Lossless compression

Computer Science · Data compression

Lossless or lossy: which would you pick?

Sort each file into the kind of compression it needs. Commit to a choice, then read the reason.

7 of 7 still to sort.

Every file needs a decision. Make it first, then read why. The reason always comes back to one question: does the exact original need to come back?

Predict, then check

A text file is compressed losslessly, sent to a friend, then decompressed. Commit before you look.

What does your friend have after decompressing it?

What you need to know

• Compression reduces the number of bits needed to represent a file, by removing unnecessary data or rearranging information more efficiently.
• A compressed file has to be decompressed before you can use it again.
• Lossless compression makes a file smaller without losing any data. Nothing is permanently removed.
• So a lossless file decompresses to exactly the original: fully reversible, with quality and exact detail kept.
• Files that must be exactly the same, like text files, need lossless, because every character has to be preserved.
• Have a goImagine a compressor changes just one character in a text file, turning 'Meet at 5' into 'Meet at 6'. Is that acceptable? Why or why not?

  No. In a text file every character must be preserved exactly.

  If even one character changes, the file is no longer correct, which is why text needs lossless compression.

• PNG is a lossless image format. It is used where the highest quality must be kept, and it supports transparent pixels.
• ZIP compresses several files into one file, and unpacking it gives you the originals exactly as they were.
• Have a goYou ZIP three files into one and your friend unpacks it. How many files do they end up with, and how do they compare with your originals?

  Three files, each exactly as it was before zipping.

  ZIP compresses several files into one file, but unpacking gives the originals back exactly, so it is lossless.

• Lossy generally gives a smaller file than lossless, but it removes data and may reduce quality.
• Choose lossless when the file must be perfectly preserved, and lossy when a slight loss of quality is acceptable for a smaller file.
• Have a goDev announces: 'Lossy always wins because it gives a smaller file, so I use it for everything.' Using the rule, when is Dev wrong?

  Whenever the file must be perfectly preserved, such as a text file.

  Lossy generally gives a smaller file but removes data, so it only fits when a slight loss of quality is acceptable.

The big picture

Compression reduces the number of bits needed to represent a file, and a compressed file must be decompressed before you can use it again. Lossless compression makes a file smaller without losing any data, so decompressing gives back exactly the original. That matters for files that must be perfectly preserved, such as text files. PNG and ZIP are lossless examples. Lossy generally gives a smaller file but removes data and may reduce quality, so the choice depends on the accuracy needed, the quality needed and the storage limits.

⚠ Watch out

Believing that a smaller file always means something was thrown away. That is true for lossy. With lossless, the file gets smaller and no data is permanently removed, so the exact original comes back when you decompress.

Memory hook

Lossless means you lose nothing. Compress it, send it, decompress it, and every bit is back in its seat.

Check yourself

A friend says 'if a file got smaller, some of it was deleted'. Explain why that is not always true, and name one file type that needs lossless.