Compression

Test yourself on Compression with exam style questions, marked instantly.

Flashcards that come back just before you would forget them.

A revision plan that remembers what you know and shows what to do next.

See which topics you have nailed and which need another look.

Quiz yourself on CompressionGet your flashcardsBuild my revision planTrack my progress

The Travel Packing Problem

You're packing for a trip. Lossless packing: You vacuum-seal clothes to remove air - they're smaller but unfold perfectly. Lossy packing: You leave your winter coat home and pack a thin jacket instead - much lighter, but not quite the same. Lossless compression is like vacuum-sealing - you can get the exact original back. Lossy compression is like leaving stuff behind - the result is good enough but some detail is permanently gone. Photos, music, videos = usually lossy (you won't notice). Documents, programs, medical scans = always lossless (need exact original)!
Two rows. Lossy compression: an 800 KB picture is compressed to 200 KB and comes back blocky, because some data is lost; examples JPEG and MP3. Lossless compression: an 800 KB picture is compressed to 420 KB and decompressed back to the identical picture; examples PNG, GIF and ZIP.

Lossy compression removes some data for good, so the file is much smaller. Lossless compression removes no data, so the original comes back exactly. The terms lossy and lossless are named by OCR and Edexcel. AQA covers two lossless methods, Huffman coding and run-length encoding.

Deep Dive: Why Compression Matters

The Storage Crisis

Remember Topic 15? An uncompressed 3-minute CD-quality song = 31.8 MB. A 4K movie uncompressed = 6,000 GB (6 TB)! Without compression:

  • Your 128GB phone would store ~4,000 photos instead of 40,000
  • Netflix would need 100× more servers
  • Downloading a movie would take days instead of minutes
  • Web pages would load in minutes instead of seconds

Compression is the invisible hero that makes modern digital life possible. Every photo, song, video, and web page you interact with is compressed.

Lossless Compression - Perfect Recreation

What is Lossless Compression?

  • Reduces file size WITHOUT losing any data
  • Original file can be perfectly reconstructed
  • Uses patterns and redundancy to make files smaller
  • Typical reduction: 30-70% smaller (varies by file type)

When to Use Lossless:

  • Text documents: Word files, PDFs, code - every character matters
  • Programs/executables: One wrong bit = program crashes
  • Medical images: X-rays, MRI scans - diagnosis requires exact detail
  • Legal documents: Contracts, evidence - must be identical to original
  • Scientific data: Research data, measurements - accuracy critical

Common Lossless Formats:

  • ZIP, RAR, 7z: General file compression/archiving
  • PNG: Lossless image format (screenshots, diagrams, graphics with text)
  • FLAC: Lossless audio (music archiving, audiophiles)
  • GIF: Simple animations and graphics (limited to 256 colors)
  • GZIP: Web compression (HTML, CSS, JavaScript files)

Run-Length Encoding (RLE) - Simple Lossless

How RLE Works:

Replace repeated values with a count + value. Instead of "AAAAAABBCC", store "6A2B2C".

Example 1: Simple Pattern

  Original: WWWWWWBBBWWW (12 characters)
  RLE:      6W3B3W (6 characters)
  
  Saved: 50% compression ratio
  

Example 2: Binary Data

  Original: 00000000111111110000000 (23 bits)
  RLE:      8×0, 8×1, 7×0 (3 pairs)
  
  Stored as: 8,0,8,1,7,0
  

Example 3: Image Data (Scanline)

  Image row (colors): RRRRRRGGGGBBBBBB (16 pixels)
  RLE:                6R4G6B (3 pairs)
  
  Original: 16 pixels × 3 bytes = 48 bytes
  RLE:      3 pairs × (count + color) = 6 bytes
  Saved:    87.5%!
  

When RLE Works Well:

  • Large blocks of same color: Simple graphics, logos, cartoons
  • Black and white images: Scanned documents, faxes
  • Video frames: Large unchanged areas between frames

When RLE Fails:

  • No repetition: ABCDEFGH → 1A1B1C1D1E1F1G1H (BIGGER!)
  • Random data: Photos, complex images - little to no repetition
  • Can make file larger if there's no pattern

Dictionary-Based Compression

How Dictionary Compression Works:

Build a dictionary of common patterns/phrases. Replace patterns with short codes.

Simple Example:

  Text: "the cat sat on the mat the rat sat"
  
  Dictionary:
  1 = "the "
  2 = "sat "
  3 = "at "
  
  Compressed: "1cat 2on 1m31r32"
  
  Original: 33 characters
  Compressed: ~17 characters + dictionary
  

Real Applications:

  • ZIP files: Use lossless compression to shrink files
  • GIF images: Use lossless compression for images with few colors
  • PDF files: Text and object compression
  • Works best: Text files, code, HTML (lots of repeated words/patterns)

Lossy Compression - Good Enough Quality

What is Lossy Compression?

  • Reduces file size by PERMANENTLY removing data
  • Original cannot be perfectly recreated
  • Removes details human senses can't detect (or barely notice)
  • Typical reduction: 80-95% smaller (massive savings!)

