Loosely based on several established techniques for recording acoustic music and ambient sound, including:
- The binaural-like stereo-compatible OSS (Optimal Stereo Signal)* by Jürg Jecklin.
- Bilateral Ambisonics**, by Zamir Ben-Hur, David Alon, Or Berebi, Ravish Mehra, and Boaz Rafaely.
- The horizontally spaced, vertically coincident 4-microphone phased array known as OCCO***, popularized by Robert Sandy “Boojum” Noyes, Jim Norman and Tony Faulkner.
Description: Two Ambisonics subarrays are centered over each side of a baffle 25cm from each other, and stereo sets of small diaphragm condenser microphones are placed further out 36cm and 79cm from each other.
This allows for unobtrusive hyper-realistic location recording from a single stand. Spot microphones are often not required, because pattern control for beam-forming or “virtual spots” can be used.
How each pair works:
- The Ambisonics microphones deliver immersive highly Jecklin-affected* bilateral** signals.
Why 25cm? Empirical testing shows an SRA (Stereo Recording Angle) of ∽81° for the Jecklin effect. OVJ considers that the combination of microphones used together with a Jecklin-type baffle should preferrably target that same SRA to avoid positional smearing.
Since the capsules on each side of tetrahedral FOAs are angled at 109.5°, they can be rotated so the left-pointing capsule on the left (5) and the right-pointing capsule on the right (12) are 25cm apart and parallel to the horizon, instead of the conventional skewed layout. The resulting B-format signals are correspondingly rotated. This method can be called "rotated for symmetry" (MFS).Arguably, the position of each FOA should be referenced to the central point between its capsules. Alternatavely, the outer capsules can be used as references for the possibility of directly usable outputs from them for real-time stereo, within the desired SRA.
- The middle pair (1, 2) adheres to traditional Jecklin OSS specs, with an inter-microphone distance of 36cm and provides hyper-realistic stereo.
Omnidirectional capsules are preferred.
Cardioids can also be used at 70.05° delivering a similar SRA with a stronger intensity component than standard OSS. We use Schoeps MK5 which can be quickly switched among both patterns.
- The outermost pair (3, 4) is accordingly spaced 79cm and provides additional spatial cues, forward gain and some front-back differentiation when using APEs.
Again, omnis are preferred but cardioids at 0° can be used if conditions are sub-optimal.
This pic shows the OVJ array being used for an opera recording.
Preliminary findings:
All pairs can be combined when phase-coherent transcoding and rotation are used for the Ambisonics arrays, as provided by Audio Brewers plugins, and the DAW correctly matches phase among all channels, but slight comb filtering can present itself at equal levels.
LF forward gain can be obtained by LPF the sum of all capsules on each side of the array, this gets rid of comb filtering and results in a balanced starting point which sounds more natural than using equalization to push up the bass range.
Hyper-realistic immersive results are available with modest processing. Suggested for Dolby Atmos within Apple Logic Pro:
Audio Brewers Advanced Decoder output from the Ambisonics arrays to all corresponding channels. The internal upscaling of this decoder effectively evades shortcomings of DAWs which lack high-order mixing.
Middle pair mostly to L and R mixed with the Ambisonics LR feeds – or not depending on the type of microphones, content and desired amount of mono-compatibility. This is an important technical and aesthetic decision.
Alternatively, L and R can come from microphones 1 and 2 and Ambisonics can be used exclusively for surround and top channels.
The outer pair can be mixed in to “openness” while mitigating effects of the baffle if needed.
Immersive reverberation based on backfacing Ambisonics beamforming or dominance processing can allow for less dependance on artificial processing.
More preliminary findings:
- The Jecklin effect is not lost if (real) spots are added to the mix, if they are carefully time-aligned to the array. Surprisingly, the same holds true for a near-field Decca tree as long as it is kept at a modest level.
- Dolby to stereo works very well. The semi-binaural cues from the Jecklin effect survive transcoding and Atmos processing, delivering an increased impression of “being there” despite the limitations of a single pair of loudspeakers.
- The experience on headphones is similar to binaural, to an extent that some might consider binaural conversion of the stereo signal optional when head-tracking is not required or possible.
- Excellent mono-compatibility, especially when a significant portion of the signal originates from Ambisonics. Lateral components mostly cancel out, unlike traditional binaural which falls horribly apart when summed to mono.
- It seems baffle absorption is critical for vocals and similar sounds in which comb filtering can be very noticeable. Software like Acon DeFilter can help mitigate this issue but can introduce artifacts of its own.
- Since we are more sensitive to horizontal than vertical positioning, vertically channelized surfaces on the baffle surface do better than pyramidal or horizontally channelized surfaces.
- When placement is less than ideal, spatial accuracy can be traded for clarity by using beamforming. If this trick goes too far you might not be able to deliver such a highly realistic immersive version but still deliver a usable stereo version.
Scientifically measuring SRA of OSS, and/or factors that determine it, in order to work at other (better?) SRA targets for inter-microphone distances and angles.
Signals from high-pedigree omnis replacing Ambisonics W signals, to mitigate shortcomings of low-priced FOAs.
Higher-order Ambisonics microphones and processing.
Ambisonics also on the outer pair.
Pointing the outer pair up and using tighter polar patterns for more Atmos-like height encoding.
Replacing the FOAs for double-MS or native “Horizontal-B” microphones.
* The “Jecklin effect” also known as Optimum Stereo Signal is a binaural-like acoustic process accomplished with a round vertical baffle between matched capsules. This delivers spatial cues from the varying intensity and frequency response depending on incidence angles, as well as time precedence caused by inter-capsule distance. Unlike dummy-head or current software-based binaural, this method is perfectly compatible with stereo loudspeakers.** This is not by-the-book Bilateral Ambisonics, but a simplified process with FOAs panned to either side within an Atmos workflow, where spatial cues from the Jecklin effect are presented to the listener somewhat separately for each ear, similar to the way traditional OSS works on standard stereo loudspeakers.*** Similarly, OVJ inherits from traditional OCCO the concept of the horizontally spaced, vertically coincident 4-microphone phased array, but not the practice of setting the inner pair to a different SRA than the outer pair.The OVJ name is inspired by Christian Amonson and his outside-in nomenclature. The V stands for “variable” to portray variability through Ambisonics.Angle and distance diagrams come from the Neumann Recording Tools app.
This content by @ignace / Ignacio Rodríguez de Rementería is open source licensed under CC BY-NC-SA 4.0 https://creativecommons.org/licenses/by-nc-sa/4.0/
Many thanks to the following friends and teachers for their help in alphabetical order: Bradford Richards, Brian Peters, Christian Amonson, Eduardo Monteiro, Eric Weber, Hudson Fair, Ian Wood, Jack Reynolds, Jake Purches, Joel Rec, John Cone, Julian J. Ludwig, Lucas Guitink, Manfred Schmidt, Mario Vila, Paal Rasmussen, Paul Fee, Stefan Kießling and many others from the excellent CMLR group on Facebook, and especially to Alejandro Cabrera, Carlos Fernández, Diego Rodríguez B., Don Booth, Esteban Zabala I., Gricelda Duarte, Helmuth Reichel-Silva, Jaime Valbuena M., Joaquín Luppi, Jorge Montesi, Jorge Sacaan M., Juan P. Quezada, J. Alberto Palacios, Julio Figueroa M., Leonard Moskowitz, Mauricio Landeros, Pablo Saavedra, Ricardo Henríquez, Romualdo Castro, Sebastián Errázuriz, Xavier León and last but not least my lovely wife Patricia Reichel for putting up with me during more than a year of research, design and testing.