Electronics Guide

Bandwidth and Gain Requirements

The required bandwidth depends on the video format being processed. Standard definition video requires bandwidth of approximately 5-6 MHz, while high-definition formats demand 30 MHz or more. The relationship between bandwidth and resolution follows from the need to reproduce rapid transitions at the edges of fine picture details.

Key specifications for video amplifiers include:

• Flat frequency response: Gain variation should be less than 0.1 dB across the video bandwidth to prevent brightness variations with picture content
• Linear phase response: Group delay variation causes different frequency components to arrive at different times, smearing edges and causing color fringing
• Low differential gain: Gain that varies with signal level causes color saturation to change with brightness
• Low differential phase: Phase shift that varies with signal level causes hue errors that depend on brightness

Video Buffer Design

Video buffers provide impedance transformation and signal isolation without voltage gain. They typically drive 75-ohm transmission lines used in video systems and must handle the full video bandwidth with minimal distortion.

Common buffer topologies include:

• Emitter follower: Simple and fast, but limited output swing and potential for crossover distortion in Class AB configurations
• Diamond buffer: Complementary emitter followers providing rail-to-rail output capability with improved linearity
• Current feedback amplifiers: Offer high slew rate and bandwidth independent of gain, ideal for video distribution
• Integrated video drivers: Purpose-built ICs with back-termination resistors, output clamping, and sync-tip clamping

Cable driving requires source termination to prevent reflections. A series resistor equal to the line impedance minus the driver output impedance provides proper matching. For 75-ohm systems with low-impedance drivers, a 75-ohm back-termination resistor is used, resulting in 6 dB signal loss that must be accounted for in gain calculations.

High-Frequency Compensation

Video amplifiers often require peaking networks to extend bandwidth and compensate for cable and load capacitance:

• Series peaking: An inductor in series with the output isolates load capacitance, extending bandwidth by approximately 40%
• Shunt peaking: An inductor in parallel with the collector resistor creates a resonance that boosts high-frequency gain
• T-coil compensation: Mutually coupled inductors provide bandwidth extension factors of 2-2.5 with flat response

Cable equalization compensates for the frequency-dependent attenuation of coaxial cables. Longer cable runs require more aggressive high-frequency boost. Active equalizers adjust their response based on cable length, often using pilot signals or adaptive algorithms.

Composite Video Signal Structure

Understanding sync separation requires familiarity with composite video signal structure:

• Sync tip: The lowest voltage level, providing the timing reference for horizontal and vertical synchronization
• Blanking level: The reference black level, typically 0 IRE in NTSC or 0 mV in PAL/SECAM
• Black level: In NTSC, slightly above blanking (7.5 IRE setup); in PAL, coincident with blanking
• White level: Maximum picture brightness, typically 100 IRE or 700 mV above blanking
• Color burst: A short burst of subcarrier reference on the back porch following each horizontal sync pulse

Sync Separator Circuits

Sync separators use amplitude discrimination to extract sync pulses from the composite signal. Since sync tips represent the lowest voltage level, a simple slicer can separate them from picture content.

A basic sync separator consists of:

• Input amplifier: Provides signal conditioning and level shifting
• Sync slicer: A comparator that detects when the signal falls below the slicing level
• Slicing level generator: Typically uses a capacitor charged through a diode to track the sync tip level
• Output driver: Provides appropriate logic levels for downstream processing

The slicing level must track signal amplitude variations while maintaining a consistent relationship to the sync tips. Common circuits use a peak detector with controlled decay to establish a reference approximately 50% between sync tip and blanking level.

Horizontal and Vertical Sync Separation

After extracting composite sync, further processing separates horizontal and vertical timing:

• Horizontal sync: Occurs at the line rate (15.734 kHz for NTSC, 15.625 kHz for PAL) with pulse widths of 4-5 microseconds
• Vertical sync: Occurs at the field rate (59.94 Hz for NTSC, 50 Hz for PAL) with pulse widths of approximately 27 microseconds, distinguished by their longer duration

Separation techniques include:

• Integrator: A low-pass filter that responds only to the longer vertical pulses, producing a pulse when vertical sync is present
• Differentiator: Extracts horizontal sync transitions; used with digital processing to count lines and identify vertical intervals
• Digital counting: Modern ICs count horizontal pulses and identify the vertical interval by the absence of equalizing pulse patterns

Sync Generation

Sync generators create the timing signals for cameras and other video sources. They must produce signals conforming precisely to broadcast standards:

• Master oscillator: Crystal-controlled at a multiple of the horizontal frequency, phase-locked to color subcarrier frequency
• Divider chain: Produces horizontal and vertical timing from the master oscillator
• Pulse shapers: Generate sync, blanking, and burst gate signals with precise timing relationships
• Genlock input: Allows synchronization to external reference, essential for multi-camera production

Modern sync generators typically use dedicated ICs or FPGAs that can be programmed for multiple video standards and provide both analog and digital timing outputs.

