Compression: Lossy and Lossless

Data files — images, audio, video, and text — can be very large. Compression reduces the file size so that files take up less storage space and can be transmitted more quickly. This lesson covers the two main types of compression: lossy and lossless.

---

Why Compress Data?

There are several important reasons to compress data:

• Reduce storage space: Smaller files take up less room on hard drives, SSDs, and other storage media.
• Faster transmission: Smaller files can be sent over the internet or a network more quickly, using less bandwidth.
• Reduce costs: Less storage and bandwidth are needed, which can save money.
• Enable streaming: Audio and video files must be compressed to allow real-time streaming over the internet.

---

Lossy Compression

Lossy compression reduces file size by permanently removing some of the data. The removed data is typically information that humans are unlikely to notice.

How Lossy Compression Works

• The compression algorithm analyses the data and identifies parts that can be removed with minimal impact on perceived quality.
• For images, this might mean reducing the number of colours or merging similar adjacent pixels.
• For audio, this might mean removing frequencies that humans cannot easily hear (e.g. very high or very low frequencies, or quiet sounds masked by louder ones).
• For video, lossy compression removes redundant data between frames and reduces detail in areas of rapid motion.

Key Characteristics of Lossy Compression

• Data is permanently lost — the original file cannot be perfectly reconstructed from the compressed version.
• Achieves much smaller file sizes than lossless compression.
• Some quality is lost, but the loss may be imperceptible at moderate compression levels.
• At high compression levels, noticeable artefacts appear (e.g. blurring in images, distortion in audio).

Common Lossy File Formats

Type	Formats
Images	JPEG (.jpg)
Audio	MP3 (.mp3), AAC (.aac)
Video	MP4 (.mp4), AVI (.avi)

Example: JPEG Image Compression

A 24-bit bitmap image of 1920 × 1080 pixels would be approximately 6.2 MB uncompressed. Saved as a JPEG with moderate compression, it might be reduced to 300–500 KB — a reduction of over 90%. The trade-off is a slight reduction in image quality that is usually not noticeable on screen.

---

Lossless Compression

Lossless compression reduces file size without losing any data. The original file can be perfectly reconstructed from the compressed version.

How Lossless Compression Works

Lossless compression algorithms find patterns and redundancy in the data and encode them more efficiently. Two common techniques are:

1. Run-Length Encoding (RLE)

RLE works by identifying consecutive repeated values and storing them as a single value plus a count.

Example:

Original data: AAAAAABBBBCCCCCCCDD

RLE encoded: 6A4B7C2D

This is much shorter — 4 pairs instead of 19 characters.

RLE works well for data with many repeated values, such as simple graphics with large areas of the same colour. It is less effective for complex images like photographs, where consecutive pixels are often different colours.

2. Dictionary-Based Compression (e.g. LZW)

Dictionary-based methods build a dictionary (lookup table) of recurring patterns in the data. Each pattern is replaced with a shorter code that references the dictionary entry.

Example:

Original text: "the cat sat on the mat"

A dictionary might store:

• Entry 1: "the "
• Entry 2: "at "

The text could then be encoded using these shorter dictionary references wherever the patterns occur, reducing the overall file size.

Key Characteristics of Lossless Compression

• No data is lost — the original can be perfectly reconstructed.
• Achieves smaller compression ratios than lossy compression (the file is not as small).
• Essential when data integrity is critical.
• Used for text files, executable programs, and situations where quality loss is unacceptable.

Common Lossless File Formats

Type	Formats
Images	PNG (.png), GIF (.gif), BMP (.bmp with RLE)
Audio	FLAC (.flac), WAV (.wav)
General	ZIP (.zip), RAR (.rar), 7Z (.7z)

---

Choosing Compression Type

flowchart TD
    A[File to compress] --> B{Can data be lost?}
    B -->|No| C[Use Lossless]
    B -->|Yes| D[Use Lossy]
    C --> E[PNG, FLAC, ZIP]
    D --> F[JPEG, MP3, MP4]
    C --> G[Original recoverable]
    D --> H[Smaller file size]

Lossy vs Lossless: Comparison

Feature	Lossy	Lossless
Data lost?	Yes, permanently	No, fully recoverable
Compression ratio	High (much smaller files)	Lower (larger files)
Quality	Some reduction (may be imperceptible)	Identical to original
Best for	Photos, music, video (where slight quality loss is acceptable)	Text, programs, medical images, archiving
Examples	JPEG, MP3, MP4	PNG, FLAC, ZIP
Can decompress to original?	No	Yes

---

Choosing the Right Type of Compression

The choice between lossy and lossless depends on the situation:

• Use lossy when file size is more important than perfect quality — e.g. streaming music, sharing photos on social media, video calls.
• Use lossless when the data must be preserved exactly — e.g. text documents, software, medical scans, professional audio production.

Exam Tip: A common exam question asks you to compare lossy and lossless compression or to recommend which type to use for a given scenario. Always explain the trade-off: lossy gives smaller files but loses data; lossless preserves all data but produces larger files. Give a specific example to support your answer.

---

Summary

• Compression reduces file size for efficient storage and transmission.
• Lossy compression permanently removes data to achieve high compression ratios (e.g. JPEG, MP3).
• Lossless compression preserves all data and allows perfect reconstruction (e.g. PNG, ZIP).
• RLE replaces repeated values with a value-count pair.
• Dictionary-based compression replaces recurring patterns with shorter codes.
• Choose lossy when quality loss is acceptable; choose lossless when data integrity is essential.

---

Run-Length Encoding in Detail

Worked Example: Compress a binary bitmap row

A row from a black-and-white logo is:

1111111100000000111111110000111111110000000011111111

That is 52 characters. Counting consecutive runs:

• 8 ones
• 8 zeros
• 8 ones
• 4 zeros
• 8 ones
• 8 zeros
• 8 ones