ZIP, 7z and why some files refuse to compress
Compressing a folder of photos saves you almost nothing, and there is a good reason for that. Which format to use, and when compressing is worth doing at all.
By Hasnain Butt Akhtar, founder and developer ·
Compress a folder of documents and it might shrink by seventy per cent. Compress a folder of holiday photos and it shrinks by two. Nothing is broken. The photos were already compressed, and knowing which of your files are in that category tells you when zipping is worth the trouble.
Why some things compress and others do not
Compression works by finding repetition. A text document repeats words, letter patterns and spacing constantly, so a program can note the pattern once and refer back to it. That is where large savings come from.
A JPEG photograph has already had this done to it, thoroughly, by the camera. So has an MP3, an MP4 video, and a modern Office document, which is a compressed archive already even though it does not look like one. There is no repetition left to find, so wrapping them in a ZIP simply adds a container around files that are as small as they are going to get.
| File type | Typical saving | Worth compressing? |
|---|---|---|
| Text, code, logs, CSV | 70 to 90 per cent | Yes, dramatically |
| Uncompressed images such as BMP or TIFF | 50 to 70 per cent | Yes |
| Databases and disk images | Varies, often large | Usually yes |
| Word, Excel and PowerPoint files | A few per cent | Only to bundle them together |
| JPEG, PNG, MP3, MP4, PDF | Nearly nothing | Only to bundle them together |
That last row is the useful one. If your archive of photographs came out the same size as the folder, that is compression working correctly, not failing.
Bundling is a reason on its own
Even with no size saving, putting files in one archive is often the right move. One file is easier to send than four hundred, transfers faster over most systems because there is no per-file overhead, preserves the folder structure, and arrives complete or not at all rather than partly.
When that is your goal, set the compression level to store or fastest. You get the single file immediately instead of waiting while the software tries to compress data that cannot be compressed.
Which format
ZIP is the universal one. Every operating system opens it without extra software, and it always will. Compression is decent rather than excellent. For anything you are sending to another person, particularly someone whose setup you do not know, use ZIP and do not think about it further.
7z compresses noticeably better, often 10 to 20 per cent smaller than ZIP on material that compresses at all, and considerably better on large text-heavy sets. Recent versions of Windows can extract it without extra software, though older systems and some phones still cannot. Use it for your own storage and backups where the saving is worth having.
RAR you will meet as something to extract rather than create, since making them requires paid software. Windows can now open them, so this is rarely a problem.
TAR.GZ appears in anything touching Linux. Worth recognising, rarely worth choosing on Windows.
Settings that make a difference
Compression level. The gap between normal and maximum is usually a few per cent for several times the processing time. Normal is the right default. Maximum earns its keep only on very large text or code sets that you compress once and keep for a long while.
Solid archives. A 7z option that treats all the files as one continuous stream, which compresses much better when you have many similar small files. The trade is that extracting a single file means working through everything before it. Excellent for archiving a finished project, poor for something you will dip into regularly.
Splitting into volumes. Both formats can break an archive into pieces of a set size, which gets a large set past attachment limits or onto smaller drives. Every piece is needed to extract, so send them all and keep them together.
Passwords, and the one that is not real protection
Both formats can encrypt, but they are not equivalent.
ZIP supports two methods. The original, sometimes called ZipCrypto or standard encryption, dates from the 1990s and can be broken quickly. The modern option is AES-256, which is genuinely strong. If your software offers the choice, take AES. If it does not say which it is using, assume the weak one for anything that matters.
7z uses AES-256 always, and it can also encrypt the list of file names. This matters more than people expect: a standard encrypted ZIP still shows anyone the names of everything inside, which for a folder called Medical or Contracts gives away most of what you were protecting.
Neither format can help with a forgotten password. There is no recovery, no reset and no support line. Put the password in a password manager at the moment you create the archive, not afterwards.
When compression is the wrong tool
If a folder of video or photographs is too large, compression will not help and you need to change the files themselves. Converting video to a more efficient format can halve it. Reducing photographs that only ever need to be viewed on a screen does the same. Both are proper conversions rather than packaging, and both lose something, so keep your originals.
Two practical defaults cover almost everything. Sending files to someone: ZIP, normal compression, no password unless it is sensitive, in which case AES. Archiving your own finished work: 7z, normal compression, solid archive on, stored alongside a note of what it contains so you do not have to open it to find out.