Which Is The Best Compression Tool For Linux?

https://www.lifewire.com/which-is-the-best-compression-tool-for-linux-4082712

Introduction

When it comes to finding file compression tools in Linux you are left with a number of different choices but which one is the best?
In this guide, I will put zip, gzip and bzip2 through their paces to see which one is best.
I have conducted a number of tests against different file types and using different settings for each tool and here are the results

Best Tool For Compressing Windows Documents

Before looking at a more detailed test I wanted to try each compression tool against a single file type so that we could see how each tool handles the file in question.

These tests have been run against the Microsoft DOCX format.
Default Settings
I have started with default settings for each program.

Tool	File Size
Initial Filesize	12202 bytes
zip	9685
gzip	9537
bzip2	10109

Best Compression
This time I have gone for maximum compression,

Tool	File Size
Initial Filesize	12202 bytes
zip	9677
gzip	9530
bzip2	10109

To make sure this wasn't a fluke I tried the same test against 2 other documents.
File 1:

Tool	File Size
Initial Filesize	14913176
zip	14657475
gzip	14657328
bzip2	14741042

File 2:

Tool	File Size
Initial Filesize	13314
zip	10814
gzip	10653
bzip2	11254

Two of the files contained text only whereas the larger file contained a lot of pages of text with lots of images and a lot of formatting.
From the first test gzip comes out on top in all categories and bzip2 is the least effective.

Best Tool For Compressing Images

This time I am going to show the results of compressing various image formats such as PNG and JPG.

In theory, JPG files are already compressed and therefore may not compress at all and could, in theory, make the file bigger.
PNG File

Tool	File Size
Initial Filesize	345265
zip	345399
gzip	345247
bzip2	346484

JPEG File

Tool	File Size
Initial Filesize	44340
zip	44165
gzip	44015
bzip2	44281

Bitmap File

Tool	File Size
Initial Filesize	3113334
zip	495028
gzip	494883
bzip2	397569

GIF File

Tool	File Size
Initial Filesize	6164
zip	5772
gzip	5627
bzip2	6051

In all cases, gzip came out on top again except for one and that was the humble bitmap. The bzip2 compression produced a tiny file in comparison to the original.

Best Tool For Compressing Audio Files

The most common audio format is MP3 and in theory, this has already been compressed so the tools may actually end up increasing the file size.
I am going to test two files:
File 1:

Tool	File Size
Initial Filesize	5278905
zip	5270224
gzip	5270086
bzip2	5270491

File 2:

Tool	File Size
Initial Filesize	4135331
zip	4126138
gzip	4126000
bzip2	4119410

This time the results were inconclusive. The compression in all cases was minimal but it is interesting that bzip2 came out the worst for file 1 and the best for file 2.

Best Tool For Compressing Video

In this test, I am going to compress 2 video files. As with MP3 the MP4 file already contains a level of compression and so the results will probably prove to be negligible in terms of how well the tools perform.
I have also included an FLV file which will not have any level of compression as it is a lossless format.
MP4:

Tool	File Size
Initial Filesize	731908
zip	478546
gzip	478407
bzip2	478042

Yet again the bzip2 format came out better than the other file types.

At this stage, it would seem that there is little difference as to which tool you use. The results are close across the board for all file types and sometimes gzip is best and others bzip2 is best and the zip command is usually there or thereabouts.

FLV:

Tool	File Size
Initial Filesize	7833634
zip	4339169
gzip	4339030
bzip2	4300295

It would appear that if you are compressing video that bzip2 is the compression tool of choice.

Executables

The last single category that I will try is executables.
As executables are compiled the code I suspect that they won't compress very well.
File 1:

Tool	File Size
Initial Filesize	26557472
zip	26514031
gzip	26513892
bzip2	26639209

File 2:

Tool	File Size
Initial Filesize	195629144
zip	193951631
gzip	193951493
bzip2	194834876

Again we see that gzip comes out on top and bzip2 comes last. For the smaller executable the bzip file actually grew in size.

Complete Folder Test

Thus far I have dealt with individual files. This time I have a folder full of images, documents, spreadsheets, videos, audio files, executables and many other different file formats.
I have created a tar file which makes it easier to compress using all of the tools available. The gzip and bzip2 commands work against single files whereas the zip command can work against folders.
By using the tar command I have created a single file which contains all of the folders and files in an uncompressed format.
I am going to monitor a number of things in this test:

• Compress using default compression settings - report results by file sizes
• Compress using default compression settings - report results by time was taken
• Compress using best compression - report results by file sizes
• Compress using best compression - report results by time was taken
• Compress using fastest compression - report results by file sizes
• Compress using fastest compression - report results by time taken

Default Compression

Tool	File Size	Time Taken
Initial File	1333084160	0
zip	1303177778	1 minute 10 seconds
gzip	1303177637	1 minute 35 seconds
bzip2	1309234947	6 minutes 5 seconds

Maximum Compression

Tool	File Size	Time Taken
Initial File	1333084160	0
zip	1303107894	1 minute 10 seconds
gzip	1303107753	1 minute 35 seconds
bzip2	1309234947	6 minutes 10 seconds

Fastest Compression

Tool	File Size	Time Taken
Initial File	1333084160	0
zip	1304163943	1 minute 0 seconds
gzip	1304163802	1 minute 15 seconds
bzip2	1313557595	6 minutes 10 seconds

Summary
Based on the final test it is clear that bzip2 is not as useful as the other 2 compression tools. It takes longer to compress the files and the final file size is larger.
The difference between zip and gzip is negligible, and whilst gzip generally comes out on top, the zip format is more common across different operating systems.
So my verdict is that definitely use either zip or gzip but maybe bzip2 has had its day and needs to be confined to history.