What Makes an Audio Format Lossy?

A look at lossy audio compression and how it affects digital music

Unhappy girl listening to the music in the city
Martin Dimitrov/E+/Getty Images

What Makes an Audio Format Lossy?

The word lossy is used in digital audio to describe the type of compression used to store sound data. The algorithm used in a lossy audio format compresses sound data in a way that discards some information. This means that the encoded audio isn't identical to the original.

For example, when you create a series of MP3 files by ripping one of your music CDs, some of the detail from the original recording will be lost -- hence the term lossy. This type of compression isn't just restricted to just audio either. Image files in the JPEG format for instance are also compressed in a lossy way.

Incidentally, this method is the opposite to lossless audio compression used for formats such as FLAC, ALAC, and others. The audio in this case is compressed in a way that doesn't discard any data at all. The audio is therefore identical to the original source.

How Does Lossy Compression Work?

Lossy compression makes certain assumptions about frequencies that the human ear is unlikely to detect. The technical term for the study of sound perception is called, psychoacoustics.

When a song for example is converted to a lossy audio format such as AAC, the algorithm analyses all the frequencies. It then discards ones that the human ear shouldn't be able to detect. For very low frequencies, these are usually filtered out or converted to mono signals which take up less space.

Another technique that is also used is to discard very quiet sounds that the listener is unlikely to notice, especially in a louder part of a song. This will help to reduce the size of the audio file while limiting the impact on audio quality.

How Does Lossy Compression Affect Audio Quality?

The problem with lossy compression is that it can introduce artifacts. These are undesirable sounds which aren't in the original recording, but are by-products of compression. This unfortunately degrades the quality of the audio and can be particularly noticeable when low bitrates are used.

There are different types of artifacts that can affect the quality of a recording. Distortions are one of the most common ones that you will likely come across. This can make drums for example sound weak without any real punch. Voice in a song can also be affected. The singer's voice may sound course and lack detail.

Why Compress Audio at All?

As you already know, most digital audio formats employ some sort of compression in order to store sound in an efficient way. But without it, file sizes would be very large.

For example, a typical 3-minute song stored as an MP3 file can be around 4 to 5 Mb in size. Using the WAV format to store this same song in an uncompressed way would result in a file size of approximately 30 Mb -- that's at least six times larger. As you can see from this (very rough) estimate, far less songs would fit on your portable media player or computer's hard drive if music wasn't compressed.