High Definition Audio Formats for Music Listening
The DVD-Audio Disc (1) - An Overview

The DVD-Audio disc - also referred to as DVD-A, is one of the new high definition multi-channel audio formats designed specifically for music listening.

In the article, we discuss DVD-A specifications, backward compatibility with DVD-Video, the use of Meridian Lossless Packing (MLP lossless encoding), and the issue of sound quality with respect to standard CD digital audio.

DVD-Audio Basics

DVD-A is a digital format designed for the delivery high-fidelity audio content on a DVD.

It supports a number of different audio configurations, ranging from single-channel (mono) up to six full-range audio channels, at various sampling frequencies and bit depth.

When compared with standard CD-DA, DVD-Audio supports far greater detail and texture, and is more lifelike than CD sound.

The main reason being that it can handle up to 256 times the sample bit depth or quantization levels supported by CD-DA. At the same time, it can also deliver anything up to four times as many samples per second. For this reason, DVD-Audio is mainly of interest to audiophiles and home-theater enthusiasts who would like to enjoy the very best in music sound.

DVD-Audio Specifications

As already indicated in the previous section, DVD-A offers many possible configurations of audio channels, ranging from single-channel mono up to six full-range channels of surround sound, at various sampling frequencies and sample rates, ranging from stereo CD-DA quality sound i.e. 16-bit/44.1kHz x 2-channel sound, up to 5.1 surround at 24-bit/96kHz sound, or 24-bit/192kHz stereo sound.

This is all possible thanks to the much higher capacity of this DVD optical medium in comparison to compact discs. This leads to far higher audio quality as a result of the higher linear sampling frequency and higher sample resolution, better spatial sound reproduction due to the additional channels, and finally, the possibility of extended running time - however, this depends on the sound quality of the recorded audio signal.

Further more, the DVD-Audio format supports additional information like a table of contents, text subtitles, still images and video clips.  Additionally, because the DVD-Audio format is a member of the DVD family, a single disc can have either two layers when delivered over a double layer DVD, (DVD-9), and even two sides that contain media when using two-sided discs (DVD-10, DVD-14, or DVD-18).

In the case of two-sided DVD titles, these would normally come with one side formatted as DVD-A and the other as DVD-Video. A common configuration includes a 'DVD-Video' zone on a DVD-Audio formatted single sided disc. In fact, many DVD-Audio titles come as a combination DVD-Audio/Video disc, with the DVD-Video portion of the DVD being used to include artist interviews and additional bonus material like music videos. 

The presence of the DVD-Video zone on a DVD-Audio formatted DVD renders the DVD-A Disc compatible with DVD-Video players. Thus, while the high-resolution, multi-channel lossless encoded audio is only playable on DVD-A hardware, the DVD-Video zone, which would normally contain also Dolby or DTS mixes, makes the disc compatible with all DVD players.

Other configurations include DVD-Audio, DVD-Video and DVD-ROM content on the same DVD disc. In reality, the possibilities are endless and may include a variety of combinations as supported by the DVD optical medium.

Just one final point about DVD-Video on DVD-A formatted discs. While DVD-A does not include region coding and therefore, is playable on all DVD-Audio players, yet the DVD-video component could be restricted by region codes in a similar manner to a DVD-Video disc.
 

The table below summarizes the many different bit-rate/sampling rate/channel combinations supported by DVD-Audio:

It should be noted that:

• The LFE channel (.1) is actually a full range channel and can be recorded at the same sampling frequency and bit-resolution as the other channels.

• Different bit rate/sampling rate/number of channels combinations can be recorded on a single DVD-Audio disc. This means that it is possible to have say a 5.1 x 96kHz/24-bit track as well as a 192kHz/24-bit stereo soundtrack on the same disc.

• Further more, in the case of a 5.1 sound track, the different channels may be grouped and recorded at different sampling frequencies and different sample resolutions.

For a detailed summary of the DVD-A specifications please refer to the table below. This table includes also the corresponding specs for CD-DA (digital audio) disc; this should help make the comparison between these two optical media much easier.

Backward Compatibility

To ensure backward compatibility with DVD-Video players, most DVD-Audio discs contain as a minimum, a Dolby Digital 5.1-channel audio track on the disc which can be downmixed to two channels by the DVD player for listeners with no surround sound setup. Some discs may also include a DTS 96/24 5.1-channel, soundtrack.

However, as already pointed out in our introduction on DVD-Audio and SACD, today, universal DVD players capable of handling DVD-Video, DVD-Audio, CD-DA, and even SACD, are relatively inexpensive and are selling at under $150.

In other words, the presence of affordable universal DVD-players is rendering backward compatibility a non-issue when it comes to DVD-Audio discs.

DVD-Audio and Meridian Lossless Packing

Audio is stored on DVD-A discs either uncompressed - using Linear Pulse Code Modulation (PCM) format, or lossless compressed using Meridian Lossless Packing technology, also referred to as MLP Lossless™ (or simply MLP).

MLP Lossless™ is licensed by Dolby Laboratories; it is the core technology of DVD-Audio, providing up to 5.1 full-bandwidth channels of pure, lossless audio playback that is bit-for-bit identical to the studio master.

MLP Lossless enables producers to encode up to six full-range channels of 96 kHz/24 bit audio, or two channels of 192 kHz/24 bit audio onto a DVD-Audio disc, using lossless encoding technology - in other words, nothing is loss during the encoding and decoding process.

Whether to use uncompressed audio or MLP Lossless encoding depends on the sampling frequency, the sample bit depth, and the number of channels, this in view that the maximum permissible bitrate supported on DVD-Audio is 9.6Mbits/s.

