Toronto AES Bulletin May 1996 Meeting Review

---

Review of SurroundView '96, May 25

In May of 1996 the Toronto AES section's annual seminar attracted about 100 participants, for an all-day grand tour of surround sound technology and its practice. Earl McCluskie, Toronto AES Chairman, welcomed us, at 9 AM, and introduced Tomlinson Holman (the man responsible for THX ), of TMH Corporation, who gave us a brief introduction, and in turn introduced the other speakers and their subject specialty.

Ian Masters, of Stereo Review, Sound and Vision, and the Toronto Star, held a session called "Surround Sound at Home: The Consumer's Perspective". He gave a brief history of multichannel sound, relating it to similar developments in music history. Over 20 million Dolby Surround decoders have been sold-- check the latest statistics at http://www.dolby.com/stats.html -- about 1 million in the Toronto area alone! The new DVD format promises to bring home discrete surround sound for those wanting the best possible sound for videos. Mr. Masters made the point that while discrete surround in the home is an exciting development, for the time being the matrixed Dolby Surround is the de facto standard. Large-screen video and related speaker placement problems can compromise HiFi sound reproduction. Mr. Masters contributed a nice 9 page "Home Theatre Basics" handout.

Brad Hohle, Dolby Laboratories, next explained that 4 channel matrix coding in a stereo pair originated from movie makers' need to reduce printing cost and time, and to avoid a separate audio medium. So Dolby Stereo was evolved, compatible with 2 channel Stereo, to give Left, Center, Right, and Surround (or ambiance/effects) channels. The center channel is encoded at -3 dB, surrounds are encoded at -3 dB and at +/- 90 degrees, from 100 Hz to 7 kHz. Over 20 million chip sets have been sold to consumers. Many TV programmes and all contemporary videos are Dolby Surround encoded. New uses on the way: radio shows, video games, CD-ROM and personal computers.

Digital, discrete 5.1 channel, AC-3 coding evolved also from movie sound, and as an audible improvement on matrix sound. It is used in digital cable systems, business (DMX), soon in DVD and HDTV, and is backward compatible with Pro Logic. New features recently added to AC-3 decoders: channel display coding, 20Hz-20kHz in all 5 channel, time delay features, dynamic range control, compression options in production. Do remember that if combining 5 channel discrete into 2 channel stereo mode, you will gain peak level, so watch for clipping. A question and answer session at the end of Mr. Hohle's presentation uncovered a problem with the greater available dynamic range: movie directors now often drop the dialogue levels, to accentuate sound effects (at near-clipping), while cinema operators drop the reference level by 3 to 10 dB, to positively avoid clipping. Meanwhile, the many people cannot hear the dialogue! Ditto for late-night movie watching at home: always fiddling the volume control to cut the peaks.

After a coffee break Tom Holman presented the evolution of stereo and multichannel sound technology. Many movie multichannel systems were experimented with from the '30s to the '50s. Eventually two channel stereo became the audio standard because of cost considerations, the LP record groove and broadcast channel capacity. In the '70's quad sound died an ignoble death due to many factors, not the least of which was a lack of understanding on the part of the industry of how to make use of the technology. Perhaps quad was doomed to fail because of the layout of the loudspeakers. Dolby rescued the technology from oblivion by turning the speakers into the L,C,R,S setup so familiar today. The centre channel was based on good psychoacoustic principles (improved clarity/intelligibility, stable imaging) and in particular the need to anchor dialogue on the screen and is now universally accepted as being crucial.

Problems in mixing for surround were covered. Matrixed surround suffers from a narrowed and unclear sound image, so mixers compensate. Some cancellation occurs in mono reproduction (as in an infamous Miami Vice episode where the gunshots vanished in mono!). Mixers: use a decoder to check your work; do not rely on meters! Mr. Holman also contributed an 11 page handout, "The Number of Audio Channels", with extensive references and descriptions of his and other psychoacoustic experiments.

John Eargle, recording engineer for Delos Records and author, followed with a primer on recording psychoacoustics, mic types and placement, and mixdown strategies. He has been recording large scale classical ensembles on an 8-channel medium for some time in an effort to future-proof his recordings. He can then mix the various channels down to the medium of choice, matrixed surround for two-channel carriers or discrete 5.1 for future formats. His standard small orchestra mic setup has an ORTF pair behind and above the conductor, flanking wide-spaced omnis at 6 dB down for more spread and as a house pair, a cardioid pair for center soloists, and a wide-spaced pair of cardioids for house ambiance. It is important to remember that directional cues rely on phase differences below 700 Hz, given reasonably similar levels, but on amplitude differences above 2 kHz. These are converted to a location cue in the brain, leading to listener-position dependence in standard stereo treble reproduction. When panning mono into stereo, the 2 kHz cancellation dip and a 6 kHz addition boost cause as much as 15 dB error in the mix and image location to be uncertain. Use spot mics on the bass section and on all percussive instruments.

Discrete 5.1 channel surround recording of acoustic sources is a new format demanding new mic techniques. Second-order (74 degree or narrower) cardioid mics are not yet available, but three at +/- 74 degree angles can, via a discrete 3 channel and L,C,R speaker set, accurately image 5 locations or channels with good listener-position independence. Mr. Eargle contributed a 10 page handout on the principles and details of sound localization, and 14 more pages from his "Handbook of Recording Engineering", on the math and practice of multichannel film and home sound.