Color Encoding Fundamentals

The NTSC, PAL, and SECAM systems encode color differently, but all rely on separating luminance (Y) from chrominance (C) information:

• NTSC: Uses quadrature modulation of a 3.579545 MHz subcarrier with I and Q color difference signals
• PAL: Similar to NTSC but alternates the phase of one color component on successive lines (4.43361875 MHz subcarrier)
• SECAM: Uses frequency modulation of two different subcarriers on alternate lines

Component video systems (Y/Pb/Pr or Y/Cb/Cr) keep luminance and color difference signals separate, avoiding the encoding and decoding artifacts of composite systems.

Comb Filtering

Comb filters exploit the relationship between color subcarrier and horizontal scanning frequency to separate luminance and chrominance without the bandwidth limitations of simple filtering:

• 1-H comb: Uses one horizontal line of delay; subtracting adjacent lines cancels luminance (which is correlated between lines) and enhances chrominance (which inverts phase)
• 2-H comb: Uses two delay lines to provide better separation by averaging over three lines
• 3D comb: Extends the technique across multiple frames, providing superior separation but requiring motion detection to avoid artifacts

Analog comb filters traditionally used glass or CCD delay lines to provide the 63.5-microsecond (NTSC) or 64-microsecond (PAL) horizontal delay. Modern designs use digital processing with A/D conversion, digital delay, and D/A conversion.

Chrominance Demodulation

Extracting color information from the chrominance signal requires synchronous demodulation using a regenerated subcarrier reference:

• Burst gate: Extracts the color burst reference from the back porch of each horizontal line
• Subcarrier regenerator: A PLL that locks to the burst and generates a continuous subcarrier reference
• Quadrature demodulators: Multiply the chrominance signal by in-phase and quadrature references to extract the two color difference components
• Low-pass filters: Remove the high-frequency products of demodulation, recovering the baseband color signals

The demodulator phase must be precisely aligned with the encoding phase to recover correct hue. User-adjustable hue controls typically shift the demodulator reference phase.

Chrominance Modulation

Encoding color information reverses the demodulation process:

• Matrix: Converts RGB inputs to luminance (Y) and color difference (B-Y, R-Y or Pb, Pr) signals
• Low-pass filters: Bandwidth-limit the color difference signals according to standard requirements
• Balanced modulators: Multiply each color difference signal by in-phase or quadrature subcarrier
• Combiner: Adds the modulated subcarrier to the luminance signal along with sync and burst

Luminance Processing

Luminance signal processing optimizes picture quality:

• Aperture correction: Boosts high-frequency luminance components to compensate for optical and sensor MTF losses, sharpening edges
• Black level clamping: Restores correct DC level after AC-coupled stages by referencing to a known black level during blanking
• Contrast and brightness adjustment: Gain and offset controls that adjust the luminance transfer characteristic
• Detail enhancement: Creates artificial edge enhancement by adding a high-pass filtered version of the signal

Display Transfer Characteristics

CRT displays exhibit a power-law relationship between input voltage and light output:

Light output = k * (Voltage)^gamma

The gamma exponent for typical CRTs is approximately 2.2-2.5. Modern flat-panel displays are inherently linear but include gamma correction to match the system standard and maintain compatibility with legacy content.

System Gamma

The overall system gamma is the product of camera gamma, transmission gamma, and display gamma:

• Camera gamma: Cameras apply a compressive transfer function (gamma less than 1) to pre-correct for display nonlinearity
• Standard gamma: Television standards specify camera gamma of approximately 0.45 (the reciprocal of 2.2)
• End-to-end gamma: The product is intentionally set slightly greater than 1.0 (typically 1.1-1.2) to provide pleasing contrast in normal viewing conditions

Gamma Correction Circuits

Analog gamma correction uses nonlinear transfer functions to implement the required power law:

• Diode-resistor networks: Multiple diodes with series resistors create a piecewise-linear approximation to the gamma curve
• Logarithmic amplifiers: The exponential characteristic of transistor junctions can approximate gamma curves
• Feedback-based shapers: Nonlinear elements in the feedback path of an amplifier create controlled transfer functions
• Multiplying DACs: Digital lookup tables drive DACs to create arbitrary transfer functions

Modern implementations typically use digital lookup tables (LUTs) that can be programmed for any desired gamma curve, including user-adjustable settings and different curves for different content types.

