within software, which can be helpful but can also lead to mistakes and more confusion. Once you have established a time code standard for a project, everything else is more straightforward.

When using the wrong SMPTE frame rate or frame count, you will notice, over time, that the audio will gradually slip either ahead or behind the video. It might take a minute to notice this drift , but it will eventually become apparent. Pay attention to the lip-sync. Determining sync problems is not as simple as going to the end of your Pro Tools session and playing it to see how bad the sync is by the end. When Pro Tools syncs up to a videotape, the SMPTE coming from the videotape machine will give Pro Tools a frame number to start on. Pro Tools will be in sync at that frame, no matter where it is, and then begin to slip from that point on. The only accurate way of determining if the sync is slipping is to watch from the beginning for at least a minute or so. With shorter program material, determining slippage is more difficult.

In any event, as I said, it is imperative that you know what the correct frame rate and count is for your current project. Try to clear up any confusion as early as possible in order to avoid a big mess later on.

To Drop or Not to Drop?

To drop or not to drop? There should only be one factor that determines whether you choose drop or non-drop: how you want your time code to relate to the real time clock. Another factor arises, though, when there is a mistake you must correct for in some other aspect of the production. I have found that some video editors prefer non-drop code for ease of calculating offsets and figuring lengths. However, with so many time code calculators available, it seems a little moot.

Many video editors want to see each frame number without any skips. It can be disconcerting for a video editor to create a ten-frame video transition that starts at 00;08;59;24 but ends on 00;09;00;06. Do the math. As there are semicolons separators, you know that these are drop-frame time code numbers, which skip two frame numbers every minute except the tenth. Counting ten frames from 00;08;59;24 but skipping frame number ;00 and ;01 you will end up at 00;09;00;06. In non-drop frame, the tenth frame would be 00:09:00:04. All modern software and equipment running in drop-frame mode will adjust for skipped frame numbers and create transitions and other edits with the correct length and position. Don't be afraid of drop-frame SMPTE. Download a time code calculator and mess around in drop-frame mode. You will get the hang of it quickly.

Once again, let me mention that 30fps drop-frame code is a misleading term that is not part of the SMPTE specification. Remember to keep the frame rate and frame count separate in your head. 29.97fps SMPTE still has 30 frames of video in each second of SMPTE time. However, each second of SMPTE time is slightly longer than a real honest second on the clock. 30fps SMPTE keeps right up with the real time clock and does not need to skip any numbers. If it did so, then it would display time that was running faster than real time. Wouldn't that be extra-confusing? Still, you'll find the 30fps drop-frame setting in Pro Tools. The purpose of including it is to allow you to correct for improper film transfers or the wrong time code on production audio tapes.

2-3 Pull-Down

When film is transferred to videotape, it must be converted from a frame rate of 24fps to the color video standard of 29.97fps. These two numbers do not have a nice, neat mathematical relationship that would make it easy to transfer frames of film to frames of video. The solution to this problem is called 2-3 pull-down . The following subsections explain how 2-3 pull-down works for film and audio transfers.

Film Transfers

Each frame of film is like a 35mm picture. It contains one image out of a series that together will make up a movie when played back at 24fps. Recall that a video frame is actually two "fields," each containing half of the total image. One field has the odd horizontal lines in the picture and the other has the even lines. For every frame of video you see, two fields have gone by that together make up the complete video frame. Fields move at 59.94fps, so that the actual frame rate is half that, or 29.97fps.

The machine that transfers film to video is called a telecine machine (see Figure 3.5). It copies successive frames of film to certain fields of video in a 2-3-2-3 relationship. The first frame of film goes to the first two fields of video. The second frame of film goes onto three fields, field one and two of the second video frame and field one of the third frame. The next film frame goes to the second field of video frame three and the first field of frame four. This goes on in the 2-3-2-3 pattern for the duration of the film transfer. This way, the film has the same basic running time on video at 29.97fps as it had originally at 24fpsbut there is a slight difference. To make this process work, the film is slowed by a very small amount (.1%) to roughly 23.97fps. Hence the term pull-down . The frame rate is pulled down to match the video rate of 29.97 using the two field, three field transfer technique. "But what does this mean for audio?" you ask.

Figure 3.5. A modern-day telecine machine from Cintel. It is used to transfer film frames to videotape, one by one, using the 2-3 pull-down process.

