Compression

What compression is and why we use it

Compression means reducing the number of bits needed to store a file. Smaller files:

• take up less storage space;
• transfer faster over a network and use less bandwidth;
• download/stream more quickly and cost less to send.

There are two types: lossy and lossless.

Lossy compression

Lossy compression permanently removes some data — usually detail the human eye or ear won't easily notice.

• The file cannot be restored to its exact original; some quality is lost forever.
• Achieves much smaller file sizes.
• Used for images, audio and video where a small quality drop is acceptable.
• Examples: JPEG (images), MP3 (audio), MP4 (video).

Lossless compression

Lossless compression reduces size without losing any data — the original can be perfectly reconstructed.

• File sizes are reduced less than with lossy.
• Essential where every bit matters: text documents, program code, spreadsheets.
• Examples: PNG (images), FLAC (audio), ZIP (general files).

Two lossless techniques you must know

Run-Length Encoding (RLE)

RLE replaces runs of repeated values with a single value and a count.

Instead of storing:

W W W W W B B W W W

RLE stores it as pairs of (value, number of repeats):

5W 2B 3W

• Works brilliantly when data has long runs of the same value (e.g. simple graphics with large blocks of one colour).
• Works badly — and can even make files bigger — when there are few repeats (e.g. a detailed photo).

Huffman coding (dictionary/frequency-based)

• The most frequently used characters are given the shortest binary codes; rare characters get longer codes.
• A frequency table / tree records which code means which character, so the file can be rebuilt exactly.
• Reduces overall size because common characters no longer waste a full byte each.

Worked example

A row of pixels reads AAAAAABBBBCCCCCCCC. Show it using run-length encoding and state the saving.

1. Runs: six A, four B, eight C.

2. RLE: 6A 4B 8C.

3. Original = 18 characters; RLE stores 3 value+count pairs — a large saving because the runs are long. ✓

Choosing the right type

Need	Use	Example
Exact copy essential	Lossless	text, code, spreadsheets
Small size, minor quality loss OK	Lossy	streaming music, web photos

Common mistakes

• Saying lossy "deletes the whole file" — it removes some detail, not everything.
• Claiming lossless gives the smallest files — lossy usually compresses more.
• Forgetting that RLE can increase size on data with few repeats.
• Mixing up examples (JPEG/MP3 = lossy; PNG/FLAC/ZIP = lossless).

Exam tips

• Learn one clear definition and one example for each type — that's often 2 easy marks.
• If asked to justify a choice, link it to the use case (e.g. "text must be lossless so no words are lost").
• For an RLE question, show the (value, count) pairs clearly and comment on whether it saves space for that data.

Key facts to remember

• Compression reduces file size → less storage, faster transfer, less bandwidth.
• Lossy removes data permanently (smaller files, some quality lost) — JPEG, MP3.
• Lossless keeps all data (perfect rebuild, less shrinkage) — PNG, FLAC, ZIP; techniques include RLE and Huffman coding.