Monday, January 30, 2006

TNL's Guide to The Music File Formats: Part Two

Part Two: MP3

This is perhaps the best well known of the music file formats. MP3 files are generally 10% the size of the original file on the CD. The compression is accomplished by using knowledge of psychoacoustics to eliminate sounds that are not heard by the human ear, or processed by the brain. For example, when we hear two sounds, we process the louder one. Therefore, during the MP3 compression process, the quieter one is eliminated.

The MP3 standard has its roots in digital audio for motion pictures. MPEG is an acronym for Moving Pictures Expert Group. The MP3 format is based on the original MPEG-1 video standard, DVD’s are based on the successive MPEG-2 standard, and some devices (like newer camcorders) use the more compressed MPEG-4 standard. The third layer of information in this standard is audio, therefore the MP3 moniker was born. While the MP3 file was invented in 1991, the first software encoder appeared in 1994, followed by the first player in 1995 (I’m wondering why folks were encoding files that they had to wait another year to play).

The MP3 is a lossy format, which means that all the data from the original recording is not present. It is commonly thought of as “near CD quality,” although audiophiles disagree and contend that the sound quality leaves much to be desired. For listening in the car, or via a portable player in anything but a silent environment, it is more than adequate for most users. It compresses audio across a range of 16 Hz to 48 kHz.

The quality of the MP3 compression can be adjusted by the bit rate, how many bits of information are contained in the file for each second of music. The most popular bit rate used is 128 kbps, which corresponds to an 11:1 compression. The maximum bit rate is 320 kbps; for comparison, a CD contains info at about 1400 kbps. Other popular bit rates are 160 and 192 kbps.

All of those numbers refer to a constant bit rate (CBR) encoding. This means that at all points in the track, there is the same amount of audio information. A further development produced what’s called variable bit rate (VBR) encoding. This specifies a minimum bit rate for the track but it’s not constant. During more complicated parts of the track, the algorithm uses more than that minimum amount to preserve the quality of the sound. This helps to create a better balance between audio quality preservation and reasonable file size. I watched the kbps readout in Winamp while playing back a track recorded at 160 kbps VBR. The track went between 160, 192, and 224 kbps on playback (with an average of 191 kbps).

Another variation on MP3 is called MP3Pro. At the lower bit rates, there was noted to be particular deficits in the high frequency sounds. By adding spectral band replication, the audio, particularly in the high frequencies was improved. It was touted that a 64 kbps MP3Pro file was equivalent to a regular 128 kbps file. Also, the files would be backward compatible with existing players, although they would play at the lower bit rate (64 kbps is roughly equivalent to FM radio). Why is MP3Pro not a common format? First, the most widely available software encoders don’t support the creation of the MP3Pro files. Second, only RCA music players will play the Pro files back. With limits in creation and playback, this gives MP3Pro a poor chance of ever achieving mainstream adoption.

For now, the one thing in favor of regular MP3 is that almost all encoders support their creation, and if a device will play back any compressed audio, it will be MP3’s. Lack of a digital rights management scheme sweetens the deal here, however this is why you can’t purchase MP3’s from online music stores. I would recommend 128 kbps to 192 kbps as reasonable compromises between file quality and file size. For my music, I currently use 128 or 160 kbps VBR for most tracks with excellent results.

Part One: An Introduction
Part Two: MP3's
Part Three: Windows Media Audio
Part Four: OGG, AAC, and a Preliminary Verdict
Part Five: Popular Software
Part Six: Conclusion & Links To Software

Back to Top

0 Comments:

Post a Comment

<< Home