Utilizing Overcrank/Undercrank Video from the Panasonic Variable Frame Rate High Definition Camera David Wiswell Group Manager Panasonic Broadcast & Television Systems Company

Utilizing Overcrank/Undercrank Video from the Panasonic Variable Frame Rate High Definition Camera David Wiswell, Group Manager Panasonic Broadcast & Television Systems Company

INEMATOGRAPHERS SHOOTING FILM USING OVERCRANK AND UNDERCRANK MOTION TECHNIQUE GOES BACK TO THE FIRST MOTION PICTURES. IN THOSE EARLY DAYS, THE FILM WAS FED THROUGH THE CAMERA BY HAND CRANKING. THE CINEMATOGRAPHER COULD SPEED UP THE FILM, ‘OVERCRANK’ OR SLOW DOWN THE FILM ‘UNDERCRANK.’ THE RESULTING VISUAL EFFECTS CREATED THE FIRST SLOW MOTION AND FAST MOTION.

C

Today, professional film cameras provide controlled overcrank and undercrank by speeding or slowing the electrical motors used to move film through the camera. Electronic television cameras have not had the ability to shoot in faster or slower frame rates. The visual effects film shooters attained by speeding or slowing the film’s rate of passage through the camera could only be emulated in the electronic domain using slow motion playback techniques most common in videotape recorders. Subtle but important visual differences arise between the VTR process and the film process. VTRs play out video for slow motion or fast motion by playing the recorded tape at variable speed and tracking the moving video tracks with a special dynamic tracking head. The VTR technique does not affect the image blur from frameto-frame and by the nature of the VTR process, frames are dropped or added as required to maintain an even frame rate output. The visual difference between the VTR and film techniques is quite visible and is important for electronic shooters to understand if they wish to create a true film-like experience. Using the film technique, the image frame exposure time can be changed in combination with the frame rate, resulting in more control over motion speed and the amount of motion blur in the individual image frames. By controlling the

-2-

film’s speed and exposure time, not only will the playback images run at modified speed when projected at 24-frames; the images themselves will be different from frame-to-frame. Cinematographers generally prefer to shoot images at 24-frames per second for a film-like look. Over the last couple of years, progress has been made toward electronic shooting with high definition camera recorders operating at 24-frames per second. Many film cinematographers have resisted changing to electronic shooting for a number of reasons but the lack of overcrank and undercrank capabilities was a major objection that, until now, could not be overcome by television cameras. Many cinematographers who are considering use of electronic high definition systems for their projects want more than 24-frame rate shooting; they want the ability to do overcrank and undercrank shooting. Panasonic has portable high definition camera/recorders with the right price and performance characteristics. We have found a way to perform overcrank and undercrank, which will lead to a broader acceptance of electronic cameras as a substitute for film shooting. A cinematographer’s willingness to consider electronic shooting rather than film is based on the premise that video images can retain a high quality “film look” through proper image handling.

Shooting in 24-frame with the Panasonic variable-frame HD camera Panasonic has created a simple yet elegant technique for overcrank and undercrank shooting using a DVCPRO high definition camcorder. The AJ-HDC27V camera is able to shoot images at variable frame rates and record the images on high definition videotape. The VTR portion of the camcorder operates as a normal 720 progressive high definition VTR recording 1280x768 images at 60 frames per second. The overcrank or undercrank effects are created in the camera section of the camera/recorder. To understand the over/under crank process, it is necessary to understand the “2:3” process of converting between 24 and 30 frame video (often incorrectly called the “3:2”.) In Figure 1, the first frame in the sequence is referred to as the “A-frame.” The A-frame sequence must follow the same pattern so that 24-frame video converted to 60-frame can always be converted

2:3 "Pull Down" A

A

A

B

B

B

C

B

C

C

D

D

D

Figure 1

back to 24 by reversing the known pattern. 720p60 is sixty frame video, if 24-frame is converted to 1080i or NTSC 525i video then the conversion is to 60 fields, not to 60 frames as shown in Figure 1. With interlace conversions the conversion will result in some frames with fields from two different 24-frames. These derived frames with fields from different 24frame images are called “jitter frames”, because of the effect seen if there is motion between two original frames. The most important thing is that the sequence remains predictable eg not broken as might happen in an edit session if the video