Gamma and Color Processing Interaction

Gamma correction affects color processing in important ways:

• Order of operations: Gamma correction should be applied to linear RGB signals before color encoding to maintain correct color rendering
• Constant luminance: Traditional encoding violates constant luminance principles because gamma correction is applied before forming luminance, causing errors with highly saturated colors
• Modern formats: Newer standards like BT.2020 specify constant luminance encoding where gamma correction follows luma/chroma separation

CCD Sensor Architecture

Charge-Coupled Device (CCD) sensors accumulate photogenerated charge in potential wells and transfer it through the device to a common output amplifier:

• Photodiode array: Photosensitive elements collect charge proportional to incident light intensity during the exposure period
• Charge transfer registers: Vertical and horizontal shift registers move charge packets toward the output
• Output amplifier: A single on-chip amplifier converts charge to voltage for each pixel sequentially
• Clock drivers: Multi-phase clocks with precise timing control charge transfer through the device

CMOS Sensor Architecture

CMOS Active Pixel Sensors (APS) include amplification within each pixel, enabling parallel readout and integrated functionality:

• Pixel structure: Each pixel contains a photodiode and typically 3-4 transistors for reset, amplification, and row selection
• Column parallel readout: All pixels in a row are read simultaneously through column amplifiers
• On-chip ADC: Many CMOS sensors include column-parallel or chip-level analog-to-digital converters
• Digital output: Modern CMOS sensors often provide fully digital output, with all analog processing integrated on-chip

Analog Signal Chain

The analog signal chain from sensor to ADC typically includes:

• Sensor output buffer: Provides low-impedance drive for off-chip connection if needed
• Correlated double sampling: Reduces reset noise and fixed pattern noise (discussed in detail below)
• Programmable gain amplifier: Adjusts signal level for different exposure conditions
• Black level clamp: Establishes correct reference level using optically dark pixels
• Analog-to-digital converter: Digitizes the conditioned signal for further processing

Clock and Timing Requirements

Image sensors require precisely controlled timing signals:

• Pixel clock: Determines readout rate, typically 10-100 MHz for video-rate sensors
• Horizontal timing: Controls row selection and horizontal blanking intervals
• Vertical timing: Controls frame rate and vertical blanking for readout of stored charge
• Exposure control: Electronic shutter timing determines integration period

CCD sensors require multiple overlapping clock phases at voltages often exceeding 10V, generated by dedicated clock driver ICs or discrete high-current switches.

CDS Operating Principle

CDS exploits the correlation between noise components in two temporally close samples:

• Reset sample: Captures the pixel output immediately after reset, containing reset level plus reset noise
• Signal sample: Captures the pixel output after charge transfer, containing signal plus the same reset noise
• Difference: Subtracting the reset sample from the signal sample cancels the common reset noise, yielding only the signal

The technique is effective against:

• kTC reset noise: Thermal noise on the sense node capacitance after reset
• 1/f noise: Low-frequency noise from the output amplifier
• Fixed pattern noise: Pixel-to-pixel offset variations (partially)
• Power supply variations: Slow drifts common to both samples

CDS Circuit Implementations

Several circuit topologies implement CDS:

• Sample and hold: Two sample-and-hold circuits capture reset and signal samples; a differential amplifier computes the difference
• Clamping: A clamp capacitor is reset to a reference during the reset interval, then follows the signal during readout
• Switched capacitor: Capacitors alternately sample reset and signal levels, with charge redistribution computing the difference
• Digital CDS: Both samples are digitized separately, with subtraction performed in the digital domain for maximum flexibility

Noise Analysis

CDS noise performance depends on the sampling interval and noise spectrum:

• White noise: CDS increases white noise by a factor of square root of 2 due to uncorrelated samples
• 1/f noise: CDS provides substantial rejection of 1/f noise, with rejection improving as the correlation interval decreases
• Optimal sampling: The sampling interval should be long enough for complete signal settling but short enough to maintain correlation

The noise transfer function of CDS has a high-pass characteristic, rejecting low-frequency noise while passing high-frequency noise unchanged.

Advanced CDS Techniques

• Multiple sampling: Averaging multiple signal samples improves SNR by the square root of the number of samples
• Delta-sigma CDS: Combines CDS with oversampled A/D conversion for low-noise digital output
• Fowler sampling: Takes multiple samples during reset and signal phases, useful for very low noise applications

Exposure Parameters

Three primary parameters control image exposure:

• Aperture (f-stop): Controls the amount of light reaching the sensor; each stop represents a factor of 2 in light level
• Shutter time (integration time): Duration of light accumulation; longer times increase exposure but may cause motion blur
• Gain (ISO): Electronic amplification of the sensor signal; higher gain increases brightness but also amplifies noise

These parameters trade off against each other and against image quality factors, requiring intelligent control algorithms to optimize the result.