When to Use Lossy:

  • Photos: Vacation pics, social media - slight quality loss unnoticeable
  • Music: Streaming, portable devices - most people can't tell difference
  • Video: YouTube, Netflix - optimized for human perception
  • General rule: Media for human consumption where exact precision isn't critical

Common Lossy Formats:

  • JPEG (.jpg): Photos - 90-95% compression, widely supported
  • MP3: Music - 90% smaller than WAV, near-CD quality
  • MP4, H.264, H.265: Video - streaming, YouTube, Netflix
  • AAC: Music - Apple's audio format (iTunes, iPhone)
  • WebP: Web images - Google's format (better than JPEG)

How Lossy Compression Works

Key Principle: Remove What Humans Can't Perceive

For Images (JPEG):

  • Human eye limitations: More sensitive to brightness than color
  • Strategy: Keep brightness detail, reduce color detail
  • Averaging: Group nearby similar pixels, store average
  • Discard high frequencies: Remove fine texture details
  • Result: 10× smaller file, looks nearly identical to most viewers

For Audio (MP3):

  • Human hearing limitations: Can't hear very high/low frequencies
  • Psychoacoustic masking: Loud sounds mask quiet nearby sounds
  • Strategy: Remove frequencies humans can't hear
  • Remove masked sounds: Quiet sounds near loud ones = discarded
  • Result: 10× smaller, sounds nearly identical to CD

Quality Levels (User Control):

  • Low quality: 90% compressed - visible/audible artifacts
  • Medium quality: 85% compressed - good balance (default)
  • High quality: 70% compressed - minimal loss, larger file
  • Trade-off: More compression = smaller file but worse quality

Lossless vs Lossy Compression

Aspect Lossless Lossy
Data Loss None - perfect reconstruction Permanent - some data removed
Compression Ratio 30-70% smaller (moderate) 80-95% smaller (extreme)
Reversible Yes - decompress to exact original No - cannot recover lost data
Best For Text, programs, medical images, legal docs Photos, music, video, web images
Examples PNG, ZIP, FLAC, GIF JPEG, MP3, MP4, AAC
When to Use When every bit matters / exact copy needed When human perception is primary concern

Real-World Example: Photo Sharing

Scenario: You take a 12MP photo with your phone camera.

  • Uncompressed RAW: 36 MB (12 million pixels × 3 bytes)
  • PNG (lossless): ~18 MB (50% compression - some redundancy removed)
  • JPEG high quality: ~3 MB (92% compression - barely noticeable loss)
  • JPEG medium quality: ~800 KB (98% compression - slight artifacts)
  • JPEG low quality: ~200 KB (99.4% compression - visible blocks)

For Instagram/Facebook: JPEG medium (800 KB) is perfect - 45× smaller, looks great on phones!

For photo printing: JPEG high (3 MB) or PNG (18 MB) - need quality for large prints

Exam Tips - Compression

Most common exam questions:

  • "Difference between lossy and lossless?" → Lossless = perfect reconstruction; Lossy = permanent data loss
  • "When to use lossy?" → Photos, music, video - for human viewing/listening
  • "When to use lossless?" → Text, programs, medical images - need exact copy
  • "Examples of lossy?" → JPEG, MP3, MP4
  • "Examples of lossless?" → PNG, ZIP, FLAC
  • "How does RLE work?" → Replace repeated values with count + value

Key comparison facts:

  • Lossless: 30-70% compression, reversible, exact original
  • Lossy: 80-95% compression, irreversible, "good enough" quality
  • Trade-off: Quality vs file size - can't have both!

RLE exam questions:

  • Always show the pattern: AAAAAABBB → 6A3B
  • Count + Value format
  • Works best when there are LONG runs of repeated data
  • Can make files LARGER if no repetition

Common mistakes to avoid:

  • Saying "lossy is always better" - NO! Depends on use case
  • Confusing lossless with lossy file formats (JPEG is lossy, PNG is lossless)
  • Forgetting that lossy is PERMANENT - you can't "uncompress" to perfect quality
  • Not explaining WHY you'd use one over the other

Quick Reference - Compression

Lossless Compression:

  • Characteristic: Perfect reconstruction, no data loss
  • Compression: 30-70% reduction
  • Reversible: Yes
  • Use for: Text, programs, medical images, archives
  • Formats: PNG, ZIP, FLAC, GIF, RAR, 7z
  • Algorithms: Run length encoding and Huffman coding (both AQA only)

Lossy Compression:

  • Characteristic: Data permanently removed
  • Compression: 80-95% reduction
  • Reversible: No
  • Use for: Photos, music, video, web images
  • Formats: JPEG, MP3, MP4, AAC, WebP, H.264
  • Strategy: Remove what humans can't perceive

RLE Quick Guide:

  • Pattern: WWWWWBBB → 5W3B
  • Format: Count + Value
  • Best for: Long runs of repeated data
  • Worst for: Random data (can make file larger!)

The Golden Rule:

If you need the EXACT original back: use lossless. If "good enough" is fine and you need small files: use lossy.

Make this stick

Signing in gives access to all PrepWise features.

PrepWise is a trading name of A3M Quality Solutions Limited, registered in England and Wales, company number 14959172. Registered office: 202 Stroud Road, Gloucester, GL1 5LA. ICO registration ZC267069.