Audio decoding: key process for the 'real' home theater

To unlock the potential of audio from video DVDs, many companies offer 24-bit digital signal processors. For example, Analog Devices provides 32-bit audio decoder along with 24-bit/192 kHz D/A converters, high-bandwidth amplifiers and volume controls. A 32-bit hardware configuration is able to decode and post-process more than one DVD algorithm, offering the design flexibility of locating the decoders in audio-video (A/V) receivers, DVD players, or even in loudspeaker enclosures.

The need for 32-bit audio decoders for 24-bit DVD audio results from the quantization noise generated by the decoder algorithm operating within the 24-bit D/A converter's noise floor, introducing audible noise and distortion. Implementing double-precision routines (involving repeated computation) and error feedback schemes to work around the noise floor problem adds DSP computational overhead by a factor ranging from 4 to 10.

For instance, to implement even simple second-order 17 Infinite Impulse Response (IIR) digital filters on 24-bit samples without resorting to double-precision math requires extra bits to execute the decoder algorithms as efficiently as possible. This points to the need for 32-bit/192 kHz decoding of 24-bit/48 kHz-encoded audio.

All DVD audio is digital audio. This means it must be sampled and converted to an analog signal to drive loudspeakers or headphones. The number of bits in the sample represents the dynamic range of the audio; the sampling frequency determines the maximum frequency range. A 24-bit sample results in a theoretical 144 dB of dynamic range. The digital audio sampling rate should be slightly higher than double the highest frequency one intends to reproduce. To reproduce up to 30 kHz, for example, the signal should be sampled at 60 kHz. Although most people cannot hear this frequency, tests have shown that harmonics and the effects of room acoustics enable some audio above 20 kHz to be heard, or at least felt. The "double" sample rates in the DVD-A specification, 88.2 and 96 kHz, are more than sufficient to reproduce ultrasonic audio frequencies.

The audio on video DVDs in NTSC broadcast television regions of the world must be encoded in at least one of two audio coding formats: PCM or Dolby Digital. Current video DVDs contain up to 5.1 channels of Dolby Digital or up to two channels of PCM and can have an additional track of DTS encoded audio. In regions using the PAL broadcast TV standard (Europe, for example), the required formats for DVDs are Dolby Digital, PCM or MPEG (and optional DTS).

The basic form of digital audio, pulse code modulation (PCM), is the same linear uncompressed digital audio format used on audio CDs, and is used to master all digital recordings. PCM on video DVDs uses up to a 96-kHz sampling frequency and up to a 24-bit sample, but only for stereo or mono. For 6-channel PCM, video DVDs can provide 48-kHz/20-bit audio. PCM can also use the same 44.1 kHz frequency and 16-bit sample as for audio CDs. The higher video DVD sampling frequency and sample size enhance the audio quality over CDs in stereo playback.

Video DVDs rely on audio and video compression to fit their source material into the space and bandwidth limitations of the DVD media. By slightly compressing the audio, space is made available for video and other features — such as multiple languages, versions and audio mixes — without sacrificing perceptible audio quality. The audio compression schemes used remove only the redundant data and the sounds that cannot be heard because they are masked by other sounds. Since the incoming audio stream is altered by the compression such that all of the original PCM data cannot be recovered on playback, it is sometimes called "lossy" compression. The result is an audio track as small as 1/15th the size of the uncompressed PCM master (6 channels x 48 kHz x 20 bits coding/384 kbits/sec = 15:1). However, the 6-channel sound — with six discrete audio channels all the way from the recording microphone to the loudspeaker — will be a noticeable improvement when compared with mono or two-channel sound.

The audio on many video DVDs is compressed with Dolby Digital, a lossy up to 5.1-channel format with compression ratio ranging from 3.4:1 to 15:1. After decoding, discrete channels emerge for left, right, left rear, right rear and center channels, with a separate bass channel for added high-impact effects. Dolby Digital uses a bit pool, which assigns bits to channels as needed, rather than maintaining a fixed number of bits per channel. It uses fixed bit-rate encoding, which on video DVDs may be set at different data rates from 64 to 448 kbits/sec, depending upon the number of channels and the audio quality of the input signal.

By reducing the audio data rate by some 90 percent, Dolby Digital makes room for a variety of features (different languages, mixes, video) and fits through all of the available digital pipelines (digital cable, digital television, digital broadcast satellite) without sacrificing significant sound quality.

Early multi-channel recording techniques, like Dolby Pro Logic Surround, encodes four channels-left, right, center and a single mono surround channel-onto two tracks of conventional stereo program sources-CDs, CD-ROMs, audio and video cassettes and regularly-scheduled broadcast TV programs. A matrix technique recovers the encoded material, but with limited channel separation and bandwidth.

More compatibility

Thanks to DSP, practically all current video DVD players and DVD-ROMs include either a stereo or multi-channel built-in Dolby Digital decoder. Thus, Dolby Digital supports Down Mixing to match the number of loudspeakers at the decoding site. Despite the number of channels encoded, Dolby Digital DVDs will play on all existing DVD players and through all existing audio systems, whether mono, stereo, Dolby Surround Pro Logic or Dolby Digital 5.1.

This backward compatibility enables all Dolby Digital video DVDs to play on earlier-generation equipment, as long as they include Dolby Digital decoders, but they must include multi-channel decoders in order to produce Dolby Digital 5.1.Users can listen to Dolby Digital encoded DVDs with their existing stereo audio systems and update them for surround sound or Dolby Digital 5.1 later.

DSP-programmable features include dynamic range control, which moderates the playback volume of material with a high dynamic range. This, combined with "dialog normalization", which smoothes variations in subjective loudness from different audio sources, creates the possibility for a personalized, "better-than-movie-theater listening experience." Many of these effects can be combined to compensate for loudspeaker deficiencies, hearing impairments, room acoustics, type of music, background noise, listening location in the room, and personal preferences.

DTS is a different up audio format. Though it supports up to 5.1-channel with lossy compression, it is compacted less than Dolby Digital and requires a higher data transfer rate. This is typically 1.509 Mbits/sec — about 3.4 times the maximum Dolby Digital data rate authorized in the DVD video standard of 448 kbits/sec. Since it takes up more physical space on the DVD disk, DTS-encoded DVDs typically do not include multiple languages. DTS compresses the six channels independently with a 3:1 ratio, removing two-thirds of the redundant PCM data and multiplexing the remaining one-third into a single bitstream.

In addition, playback effects and enhancements can be added to the decoded audio. For example, Lucasfilm THX provides "Bass Peak Management" and "Loudspeaker Time Synchronization" to enhance the accuracy in the variable acoustics of home environments, re-equalization to smooth the sound for listening in a home theater, and "Timbre Matching" to reduce the perception of a change in timbre as signals cross from front to rear surround speakers. THX Surround EX enables Dolby Surround EX encoded movies to be reproduced in a home theater with a third rear surround channel.

Sony has designed theatrical Digital Cinema Sound, which duplicates the multi-channel experience possible at one of the Sony Pictures Entertainment dubbing theaters for a home environment with as few as two stereo speakers.

Essentially, video DVDs enable the audio to have greater realism and sense of envelopment by including 3-D sound. Certainly, this offers new opportunities for composers, recording artists, sound designers and creative sound mixers. They can enhance sound reproduction by taking advantage of the DVD storage capacity (17 Gbytes for up to 8 channels of multiplexed audio), resolution (up to 24 bits), sampling rate (up to 192 kHz) and data rate (up to 6.14 Mbits/s for audio or 10 Mbit/s for audio and video).