Film - A Practical Guide to Video and Audio Compression - page 46

The width, height, and color depth dictate how much information you have to deal
with per frame. Some of your pre-compression processing, such as cropping and scaling,
is going to be based on what you know about these values.
The frames per second (fps) value tells you whether any time-based correction or
pulldown is necessary to introduce additional frames for your target output platform.
The scanning method indicates the complexity of any motion-based processing. Raw
film can be converted to interlaced or progressive scanning quite easily. Video originates in
an interlaced format, and that does not conveniently convert to progressive scanning with-
out some complex interpolation. This interpolation is difficult because it takes place in
space (X and Y) and on the time axis (between successive fields). The interpolation is much
easier if the video was created from a film format with a telecine process. This will become
apparent after the discussion of how pulldown works.
Film formats have picture information at the highest resolution. Frame rates are
generally stable across all sizes at 24 fps. Some older home-movie footage may have been
shot at 18 fps to economize on film stock. This text will work down from the highest qual-
ity to the lowest, making some simple calculations about potential digitized image sizes
on the way.

4.5 Film Grain and Scanning Resolution

During the research for this topic, a number of conflicting dimensions for film formats
were discovered. Film grain is such a variable attribute that scanning resolutions will vary
widely.
You may choose to scan at a higher or lower resolution in your own projects. Simply
recalculate the storage density equations, which are straightforward. Here is an example:

Given:
W = Imaged area width in mm.
H = Imaged area height in mm.
D = Dots-per-millimeter (dpmm) resolution selected.
F = Frame rate in fps.
T = Duration of the movie in minutes.
Z = Bytes per pixel (RGB 8 bit = 3).
GigaBytes per movie:
Disk space required = W * H * Z * F * T * 60 * D 2 / 2 30

Having decided on a scanning resolution, the frame size in pixels is then computed.
The storage sizes for our capacity planning are derived from those values.
A resolution of approximately 115 dots per millimeter (dpmm) is considered a rea-
sonable compromise for a maximum scanning density for the purposes of calculating
image sizes. This is why:
An IMAX movie was produced by the Massachusetts Institute of Technology in 1996
and worked on at a resolution of 4K in the horizontal axis. This was considered to be an