Stay in touch…

Blog

Read the latest Bitstream

RSS Feed

LinkedIn

Look for us at LinkedIn

Twitter

Follow us on Twitter

Studio Systems Magazine

An article for Studio Systems, published in 2000.

DVD-A — The Story Continues

The consumer audio landscape is changing, and high fidelity is no longer the sole purview of the well to do audiophile. The DVD family has begun to supersede older distribution media and, with DVD-Audio on the scene, the Compact Disc could be eased off it’s throne.

With compelling benefits such as delivery of uncompressed multichannel audio, visual accompaniment and text services, DVD-A shows great promise. In it's basic form, it is physically identical to it’s older brothers, DVD-Video and ROM. The logical data structures are extended to accommodate it’s unique requirements. An Audio Manager, containing visual menus, organizes a series of Audio Title Sets, with multiplexed Audio Objects (AOBs) within. Within an AOB, there can be one or two audio streams, raster still images including color stills and subpictures as well as text files.

Highly Logical

The overall logical volume of a DVD-A disc starts with the Audio Zone, followed by the now familiar Video and Other (ROM) Zones. Within the Audio Zone is one Album per physical side. Within Albums, there are “Groups” which reference Titles containing individual Tracks. Each Album can contain up to nine Groups. As with CDs, a Title can contain up to 99 Tracks and each Track can contain up to 99 indices.

Index ➔ Track ➔ Title ➔ Group ➔ Album

The Group construct is a play list approach to organizing Titles, allowing Titles to be associated in more versatile ways than CD. For instance, one Track could be referenced by several Titles and Groups. For instance, in a retrospective release highlighting a musician’s career, this would allow the author to define one Group with all tracks that include the star performer playing with one collection of side men and another Group with a different collection of backing musicians. The consumer could play each Group, and hear the tracks that illustrate particular stylistic influences on the artist’s performances. Another example would be to simply collect all tracks of a particular type, such as instrumentals, into their own Group. A final example would be to “share” the same short introductory track, an audio logo for instance, between two Titles, thus allowing that track to be stored only once on the disc.

A Plethora of PCM

Concerning the aforementioned PCM audio variations, from 1 to 6 channels of audio are permissible, though at the highest sample rates, channel count is limited to stereo, again due to data rate restrictions. Here’s a breakdown:

Channel Group 1: 1 to 4 channels† — 16, 20 or 24 bits Channel Group 2: 0 to 3 channels — word length ≤ Group 1
44.1 kHz sample rate 44.1 kHz sample rate
48 kHz sample rate 48 kHz sample rate
88.2 kHz sample rate 88.2 or 44.1 kHz sample rate
96 kHz sample rate 96 or 48 kHz sample rate
176.4 kHz sample rate none
192 kHz sample rate none

 

† – Note that there can be a maximum of 6 channels except at 176.4 and 192kHz, where there can be only 2 channels

Looking at the above table, don’t confuse groups of tracks with channel groups, which facilitate versatility during playback: only one channel group or both can be reproduced.

Images

First, still images are available, being multiplexed on disc with their associated audio. The MPEG-2 Intra-frame file can, optionally, be accompanied by a subpicture. Transitions between stills include cuts, fades, dissolves and wipes, though this feature has yet to show up in players. Two display modes are available:

1. A slide show mode where the stills are displayed as they are retrieved from disc along with the audio. Sequential slide shows are buffered material, and are organized into an ASVU or Audio Still Video Unit data structure. ASVUs must be under 2 MB in size as they are loaded into buffer memory when the track is called and played out of memory, not off disc. During authoring, you can specify a presentation duration for each sequential slide.

2. A browsable mode where up to 19 still pictures are pre-loaded into buffer memory before the audio starts. Their display is under either user or program control, random or ordered.

In addition to still images, full motion video (FMV) is also available as in a DVD-V title. The difference is that, when FMV is employed, you are limited to 48 kHz sampled multichannel audio because one third of the total bandwidth is occupied by the video. If you want higher fidelity, you have to be satisfied with still images.

Bandwidth Reigns

Its purely a bandwidth limitation issue. The disc is capable of safely delivering only a 9.6 Mbps data rate and all output comes from that thin stream. But, six channels of 96 kHz audio exceed that rate. Bit of a problem…To the rescue comes Meridian Audio Limited, known for their work on Ambisonics, ultrafidelity consumer audio products and a “High-Quality Audio Disc” proposal to the WG-4, Meridian’s Meridian Lossless Packing (MLP) provides at least 4 bits of versatile data rate reduction without any loss of information. On average, 5 to 9 bits of rate reduction can be expected. In addition, decoding MLP’d data requires only modest processing power, assuring affordable consumer products.

Getting Down To Work

Since most all professionals can now afford multichannel 48 kHz/24 bit systems, its not too difficult to get into “surround” production. If 96 kHz sampling is available to you, so much the better as the master recording will be higher resolution and thus, more “future proof.” If you’re skeptical as to the value of oversampled data, some critical listening should prove the worth of those extra bits. Make your own judgements, rather than relying solely on someone else’s opinion.

