The Model 448S Video Frame Averager, Video Frame Grabber or Video Freeze Frame creates a freeze frame picture image, upon request, representing the average of a selected number of subsequent frames. This function is most commonly used to reduce dynamic noise, such as "snow", in a scene where there is relatively little or no motion.

In its "live" mode, the input signal will pass to the output. When the Frame Averager is flipped to its "average" mode and its "start" button is pressed, it will grab or capture a selected number of subsequent frames or fields of video from the input and provide their average value at its output continually until it is given a command to average a new series of frames. The number of frames is selectable in powers of 2 by a series of four switches on the unit's front panel. In "Frame" Mode, resolution is 720H X 480V pixels at 29.97 frames per second. In "Field" Mode, resolution is 720H x 240V pixels at 59.94 fields per second. Video inputs to the Video Frame Averager from analog tape must be time base corrected.

The Colorado Video Model 449 Dynamic Video Noise Filter / Reduction System or Video Noise Reducer can filter or reduce dynamic video noise (like "snow" or "static") out of a monochrome video signal by essentially averaging frames over time. This type of video noise filtering or reduction is best suited for signals with rapidly changing noise in pictures of relatively static (unchanging) subjects.

The Model 449 Dynamic Video Noise Reduction System provides a running quasi-average, or temporal filtering, of the video signal at its input. More precisely, the value for any given pixel in the unit’s output is determined by the equation:

where Pn is the value of the pixel in the current output frame, Pi is the value of the pixel in the current input frame, Pn-1 is the value of the pixel in the previous output frame, and t is a user selected time constant. This equation is solved for each pixel at the normal video frame rate of 30 times per second. As the user selects larger values for t, the exponentially decaying “influence” of a given pixel’s value in a particular input frame will persist over a longer period of time in subsequent output frames. The Model 449 allows values of t to range from zero (binary 0000) to 15 (binary 1111), making possible temporal filtering of from one to 32,768 (in steps of powers of two) frames on a running basis. The Model 449’s 16-bit grayscale resolution provides 65,536 shades of gray for the filtering process.