NOTE

TELECINE TRANSFERS: 2-3 OR 3-2

Telecine transfers typically start with two then three fields of video per frame of film. Telecine machines are capable of beginning with three video fields then two, but this is atypical. You will often hear the transfer process incorrectly called 3-2 pull-down. Unless specified prior to the film transfer, the process will begin with 2 fields of video for the first frame of film, then 3 fields for the second frame of film, and so on.

Pull-Down Audio Sample Rates

Let's say that during a film shoot at 24fps, the production audio is recorded directly to a DAT machine at 48kHz with time code running at 30fps. When it comes time to transfer the film to video, the audio will be transferred to the same videotape. Since the film speed must be slowed down slightly to match the video frame rate, so too must the audio. The sample rate will slow from 48kHz to 47.952kHz (also called 48k pull-down) and as a result, the time code recorded at 30fps will magically wind up at 29.97fps. Amazing, isn't it? This way, the audio time code will line up exactly with the video time code while transferring. The downside is that a sample rate of 47.952kHz cannot be transferred digitally without going through a sample rate converter (See Figure 3.6). Most film transfer houses will not do this. They will make an analog copy of the audio from the DAT to the VTR.

Figure 3.6. The Z-Systems z-8.8a router and sample rate converter. It can take incoming digital audio at one sample rate and convert it to another in real time. It also functions as a very handy digital router for my studio setup.

Sometimes, the audio is not transferred during the film transferthere might be problems with the production tapes or perhaps the transfer facility does not have the capability. Either way, there will be times when you have to retrieve production audio from the original DAT tapes used on the shoot. If you try transferring takes directly into Pro Tools without accounting for the speed difference made during film transfer, your audio will be out of sync with the transferred video. There are several ways to compensate for this speed difference. Some of these methods require additional hardware, such as Digidesign's Sync IO, USD, or a third-party synchronizer. But you can compensate for speed difference without hardware using only the Pro Tools software, including Pro Tools LE. More on that in Chapter 5, "Recording to Picture."

The basic method of transferring audio from a film shoot to video transfers of that film involves slowing (pulling down) the source tape by .1% using a hardware synchronizer, and then recording the audio into Pro Tools using either analog connections or a real time sample rate converter such as the one found on the HD 192 I/O interface. The analog connection has the disadvantage of being a generation loss. The sample rate converter is a better option but is only available in the 192 I/O interface or in an outboard sample rate converter such as the Z-Systems z-8.8a shown in Figure 3.6.

Pull-Up Sample Rates

Some portable DAT machines have the ability to record faster than the standard sample rate of 48kHz. Recording at 48.048kHz (called 48k pull-up) during filming gives you the advantage of transferring digitally later during a film transfer. When 48.048kHz is pulled down for the film transfer, it winds up back at 48kHz. This way it can be digitally recorded onto the VTR along with the film. The time code frame rate still is recorded at 30fps, so it will pull down to 29.97 when played back at the normal 48kHz sample rate.

The above scenario holds true for projects that are shot on film but that will remain in video format throughout the rest of production and delivery (such as the made-for-TV movie that never shows in theaters). If the project will eventually be shown in theaters on film, then a slightly different approach might be necessary. Audio that will eventually be used for a film release will have to do one of two things: It must either go through the reverse process of 2-3 pull-down (see Figure 3.7) to return to the film rate of 24fps (to reverse the 2-3 pull-down process, it is necessary to "pull-up" the sample rate of the finished audio to match up with the original film speed of 24fps) or it must be edited using 24fps as a reference and never change speed.

Figure 3.7. 2-3 pull-down block diagram.

24p HD Video

HDTV is bringing a whole new technology to video. HD cameras, such as the one shown in Figure 3.8, are able to record in a format known as 24p (24fps Progressive Scan) that allows video shot with HD cameras to look more like film than traditional video cameras using interlaced images at 29.97fps. In order to transfer these HD images to regular video, you go through the 2-3 pull-down process. This can be done after the shoot, though some cameras can use a special frame rate of 23.976fps to record initially. Recording at 23.976fps allows a hardware unit to translate the HD image into a NTSC video image running at 29.97 in real time. Directors can record simultaneously to HD video and regular NTSC video, allowing them to view the shots on less expensive NTSC equipment.

Figure 3.8. A Panasonic AJ-HDC27 Variable Frame Rate 16:9 HD Cinema Camera capable of recording at 23.976fps.

These cameras can record multiple tracks of either 16 or 24 bit 48kHz in order for the audio to be in sync.

[ LiB ]