All video and indeed movie film is based upon a conjuring trick – an optical illusion that makes our brain think that a series of still pictures represents a moving scene. It is helped to this end by two physiological traits that we all have.
The first is technically called the flicker fusion threshold (FTT) and represents the frequency at which we stop seeing a flashing light as flicker and instead see it as a steady illumination. The actual threshold rate for an individual depends on many factors, principally the environment, the nature of the source, the angle of view and the experience and fitness of the subject. For most subjects it will be in the range of 30-40 Hz, but could be much higher in certain circumstances. The average for the population is known as the Median rate. It is fair to say that a great deal is known about FTT since it is used widely in physiological experiments (presumably it is easy to measure) but it is also fair to say that very little is known about the actual mechanism other than it is directly related to persistance of vision and has some obvious evolutionary connections like the ability to see movement in our peripheral vision. Basically – we just make use of it and live with the consequences. Unfortunately because the effect is so variable, there will be circumstances where the resultant illusion partially breaks down. This is inevitable and we have to accept that it is an imperfect ‘art’ or ‘science’.
At least the FTT is a relative ‘constant’ compared to our ability to compute a series of still images into a moving scene. This ability is far more complex and variable, and directly related to the actual material concerned. We can compute a perfectly satisfactory moving scene with as little as 10fps whereas a perception of ‘judder’ can persist right up to the FTT and even beyond with the ‘wrong’ material.
Historically, in an attempt to minimise film stock costs, cinematographers used hand cranked speeds related to the subject material. Faster action required ‘over-cranking’ to smoothly capture the motion. Typical speeds were about 16-18fps, which was fine for most scenes (and was also used on later domestic formats). It was the advent of the ‘talkies’ in the ‘20s that raised the frame rate to 24fps in order to improve the sound quality. This rate remains the same to this day of course.
Projecting movie film at 24fps results in unacceptable flicker, so each frame is shown twice or three times by using a shutter to increase the effective flicker to 48 or 72Hz. The former is rather marginal for the general population which is why low cost cinemas (typically using 48Hz) were called ‘The Flicks’ in my youth!
So to re-iterate, shooting at 24fps gives acceptably smooth motion for much of the population, projecting at 48 or even better 72Hz gives acceptably low flicker. To emphasise – SOME folks (like me) will still see juddery movement and SOME folks will still see flicker. These speeds are directly related to media costs and are necessarily a compromise. Such compromises are enacted throughout the industry and continue very actively in the development of video. The cost of providing a particular media system is a very important factor!
And so to the invention of television. For several reasons it made sense to make the shooting rate the same as the mains electricity, namely 60Hz in USA and 50Hz in Europe; this was once the real practical systems evolved after the 2nd World War. These rates gave both satisfactory flicker and movement performance.The problem was that with enough ‘lines’ to give acceptable resolution (we’ll look into this later), the 50/60fps shooting rates took unacceptable amounts of ‘bandwidth’ or media space and therefore cost. An engineer at RCA, one of the principle developers of television, came up with the idea of interlace. This basically involves halving the line rate per frame, but vertically offsetting alternate scans so that the resultant frame lines form a combined image that has the full resolution. The two scans are known as fields making up a frame. Fortunately, we are not very sensitive to resolution on moving images, helped by the motion blur imposed by most image sensors. Critical viewers will notice more stair-stepping type ‘aliasing’ on sloping edges, particularly at high contrast, but most viewers will not be bothered by this. Perhaps more noticeable is that any extreme sharpness in the vertical direction will result in a big difference at edges in the content of the two fields, and this difference will be presented at the frame rate of 25/30Hz which will be seen as flicker. The effect is known as interlace flicker , also known amongst engineers as twitter, and due to the nature of camera design it has not presented a big problem with natural pictures but plays havoc with graphics and captions, but can be overcome with suitable design and filtering. For most viewers, interlace is a fairly transparent technique that has served the industry well.
Motion film has always been (of course) a whole frame system, and when translated to video (in a Telecine machine) it produces (in PAL) what we call a progressive scan where the two fields of a frame are shot at the same time. In NTSC it is more complicated because the film has to be converted to 30fps by using 3:2 pull-down. Either way, the result is a less smooth motion rendering without any of the artifacts of interlace. A lot of production people like to shoot video with a similar ‘look’, probably more in a hope of association with Hollywood production values rather than any real technical added value. Cameras that shoot in ‘progressive’ mode, have thus become popular.
