Unit
- Hue, Saturation, and Luma: Understanding these three attributes is the cornerstone of color grading. Hue defines the color family (red, green, blue), saturation controls intensity, and luma represents brightness. Adjusting each independently allows precise tonal control without affecting the others.
- Additive Color Model (RGB): Displays use red, green, and blue light channels. Combining them at full intensity produces white. Grading operations manipulate these channels to shift color balance. For example, reducing blue increases yellow, a common technique to warm an image.
- Complementary Colors: Pairs opposite each other on the color wheel (blue/orange, red/cyan, green/magenta). Applying a complementary color scheme creates visual contrast. In skin tones, a slight orange-blue separation enhances depth while keeping natural hues.
- Color Temperature: Measured in Kelvin, temperature describes the warmth (orange) or coolness (blue) of light. Daylight is around 5600K, tungsten at 3200K. Grading setups often involve white balancing to neutralize unwanted color casts before applying creative adjustments.
- Color Spaces and Gamuts: Rec. 709 is standard for HD video, DCI-P3 for cinema, Rec. 2020 for UHD. Each defines a specific range of colors (gamut). Grading operations must be performed within the appropriate color space to ensure accurate reproduction across devices.
- Histogram and Waveform: Scopes display image luminance distribution. The waveform shows brightness from left to right (black to white). A balanced waveform typically has information spread across the range without clipping at extremes, indicating good exposure.
- Vectorscope: Plots hue and saturation on a color wheel. Skin tones fall along the yellow-red line. The vectorscope helps identify color casts and ensures skin tones remain natural. Over-saturation pushes signals past the target box, indicating illegal colors.
- Perceptual Uniformity: Human vision is more sensitive to changes in darker areas. Gamma correction, often applied to video signals, distributes tonal information non-linearly. Understanding this helps predict how adjustments in shadows affect the final image.
- Lift, Gamma, and Gain Controls: These three primary adjustments target shadows, midtones, and highlights respectively. Lift affects dark regions, gamma controls the middle range, and gain modifies bright areas. Separating these zones allows precise contrast sculpting without shifting the entire tonal curve.
- Qualifiers or Keys: Grading tools often include hue, saturation, and luma qualifiers that isolate specific color ranges. By keying on a skin tone or a sky region, you can adjust only that area. This selective approach avoids unwanted changes to the rest of the image during refinement.
- Power Windows and Masks: Geometric shapes or custom splines create localized grading adjustments. For instance, a circular window can brighten a subject’s face while leaving the background untouched. Feathering and tracking ensure the mask follows movement and blends naturally with surrounding pixels.
Code Example
Python
import cv2
import numpy as np
# Load image in BGR format
img = cv2.imread('frame.png')
# Convert to HSV for perceptual adjustments
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
# Increase saturation by 20%
hsv[:, :, 1] = np.clip(hsv[:, :, 1] * 1.2, 0, 255)
# Convert back to BGR
adjusted = cv2.cvtColor(hsv, cv2.COLOR_HSV2BGR)
cv2.imwrite('graded_frame.png', adjusted)This Python script demonstrates basic color manipulation using the HSV color space. Converting to HSV allows independent control of hue, saturation, and value (luma). The saturation channel is scaled by 1.2, then clipped to typical 8-bit range before converting back to BGR for saving.
Quick Revision
- Color grading relies on hue, saturation, and luma adjustments.
- The RGB additive model underpins digital color; complementary pairs create contrast.
- Color temperature and white balance neutralize casts before creative work.
- Scopes like waveform, histogram, and vectorscope guide objective grading decisions.