Header image  
by Ralph Glasgal
line decor
Home Tutorials Technical Papers Kudos and Pictures Demos Bio Free Ambio
Glossary Software
Rec Engineers' Corner FAQ/Forum Links Contact us
line decor

Audio Engineering Society
Convention Paper

Scalable Tri-play Recording for Stereo, ITU 5.1/6.1 2D, and Periphonic 3D (with Height) Compatible Surround Sound Reproduction - Page 5
Robert E. (Robin) Miller III, FilmakerStudios, Bethlehem, Pennsylvania 18018, USA


Produced as on page 4, the transformed 3D/2D release plays directly and uncompromisingly in 2D stereo or 5.1/6.1 (5.0/6.0 in absence of a 0.1 LFE media channel). No decoder or other alterations are needed for an ITU-standard home theater until 3D capability is desired. The release is "3D-ready" until that time.

7.1. 3D Reconstitution

When the home theater owner chooses to install 3D, speakers and power amplifiers must be added, as below. Reconstitution of 3D is performed by a decoder (e.g. DSP chip in a receiver now in prototype form) that applies the inverse of the transformation matrix chosen for the recording. One of six or more reconstitution matrices may be automatically selected if implemented in metadata in mastering the DVD-A. Six reconstitution matrices are in Fig. 11 & 12.

Fig. 11.
 4x4 matrix {P} (lightly shaded) reconstitutes six PerAmbio 3D modes i, j, & k without loss for fully periphonic 3D surround.

Fig. 12.
  4x4 matrix {Pí} reconstitutes six PerAmbio 3D tilted modes ií, jí, & kí without loss and restores any tilt for periphonic 3D replay (+30° shown).

7.2. 3D Speaker Layouts

PerAmbio played in 3D is scalable from 10 to 14 to 26 speakers. At this point, 6 channels of reconstituted 3D become 10 or more speaker feeds (plus redirected bass to a subwoofer). Amplifier power for the two front channels may be 100~200W; 8 or more ambience channel power amplifiers may be 50~100W.

Two front speakers must be of high quality and precisely placed for optimal imaging of staged sources. All other ambience speakers may be satellite-quality, as power requirements are less and pure height cues are >1kHz [18]. 8 ambience speakers is the minimum, for which isotropic positioning is ideal (pair-wise opposite is a good second choice); 12 or 24 speakers may be placed randomly and the decoder told where they are in degrees of azimuth and elevation.  This is similar to the case where hall impulse responses are convolved to generate signals for an arbitrary number and placement of surround speakers.  This method can always be used if the recorded ambience (Ambisonic or otherwise) is not provided on the disc.

Isotropic choices for 8 speakers that perform well in the author's tests are a cube, bi-square (Fig. 2, 7, 13), and a modified bi-square after Barbour [18] with UL,UR raised to 60-degrees and DL,DR raised to 30-degrees to be pair-wise opposite and easier to position in domestic settings. 3D replay preserves timbre for all audience members who are not near any speaker. However, this distance may be less, hence the listening area larger, than for ITU 5.1/6.1 as 3D ambience is at lower levels. For serious music listening, all sit in the median plane for added stage localization from the Ambiophonic front pair.

Fig. 13. 8x4 Ambisonic matrix (shaded) for bi-square array (after Gerzon).

7.3. Playing legacy ITU 5.1 using 3D Layout

Once a 3D layout has been installed, listening to the userís collection of stereo and ITU 5.1 recordings is a matter of moving back 26% of the speaker diameter (Fig. 14), with adjustments by the receiverís DSP for different levels and delays. Center feeds both FL, FR with minimal comb filtering as they are closely spaced.

Fig. 14.
  PerAmbio 3D bi-square layout accommodates audience of six using 10 speakers, plus subwoofer(s). For best 3D music, two listeners sit in the median plane. For 5.1, viewers sit back 26% of the speaker diameter, where angles meet ITU standards (DSP changes levels/delays).


At AES114th Convention March 2003 Amsterdam, The Netherlands, attendees heard PerAmbio 3D/2D reproduced on 5.1 (courtesy Meridian Audio) and 6.1 (courtesy Genelec) systems of high quality. The demonstration of a Vivaldi guitar concerto began in 2-channel stereo and switched midway to 5.1/6.1 transformed from 3D. During the stereo portion, attendees could easily hear that moving slightly off the median plane caused the guitar to toggle from a phantom center to the nearer speaker. Upon switching to transformed PerAmbio 2D, two major improvements were observed. First, the center guitar image stabilized and attendees were able to walk freely in the listening space (except near a speaker). Second, the tonality of guitar and orchestra became natural.

In 3D, the author led the Ambiophonics Team in a multi-format comparison demonstration at the AES 24th International Conference June 2003 in Banff, Canada [ 22]. Switching simultaneously made recordings among 19 speakers, approx. 100 attendees compared stereo, ITU 5.1, Ambiophonics 2.0 with hall convolution [ 23], PanAmbio 4.0, and PerAmbio 3D. OCT [ 9] was used for ITU 5.1/6.1 and the system described here and in prior papers [ 6, 7] for PerAmbio 3D

In Banff, George Massenburg told the author "PerAmbio wins hands down-you really make your point [about the importance of height/3D]."  Jean Marie Geisjen: "I like PerAmbio best, but 5.1 was close." Tomlinson Holman: "I liked 5.1 best for timbre but PerAmbio second because of height effect."  David Griesinger: "Nice sounding 5.1, but in the sweet spot Ambiophonics does what nothing else can - you are there."  David Josephson: "PerAmbio is most impressive; 5.1 second."  Jim Barbour: "The addition of [PerAmbio] is a great idea...and sounded great...imaging with crosstalk cancellation is sensational."

While these listening ìtestsî are informal, in both Banff and the author's studio in Bethlehem PA USA, the response upon hearing PerAmbio 3D is typically wide-eyed and smiling, if not a jaw-dropping. Pianist Fr. Se·n Duggan: "It puts you in the middle of the performance."  Oboist Lisa Kozenko: "Finally itís my sound after all these years."  Audiophile James Robertson: "I can't tell which speakers are playing." and "It doesn't sound like THIS [control] room, it sounds like THAT [studio] room!"  Composer Howard Moscovitz: "This is the way [surround] is going to be."  Conductor Donald Spieth: "I want it; when can I have it?"

ANOVA testing is planned to confirm results.

<< Previous Page | Next Page >>

Article Pages 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8

^ Back to Top ^