In other words, if the channel/resolution combination exceeds the maximum bitrate supported by DVD-A, lossless compression becomes mandatory. In uncompressed modes, it is possible to get up to 96kHz/16-bit or 48kHz/24-bit in 5.1 mode, and 192kHz/24-bit in stereo mode. MLP encoding is necessary to record 5.1 multi-track audio in 88.2kHz/24-bit, 96kHz/20-bit and 96kHz/24-bit.

DVD-Audio Sound Quality vs CD-DA

As indicated earlier on in our discussion, DVD-A does provide far better sound quality than CDs as a result of the higher resolution supported by the DVD-Audio format.

But to better appreciate this whole issue of better audio quality, it is important to first have an understanding of why a higher sampling frequency and a deeper sample bit-resolution lead to better sound.

Sampling Frequency

A digital representation of an analog signal is as faithful to the source as much as the digital to analog conversion is able to follow the original raw analog waveform.

In an ideal world, a recording would yield an exact replica of the original signal. However, when it comes to a digital representation, the original source is represented by a series of samples taken at regular intervals - depending on the sampling frequency; therefore, any data in between the samples is lost. Thus, opting for a higher sampling rate would enable the digital representation to follow more closely the signal source.

The diagram below compares the higher quality DVD-A digital representation of a 10kHz signal using a sampling rate of 192kHz, as against the 44.1kHz sampling resolution of a CD-DA output.

CD DA vs DVD-Audio: A digital representation of a 10kHz analog signal

At this point, many would argue that if the human hearing is limited to around 20kHz, then the 44100 samples per second supported by the CD-DA format should pose no problem to handle the full range of audible frequencies, as these would yield a maximum frequency response of over 22kHz.

This reasoning is based on Nyquist sampling theorem which states that for a band-limited signal, the highest frequency response is equal to half the sampling frequency. By band-limited, we understand a signal whose Fourier transform or power spectral density, is zero above a certain finite frequency.

If so, then... why do we need the higher sampling frequencies of the DVD-A format (and SACD) in order to support an even wider frequency response?

In the real world, the acoustical spectrum of many sounds - irrespective of whether these are originating from a musical instrument or otherwise - often extend far beyond the range of frequencies audible to the human ear.

Studies show that though the human hearing is generally very poor above 20,000kHz, yet it is very sensitive to phase distortions. Now, filtering off the ultrasonic frequency component part of the signal would produce considerable phase and amplitude distortions in the audible part of the acoustical spectrum.

You may be tempted to point out that even if we were not to filter these ultrasonic frequency components, the speakers in the playback chain would do it in any case. This is correct, but there is a bit of a difference here, as we will further explain.

Well-designed speakers exhibit a rather gradual sloping signal falloff characteristic at the high frequencies - mostly dependent on the tweeter design. This leads to fewer phase distortions than those resulting from the steep amplitude-frequency response falloff characteristic of CD-DA signals.

Thus, while sampling frequencies at twice the maximum supported by human hearing is a minimum requirement, twice is not fast enough to accurately capture the characteristics - in particular - of the high frequency components in an audio signal. 

The audible effects may be subtle, but undesirable. Aware of the human hearing sensitivity to phase distortions, it is only logical to presume that considerable decrease in the level of such distortions would result in systems with a wider signal spectrum.

Therefore, a higher sampling frequency offers far better accuracy in reproducing the high frequency audio information of an audio signal.

Sample Depth (word length)

The sampling rate is just one aspect out of the two that defines the resolution of a digital signal. Directly related to the sampling rate is the sample depth or sample bit-resolution - often referred to in PCM terminology as 'word length'. 

Each sample of the analog signal is characterized by the magnitude of its voltage value, which in itself can take an infinite number of values. This is not the case with its digital representation, which can have only a finite number of  values depending on the number of bits used in the data word (a series of bits) that will eventually represent that sample value. 

The more bits in the data word, the wider the range of analog voltages that the digitized sample can represent accurately; more bits also means finer gradations of values in that range - thus enabling the digitized sample to follow more closely the actual value of the analog signal.

This also means that the more bits representing a sample, the wider will be the difference between the softest and the loudest sounds that the system can handle. This is referred to as 'dynamic range' and represents the difference between floor noise and the loudest sounds.

The CD format supports a sample length of 16bits, leading to a dynamic range of about 96 dB (decibels). DVD-Audio supports various sample word lengths - including 20-bit and 24-bit audio. 20-bit yields around 120 dB while at its 24-bit maximum quality, DVD-Audio supports a dynamic range of approximately 144 dB.

With today's state-of-the-art audio system components supporting a signal-to-noise ratio of around -120dB, the 144dB is only theoretical in that the limiting factor at this point would turn out to be the noise level inherent in the electronics - mainly due to thermal noise. This means that at present, the 24-bit sample depth supported DVD-A is more than adequate.

Whether you will be able to hear the difference in sound quality between high resolution audio from a DVD-A and standard sound from a CD-DA, is all a matter of the type of audio playback setup that you have. The difference is there and is significant - just try to play the same piece of music on CD and then switch on to DVD-A.  If you have the appropriate audio playback setup, once you hear the sound from a DVD-Audio disc, you will surely be amazed at what you have been missing.

More on the DVD-A Disc will follow soon.	Last updated on: 10th February 2007
Subscribe to our free e-zine to remain abreast with the latest additions. Click here for details.

 

  ...Part 2: DVD-Audio Player Connectivity Requirements Playback setup and Content Protection

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

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