It’s probably THE single most common source of confusion I see in discussions of audio.
People say things like:
“The compression on YouTube really kills the dynamics”
“To get a good encode the music needs to be compressed really hard”
“I hate the sound of compression. mp3s sound really squashed”
Now, all of those statements are based on real opinions about music and sound quality, but they’re all also horribly confused.
Which is understandable, because they’re all talking about compression – but in audio, we commonly talk about two completely different types of compression !
They do different things, they have different purposes, and they have different effects on the sound. But people still refer to them both as “compression”, without saying which one they’re talking about. Sometimes it’s obvious from the context – but often, it’s not.
So, this my infographic is my latest attempt to help people sort out the difference – if you know someone who might find it helpful, please share !
(Click on the image above to see a higher-resolution version, or to download a PDF copy, click here)
And if you want to dig this topic in more depth, here’s something I wrote a few years ago which explains the difference using sponges.
The Gory Details
OK, so you already get it – data compression affects file size, but not dynamics. Dynamic compression affects dynamics, but not file size. And they both affect the sound, but in different ways.
High-quality data-compression can sound almost identical to the original source, while using far less space and bandwidth. But some encoders, codecs and data-rates can suck the soul out of the music, rendering it subtly cold, lifeless, edgy and two-dimensional – or even blatantly distorted, with added ultra-sonic birdies for good measure.
Whereas great dynamic compression can enhance almost every aspect of a recording, adding punch, power, impact, consistency, density and warmth. Just for starters. But inappropriate or clumsy over-compression can also suck the life out of the music, robbing it of almost all the same attributes, or even blatantly distorting the sound.
And even then we’re not done, because the two types do interact in some subtle ways.
Compression plus compression
Excessive dynamic compression actually makes it harder to get a great-sounding data-compressed encode, contrary to popular belief, because the encoder struggles to decide what’s important mjusically when everything is at full tilt the whole time.
And data-compresion can seem to affect the micro-dynamics of the music, by changing the peak level of the reconstructed waveform as a side-effect of the encoding & decoding process. And the more heavily dynamically compressed and limited the source, the more noticeable this effect is. It has no audible effect, though – except perhaps adding extra clipping distortion on playback systems that don’t have enough headroom to deal with the higher peaks.
And because almost all online streaming services use data compression plus loudness management, it’s easy to be fooled into thinking they’re somehow affecting dynamics, too – since really “loud” music seldom sounds anywhere near as impressive when it’s reduced to the same playback loudness as everything else.
Find the sweet spot
Luckily though, the solution to both these complications is straightforward.
Always leave at least 1 dB of peak headrooom, and then find the loudness sweet spot for your music -where you have the perfect balance of loudness and dynamics.
It won’t get turned down online, it’ll encode cleanly to mp3, AAC and other lossy data-comressed formats, and it’ll sound great – maximising the potential for punch, power and impact.
Compression is your friend (both kinds!) provided you understand how it works, and how to get the best out of it.