Audio, video, and image files are so large that they effectively can’t be used on the internet, or in portable devices like cameras or iPods, without somehow making them smaller. And while there are ways to compress files without losing any of the data (think ZIP files), most of the formats we use today, MP3, JPEG, GIF, MP4, AVI, etc., use lossy compression. A song saved in the MP3 format, for example, throws away some of the audio but does it in such a way that most people don’t notice the difference.
But repeatedly compressing the same file will cause it to lose more and more of the actual data—sort of like making a photocopy of a photocopy. This is why I don’t use Audacity to edit MP3 files, because it first converts them to the WAV format for editing then back to MP3 for saving.
The linked article shows what happens to a video that was re-saved a thousand times. The author created the original video, uploaded it to YouTube, downloaded it, uploaded it again, and so on until he had a thousand uploads. The final video very clearly shows the cumulative effect of lossy compression.
Thanks to Mike Primm for this link.