I want to learn how to highly compress a file. I have downloaded files whose compressed size was 4 GB, and after extracting they became 10 GB. Which software do I need to use for such high compression? Will I lose quality of the data by compression? If you direct me to WinRAR or 7-Zip, please tell me how can I highly compress a file because I have tried compression with it, but the size was reduced just by only a few MBs.

  • 1
    Which program did you use to decompress the 4gb file?
    – akira
    Mar 3 '14 at 15:01

There are mainly two types of compression : lossless and lossy. When you use softwares like WinRAR and 7-Zip, they use lossless algorithm. When you use software ripping DVD, encoding audio/video, etc… they generally use lossy algorithm. So it depends on what data you're dealing with. If it is unknown binary data, you'll surely use lossless algorithm. If it is audio/video, you'll use lossy algorithm (except if you don't want to loose any quality).

Lossy compression is surely more efficient regarding size of data than lossless.

That say, in your case you want lossless compression. I think that the best compression ratio would be obtained with LZMA algorithm in case it is unknown binary data. WinRAR doesn't support it as far as I know for compression, but 7-Zip does, through the use of the 7z container. When you make an archive with 7-Zip, select the 7z archiving format, then the LZMA (or LZMA2) compression algorithm, and then you'll be able to choose between some presets, from poor ratio but quick compression using less resources to best ratio but slow compression using more resources.


How much a file is compressed depends largely on what type of file it is. Generally speaking, text files can be highly compressed (even 90% or more) whereas some file formats like images are already highly compressed in their existing form and adding them to a compressed archive will do little or nothing to further decrease their size.

Having said that, applications like 7-Zip usually have settings where you can choose the level of compression. Using 7-Zip as an example, when you create an archive you can select from six different levels of compression ranging from Store (no compression) to Ultra (smallest file size). Higher compression requires more time to process, so if you're in a hurry or saving every last byte isn't critical, choosing something like Normal is going to give you acceptable results. Since your question is about how to get the most compression, you'd probably be happier choosing Ultra. But again, some files can be compressed a lot more than others and even choosing Ultra when trying to compress an image file may not save you more than a few bytes of space, if that (depending on various factors including original file size).


The size was reduced just by only a few MBs.

Sounds like the data is already compressed. Compressing already compressed data will only save you very little space.

Please tell me how can I highly compress a file because I have tried compression with it.

Using 7-zip or winrar should work fine. When compressing with Winrar or 7-zip, you have the option to use better compression at the cost of it taking longer.

Will I lose quality of the data by compression

When using lossless compression (think 7-zip/winrar) then, no. Compression algorithms like MP3 or jpegs do loose quality when compressing.

  • i am under the impression that OP told 7zip to shrink the 10gb file again.
    – akira
    Mar 3 '14 at 15:00

This is highly dependant on the content of the file(s) you want to compress. Lossy compression usually gives smaller files due to the fact that it throws information away (lossy).

Find out about good comrpessors and how to use them and in what situations:



There is now magic way to turn 10 GB into 4 GB. If you have a 10 GB file consisting of 10,000,000,000 repetitions of the letter A, it can be compressed into a few hundred bytes. If it's 10 GB of random data, you won't be able to compress it at all.

Not the answer you're looking for? Browse other questions tagged or ask your own question.