Bandwidth for streaming podcasts is expensive, and using it up with podcast audio that is too quiet, then too loud, then too quiet again is a giant waste of money. Fear not! Just toss in a little bit of compression and you can save money, conserve bandwidth, and please the ears of your listeners all at the same time.
But… audio compression, or file compression? Wait, what’s the difference? I’m already exporting to .mp3 files, and they’re compressed, right? Yes, but that’s file compression. Audio compression is something entirely different, and is just as important. Knowing when you have enough compression of both kinds is also important, and knowing what they are and how they affect your podcast is critical.
Nearly everyone that knows how to use a computer is familiar with file compression. You’ve seen it and its effects more times than you can count. Ever had a .zip file? Or a .jpg file (hint: you’re looking at some on this site)? These file extensions are indications of file compression protocols. Podcasts that use .mp3 files also use compression. Each of these compression types offers something different, and specialized, and there are still hundreds (thousands?) of other file compression types around.
Delivering audio in a .mp3 file allows files to be much smaller than other uncompressed formats. But this comes at a cost. The mp3 format uses a lossy compression algorithm: a means of compression that reduces the audio quality in order to reduce the amount of information being stored in the file. Smaller files come from less information being stored.
How Much File Compression?
The effect of a .mp3 file saved at a very low bit rate is terrible sounding audio in some cases, but not in others. This is what makes compression somewhat tricky.
If you’re recording human voices for a podcast, for example, you can get away with a much lower quality file (lower bit rate). There isn’t a lot of dynamic range of content, or overlapping content in a human voice – even with multiple people speaking at the same time.
When recording musical instruments, however, low quality audio becomes apparent very quickly. If you choose a low bit rate for music, high pitch sounds become distorted in odd ways and low pitch sounds become muffled and muted. Mid range sounds can do ok, still, but get mixed up with each other when multiple instruments are in the same frequency range. Music demands a higher bit rate, but not alway the highest.
Avoid “A Copy Of A Copy” Syndrome
One thing you also need to be aware of: if you’re importing existing audio files in to your screencast, be sure they are high quality to start with.
Why, you might ask? Have you ever made a photocopy of a photocopy of a photocopy? Compressing already compressed audio has the same effect. Starting with high quality sound will ensure the file version sounds as good as possible.
Audio compression can also have a dramatic effect on the quality of audio, but in a very different way.
If you’ve ever heard a song or a podcast or a radio station where literally everything came out at exactly the same volume level – not matter whether the person was wispering, or the music was supposed to be subdued and quiet – you’ve heard the ill-effects of over-compressing audio.
On the opposite end of the spectrum, is classical music. If you’ve ever tried to listen to classical music in a car or on public transport systems, you’ve probably been frustrated by the near impossible to hear quiet moments followed by ear-splitting loud moments. This is the effect of no audio compression at all.
Controlling Signal Levels (Volume)
Audio compression is a means of flattening the signal level for the audio track(s) being compressed. It makes the quiet parts louder so you can still hear them, and makes the loud parts quieter so they don’t blow your ear drums out.
Too much compression makes everything sound flat. It removes all of the dynamics of the audio, preventing the natural ebs and flows of volume differences. It tends to make things sound over-produced. Radio stations tends to use a ton of audio compression to compensate for road noise and other ambient noises in the typical radio lisening environment: transportation. Classical music, on the other hand, tends to opt for purity and intent of the conductor and orechestra, leaving off any audio compression of any kind. Because of this, classical music is usally best experienced in quiet places.
Audio Compression And Podcasts
You need a minimum amount of audio compression in your podcasts – whether it’s a music podcast or a spoken / voice podcast. Without it, your podcast will simply come out too quiet most of the time, with sudden and bothersome loud moments. But at the same time, you don’t want to over-compress your audio. This tends to create the appearance of yelling all the time, because there are no dyanmics or variance in your volume levels.
Learning Through Practice
There are no right and wrong settings for everyone, when it comes to audio compression, though. Each persons voice and speech patterns are unique, creating a need for variance in how much compression is used. You’ll want to practice some with different levels of compression to find out what works best for you.
Where Do I Go From Here?
The topics of what settings to change for what purpose in both audio compression and file compression are worthy of their own blog posts. Getting the high level idea of what these things are and how the affect your podcast is an important step in the right direction, though. Without this basic understanding, having all the knowledge of what settings to change won’t mean much.
For tips and advice on audio compression settings, check out Scott Burton’s post on audio compression. This is a great starting point with advice on what the different settings for audio compression will do. This is a post that deserves your attention – one that I wanted to write, but would rather share Scott’s quality work.
For info on file compression, bit rates and the effects that they have on your .mp3 file exports, check out the post on MP3 bit rates, file size, and audio quality. In this post, I export the same podcast episode using 10 different bit rates (5 constant, and 5 variable) and look at the results.