therockman
02-25-2004, 10:59 AM
Digital maximization and digital compression are two forms of interrelated digital distortion that is applied intentionally to the music stored on compact discs. The technical ability to "crank up the volume" (digital maximazation) and to level out the dynamics (digital compression) results in a louder disc that plays loud all the way through, without the dynamic range normally associated with the natural reproduction of music. It is easy to understand the technical aspects of these two intentionally applied distortions if one knows a little bit about digital mastering for the compact disc.
__First of all, all music begins life as an analog signal. When music is converted into a digital stream of information ready to be transferred to compact disc, the music has to undergo a A/D (analog to digital) conversion. This A/D conversion requires that the original analog signal that represents music undergo a process called "sampling" and "quantization". Sampling and quantization are the cornerstones of digital audio and are actually distinctly different but interrelated processes which work to covert music from an analog signal into a series of numbers.
The analog signal that represents music exists as a voltage that varies over time. The faster the signal varies over time, the higher the audio signal's frequency, (this is represented as "time information"). The greater the amplitude of the signal, the louder the signal is. Sampling and quantization is the process of producing a series of binary numbers, called "words," that represent the analog waveform. An A/D converter encodes the signal's time information by sampling the analog audio signal at discreet time intervals and assigning a numerical value to this information. The amplitude information is encoded by generating a number that represents the analog waveform's amplitude.
The combination of this "time information" and the amplitude information is stored as a binary word in the digital realm. Thus, when the binary words are converted back into voltages, the analog audio waveform is roughly reproduced.
Quantization is the most important factor of the A/D conversion process that allows an engineer to digitally maximize and compress the audio signal. It has to be remembered that quantization is simply the generation of a binary number that represents an analog waveform's amplitude. In other words, the binary number that is generated through the process of quantization is a digital representation of the audio signal's analog voltage when the sample is taken. It is easy for an engineer to increase the binary number that represents voltage, thus increasing the amplitude of the audio signal. This process is called "digital maximazation." The process of adding a like amount of digital amplitude information to the entire digital sample (a song for instance) is called "digital compression." By subjecting a digital sample of recorded music to the interrelated process of digital maximazation and digital compression, an engineer ends up with a louder piece of recoded music that plays loud all the way through. This final product is now suitable for radio playback because the artist will now have the loudest song on the radio.
The only problem with the final product is that now the music doesn't have a natural waveform. A sine wave should roll naturally, with nice round tops and bottoms. Through the process of digital maximazation and digital compression, the engineer creates a waveform that is clipped at the peaks (top and bottom) and our ears perceive this clipping as distortion. This is the reason that most "older people" find new music irritating and they get tired listening to new music real quickly. It is not necessarily the music itself that is irritating them, it is the way it is reproduced.
So there you go. This is the story of digital maximazation and digital compression. Even though an audiophile has invested a great amount of time and money into their system, a new CD will still sound terrible simply because of the distorted waveforms that we are reproducing through our systems.
__First of all, all music begins life as an analog signal. When music is converted into a digital stream of information ready to be transferred to compact disc, the music has to undergo a A/D (analog to digital) conversion. This A/D conversion requires that the original analog signal that represents music undergo a process called "sampling" and "quantization". Sampling and quantization are the cornerstones of digital audio and are actually distinctly different but interrelated processes which work to covert music from an analog signal into a series of numbers.
The analog signal that represents music exists as a voltage that varies over time. The faster the signal varies over time, the higher the audio signal's frequency, (this is represented as "time information"). The greater the amplitude of the signal, the louder the signal is. Sampling and quantization is the process of producing a series of binary numbers, called "words," that represent the analog waveform. An A/D converter encodes the signal's time information by sampling the analog audio signal at discreet time intervals and assigning a numerical value to this information. The amplitude information is encoded by generating a number that represents the analog waveform's amplitude.
The combination of this "time information" and the amplitude information is stored as a binary word in the digital realm. Thus, when the binary words are converted back into voltages, the analog audio waveform is roughly reproduced.
Quantization is the most important factor of the A/D conversion process that allows an engineer to digitally maximize and compress the audio signal. It has to be remembered that quantization is simply the generation of a binary number that represents an analog waveform's amplitude. In other words, the binary number that is generated through the process of quantization is a digital representation of the audio signal's analog voltage when the sample is taken. It is easy for an engineer to increase the binary number that represents voltage, thus increasing the amplitude of the audio signal. This process is called "digital maximazation." The process of adding a like amount of digital amplitude information to the entire digital sample (a song for instance) is called "digital compression." By subjecting a digital sample of recorded music to the interrelated process of digital maximazation and digital compression, an engineer ends up with a louder piece of recoded music that plays loud all the way through. This final product is now suitable for radio playback because the artist will now have the loudest song on the radio.
The only problem with the final product is that now the music doesn't have a natural waveform. A sine wave should roll naturally, with nice round tops and bottoms. Through the process of digital maximazation and digital compression, the engineer creates a waveform that is clipped at the peaks (top and bottom) and our ears perceive this clipping as distortion. This is the reason that most "older people" find new music irritating and they get tired listening to new music real quickly. It is not necessarily the music itself that is irritating them, it is the way it is reproduced.
So there you go. This is the story of digital maximazation and digital compression. Even though an audiophile has invested a great amount of time and money into their system, a new CD will still sound terrible simply because of the distorted waveforms that we are reproducing through our systems.