Digital Music

A certain thing made me catch a glimpse at the world of digital music, in which I have been a stranger so far.

Everyone knows that CD = Compact Disk is digital. Philips and Sony defined CD-DA = CD Digital Audio specification or "Red Book" in 1980. The audio signal waveform is sampled at 44.1 kHz, or 44.1 thousand times a second, to generate a series of digital data in 16 bits or 2 bytes each. A CD of 74 minutes is, therefore, packed with the net amount of data of
44.1k x 2B x 2 Stereo x 60sec x 74min = 783 thousand B = 747 MB
and some more control data. Roughly, 1 minute of music takes 10 MB.

When a music CD is read into a PC, a format called WAVE or ".wav" is used. Microsoft and IBM jointly defined its specification. Windows after 95 support it. A control header, capable of designating the sampling rate and the sample bit length among others, is attached to the CD-DA data. When a CD is read and a WAVE file is generated with the sampling rate and the sample bit length unchanged, 1 minute of music also takes 10 MB.

When a music is handled on a PC, a file in MIDI = Musical Instrument Digital Interface or ".mid" is also used. As its name suggests, it is a collection of instructions to instruments connected to the PC. If a simulated sound sources are used instead of instruments, a mere PC can play music. A MIDI file is smaller, because it consists of a series of instructions such as "Play a sound "C" for 0.1 second in piano", but can not handle recorded music.

These years, a format called MP3 in ".mp3" is prevailing. When an ISO standard MPEG-1 for compression of video and audio was defined, three audio compression standards, Layers 1, 2 and 3 were included. The layer 3, or in short MP3, achieved the highest compression rate, and the compressed data are as small as 1/10 - 1/12 of WAVE. Roughly 1 minute of music takes 1 MB. The human sense of hearing is such that, when a dominant frequency component is heard, weaker frequency components near-by are masked and can not be heard. Those could be abbreviated. Also somewhat coarse sampling can be made feasible since noises generated by sampling mechanism could be cleverly distributed within the masked ranges. This theory has enabled high compression rate and has become the basis for the compression techniques after MP3. Even if a modem of 28.8 kbps works in the full speed, downloading music of 1 minute = 10 MB takes 45 minutes. With 1 minute = 1 MB, it takes only 4.5 minutes. In order to download favorite music, people can wait for several times the music length. Hence digital music has swiftly flourished on the Web. Its forerunner, Napster, provided a platform on which members can exchange MP3 music. Since Napster didn't provide music, but only the platform for individual users of music, it doesn't directly infringe the copyright law. But the music industry sued Napster for big damages caused in reality by it, and mass media highlighted it.

AAC = Advanced Audio Coding was created in 1997 as an audio compression standard associated with MPEG-2 through the efforts by Sony and Dolby among others. It supports multiple channels of audio such as "surround", with data 30% smaller than MP3. In Japan, BS-DBS adopted it and CS-DBS added it, making it the main stream of the audio standard for the digital broadcasting.

On the other hand, Microsoft announced in 1999 WMA = Windows Media Audio, which was claimed to use only one half amount of data of MP3 and to be convenient for protection of copyrights. Since 2000, music devices supported it and so did Napster.

Lately provided are portable CD players capable of playing CD's with compressed music, as well as regular CD's. One can carry around such a CD player with a single CD, on which 10 regular CD's are compressed. I bought an IC memory player named "Rio 800". It supports MP3, AAC and WMA and has 128 MB of memory, which can accommodate more than 2 hours of music. The attached software can create WMA files from CD's or download music from the Web, and pack them into Rio 800. I thought first that all the music on the Web was noisy music, but eventually found that it was not the case. There are portals like "www.listen.com" from which all genres of music can be sought for and downloaded. There is also a portal operated by a company called "RioPort" in affiliation with Rio. For instance, Schubert Winterreise Fruhlingstraum, 4:15 minutes long, is in 4 MB and downloaded for $ 0.99. Tango de Milo is 4:31 minutes long and in 4 MB but this is for free. But you see, when I download those into my home PC through a modem, I pay the price for the music and almost the same amount of money to NTT in the phone bill. We must suffer from this disadvantage in the country under development in communication.

Sony provides stereos capable of connection to PC's so that downloaded music can be played. I asked "Any models from other than Sony ?" The reply was interesting: "Sony is an integrated manufacturer. But others are mere audio manufacturers and can not develop such a thing."

At any rate, I was marveled at expanse of the completely new world.

End