In part 2 of this article, we looked at how a Pulse Code Modulation (PCM) signal is created from the analog representation of the audio signal (graphic 1) to create CDs and DVD-Audios. One of the limitations of this system is that the highest frequency allowed to pass through the analog to the digital converter is one half the sampling rate (usually called the Nyquist frequency). We therefore have to filter the signal at the input to eliminate all high frequency information before it reaches the converter using an anti-aliasing, low-pass filter. Typically, however, we would like to have as much high frequency content as possible, so this filter is usually designed to let in all of the frequencies up to the Nyquist frequency, and it eliminates everything above it.

Once upon a time, these filters were designed and built as analog devices--complicated (and therefore expensive) designs that used specific arrangements of resistors, inductors and capacitors that magically filtered the signals as required. Unfortunately, more often than not, the filters tended to have detrimental effects on the quality of the surviving audio signals. As digital systems improved, it was decided that this filtering could be done by a computer using Digital Signal Processing or DSP. So, instead of passing the signal through an analog, anti-aliasing filter and performing the level measurements directly on its output, a new, cheaper, and higher-quality system was created.

In this system, which is used in almost all digital recording devices today, the analog signal (graphic 2) is first converted to a digital stream of bits using a method other than PCM. This method, known as Sigma-Delta Modulation (SDM) uses a single bit alternating very rapidly between Low and High (0 and 1) to describe the level of the analog audio signal. Unlike a PCM waveform which looks like a staircase simplification of the wave, the result of the S-D (igma-delta) converter is an odd looking square wave (graphic 3) where the average of a number of adjacent bits in the stream is the level of the original waveform. Therefore, when the analog level is high, there are more 1’s than 0’s in the digital stream. When the analog level is zero, there is an equal number of 1’s and 0’s.

In a PCM system, this S-D stream is subsequently filtered using DSP and then converted into a PCM representation before it is stored on DVD-Audio or on a CD. However, a couple of years ago, some engineers started asking themselves why it was necessary to do the conversion to PCM at all--why not just store the stream of bits from the - Modulation converter? The result of this questioning is what we now call irect Stream Digital or DSD--the system that is used in Super Audio Compact Discs (SACD). The signal that is stored on the disc is the stream of single bits that would normally be used as the raw materials from which a PCM signal is created.

The advantage of this system is that the severe filtering that’s needed to create the PCM conversion is no longer required--a much gentler filter will do nicely, thus having much less negative impact on the audio signal we’re trying to record. In addition, there is the minor technical advantage that the system uses 1 bit (rather than CDs 16 bits or DVD-Audio’s 24 bits) to describe the level of each sample. The technical disadvantage is that a DSD signal must have a much higher sampling rate than a PCM signal. In fact, in an SACD, the sampling rate is 64 times that of a CD--a rate of 2 822 400 samples per second (or approximately 2.8 MHz).

Similar to its DVD-Audio competitors, the SACD was also designed with the consumer and the retailer in mind. The manufacturers realized that people still hold a grudge against record companies for making them buy CDs to replace their vinyl--and probably woundn’t be interested in repeating the procedure to replace our compact discs. Consequently, SACDs are capable of “backwards compatibility.” This means that not only will you be able to play your CDs in your SACD player, but in many cases, you will also be able to play your SACDs in your old CD player. This is possible because an SACD can have two separate layers of information--much like a two-layer cake. The top layer contains the information that is read by a CD player--in fact, as far as the CD player is concerned, it isn’t even aware that an SACD is being played. However, this layer is semi-transparent, and it allows an SACD player to read the layer below by looking through the top one. The bottom layer can be used to store over six times more information than the top layer. This “high density” layer can contain two separate versions of the same recording, a stereo (2-channel) version, and a 6-channel surround sound version in addition to a small amount of extra graphics, pictures, video or text. Also, since most CDs today contain a PCM signal which is the result of a converted DSD stream, the CD layer on an SACD can be created using the DSD recording for the high-density layer, making it simpler for the recording engineers and equipment manufacturers.