Audio compression refers to two completely different processes that share a name. One shrinks audio files so they take up less storage space and stream faster over the internet. The other controls volume levels during music production and broadcasting. Both are called “compression,” which causes endless confusion, but they solve different problems and work in entirely different ways.
Two Types With the Same Name
Audio data compression (sometimes called file compression or codec compression) reduces the size of a digital audio file. Uncompressed CD audio requires about 1.4 megabits per second of data, which means a single album can easily exceed 600 MB. Compression algorithms shrink that footprint by removing or reorganizing data, making it practical to store thousands of songs on a phone or stream music over a cellular connection.
Dynamic range compression is a volume-processing technique. It narrows the gap between the loudest and quietest parts of an audio signal, either by turning loud sounds down, quiet sounds up, or both. This is the compression music producers, podcasters, and broadcast engineers use when mixing audio. It has nothing to do with file size.
If you searched “what is audio compression” because you’re curious about MP3s, streaming quality, or why files come in different formats, you’re looking at data compression. If you’re learning about music production or wondering why modern songs sound louder than older ones, you’re looking at dynamic range compression. Both are covered below.
How Data Compression Shrinks Audio Files
Raw digital audio is essentially a long list of numbers representing the shape of a sound wave, sampled tens of thousands of times per second. That’s a lot of data. Compression algorithms reduce it using one of two strategies: lossy or lossless.
Lossy Compression
Lossy compression permanently removes parts of the audio data that the algorithm determines you’re unlikely to hear. It relies on a branch of science called psychoacoustics, which studies how human hearing actually works. One key principle is auditory masking: when a loud sound and a quiet sound occur at the same time in a similar frequency range, your brain can’t perceive the quiet one. The compression algorithm identifies these masked sounds and discards them, significantly reducing file size without an obvious change in what you hear.
MP3 and AAC are the most common lossy formats. At a bitrate of 128 kbps (kilobits per second), an MP3 file is roughly one-tenth the size of the original uncompressed audio. Bumping up to 256 kbps or higher brings the sound closer to CD quality while still offering major size savings. YouTube, for example, recommends 128 kbps AAC for uploaded audio.
The tradeoff is real, though. Once data is removed, it’s gone. You can’t “uncompress” a lossy file back to its original quality. And when compression is too aggressive or the psychoacoustic model misjudges what’s audible, you get compression artifacts: audible flaws like metallic ringing, a hissing quality, pre-echo (a faint ghost of a sound appearing just before the actual sound), or an overall “underwater” feeling. These are most noticeable at very low bitrates or on recordings with complex, detailed sound.
Lossless Compression
Lossless compression reduces file size without discarding any audio data at all. When you decompress the file, you get back a perfect, bit-for-bit copy of the original. Formats like FLAC, ALAC (Apple Lossless), and WavPack use this approach. They work by finding patterns and redundancies in the data and encoding them more efficiently, similar to how a ZIP file compresses a document.
The downside is that lossless files are substantially larger than lossy ones. A FLAC file is typically 50 to 70 percent of the original uncompressed size, compared to an MP3 at 128 kbps, which might be closer to 10 percent. For casual listening on a phone with limited storage, the difference between a 320 kbps MP3 and a FLAC file is difficult for most people to hear. But for archiving, professional audio work, or listening through high-end equipment, lossless formats preserve every detail.
Common Audio Formats at a Glance
- WAV / AIFF: Uncompressed formats. Full quality, very large files. Used in professional recording and editing.
- FLAC: Lossless compression. Full quality restored on playback, roughly half the size of WAV. Popular for music libraries and audiophile streaming tiers.
- ALAC: Apple’s lossless format. Functionally similar to FLAC but designed for Apple devices and iTunes.
- MP3: Lossy compression. The most widely recognized audio format. Good quality at 256 to 320 kbps, noticeable degradation below 128 kbps.
- AAC: Lossy compression. Generally delivers better sound quality than MP3 at the same bitrate. Used by Apple Music, YouTube, and many streaming services.
- OGG Vorbis: Lossy, open-source alternative to MP3 and AAC. Used by Spotify for streaming.
Why Data Compression Matters for Streaming
Without compression, streaming music would be impractical for most internet connections and mobile data plans. Uncompressed stereo audio at CD quality eats through 1.4 megabits per second. A three-minute song would be about 30 MB. Multiply that across a day’s listening, and you’d burn through gigabytes of data quickly.
Lossy compression at 256 kbps drops that same song to around 5 or 6 MB. Streaming services use this math constantly, balancing sound quality against bandwidth. Most platforms offer a standard tier around 128 to 160 kbps and a high-quality tier at 256 kbps or above. Some now offer lossless tiers for subscribers who want full CD-quality or better, though those streams use significantly more data.
Streaming platforms also apply loudness normalization, adjusting all tracks to a consistent volume so you don’t get blasted when a new song starts. The broadcast standard targets a level of -23 LUFS (a standardized loudness measurement), though individual streaming services each set their own targets. This normalization is separate from compression but is part of the same ecosystem of processing that shapes how audio reaches your ears.
How Dynamic Range Compression Works
Dynamic range compression is an entirely different concept. Instead of reducing file size, it controls the volume envelope of audio in real time. A compressor watches the signal level and, when it crosses a set threshold, reduces the volume by a specific ratio. If the ratio is 4:1, a sound that exceeds the threshold by 8 decibels will be reduced so it only exceeds it by 2 decibels.
Engineers control this process with a few key settings. The threshold determines the volume level at which compression kicks in. The ratio determines how aggressively the volume is reduced. Attack controls how quickly the compressor responds once the signal crosses the threshold, and release controls how quickly it lets go once the signal drops back down. Fast attack times clamp down on sudden peaks like drum hits. Slow attack times let the initial punch through before compressing.
Used carefully, dynamic range compression is one of the most valuable tools in audio production. It keeps a vocalist sitting evenly in a mix, prevents a bass guitar from overwhelming quieter instruments, and ensures a podcast stays at a comfortable listening level even when the host moves closer to or farther from the microphone. Broadcast television and radio rely on it heavily so that commercials, dialogue, and music all play back at roughly the same perceived loudness.
The Loudness War
Dynamic range compression becomes controversial when it’s pushed to extremes. Starting in the 1990s, record labels and mastering engineers began compressing and limiting albums more aggressively to make them sound louder than competing releases. This trend, often called the “loudness war,” resulted in music where quiet moments barely exist. Instead of having natural variation between soft verses and loud choruses, everything sits at one intense volume level.
The cost is real. Heavy compression reduces musical dynamics, flattens the emotional arc of a performance, and frequently introduces distortion. Loud drum hits that should punch through the mix get squashed flat. Sustained notes lose their natural decay. Listeners experience fatigue more quickly because there’s no dynamic contrast to give their ears a break.
Loudness normalization on streaming platforms has helped counteract this trend. When every track gets adjusted to the same target loudness regardless of how aggressively it was mastered, there’s less incentive to compress everything to the ceiling. A heavily compressed track and a dynamic one will play at similar perceived volumes, but the dynamic version will often sound better because it preserved its natural range.