General Compressibility Of Various File Formats



Has anyone come across list of which file formats generally compress well (in general algorithms)?

Generally, text files .txt compress well, and media-files (.mp3, .avi, etc) do not compress well.

I'm looking for a list (or database) of file formats so I can answer the question, "Will a general, lossless algorithm compress a <.xxx> file?".

EDIT: Of course by nature, random data, media files, and compressed data do not compress well. That is not my question.

Question: Where can I look up descriptions of file-formats, specifically to discover if the format already contains compression?

AutoCAD Drawings, Various Microsoft Office formats, Oracle Database .DBF files, Adobe PDF files, Outlook PST files, etc


Posted 2012-09-27T17:18:53.710

Reputation: 24 804

Question was closed 2012-09-27T21:55:26.020

Right now, your question is asking for anecdotes, which makes it unfit for Super User. Of the most common compression formats, LZMA generally achieves the best compression, but it's also the slowest algorithm. Pseudo-random and already compressed data cannot be compressed by any of these algorithms. – Dennis – 2012-09-27T17:27:43.203

1in general all data with lots of common patterns compress very well, data with noise does not. files which are compressed already fall into the latter category of files. – akira – 2012-09-27T17:54:04.230

In general, files that are already compressed to some degree will not compress well. Most standard audio and image formats are compressed. Those that aren't are not generally used for internet operations because they take so much longer to transfer. – Daniel R Hicks – 2012-09-27T17:55:27.880

See edit to question. – Steven – 2012-09-27T18:37:44.800



From what I can understand of your question, it sounds like you have tried to compress an already compressed file. File formats such as .png, .jpg, .avi, pdf, .mp4 and many others may already have their data in a compressed format. Attempting to compress such files will not significantly reduce their size.


Posted 2012-09-27T17:18:53.710

Reputation: 45 216

See edited question. – Steven – 2012-09-27T18:38:50.867