-3-

is cut in the sequence’s middle. As long as the sequence is not broken, the original 24-frame video can be recovered. If the variable frame rate camcorder is recording 24-frame per second video a standard 2:3 sequence will be created to fit the 24-frame video onto the VTR’s 60 frame recording system. The system marks each new 24-frame image by recording a timecode user bit for each frame. Each time a new video frame is acquired by the camera, it is marked as an “Active Frame.” The video is recorded on the 60-frame VTR using the standard 2:3 sequence as seen in Figure 1. If a frame is repeated just for the sake of the recording process, it is not marked as an Active Frame. Instead it is marked as an Inactive Frame that can be dropped if necessary. The camcorder’s VTR records video with a 24frame or irregular pull-down sequence at 60frames per second. A record of each new image frame is always recorded in the VTR’s timecode User Bits, thereby marking the sequence of new and repeated images. 24 fps Later in the post-production process, the User Bits will recovering the 60 fps facilitate images at the intended frame D rate. The marked Active Frames can be assembled into a 24, 30 or 60-frame per second stream and scaled to a standard image format, 1920x1080, 1280x720 or 720x486. The process of recovering 24-frame video from the recorded multiple frame rates are relatively straight forward if the video is shot at an even 24frames per second. The camcorder records standard 24-frame video with 2:3 pull-down added. The 2:3 record sequences start on timecode “00.” The timecode recorded is non-drop timecode so every fifth frame is an Active Frame. The first frame in the 2:3 sequences are is called the A-frame; the A-frame is the most

import frame for the decoding sequence. The tunity for creative visuals. Referring to Figure 3, 2:3 sequence repeats every five frames so there and using 24 frames as the example display is an A-Frame every five frames starting at 00, rate; the effect of over or under cranking the then 05, 10, 15 and so on. Once you know which frame is "Undercrank" the A-frame, you can always accurately decode the images A B C D back to the original frame sequence. A number of devices are available for extrac24 fps tion of a standard 24-frame, 2:3 sequence in high definition and standard definition.

A

B

C

D

E

F

G

H

Shooting in Variable Frame Rate Mode "Overcrank" Active frame IDs Over/undercrank is a little more complex to recover Image frames in "Undercrank" mode are longer than 24-frame image times because the frames are not and "Overcrank" images are shorter than 24-frame image duration. exactly 24-frame and are not Every active frame is given an ID and recorded at 24-frames per second. as easily predicted, see Figure 2. Panasonic has demonstratFigure 3 ed a processor that automatically extracts the “Active Frames” and drops the video is skipped or repeated frames. The Inactive Frames that are simply repeats of Display rate could be 30-frames with over or under crank speeds above or below 30. The visual Irregular "Pull Down" effects are the same, a uniquely created affect that A A A A B B B C C C creative video producers will 60 fps find very useful. The Panasonic processor ingests 720p60 video with User Bits over a standard HD-SDI or SDTI input. The SDTI input can be used to directly record the compressed HD video from the VTR without decompression. Recording compressed images prevents any unnecessary compression losses from decompression and recompression. The images are stored on disc in their original compressed state exactly as recorded in the camcorder. The output video can also be transferred over SDTI or SDI if desired.

X fps

A

B

C

Figure 2

Active Frames. Active Frames are marked with the User Bits as previously described. If the video frames are not extracted and assembled into new frame rate stream with sped or slowed motion, then a different and unique visual effect is created. The actual visual effect from over or under cranked images can be seen as a motion effect that visually offers another oppor-

-4-

Only original video frames are stored on disc, thereby reducing the cost of storing high definition footage on the processor’s hard discs. Any repeated frames necessary to produce the output video can be added at the output without the processor storing repeated frames.

ed that detects the User Bit coding and extracts the desired active frames then builds them into an AVI file for play out or editing at the required frame rate, 24, 30 or 60.

Twenty-four frames will be the most common frame rate used for the over/undercrank video. The processor will record each marked Active The Panasonic processor also operates at 30Frame in an AVI file on the internal hard discs. frames. Editing video originated at 24-frames in After the video file is built, it will be read out at the a 24-frame system is highly desired because desired frame rate with appropriate frame you avoid all the problems inherent in the editing repeats for the output video frame being selected. process if a pull-down conversion is used. Once the video is played out at the user’s choice of If over/under crank is not used, the process of frame rates and image formats, the editing and recovering 24-frame video can be real-time post process becomes very conventional. The recorded in real camera running time on an exterat higher or lower nal VTR or speeds than the Variable Frame Rate Camera recorded to the final video frame internal disc rate has no affect drives for later on the edit and play out. The conpost production Recording on standard DVCPRO HD 720p60 video tape version process process. can be accomplished in real This is the dawn 1080p24 or 1080i time to 1080p24, of new era in cinSDTI HD SDI 1080i30 or ematography, an 720p60 output AJ-HD150 DVCPRO HD VTR era that utilizes Frame Rate Converter AJ-HD3700 D-5 HD VTR video. These are both film and standard HD electronic camD-5 HD master tape 24p or video formats that eras. Now, high other HD/SD format can be recorded definition televion a D-5HD massion cameras can Figure 4 ter videotape or provided cinesent directly into a post-production editing matographers with the versatility they were process. An example of a basic system using accustom to in film, but now the Panasonic AJthe processor is shown in Figure 4. HDC27V, with its variable frame rate capabilities, provides them the invaluable flexibility and Editing of Variable Frame Rate Video creative freedom they desire. The process of extracting the marked frames can also be accomplished using a non-linear editing system. Again, frames that will compose the final stream will be extracted from the 720p60 videotape recorded in the variable frame rate camcorder. Once the video is stored in the editing system, a simple process is need-

-5-