High Definition Audio Formats for Music Listening - SACD (2)
Direct Stream Digital Technology

Direct Stream Digital - or DSD recording/playback technology, takes a completely different approach to existing PCM based systems, to deliver the high-resolution stereo or multi-channel audio found on the Super Audio CD disc.

Direct Stream Digital - the Technology behind SACD

This is fundamentally different to the pulse code modulation (PCM) technology used for audio CDs and DVD-audio.

DSD encodes audio data using 1-bit samples taken at an exceptionally high sampling rate of 2,822,400 samples per second.

At the same time, this high sampling rate gives Direct Stream Digital based encoding an extended 'practical' frequency response from DC to close to 100kHz - even though most SACD compatible players would limit this to around 50kHz to 80kHz.

DSD also makes use of strong noise shaping quantization techniques in order to push the 1-bit quantization noise up to ultrasonic frequencies. This leads to a quoted dynamic range within the 20Hz to 20kHz band of 120dB. However, this noise shaping mechanism implies that DSD's dynamic range decreases quickly at frequencies over 20kHz due to the rising noise floor just above the audible range. This is completely different from PCM based encoding, which yields a constant dynamic range over the full supported frequency band.

This noise issue also explains why unlike PCM systems, Direct Stream Digital based audio has a much lower frequency response than the Nyquist frequency limit - which in the case of DSD is 1.41MHz. (According to the Nyquist sampling theorem, the highest frequency response for a band-limited signal is equal to half the sampling frequency).

The use of noise shaping techniques that push noise towards the higher frequency end of the spectrum makes it necessary for SACD players to use low-pass filters to reduce noise in the ultrasonic range - thus limiting the high frequency response.

In most simplistic terms, the process of creating a Direct Stream Digital signal theoretically requires nothing more than a 1-bit delta-sigma A/D converter but without the decimator (used to convert the 1-bit bitstream into multi-bit PCM), to record directly the 1-bit signal; playback would then require just a low-pass filter to reconstruct the original analog waveform.

This is a typical example of 'simpler the better'. The main problem with standard PCM technology is the use of steep (brick-wall) filters to block frequencies above half the sampling rate, as well as the introduction of re-quantization noise for its decimation (down-sampling) and interpolation (up-sampling) digital filters. These problems limit the actual fidelity of the reproduced audio. Direct Stream Digital simply does away with all the filters and records the audio directly in its 1-bit, 64x over-sampled delta-sigma modulated form.

Technically, things are a bit more complex in that it is close to impossible to carry the necessary digital signal processing in the studio e.g. balancing, equalization, etc., in the digital domain in a 1-bit environment. For this reason, a 'DSD-wide' variant was developed for studio applications, which uses an 8-bit word per sample rather than 1-bit. This is somewhat similar to PCM encoding except that the process still relies heavily on noise shaping towards the ultrasonic range. This DSD-wide signal is then down-converted to regular 1-bit DSD for SACD mastering.

Direct Stream Digital vs Pulse Code Modulation

In view that DSD and PCM are conceptually different, it is not possible to make a direct comparison between the two. However, an actual physical analysis shows that DSD is comparable in signal quality to a PCM format having a bit depth of 20-bits and a sampling frequency of 192kHz.

This yields a dynamic range of 120dB between the loudest sound and noise floor level; in contrast, DVD-Audio supports a dynamic range of 144dB for 24-bit/192kHz PCM audio. Further more, as indicated earlier on, this 120dB range is only supported within the 20Hz to 20kHz audible frequency range due to the use of noise shaping techniques that pushes noise in the ultrasonic frequency range.

On the other hand, PCM does not provide anywhere near the same high sampling rate of 2.82MHz as with Direct Stream Digital recording. This exceptional high sampling rate leads to a digital representation that is remarkably close to the analog waveform. On the contrary, the highest frequency in which PCM is being edited in studio applications is 352.8kHz - prior to being down-converted to 192kHz for use in DVD-Audio discs.

Direct Stream Transfer (DST)

Directly related with SACD and Direct Stream Digital technology is the use of Direct Stream Transfer (DST) lossless compression scheme, in order to reduce space and bandwidth requirements of DSD audio.

Direct Stream Transfer compression is compulsory for multi-channel audio and optional for stereo soundtracks. DST can achieve a compression factor of between two to three - thus enabling a SACD disc to contain up to 80 minutes of high resolution multi-channel audio, or up to 100 minutes of 2-channel high resolution stereo sound.

More on Direct Stream Digital Technology will follow soon.	Last updated on: 4th March 2007
Subscribe to our free e-zine to remain abreast with the latest additions. Click here for details.

 

  ...Part 3: SACD Playback Setup and Content Protection

...back to DVD-A/SACD Section Index

 Copyright © - Practical-Home-Theater-Guide.com  All rights reserved.