Triple Side Technique for Stereo Audio Recordings powered by AI for vlogs and podcasts
The sound recordings are an essential part of any modern multimedia endeavour when commencing interviews, vlogs, podcasts, conference calls captured into mono audio channel, or when performing instrumental and location stereo recordings, or when capturing or generating sound sources into complex acoustic environments using immersive mic techniques. Tri-side technique provides a flexible multichannel pattern to record a relatively wide but accurate stereo sound image with a warm sense of ambience preserving the basic mono compatibility and the ability to be automated and controlled by adaptive AI-based software.
The term stereo mic technique refers to a routine for obtaining multichannel and coherent sonic image of particular recording, using two or more microphone capsules. There are four general types of mic techniques used for stereo recording. Each technique is using particular set of microphones and provides specific sound image with advantages and disadvantages compared to one another:
- Coincident pairs – consists of two directional mics mounted so close nearly touching their diaphragms under particular angle predefined by the mic technique. These techniques have good but narrow stereo imaging and the signals are mono-compatible.
- Near-coincident pairs – consists of two directional mics set up with some distance between their diaphragms again in particular angle specific for the mic technique of this type. These techniques have accurate and wider stereo imaging with better ambience than coincident pairs mic techniques.
- Spaced pairs – consists of two matched-pair (almost identical) mics positioned at several meters apart straight ahead toward the sound scene. These techniques have diffused off-center stereo image and provide great sense of ambience; however, the stereo spread could be more or less exaggerated with possible phasing issues.
- Baffled pairs – consists of two matched-pair mics located at a fixed angle and distance from each other with a barrier between them. These techniques are designed to emulate the natural spread of the stereo signal from the sound sources to the human head and ears. The stereo imaging is not wide as spaced pairs but sharp with excellent low-frequency response. However, these techniques require special equipment, absorbent materials with peculiar shapes for the barrier, such as Jecklin Disk, spherical forms or even dummy human head.
There are several specific mic techniques and methods for each of the mentioned four general types of recordings. Detailed specification and explanation of these various mic recording techniques is out of the scope of this introduction. Instead, for the purpose of this article the following stereo recording challenges shall be considered for comparison:
- Achieving accurate stereo image;
- Recording wider stereo image for better sense of ambience;
- Preserving the mono-compatibility;
- Multi-channel encoding and post-recording decoding;
- Overall flexibility and environmental adjustability.
By definition, the distance, positioning and angle between the microphone capsules are predefined by the respective mic technique. In most cases, the configuration parameters are fixed. For example, ORTF near-coincident technique requires two cardioid microphones with capsules positioned at 17 cm in a plane at angle 110°; NOS near-coincident technique – capsules positioned at 30 cm in a plane at angle of 90°; X/Y coincident technique – capsules positioned one onto another in a plane at angle of 90°, Blumlein coincident technique – two bidirectional microphones in the same point and angled at 90° to each other, etc.
On the other hand, the spaced pairs microphones, such as A/B and Decca Tree techniques provide excellent wide stereo image with great ambience but with phase signal and cancellation issues. The Spaced Cardioids is another spaced pair which allows to be extended with third center positioned mic in order to mitigate the stereo image exaggeration. Nevertheless, all spaced pairs techniques are fixed and always positioned straight ahead to the sound scene.
Some recording techniques require additional post-recording decoding into the digital audio workstation (DAW). For example, the Mid-Side (M-S) coincident pair technique requires the side-channel to be duplicated, inverted to negative signal phase, panned and merged to the mid-channel in order the final stereo effect to be created.
Furthermore, the stereo recording techniques are designed to use set of paired microphones routed to the left and the right stereo channels. Additional flexibility to this traditional approach could be achieved by using multi-channel encoding into poly-wave files (BWF audio file format) during the recording process. Unfortunately, only few stereo recording techniques, such as Decca Tree and Spaced Pair with center mic, can take advantage of this technology.
Obviously, there is no universal technique to perform excellent stereo recordings in all possible occasions. Within this article, however, is discussed a stereo recording method that can achieve mitigation of the disadvantages and acceptable balance between the spaced and near-coincident pairs techniques, using their benefits to provide relatively wide multichannel recording with improved sense of ambience, mono-compatibility, flexible post-recording and some automation and adjustments of the recording pattern to the acoustic environment.
The triple side (Tri-Side or T/S) method for stereo recording should be considered from two perspectives – as a microphone recording technique, and as a production method for preparing the sound-stage recording, to perform actual sound recording and finally for commencing post-recording actions.
In general, the T/S technique consists of three condenser microphones with cardioid capsules symmetrically positioned in a plane. The center microphone is positioned straight ahead toward the sound scene and is used to focus and calibrate the entire recording triad of microphones.
The left and right-sided microphones are positioned at a non-fixed calculated angle defined by their cardioid polar pattern characteristics and in accordance with the physical position of the microphone array within the acoustic environment. The side microphone capsules must be a matched pair. The center microphone, however, may have different polar pattern including subcardioid, supercardioid or even hypercardioid pattern. Nevertheless, the subject of findings of this article will assume that the configuration of center microphone capsule is identical to the side mics.
The AI automation with T/S shines when real-time measurements and adaptation to the surrounding environment is required. The assumption is that this method requires the recordings to be commenced with modern multitrack audio recorder into poly-wave file and/or into multiple mono-wave files. AI continuously monitors and controls the high-pass filters (HPF) in order to avoid low-frequency rumble and when used must be equally applied to all audio tracks. Furthermore, AI controls all audio channels to be kept with no limiters and phase delays, as well as it allows the left and right channels to be linked into stereo pair dynamically and therefore any manual panning shall be avoided.
The ultimate goal of AI powered T/S recording is to improve automatically the sense of ambience in accordance with the audio surroundings and to keep the flexibility to capture the microphone array into three separated mono audio tracks – one for the center, and two for the left and right sides.
The triple side method for stereo recording provides flexible multichannel pattern to record a relatively wide but accurate stereo sound image. Three key parameters allow to control the recording coverage area and the level of ambience, preserving the basic mono compatibility during the post-production and preparation of the final audio product.
T/S method has a converged mathematical model suitable for software automation and small footprint deployments, such as IoT, embedded firmware solutions and robotics for intelligent microphone adjustments and sound-staging placement.