Digital Video Processing: Sampling, Compression, and Sensors

Signal Sampling and Digitization

Signal A is continuous in time; therefore, digitization must occur at fixed intervals. According to the Nyquist-Shannon sampling theorem, the signal must be sampled at a frequency at least twice the maximum frequency of the signal.

DCT: Discrete Cosine Transform

The Discrete Cosine Transform (DCT) is a mathematical operation used for image compression by decomposing coefficients. In this process, redundant points are reduced toward zero.

IT Sensors: CCD Technology

The CCD (Charge-Coupled Device) is a type of Interline Transfer sensor. These image sensors combine vertical columns with transfer registers to capture light efficiently.

Gamma Correction

Cathode ray tubes (CRT) in TV receivers do not exhibit a linear relationship between input voltage and output luminance. Consequently, cameras apply Gamma Correction to compensate for this non-linear gap.

Vectorscope Analysis

The Vectorscope is a measurement tool for TV signals used to analyze chrominance components (U and V). It measures:

• Chrominance levels
• Phase and amplitude of colors
• Gain and phase adjustments

It provides a visual representation of color bars and phase modulation, where the vector length indicates saturation and the angle indicates hue.

Digital Quantization

During video signal digitization, each sample is assigned a finite memory capacity based on the number of bits on the Y-axis. These values represent specific tones. The difference between the input voltage and the output voltage creates quantization error. Increasing the bit depth reduces this error.

Video Compression: FBS and Frame Types

For transmission error detection and correction, images are divided into macro-blocks. Synchronization data and error correction bits are added to create digital frames. Compression is achieved using three frame types:

• I-frames (Intra-coded): Reference images that provide the basis for compression.
• P-frames (Predicted): Contain information based on previous I or P frames, allowing for higher compression.
• B-frames (Bi-directional): Formed by comparing information from both preceding and following reference images, providing the highest level of compression.