canal.cl / grabación: The Omni Variable Jecklin Microphone Array

Please be aware there is a newer version of this post is available: The OVJ array v2.

The OVJ microphone array is loosely based on several proven techniques for recording acoustic music and ambient sound, including:

The horizontally spaced, vertically coincident 4-microphone phased array known as OCCO, popularized by Robert Sandy “Boojum” Noyes, Jim Norman and Tony Faulkner.

The binaural-like OSS (Optimal Stereo Signal) “Jecklin Disk”, by Jürg Jecklin*.

Bilateral Ambisonics***, by Zamir Ben-Hur, David Alon, Or Berebi, Ravish Mehra, and Boaz Rafaely.

OVJ array viewed from the back

Two tetrahedral FOA (First Order Ambisonics) subarrays are centered over each side of a baffle, and two matched stereo sets of SDCs are placed closer in and further out.

How each pair works:

The innermost pair (1, 2) with an inter-microphone distance of 30cm or less between provides intense spatial cues based mostly on direction-dependant intensity and frequency response caused by the baffle, as well as some time precedence. This is known as the Jecklin effect.

We have empirically determined an SRA (stereo recording angle) of aproximately ∽81° for the Jecklin effect, therefore that is the target SRA suggested as a starting point for all inter-microphone distances and angles.

Omni capsules are preffered. Cardioids can also be used, becoming “baffled NOS” with similar SRA consisting of a stronger intensity component than either standard NOS or Jecklin OSS.

In either case, panning and mixing with outer pairs can smooth the effect if too strong, and automatic equalization such as Acon Digital DeFilter can compensate for comb filtering, should reflections from the baffle be too noticable.

The outermost pair (3, 4) can be spaced accordingly at 79cm for a similar time-based SRA.

Again, omni capsules are preffered. Cardioids can be used if acoustic conditions are sub-optimal.

This pair adds position cues beyond the limits of OSS and some front-back differentiation when using APEs or cardioids.

Pulling this pair in a little closer can deliver a Tony Faulkner-styled sound, where the outer pair has a wider SRA than the inner pair.

The FOAs deliver immersive Jecklin-affected signals:

We position the FOA arrays so the left-pointing capsule on the left (5) and the right-pointing capsule on the right (12) are 24 to 36cm apart.

Ambisonics dominance processing and post-transcode panning can control left-right separation and increase front-back discrimination if needed.

♥ Better together:

All pairs can be combined within (and without) immersive workflows when phase-coherent transcoding and rotation are used, as provided by Audio Brewers plugins.

Some additional forward gain can be obtained with each added pair of capsules, especially for LF.

As usual with multiple microphone pairs, an important caveat: SRA doesn't tell the whole story. On-location monitoring and adjustment of distances and angles might be warranted.

Nevertheless, values for coherent SRA seem to work very well, therefore the default configuration of the array can be useful when time or space for optimized positioning are unavailable.

Realistic immersive results are possible with modest processing. Suggested for Dolby Atmos:

inner pair to L and R,
outer pair -12dB to L, R, Lm and Rm ,
transcoded signals from the FOAs to all corresponding channels, except -12dB or more attenuation on L, C, R, Lm and Rm.

Dolby folddown to stereo works very well, realism is surprisingly preserved despite the limitations of a sigle pair loudspeakers. The binaural-like cues from the Jecklin effect also provide an impressive experience on headpphones. Personalized binaural processing becomes optional (but is of course needed for head-tracking).

Furthermore, thanks to a significant portion of the signal originatating from Ambisonics, the result is very mono-compatible, unlike traditional binaural, beacuse lateral components are cancelled out.

Alternatively, Jecklin-affected stereo is available by using capsules 5 and 12 and adding the low-passed sum of the other capsules from each FOA for left and right respectively. This avoids aliasing from the inter-capsule distance and can be useful when low latency, reduced visibility or increased portability are needed, and multichannel immersive is not.

Notes:

We initially miscalculated some angles and the FOAs were running at a different default SRA compared to the other pairs (66° versus 81°). The OVJ v2 corrects this by swapping inner pair and FOAs positions.

* The “Jecklin effect” is a binaural-like acoustic process accomplished with a baffle between (usualy omnidirectional) matched capsules. The standard diameter for the baflle is 35cm. Unlike dummy-head or current software-based binaural, this process invented by Jürg Jecklin is perfectly compatible with stereo loudspeaker playback.

** MFS stands for “modified for symmetry”. FOA microphones and the resulting B-format signals are rotated, resulting in a symmetrical physical layout in the left-right (front) axis instead of the conventional skewed layout… probably a good thing since human ears are usually at the same height. What were Gerzon and Craven thinking?

*** We do not claim to be using by-the-book Bilateral Ambisonics. Ours is a simplified process with FOAs panned to either side of the “bed” in an Atmos workflow, so spatial cues are presented to the listener mostly seperately for each ear, thus enhancing realism in combination with the Jecklin effect.

The OVJ name is inspired by Christian Amonson and his outside-in nomenclature. The V stands for “variable”, both to portray variability through Ambisonics and because the Jecklin effect can be heightened or softened using more of the inner or outer pairs respectively.

This content by @ignace / Ignacio Rodríguez de Rementería is open source licensed under CC BY-NC 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, Julian J. Ludwig, Lucas Guitink, Manfred Schmidt, Mario Vila, Paal Rasmussen, Paul Fee, Stefan Kießling and many others from the CMLR group on Facebook, and especially to Alejandro Cabrera, Carlos Fernández, Diego Rodríguez B., Don Booth, Esteban Zabala I., Gricelda Duarte, 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, 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.

Also available in Spanish.