<<
Home , 24p

(HDVP I Misinformation ) Pulldown Revisited: 60i Or ? t e u

Why 24 fps from 60i just isn’t the same as from o l l i a

By C.R. Caillouet C 24p native . R . C

If a camera such as Panasonic’s This sequence of images illustrates the AJ-HDC27 VariCam records at 60 fps, results of 60 fps and 24 fps capture and playback why should I capture at 24 fps? Surely variations. The image at top is the test object at I can extract 24 from 60 and keep that rest, a small airplace attached to a pendulum, 60 fps original in case I need footage captured as it travelled across the screen . All in slo-mo, right? other images are composites of frame sequences. Yes, you could do that—but you There is some variation in the width of the may not really want to. pendulum bar in the capture, but the true 24 fps The answer here requires us to capture shows distinctly wider bars because of bend our minds a bit around what’s additional motion blur. The second and third really happening when we capture images show 60 and 24 fps capture. and view images. It also requires us to The fourth image shows the irregular playback think about human conditioning and caused by 24->60 conversion using 2.3 pull-down. adaptation. The fifth image shows the additional irregu - When we capture in a camera, we larity that results from a 24 fps extraction from set the to give us a consis - 60 fps original footage, followed by a 2.3 pull- tent time between frame captures, down conversion back to 60 for standard televi - sion distribution. This does not take into account then we set the exposure time with any problems from the use of . the shutter to control the amount of time of the capture. The shutter-open time can vary would blur across half the distance, from the fastest time that the camera then jump to its starting position in the can handle to very close to the time next frame. (The exposure time can between frames. happen at the beginning or end of the We can select a value between the frame time but the effect is the same.) two shutter limits based on the amount The jump is referred to as motion of exposure time, the amount of blur in strobing and is more noticeable for the image and the amount of strobing shorter exposure times at the same that we desire in moving objects. frame rate (narrow shutter angle in The range of impressions from terms) and for faster panning maximum blur to maximum strobing speeds. As it becomes more notice - allows a filmmaker to affect the emo - able, it’s often referred to as stutter. tions of a viewer. In this manner, one Motion strobing (or stutter) is a peri - can emphasize motion with blur or odic effect that’s regular in occurrence. make the viewer uncomfortable with I’ll use the term “judder” to refer to a lots of strobing. type of strobing that’s not necessarily I guess this is as good a time as any regular. We’ll get there in a minute. to clarify some terminology. When we play back captured Motion strobing is the effect images, we can play them back at the caused by limiting the shutter open - original frame rate or we can duplicate ing time to less than the entire frame the frames and flash the same image time. Think about a camera panning (Cont’d on page 92) across a flagpole. If the shutter were open for the length of a frame, which “Misinformation” is a joint effort between HDVideoPro and the Sachtler Academy . a video camera can come very close The Sachtler Academy is dedicated to promoting open knowledge exchange to, then the pole would blur across among production professionals worldwide. Initiated by renowned camera support the entire frame. manufacturer Sachtler, the Academy offers a nonpartisan venue by which cinematographers and videographers can hone their talents, discuss techniques If the shutter were only open for and stay updated on technical advances from various manufacturers. To find out half of the frame time, then the pole more, visit www.sachtler-academy.com/ and www.sachtler.us .

August 2008 I 25 MISINFORMATION (Cont’d from page 25) each vertical pass), then some of the video frames contained complete film two or three times for one frame time. frames and some contained portions of The reason for repeating frames has two adjacent film frames. to do with two reactions in our brains: If the viewer saw the video sequence First, there is a minimum played back with the two/three cadence, required to convince us that we are see - then motion in the film scene was mod - ing continuous motion. Depending on ified by a new irregular judder that the type of motion, that minimum rate occurred 12 times a second, causing the might be between 12 and 24 frames per flag pole in the above example to alter - second (fps). nately move normally and slow down. The minimum rate doesn’t necessari - Again, human adaptation plays a role ly impart a feeling of realism to us, but in our perception of the judder effect. it does make us think that the object Those of us who grew up watching film motion is believable. A higher frame transfers to 30-frame video think that rate might give us a more realistic judder is normal for on video. But impression, but 24 frames works for those folks who grew up in 50 fps coun - many applications. tries find pulldown annoying. There’s also a limit called the flicker So what does this have to do with fusion threshold. If still pictures are shooting at 60 fps and extracting 24 flashed on a screen and removed quick - frames in each second? The short answer ly, we perceive the flicker up to 50 or 60 is that none of us are adapted to watching flashes per second. Our perception of motion captured and displayed that way. flicker is affected by conditioning, so In the pulldown discussion above, the people who grow up in 50 fps countries original frames were captured at even tend not to notice flicker in 50 image- intervals and then displayed at uneven per-second sequences, whereas those of intervals, and we eventually got used to it, us in 60 fps countries are bothered by 50 so it doesn’t affect, at least, some of us. flashes per second. Flicker is also more But if we capture a sequence at 60 fps noticeable at the higher brightness of and alternately drop one or two frames television displays, compared to film. to get to 24 frames a second, we’ll have Film projectors deal with the discrep - effectively captured at uneven intervals. ancy between the minimum acceptable If we then play that sequence back at frame rate and the flicker threshold by a multiple of 24 fps, we’ll preserve the doubling or tripling the display rate uneven capture interval. If we display it with repeated frames. as a pulldown stream, we introduce dif - Repeated frames are regular effects; ferent irregularities at capture and dis - they don’t affect the spacing of each play, thereby causing multiple judder unique frame in time. But if we try to con - effects to be cascaded. vert from film’s 24 frames to video’s 60 In either case, we introduce motion fields or frames, we either have to do some artifacts that no one is adapted to, and mighty sophisticated motion analysis and the results on viewer emotions are prediction or we need to use a cadence or unpredictable. The effects are more evi - irregular sequence of repeated images. dent on big screens. The technique of pulldown was In addition, capturing at 60 fps lim - 1 developed in film-to-video transfer its us to ⁄60 -second shutter. If we were devices in the early days of television. A shooting at 24 with a 180 degree shut - 1 television camera scanned a ter, we would have ⁄48 second. Many alternately for two or three vertical shooters are accustomed to stretching scans, then the film was “pulled down” the shutter to above 200 degrees (longer 1 during a blanking period, and the next than ⁄48 sec) to compensate for the film frame was exposed to the camera. absolutely crisp all-on/all-off shutter of 1 The result was a continual video the video camera. Going to ⁄60 sec takes sequence of first two and then three you in the other direction (144 degrees.) 1 repeated images of the film frames. If the And you lose at least ⁄3 stop. video system used interlaced frames (half Bottom line? Just because you can, of the lines in the image are scanned in doesn’t mean that you should! HDVP

92 I HD VideoPro