User's Manual

GUIDE TO INSTALLATION AND OPERATION
6 | JAZZ-800
1.3 Applications
Up-/Down-conversion
In situations where the highest image quality up- or down-conversion is necessary, Miranda’s JAZZ-800 uses
proprietary PixelMotion™ De-interlacing, 3:2 Detection, and Detail Enhancement combined with a powerful
array processing module, provides the best image solution.
The first step in the conversion process is to identify whether the input material is video or film originated.
The identification process happens in a fully automatic mode and selects either PixelMotion De-interlacing for
video based material or 3:2 Detection for film based material. The goal is to apply the appropriate filter in
order to recover the full vertical resolution of the input material.
PixelMotion de-interlacing produces perfect progressive frames in preparation for further processing. The
processing aperture is adjusted on a pixel-by-pixel basis, which preserves all of the detail of the original
interlaced image. Additionally, the filter eliminates “jaggies” in the output image, providing well-defined edges
on objects and producing the sharpest images possible.
3:2 Detection recognizes the redundant fields inserted by the telecine during the conversion of film to video.
This advanced 3:2 pull-down filter avoids frame rate conversion artifacts and provides the highest vertical
resolution and motion quality.
Since the quality of the video de-interlacing is so high, there will be no difference visible in the vertical
resolution seen with film originated material and that of the de-interlaced video originated material.
Once the image has been de-interlaced and up-converted, Detail Enhancement can be applied to the image
to further sharpen and enhance the output. This filter adds an additional level of image quality by helping to
better define the detail in the up-converted image. This is achieved by detecting the edges of objects and
adjusting the contrast ratio around those objects to help separate them from the background.
Frame-rate Conversion
Moving images exist in three dimensions. Firstly, in the horizontal direction, images are made up of
individual pixels. Secondly, in the vertical direction, they are made up of the lines contained in the field or
frame. The combination is referred to as the spatial domain. Thirdly, the motion image is composed of a
number of fields or frames per second, which is referred to as the temporal domain.
The process of frame-rate, or standards, conversion is a form of sample rate conversion in two or three of the
above dimensions. It consists of expressing moving images sampled on one three-dimensional sampling
lattice to a different three-dimensional lattice.
A process called interpolation is used to convert between these various spaces. Interpolation is defined as
computing the value of a sample, or samples, which lie outside the sampling matrix of the source signal. In
other words, it is the process of computing the values of output samples that lie between the input samples.
Per-pixel processing allows the processing aperture to be adjusted on a pixel-by-pixel basis, which preserves
all of the detail of the original image.