Tried and true methods of panning and synthesizing reverb can transform monaural or stereo material into something different, if not necessarily better. If possible, start with multichannel recordings rather than repurposing existing 2-channel work, even if that means 4 channels rather than 5. The center channel can be derived by summing the left and right front in post–production. Since both a stereo and multichannel version can be stored on a DVD-A, it's a good idea to create both mixes during production. If the budget won’t permit creating a separate stereo mix, make sure you determine and document downmix coefficients, each channel’s amplitude and phase, during mixing so they can be entered in at the premastering stage.


Figure 1: MLP options in Sonic Solutions’ OneClick DVD-A authoring


Figure 2: SonicStudio HD set to feed content to OneClick [click to zoom in]

Also note the average sound pressure level since, for backward compatibility to DVD-V, its a good idea to include a Dolby Digital version on disc. As Dolby says, “It is crucial to set Dialog Normalization correctly to avoid peak overloads and unwarranted application of Dynamic Range Compression.” Also, please document your work as much as possible. DVD projects are much more complex than CD jobs and complete labeling and documentation will earn the respect of anyone further along in the production cycle.

A wide variety of small monitors suitable for 5 channel installations are available. These range from film sound–oriented products from JBL and Miller & Kriesel to more neutrally voiced products from B & W, PMC or Dynaudio. Also, don’t discount audiophile offerings that would fit nicely in a mastering environment, such as the NHT, Divergent Technologies or Paradigm families. Power handling, not the forte of audiophile speakers, is a bit less important in a multichannel installation since there are more transducers aimed at you, making the overall sound pressure greater than an equally powered stereo installation.

Transporting your work to a mastering facility greatly depends on which formats are supported. Common choices are DTRS with a Prism or other “bit splitter,” AIFF archives in Retrospect format and SonicStudio native archives. The last choice makes some sense as Sonic Solutions is the only vendor presently selling premastering tools for the DVD-A format. For most high end mastering houses, this is both convenient and unfortunate. Since Sonic has been a champion of the DVD and SACD formats and pioneer tool vendor, they have helped to define the standard. Unfortunately, their progress in providing a comprehensive production suite that’s easy to use has been slow but an integrated tool suite is due by the 4th quarter of 2000. Currently, most of the premastering process is mundane, with textual script interfaces and little or no integration of the mastering, multiplexing, formatting and disc image creation tools. Also, other routine implements for replication, such as disc verifiers, have all been developed in–house. Its interesting that many of the format’s features do not yet appear in authoring tool controls.

Just Getting Started

Its too early to tell whether the launch will be successful. As I see it, a critical imperative is the introduction of true “universal” players that can handle most all-optical release formats, including DVD-R and SACD. Two years in and there is still little progress in improving public awareness and promoting the benefits of this new member of the DVD clan. Though the format is attractive from an engineering and audiophile point of view, it remains to be seen whether consumers will perceive any benefit from the higher fidelity provided by the longer word lengths, higher sample rates and multichannel sound. Some would say that HDCD, DTS and Dolby Digital have already addressed these issues and DVD-A is forcing a weak format onto an audience that is satisfied with Compact Disc and DVD-V technology. One thing is certain: in order to benefit from whatever the public embraces, you had better be ready to handle it in your studio environment. So, study the technology, request real-world demonstrations of the available production suites and, adapt to the changing audio landscape!

Sidebar

With other members of the DVD family surging ahead in marketplace acceptance, DVD-A seems to be having a bit of a problem. No sooner was it scheduled to open up multichannel audio applications to a wider audience than it was shut down by developments in the Linux Open Source community.

What followed was retrenching by both the DVD and the intellectual property protection communities. Heres a brief DVD-Audio time line…

January 1996 – The DVD Forum’s Working Group 4 (WG-4) meets to formulate a new standard for DVD-Audio.

November 1997 – The WG-4 releases a draft specification to members for review.

March 1998 – The WG-4 conducts proof–of–concept shows in New York and Los Angeles for the industry and press, demonstrating what will become DVD-Audio technology. Plans call for a DVD-A specification to be delivered in May, later revised for June.

May 1998 — The DVD Forum issued Version 0.9 of the DVD-Audio Specifications.

February 1999 – The DVD Forum approves Version 1.0 of the DVD-Audio Disc specifications, the fifth member of the family after DVD-Video, ROM, RAM and DVD-R.

March 1999 – Masushita Electronics, together with Toshiba Corp., Intel Corp. and IBM, announce the CSS2 copy protection scheme. The four technology companies chose digital watermarking and encryption technologies to reduce pirating of music on DVD-Audio. To placate home recordists, the companies agree to allow one copy from each disc.

November 1999 – Sonic Solutions announced that it has created, along with Panasonic, Sonopress (BMG Storage Media), Sony Music, Universal Music and Warner Music Group (WMG), the first DVD-Audio test discs.

November 1999 – DeCSS spreads across the Net, allowing anyone with a bit of tech savvy to decrypt DVD-Vs.

December 1999 – Masushita announces that there will be a delay of at least six months in the roll out of DVD-Audio players.

June 2000 – Masushita announces plans for a July roll out of DVD-Audio players in the US. Impressive consumer models, such as their model CQ-DVR909U combination DVD-A/V player and digital radio receiver, begin to appear in the marketplace. Now things start to get very interesting…

Bio

Oliver Masciarotte, based in Northern California, is a technical facilitator for New Media production. For more information on the DVD standard and production practices, visit http://www.seneschal.net/ for a copy of this article and links to related sites and useful information.