You’ve probably noticed that this book has spent a lot of time dealing with algorithms that are lossless in nature. That is, the decoded version of the data is bit-identical to the source version of the data.
However, most of the content you’re really worried about in the day-to-day operation of your application is compressed by lossy compressors. Things like images, sounds, and video contain far more information than the human visual and audio systems can process—or need—to fully enjoy the experience. Lossy compression formats get rid of those extra bits.
Lossy compressors are typically applied to the data first, to reduce its dynamic range, in preparation for further lossless compression.
Let’s be clear about this: there is an unlimited number of lossy compressors out there, depending on your data type, needs, and how much error your users are willing to tolerate. In reality, it’s one of the most fertile grounds for data compression science because there’s just so much left undone there.
So, why didn’t we talk more about lossy compression algorithms in this book?
Because...well...that’s a different book.