A half-hour sandwich lunch was next providing much needed refueling, and of course an opportunity to meet with the speakers and other attendees for discussions.

Tomlinson Holman then analyzed the question "How Many Channels?" The three main issues are psychoacoustics, physical acoustics, and the composer's intent, to deal with three types of sound: direct, reflected with directionality, and amorphous reverb. The brain is capable of 1 degree acuity in front, much worse from the sides, rear, and up or down. Therefore, it makes sense to concentrate on the front. Michael Gerzon suggested that for true realism about 1 million channels of sound ought to do the trick. Holman's point was that we should try not to lock ourselves into formats which limit future developments (as happened with two-channel stereo for many years). For the time being the state of the art in home theater requires at least five channels. More channels is better but we hit a point of diminishing-returns at eight.

With DVD-Audio we have a bit budget. Holman argued that 16 bits @ 48 kHz is sufficient as a delivery medium to consumers. We have reached the reasonable limits in frequency response and in dynamic range, but not directivity. The selling point for consumers for a possible DVD-Audio format will not be more dynamic range or higher frequency response but more channels. Several groups are attempting to define an audio only standard for DVD. Tom Holman and John Eargle are co-chairing a committee on this topic for the AES.

Demonstrations were next and took place in three different rooms in our venue, Roy Thomson Hall: the large Rehearsal Hall (room for a full orchestra), and two meeting rooms (about the size of large living rooms). All three setups had several thousand Watts of power on tap, to drive 5.1 channel discrete and/or matrixed surround systems. Attendees broke up into groups of about 30 and rotated with 40 minute sets in each room.

The Toronto AES executive is very grateful to Mr. Paul Barton of PSB for outfitting two of the listening rooms with extraordinary collection of loudspeakers, amplifiers, and other equipment. For example the Rehearsal Hall had 5 PSB Stratas Golds, 3 PSB subwoofers, 4 dipoles, many NAD power amps, and an audiophile unit from EAD which handled Dolby Digital (AC-3) and Dolby Pro-Logic decoding.

Three presenters worked in the Rehearsal Hall. Brad Hohle, Dolby, demonstrated the audible differences between Dolby Pro Logic and Dolby Digital using specially chosen music. AC-3 was not susceptible, as you might imagine, to location shifts by loudness shifts, and had full-range sound in 5 channels. Brad acknowledged that mixing for matrix is much trickier, and that the depth, spaciousness and clarity is better in discrete. Sub and special-effects require about 20 KW in a large cinema, 2 KW in a small hall.

Tomlinson Holman demonstrated some of his favourite sounds from several different media. Very impressive and clean. Imagine this setup in a living room 10 times smaller in volume!

John Eargle played some of his favourite recent recordings in hall one, including a superb Tchaikovsky 1812 Overture with orchestra and choir which he mixed from DA-88 on the fly. This demonstration prompted a very positive reaction from the audience. SurroundView attendees are still talking about this one! Mr. Eargle made an argument for using full bandwidth direct radiator surround speakers for music recordings (as opposed to dipole surrounds). Because of the abundance of speakers supplied by PSB he was able to demonstrate the difference in sound by switching loudspeakers!

Two presenters operated out of Room 2.

Ian Masters demonstrated some audio spectaculars using only matrixed surround from S-VHS and laserdisc sources through a Pro Logic decoder (Lexicon). The results were very satisfying, and a good illustration of just how good matrixed consumer surround sound and video can be right now.

Neil Muncy used the same setup to demonstrate some very good quality TV audio and video from S-VHS tape, showing the channel separation on a display of his invention. This was followed by some of the worst samples taped off cable in the Toronto area. The best audio was very good but the worst had terrible problems from fake stereo, or from misadjusted cable and/or TV studio equipment. Mr. Muncy has been on a personal campaign to educate the broadcasters (he is responsible for many improvements already) and suggested that whenever we encounter poor quality we should complain!

Paul Smith, one of the developers of DTS Digital Surround Sound, presented his technology in Room 3. DTS is well known as a theatre sound format but is now pushing into the realm of the music industry. Mr. Smith brought with him brand new remixes of songs from artists such as Seal, Boyz II Men, and jazz groups. The music is mixed to 5.1 discrete channels (up to 24 bits) and delivered to DTS for encoding. The algorithm is optimized for music reproduction and is able to code all the channels on a special CD. The CD can be played on any CD player with a digital output which is used to feed the DTS decoder. Listeners were very impressed with the results.

Following the demonstrations attendees gathered again in the Rehearsal Hall for a panel discussion and Q&A featuring the day's presenters. The lively session capped another exciting day long event for the Toronto Section.

Presenters and attendees alike commented on the day's good mix of the practical and the theoretical and the very high level of the presentations.

Thanks to SurroundView co-chairs Earl McCluskie, Peter Cook, and Glenn Specht and others on the executive for their work in organizing the event; Paul Barton and PSB for the use of their marvelous equipment; and to all the presenters for making SurroundView'96 such a success.

Review by John Fourdraine

Web ed's note: Check back again for pictures of SurroundView96.

| Preview of this Meeting | Back to the Nov-Dec 96 Issue |

| Toronto AES Home | Current Bulletin | Future Meetings | Back Issues | Search |