Representing Images

Calculating File Size

• To determine the raw storage requirement of an image in bits, use the formula: $\text{File Size (bits)} = \text{Width} \times \text{Height} \times \text{Color Depth}$
• To convert this value into bytes, divide the total bits by 8. Further conversions to Kilobytes (KB) or Megabytes (MB) require dividing by 1024 (or 1000, depending on the standard used).

Run-Length Encoding (RLE)

• RLE is a lossless compression technique that reduces file size by identifying sequences of identical pixel colors. Instead of storing each pixel individually, it stores a 'data pair' consisting of the count and the color value (e.g., '15 Red' instead of 'Red' repeated 15 times).

Metadata

• Metadata is 'data about data' stored within the image file. It provides the computer with essential instructions on how to reconstruct the image, such as the width, height, color depth, and resolution.

• Resolution vs. Quality: While increasing resolution generally improves quality, it cannot add detail to an image that was originally captured at a lower resolution; it merely makes the existing grid finer.

• Raw vs. Compressed: A raw bitmap stores every pixel's data explicitly, whereas compressed formats like RLE use mathematical patterns to reduce redundancy without necessarily losing visual information.

• Unit Awareness: Always check if the question asks for the answer in bits, bytes, KB, or MB. Forgetting to divide by 8 to reach bytes is the most common source of lost marks.

• The $2^n$ Rule: Memorize powers of 2 (2, 4, 8, 16, 32, 64, 128, 256) to quickly determine the number of colors supported by a given bit depth.

• Metadata Necessity: If asked why an image cannot be displayed from a raw bitstream, the answer is usually the lack of metadata. Without knowing the width and height, the computer cannot know where one row ends and the next begins.

• RLE Efficiency: Remember that RLE is only effective for images with large areas of uniform color. It may actually increase the file size of highly detailed, noisy images where colors change every pixel.