Compression - part 1

Compression is a great tool! When used during mixing and mastering especially, it has many uses. But during airplay it’s very rarely beneficial, especially when you have no idea what you’re doing…

A land of confusion

But when talking about compression, the first thing we need to define is what type of compression we’re going to look at. Because when it comes to audio, there are 2 types of compression that people might talk about. Welcome to the land of confusion! Hopefully, I’ll be able to help clear things up a little bit…

The first type of compression is the one we’re going to look at in details. It is the one used during mixing, mastering and also during airplay. It affects the audio directly, and you might see it referred to as dynamic range compression. Another term that we’re going to see used for compression is limiting (or even brick-wall limiting), which is nothing else but audio compression with extreme settings.

File compression

The second type of compression that you might hear about is file compression. This is the difference between a .wav (or .aif) file and a mp3 for example.

There are various types of file compression, some are lossless (because they will not affect the sound in the end, no information will be lost because these formats will be de-compressed when played) others are lossy (some information is lost during the compression process).

Think of lossless as a zip file. It is a compressed file, but you can always decompress it and get the contained files intact after the process. Lossy compression though will remove some information based on clever algorithms that analyze the sound to get rid of whatever is deemed non-essential to reproduce it. It’s based on the physics of how we perceive sound and what frequencies are more important than others, and on various other factors. How much the sound is compressed with lossy compression depends on the bitrate per second, measured in kbps (Kilo Bits Per Second), the maximum for mp3 being 320kbps, which is almost (but not quite) lossless.

So, in audio land you can have:

  1. raw files (not compressed at all), like .wav or .aif 
  2. lossless files like .flac or .ogg 
  3. lossy files like .mp3 or .aac

Although lossy compression affects the sound (and the lower the bitrate the more it will), this is not what we’re going to look at. The reason being that most radios will play at a rate of 128 kbps or 192 kbps (some use 64 kbps which is hardly listenable), and although of course this means a loss in quality compared to raw files (for 128 kbps it can mean as much as 90% of the initial information lost), it is bound to the bandwidth they have, that bandwidth itself being based on how much they pay and how many listeners the stream provider can support at that rate. So, in short, there’s not much they can do about it…

What online radios can work on to improve the quality of their sound is the first type of compression, which is audio compression (and limiting). So, this is mainly what we’re going to examine in detail, in the hope that it will give everyone a clue as to what they hear and whether too much compression is damaging it… 

Tune in next week to start diving into the wonderful world of audio compression!

4 comments

  • Jon Magnusson

    Jon Magnusson Stockholm

    Looking forward to read your take on it! Great to see we seem to stand on the right side in the loudness war!

    Looking forward to read your take on it!

    Great to see we seem to stand on the right side in the loudness war!

  • Ghostly Beard

    Ghostly Beard

    Thanks for visiting and your comment Jon! The loudness war is unfortunately still very present and especially on indie radios, where the quality of the sound is rarely as good as it should be. As I said here, compression is a really great tool, but it's too often misused mostly because it's misunderstood. My aim is to try and make people understand what it is, when to use it and when NOT to use it.

    Thanks for visiting and your comment Jon! The loudness war is unfortunately still very present and especially on indie radios, where the quality of the sound is rarely as good as it should be. As I said here, compression is a really great tool, but it's too often misused mostly because it's misunderstood. My aim is to try and make people understand what it is, when to use it and when NOT to use it.

  • John Robson

    John Robson Redcar, N York's. UK

    Interested to see some more on audio compression. I just tend to flick through a bunch of presets until I find something that just makes it all sound more "solid". I just use the native FX in Cakewalk & seem to do OK, but I'd love to get the kind of results that LANDR's process gives. I think most of what they do is based on compression.

    Interested to see some more on audio compression. I just tend to flick through a bunch of presets until I find something that just makes it all sound more "solid". I just use the native FX in Cakewalk & seem to do OK, but I'd love to get the kind of results that LANDR's process gives. I think most of what they do is based on compression.

  • Ghostly Beard

    Ghostly Beard

    I will go into the details of some of the parameters and will try to provide audio examples, but I might not get into details of usage in a DAW, that would go way too far for that series of articles that is supposed to be understandable by all. Hopefully some of it will be useful though. One advice I would give you to learn what compression does is to work on simple sounds like a drum loop, and push the parameters to their maximum, learn how they each of them affect the sound differently, learn to recognize what they do, then back up until the effect is felt but not heard.

    I will go into the details of some of the parameters and will try to provide audio examples, but I might not get into details of usage in a DAW, that would go way too far for that series of articles that is supposed to be understandable by all.
    Hopefully some of it will be useful though.
    One advice I would give you to learn what compression does is to work on simple sounds like a drum loop, and push the parameters to their maximum, learn how they each of them affect the sound differently, learn to recognize what they do, then back up until the effect is felt but not heard.

Add comment