.mp3 Demystified

by | Mar 31, 2017 | Technology

Most of your voice-acting pros will know about this, but it’s never too late to learn something new, or even brush-up on something “old”.

We all use .mp3, but do we really know how this marvel of audio compression works,  It’s mostly credited to Fraunhofer (and others) (also HERE & HERE)…and is a file form that offers the best of both worlds.  small file-size, and quality that is acceptable to most listeners.

A question came up within an online VO forum from someone editing with Audacity.  The question had to do with the .mp3 parameters.

Here’s the query:
“… when I save the files as an MP3, it give me options…of how to  save….I usually pick “128kps – 44,100″  What is the 128kps number?  How is it different than the bit rate and sample rate?…”

The answer was simple, elegant, and helpful.  Here it is:

MP3 is a “lossy” compression format — that is, it produces smaller files, at the cost of discarding some amount of information that’s present in the original.  It’s designed to do this while sounding more or less the same to the human ear.  (Similarly, JPEG is a lossy compression format for photographs — it throws away some informationwhile looking roughly the same to the eye.  In contrast, GIF and PNG are “lossless” formats; they preserve all the information.  Some corresponding lossless audio formats are WAV and AIFF.)

When you have a lossy compression format, you can usually choose the degree of tradeoff — you control about how much information gets discarded.  The “128kbps” means “128 kilobits per second”, and it’s how much data is encoded.  If you choose a larger number, your MP3 files will be larger, but they’ll contain more information and theoretically sound closer to the original.
MP3 has an additional twist — you can choose constant bit rate (CBR) or variable bit rate (VBR).  Constant bit rate, as the name implies, encodes the same amount of data in every second of the file.  Variable bit rate attempts to adapt to the audio stream, saving less data when things are quiet or simple (less information to encode), and more when there’s intricate sound to reproduce.  The tradeoff is that it requires more processing power to create or reproduce.
So the bit rate and sample rate apply to the *raw* audio that Audacity is recording and manipulating; when you export to a 128kbps MP3 (which is pretty typical), you’re distilling that information down to the essentials.


Have a great weekend!




Share This