Chapter 1: The Basics of Audio Compression

MP3 compression has been around for quite some time, with its origins tracing back to the late 1990s. Despite its widespread adoption over the last 25 years, a limited number of individuals grasp the intricacies of MP3 compression and what makes it almost exceptional, yet not entirely satisfactory.

Let’s delve into what compression really means. In general, compression indicates a decrease in size or capacity. When discussing audio, “compression” can refer to minimizing the dynamic range between the loudest and quietest sounds. However, in the context of MP3s, it pertains to reducing file size through data compression techniques. For the remainder of this discussion, we’ll focus on this specific interpretation of compression.

How is audio data compressed? The primary aim is to lower the file size for storage or transmission. There are two main categories of file compression: LOSSLESS and LOSSY.

Lossless compression reduces the data size by substituting repetitive data with shortcuts that denote where the original data can be located. For instance, consider the following data sequence:

0123456789 80974234 23498097543 5034987 0123456789 20923798 0123456789 09878 230975345 0123456789 45938456487 0123456789

In this example, the sequence “0123456789” appears multiple times and can be replaced with a reference pointing to its initial occurrence. Here’s how this process looks:

[0123456789]($1) 80974234 23498097543 5034987 $1 20923798 $1 09878 230975345 $1 45938456487 $1

The brackets “[ ]” indicate the data pattern, while “($1)” signifies that whenever we encounter “$1,” it refers back to the original data in brackets. This method effectively decreases the data size by around 20% (from 120 characters to 94 characters), illustrating how ZIP file compression functions. The encoder identifies patterns in the data, replacing segments with shortcuts that can be restored later, hence the term “LOSSLESS.”

Conversely, “Lossy” compression involves permanently discarding data deemed non-essential.

To visually represent lossy compression, consider the following image of a clear blue sky:

In this pristine image, every color is preserved. Now, let’s look at a compressed version:

This version shows a significant reduction, allowing only six colors. The wide range of colors in the original has been greatly minimized. However, this level of compression is easily noticeable. The challenge lies in identifying a compression level that reduces file size without compromising quality too drastically.

In this case, allowing 30 colors still adequately showcases the image's beauty. However, when we apply this concept to more colorful images, the challenge becomes apparent.

Due to the vast array of hues, the image with 60 colors still shows signs of compression. Notably, the gradient at the bottom reveals “steps” in color. Increasing the color count to 120 results in less noticeable compression.

Ultimately, the key lies in determining which data can be sacrificed. In images, we can often compromise on size, sharpness, and color variety. In audio, various techniques are employed to diminish data accuracy. The main methods include eliminating quality in quieter sections of the audio and removing frequencies that are inaudible to most listeners.

In MP3 compression, frequencies above 15,000 Hz are generally discarded, while low-volume sounds are encoded at a significantly lower file size. Although present, the quality resembles that of tape recordings rather than CDs. By removing non-essential audio (such as frequencies above 15,000 Hz) and lowering the quality of quiet segments, and subsequently applying lossless compression, the final audio file size is reduced to approximately one-tenth of the original. A CD-quality audio file usually occupies around 10 megabytes per minute, while MP3 files typically take up about 1 megabyte per minute.

This process results in the loss of subtle details that are often unnoticeable to the average listener. For audiophiles, however, it may not be entirely acceptable. Intriguingly, streaming platforms have relied on MP3 quality audio for years without significant complaints. It’s only in recent times, with the advent of improved broadband, that lossless playback has begun to gain traction on services like Apple Music and Spotify.

The pressing question remains: have you noticed the difference?

The first video, titled "AM I CRAZY?!? MP3s Sound GREAT | MP3 vs WAV vs AAC," explores the sound quality differences between various audio formats, shedding light on why MP3s might sound impressive yet fall short of perfection.

The second video, "Why your mixes sound thin and weak (probably)," addresses common pitfalls in audio mixing and how to achieve a fuller sound, which ties back to the principles of audio compression discussed earlier.

Did you find this article informative? If so, please consider giving it a clap! Also, feel free to follow my content on Medium.