Sight vs. Sound

The Core Challenge of Film Sound Work


Film sound preservation is different from film image preservation in one very important way:  the image on the film is comprised of discrete still images that, when used along with a shutter, simply fool the brain into perceiving smooth motion.  However, the audio on the film is continuous and requires continuous movement to properly convey the same level of realism and quality as the picture.  This difference between motion picture images and sound can be seen in the above picture of a composite print.  

Just continuous movement is not enough for good sound, though.  While the eye/brain combination has a very coarse perception of images over time (24 frames per second with 48Hz shutter is sufficient), the ear/brain combination is mind-bogglingly sensitive to sound in the time domain.   This extreme sensitivity allows the enjoyment of high fidelity sound, but its real purpose is to enable humans to accurately locate the position of sounds around us.  The more accurately we can pinpoint danger, the more accurately we can plot a course to run in the opposite direction.  

As a result of our acute sensitivity to auditory timing cues, high quality sound reproduction must be done with extremely steady fluid motion.  The importance of precise and steady movement is a cornerstone of good sound quality in any format, and steady movement has been the subject of much of the R&D on sound transports over the last century.     Slow speed errors are be heard as wow, slightly higher frequencies are heard as flutter, and the highest frequency disturbances are heard as “harshness” or “roughness” in the audio.  Our ears are so sensitive that they can even pick up on delays between frequencies in program material.  However, this type of error is caused by audio electronics design rather than mechanical transports.

Since high-quality sound necessitates real-time transfer with specialized transports, sound requires a very different and more demanding approach to motion than picture.  Picture capture and restoration is simplified by that fact that work can be done frame-by-frame in steps. Although it is now possible in the digital domain to stitch together digital “snapshots” of a sound track as is occasionally done with soundtrack scans from a picture scanner, this is done as a convenience and is not an acceptable solution for high-quality sound reproduction.  In fact, professional projector soundheads from the early 1930s can typically offer better speed consistency than this high-tech stitching approach. 

Perhaps the most important point to make about the continuous nature of film sound is that it affects many other seemingly unrated sound quality variables.  For example, cleaning a few frames of mag track so it looks good to the eye is not terribly difficult, but cleaning an entire continuous roll so that there are no friction differences that will affect the smoothness of motion is a wildly more complicated endeavor. Additionally, variables like the perf-to-perf relationship of splices, which are not particularly important in modern picture scanning, are critical for smooth real-time transfer of motion picture sound elements.  As specific problems with sound elements are discussed here and elsewhere, always keep in mind that the importance of continuous smooth motion of sound adds a layer of complexity to deterioration solutions that can be largely sidestepped with similar picture solutions.