Close-Microphone Techniques for Capturing Individual Instruments in Ensemble Performances: an Acoustic and Perceptual Perspective
Close-Microphone Techniques for Capturing Individual Instruments in Ensemble Performances: an Acoustic and Perceptual Perspective
Project Report
Published: April 14, 2025
Authors
Ying-Ying Zhang (McGill University) and Jithin Thilakan (Hochschüle für Musik Detmold)
Introduction
Accurately capturing individual sound sources when recording a multi-musician performance requires a crucial balance between obtaining “acoustically clean” signals from each instrument and preserving musicians' visual and auditory feedback from co-performers and the room acoustic environment. This balance is crucial for research on orchestral and ensemble sound, particularly in studies on musical blending between instruments in ecological conditions, as well as applications in virtual reality orchestra simulations and music recording techniques.
When it comes to individual instruments, the direct approach to capture acoustically clean signals is to record them in an anechoic environment, by electronically providing room acoustic feedback to performers if needed. On the other hand, capturing ecologically valid and acoustically clean signals of multiple instruments of an ensemble remains challenging due to the multitude of factors involved. These challenges can be addressed through three methods. (1) recording each instrument of the ensemble separately in an anechoic chamber while providing external audio and visual cues [1, 2], (2) recording each instrument individually during a joint musical performance by situating the entire ensemble or orchestra in an anechoic environment [3, 4], or (3) employing close-microphone techniques to capture individual instruments during a performance in a traditional performance setting [5].
Previous research highlights that a musician's individual performance is significantly influenced by factors such as visual and audio feedback from the co-performers [6, 7], joint performance strategies [8, 9], and the room’s acoustic feedback [10, 11]. Therefore, recording the sound sources individually or as an ensemble in a reflection-free environment, even with artificial visual and auditory feedback, may not necessarily preserve the intrinsic and natural representation of the orchestral/ensemble sound. Conversely, recording individual sound sources using close-miking techniques offers a viable approach that inherits the qualities of joint performances by incorporating collaborative performance strategies, audio-visual feedback between musicians, and room acoustic feedback into the recordings. This makes it the optimal choice for musical performance research, particularly for investigating the source-level blending between instruments in realistic joint performances under in-situ acoustic conditions. However, the presence of cross-talk between sound sources due to spatial proximity, noise from the musician and instrument, and room acoustic feedback in reverberant environments are the main challenges of this method, potentially degrading the quality of individual instrument recordings. Therefore, both subjective and objective evaluation of the quality requirements of the close-microphone recordings of instruments in a joint performance is necessary to validate their applicability in musical performance research and music recording techniques. This module presents pilot experiments investigating the acoustic and perceptual quality aspects of a specific close-microphone recording technique through the recording of singers in a choir performance.
Pilot Study on the Objective and Subjective Quality of Close-mic Recordings
A preliminary investigation on the applicability of a potential close-microphone recording technique in choral singing was conducted at Hochschule für Musik Detmold. The solo and joint performances of two singers (one male and one female) from a choir were recorded at Detmold Concert House under varying acoustic conditions, using two different microphones: a studio-quality Neumann TLM 127 cardioid microphone and a Sennheiser KE4 capsule measurement microphone. While the TLM 127 microphones were positioned in front of the singers in a conventional setup, a novel method was tested with the KE4 capsule microphone, where the SPL calibrated KE4 microphone was attached above the nose tip and directed toward the mouth to explore its applicability for isolated source recording. The study explored the variations in the close-mic recording outputs with the changes in the acoustic environment surrounding the singers by adjusting the spacing of singers, incorporating acoustic panels, and modifying room acoustics using the concert house’s artificial reverberation system. A total of twelve conditions were examined by systematically altering three key factors: (1) the singer-to-singer distance (1 m and 2 m), (2) the placement of absorbers (no absorbers, individual enclosures around the singers using absorbers, and absorbers placed on the floor), and (3) the reverberation time (natural RT60 of 1.6 s and an artificially extended RT60 of 3.2 s). For each acoustic condition, sound samples were recorded with the singers performing both individually and together using these two different sets of microphones, and used for the analysis.
Figure 1: Close microphone quality recording experiment at Detmold University, showing a close spacing with no barrier (left) acoustic baffling (center) and a close-up of the nose microphone (right).
Sound samples were prepared from the recordings of each acoustic condition, with one singer performing (either male or female) while the other remained silent. These samples were then used to objectively analyze the mic bleed caused by the singer at the silent co-performer’s microphone for both microphone techniques. The results suggest that the KE4 capsule microphone consistently exhibits lower mic bleed between the two singers compared to the Neumann TLM 127 microphones, demonstrating this reduction across varying room acoustic conditions and singer separations with at least 10 dB lower mic bleed between the KE4 capsule microphones. Analyzing the SPL calibrated KE4 capsule microphones shows that, while certain acoustic conditions shows a reduction of almost 40 dB from the singer's mic to the silent co-performer’s mic, even under highly reverberant conditions with 1-meter separation distance (3.2 s RT without absorbers), they maintain almost close to 30 dB difference in sound pressure level between these capsule microphones (the audio samples for this condition are provided below). This significant signal-to-noise ratio supports the applicability of these close-mic recordings.