Light Metering

AEC systems measure scene brightness using various metering strategies:

• Center-weighted average: Emphasizes the center of the frame, assuming the main subject is centered
• Matrix/evaluative metering: Divides the frame into zones, analyzing patterns to identify the subject and set appropriate exposure
• Spot metering: Measures a small area, typically the center, for precise control in difficult lighting
• Highlight priority: Exposes to preserve detail in the brightest areas, preventing highlight clipping

Modern cameras often use both luminance and color information from the image sensor itself, analyzing pixel statistics to determine optimal exposure.

Control Loop Design

AEC systems typically implement closed-loop control:

• Target brightness: A reference level representing the desired average or weighted image brightness
• Error calculation: Compares measured brightness to target, computing required adjustment
• Loop filter: Smooths adjustments to prevent hunting and provide appropriate attack/decay behavior
• Parameter allocation: Distributes required adjustment among aperture, shutter, and gain according to priority rules

The control loop must balance responsiveness with stability. Fast response handles rapid lighting changes but may cause visible pumping on normal content. Hysteresis prevents oscillation at boundaries between exposure settings.

High Dynamic Range Considerations

Scenes often exceed the dynamic range of the sensor, requiring exposure trade-offs:

• Highlight headroom: Some exposure latitude is reserved to handle specular highlights without clipping
• Shadow noise: Underexposure to protect highlights increases noise visibility in dark areas
• Multi-exposure HDR: Some systems capture multiple exposures and combine them for extended dynamic range
• Local tone mapping: Compresses dynamic range while preserving local contrast for viewing on limited-range displays

Color Temperature and Illuminants

Light sources have characteristic color spectra described by color temperature (for thermal sources) or spectral distribution:

• Tungsten/incandescent: Approximately 2800-3200K, with strong red and weak blue content
• Daylight: Approximately 5500-6500K, relatively neutral with slight blue bias
• Fluorescent: Variable depending on phosphor type; often has discontinuous spectrum with peaks at certain wavelengths
• LED: Variable depending on design; may have gaps in spectrum that cause metamerism issues

White Balance Adjustment

White balance applies different gains to the color channels to compensate for illuminant color:

• RGB gain adjustment: Scales each color channel independently to achieve neutral reproduction of gray objects
• Typical range: Gain adjustments from 0.5 to 2.0 cover most lighting conditions
• Preset modes: Fixed gain settings for common illuminants (daylight, tungsten, fluorescent, etc.)
• Custom white balance: User captures a reference white or gray card to set gains for current conditions

Automatic White Balance

AWB automatically estimates scene illumination and sets appropriate channel gains:

• Gray world assumption: Assumes the average of all scene colors should be neutral gray; adjusts gains to achieve this
• White patch detection: Identifies the brightest points in the image, assuming they are white or specular reflections
• Illuminant estimation: Uses statistical analysis of color distribution to estimate likely illuminant
• Scene analysis: Examines color patterns and context to identify illuminant type

AWB algorithms must handle challenging cases like dominant single colors, mixed illumination, and intentionally colored scenes.

White Balance Circuit Implementation

Analog white balance circuits provide independent gain control for each color channel:

• Multiplying DAC: Digital control word sets gain applied to analog color signal
• Variable gain amplifier: Voltage-controlled amplifier with separate control for each channel
• Programmable gain stage: Switched resistor networks provide discrete gain steps

Modern systems typically perform white balance in the digital domain after A/D conversion, offering greater precision and flexibility for nonlinear corrections and mixed-illuminant handling.

Signal Integrity

Video and imaging circuits are sensitive to signal integrity issues:

• Ground loops: Can introduce hum bars or banding in the image; use proper grounding techniques and isolation
• Crosstalk: Coupling between color channels or adjacent signal paths causes color errors or ghosting
• Power supply noise: Appears as vertical bands or brightness variations; requires careful decoupling and filtering
• EMI: External interference can create herringbone patterns or other artifacts

Timing Accuracy

Video timing must meet strict tolerances:

• Sync accuracy: Timing errors cause horizontal or vertical position shifts
• Sample clock jitter: Causes noise and reduced effective resolution in A/D and D/A conversion
• Color burst phase: Phase errors cause hue shifts; 1 degree of phase error equals approximately 1 degree of hue error
• Delay matching: Luma and chroma paths must have matched delay to prevent color fringing

Testing and Measurement

Video system testing requires specialized equipment and test signals:

• Test pattern generators: Produce color bars, multiburst, and other standard test signals
• Waveform monitors: Display signal amplitude versus time for checking levels and timing
• Vectorscopes: Display chrominance phase and amplitude for color accuracy verification
• Resolution charts: Physical test targets for measuring optical and system resolution