| |
Digital Audio on a Windows PC
Most PC based hardware uses a technique called Pulse Code Modulation (PCM) to store
music data. Electrical signals are recorded as digital values to recreate an analog signal.
Sound waves, such as those from a microphone, are turned into numbers by a circuit called an Analog to Digital
Converter (ADC) and then saved in computer memory. The numbers that create any given
sound, or music, can be saved in a file on the computers hard drive and then be
manipulated very easily. When a file is played, a Digital to Analog Converter (DAC)
converts the numbers back into analog electrical signals. Audio equipment then amplifies
these signals and sends them to speakers or headphones, where you hear them as sound.
The common audio file format for PCM audio on a Windows is called a “wave” file,
identified by the WAV file extension. However, many other common formats
exist. AIF for example, the Audio Interchange File Format, is used for high quality audio on Windows and also used on Macs.
More important that the file extension, audio has standard characteristics.
These characteristics are SAMPLING RATE, BIT LENGTH , NUMBER OF CHANNELS, and COMPRESSION.
SAMPLING RATE
When recording a sound, the ADC must select and measure a sample (a discrete instant in
the sound wave), and store its amplitude (a measure of its loudness) as a number in your
computer. To capture a sustained interval of sound, this process must be repeated very
rapidly. The sample rate is the frequency with which the sampling process occurs.
Sample rates are generally expressed in Kilohertz (kHz), or thousands of cycles per
second. What is commonly referred to as “CD quality” requires a sampling
rate of 44.1 kHz. Since most FM Radio stations trim the size of their signal,
“FM radio quality” music is associated with the sampling rate of 22.05 kHz.
BIT LENGTH
The amplitude of each sample is expressed as a number and, like all values stored in a
computer, these are binary numbers. The electronic “digits” that represent
numbers are called bits. Bit resolution (also called “bit length” or
“sample size") is the number of binary digits that make up each sample. Eight
bits can represent a range of values from 0 to 255, while sixteen bits can represent
values from 0 to 65,535. The larger bit length allows audio to be reproduced at a
higher quality level.
NUMBER OF CHANNELS (MONO, STEREO, 3.1, 5.1, 7.1)
A STEREO signal consists of two streams of data working together, one assigned to the
left channel, the other to the right. Therefore, at any given sample rate and bit length,
a STEREO file requires twice as much data as its MONO equivalent. A MONO signal is one
stream of data for one audio channel. When you play a MONO track on a STEREO output, the track data is reproduced equally on both speaker channels. 3.1, 5.1, and 7.1 are descriptions of three, five, or seven discrete audio channels and the .1 indicates an audio data channel for the bass frequencies.
COMPRESSION
To increase the speed of processing of large amounts of data, or to shrink the hard
disk space necessary to store large audio files different types of compression may be
used. Before the data is stored, some of it is removed and, on playback, this data
may be artificially restored. In the case of MPEG files, for example, hearing models are used to discard some audio completely. Depending on the compression scheme used, the level of compression, and the original
quality of the audio file before compression, the music played back can sound slightly (or
greatly) different than the original sound. Click here for more detailed information on Audio Compression.
The Sizes of Audio File Formats
The size of a file is a constant calculation based on sample rate, bit depth, number of
channels, and compression. Remember that a musical dead space takes
an equal amount of space as a music file with data in it. Considering all audio charactaristics, and modifying for compression, smaller source files represent a lower maximum quality level of the audio data.
Any size less than full size is a compromise, the question is how much do you want to compromise.
Sizes for an uncompressed 16 bit, 44.1 kHz PCM WAV music file:
Size in Bytes |
1 minute |
2 minutes |
3 minutes |
Mono
Stereo |
5,168 (5.2 MB)
10,336 (10.3 MB) |
10,336 (10.3 MB)
20,672 (20.6 MB) |
15,504 (15.5 MB)
31,007 (31 MB) |
Calculations are based on the "correct" value of a 1024 bytes/KB - NOT the rounded value of 1000 bytes/KB
|