A perceptual evaluation of the KE4 capsule microphone recordings was carried out among expert listeners with musical and technical ear training, mainly Tonmeister students at the Hochschule für Musik Detmold. They reported a high signal-to-noise ratio and satisfactory quality, along with significantly reduced cross-talk from neighboring sources. Additionally, they confirmed the recordings' quality and suitability for further mixing processes.
Experiment with the Shaar Hashomayim Synagogue Choir
As part of a series of experiments done in the Multimedia Room (MMR) at McGill University’s Schulich School of Music, we applied this microphone technique to a small vocal ensemble in order to examine its efficacy in an acoustic environment with a setting closer to real performance.
The Shaar Hashomayim Synagogue Choir, a religious men’s choir of twelve voices plus one soloist, participated in the experiment. Unlike the singers in the Detmold experiment, choir members would be actively listening to each other in order to blend, and we did not use any physical baffling between performers. While baffling created more isolation between singers during initial testing, this change was made to bring the experimental protocol into a more natural performance environment. The Cantor (soloist) was physically separated from the choir by 2 meters, in order to mimic their spacing were they singing in synagogue, but other choir members were seated relatively close to each other in a semi-circle opposite the conductor.
The choir was recorded using Scheops MK4s in a spaced AB configuration, a Neumann KU100 in the center of the musicians, and the channel-based and Ambisonics setup described here.
Figure 2: Close microphone on the Cantor of the Shaar Hashomayim Synagogue Choir.
Figure 3: Shaar Hashomayim Choir in rehearsal.
The choir arrived to conduct their typical rehearsal, during which we experimented with different acoustic environments via the Meyer Constellation system. In the end, they chose an acoustic setting that they felt best suited their performance, and we recorded two performances of liturgical work, an example of which is provided below.
As you can hear from this excerpt from Shema Yisrael, some microphone signals are very isolated, while others. While some microphone signals have a completely anechoic characteristic, others also captured a neighboring voice. In the future, experiments with placement and spacing could help eliminate these issues. Despite this, there is no prevalent room characteristic or reflection in any of the signals individually. However when all anechoic sources are mixed together, a perception of space can be heard in the slight timing delays between performances.
Works Cited
[1] Pätynen, J., Pulkki, V., & Lokki, T. (2008). Anechoic recording system for symphony orchestra. Acta Acustica united with Acustica, 94(6), 856-865.
[2] Gomes, O. C., Lachenmayr, W., Thilakan, J., & Kob, M. (2021, September). Anechoic multi-channel recordings of individual string quartet musicians. In 2021 Immersive and 3D Audio: from Architecture to Automotive (I3DA) (pp. 1-7). IEEE.
[3] Witew, I. B., Paprotny, J., & Behler, G. K. (2006). Auralization of orchestras in concert halls using numerous uncorrelated sources. In Proc. of the Institute of Acoustics, 28(2), 293-296.
[4] Böhm, C., Ackermann, D., & Weinzierl, S. (2021). A multi-channel anechoic orchestra recording of Beethoven’s Symphony no. 8 op. 93. Journal of the Audio Engineering Society, 68(12), 977-984.
[5] Postma, B. N., Poirier-Quinot, D., Meyer, J., & Katz, B. F. (2016, June). Virtual reality performance auralization in a calibrated model of Notre-Dame Cathedral. In Proc. of Euroregio, Portugal.
[6] Daugherty, J. F. (2003). Choir spacing and formation: Choral sound preferences in random, synergistic, and gender-specific chamber choir placements. International Journal of Research in Choral Singing, 1(1), 48-59.
[7] Ternström, S. (1999). Preferred self-to-other ratios in choir singing. The Journal of the Acoustical Society of America, 105(6), 3563-3574.
[8] Lembke, S. A., Levine, S., & McAdams, S. (2017). Blending between bassoon and horn players: an analysis of timbral adjustments during musical performance. Music Perception: An Interdisciplinary Journal, 35(2), 144-164.
[9] Goebl, W., & Palmer, C. (2009). Synchronization of timing and motion among performing musicians. Music Perception, 26(5), 427-438.
[10] Bolzinger, S., Warusfel, O., & Kahle, E. (1994). A study of the influence of room acoustics on piano performance. Le Journal de Physique IV, 4(C5), C5-617.
[11]Gari, S. A., Kob, M., & Lokki, T. (2019, September). Analysis of trumpet performance adjustments due to room acoustics. In International Symposium on Room Acoustics (pp. 65-73).