Note: Ce document est une exclusivité web, disponible uniquement sur ce site.
Optimization of microphone positions
These four videos illustrate the method presented in the article to optimize microphone positions. In these videos, all 37 notes of the sheng are played chromatically as long notes, with little attack and dynamics as constant as possible. The mono downmix then reveals possible balance mismatch and tone color distortion problems, which may be the result of suboptimal positioning of the single microphones, or destructive interferences when multiple microphones pick up the sound of a single pipe.
In each of these examples, the microphone system is made of:
- four supercardioid miniature microphones (DPA 4099) positioned around the instrument with little but sufficient (5—10 cm) distance.
- a fifth omnidirectional miniature microphone (AKG C417) on the top of the instrument to capture the sound emitted by the highest central pipes.
The sheng is represented in a schematic top view; the drawback of this view is that it hides the vertical positioning of each microphone, which is of major importance. However, this information can be found in the article for the second setup (videos 3 and 4). In all four examples, a red circle highlights the note being played, and the energy for each microphone is represented by green circles with varying intensity.
Video 1 — First position, mono: This example corresponds to the initial positioning of the 5 microphones, downmixed in mono. This video reveals, among other things, balance problems between the pipes: G#3, B3, A4, and C#5, among others, are too loud; several notes on the right side (C6, D#6, F#6, among others) are too soft because the sound holes are too far from the microphones. Several notes are present in two microphones (they would then be positioned in a multichannel setup as virtual stereophonic sources, see article for more details). Some (A#5 for example) mix well in mono, but for others (for instance A3, C4 and A5) the mono mix leads to tone color distortion and loss of presence.
Video 2 — First position, binaural: This is the same example mixed for headphones through binaural processing (without additional processing). The balance problems noted in the mono version are still there, but tone color distortion for the notes present on several channels is significantly reduced since a binaural presentation implies less mixing between channels than in the mono case. A pure multichannel presentation of the same material, where each microphone is assigned to a given loudspeaker, completely solves this latter issue.
Video 3 — Second position, mono: This shows the same experiment using a second positioning of the 5 microphones, downmixed in mono. The microphones are more regularly spaced around the sheng, to better cover the right-hand side, and the front-right microphone (at the bottom right in the video) is farther away from the G3 pipe. Although some balance issues remain (G3 and F4 are now too loud, for instance), the register is more consistent (though it still has problems, for instance D#4), especially for the right-hand side. This is also true for tone color, even if some notes suffer from distortion due to destructive interferences (for instance G3, F4).
Video 4 — Second position, binaural: This is the same example as the previous one, but mixed for headphones through binaural processing (without further processing). Here again, a binaural presentation solves many tone color issues. The result is not completely optimal, but is definitely heading in the right direction.
The following examples are based on a piece of music by Wu Wei for solo sheng, The Fragrance of Cherry Blossoms 3, recorded in a small studio. The microphone setup is the same as for the previous part, i.e. four supercardioid miniature microphones around the instrument and a fifth omnidirectional miniature microphone to the top.
Video 5 — Short solo improvisation: This video is mixed for headphones (binaural presentation) without any further processing; neither equalization nor reverberation. It begins with a descending chromatic scale with octave jumps, which allows the listener to better feel how the pipes are positioned in space and reveals how far from each other pipes with the same chroma but different octaves may be. The second part is an improvised transition to the beginning of Wu Wei’s piece, The Fragrance of Cherry Blossoms.
Video 6 — The Fragrance of Cherry Blossoms (excerpt in studio): Here another excerpt of Wu Wei’s piece, The Fragrance of Cherry Blossoms, mixed for headphones (binaural) with slight equalization as well as reverberation, and using the same setup as in videos 3 and 4. The four side microphones are assigned to four virtual sources around the listener, and the fifth (above) to a front source. The purpose of the reverberation here is to put the sound sources farther away and start to shape a (very rough) space, which in this case sounds bigger than it looks on the image (contradiction between audition and vision). This example reveals the inherent polyphony of the piece: voices are not only distinguished by rhythm and register, but also in space.
More elaborated musical examples
Those sound excerpts are taken from real musical situations (one during a first residency in GMEM/Marseille, the other one during a concert in Amsterdam). In both situations, the Sheng is not in solo, but in trio (first example) or duo (second example) with electronic music performers.
The microphone setup is now more elaborated, using overall 7 to 9 close microphones:
• Four supercardioid miniature microphones (DPA 4099) positioned at little distance (ca. 5 cm) around the instrument
• One omnidirectional miniature microphone (DPA 4060) on the top of the instrument to capture the sound emitted by the highest central pipes.
• Two miniature contact microphones (AKG c411, only one used in Amsterdam — example 8) attached to the front left and front right of the instrument, easily accessible to the left and right index, which intentions are:
1. To reinforce (when desired) the natural percussive sounds of the leverage keys of the instrument, and
2. To provide extra percussive possibilities to Wu Wei by directly hitting them.
• One omnidirectional microphone (DPA 4061) inside the wind chest to capture the air flow at the source, which contains a significant amount of low and low-mid frequencies (<250 Hz) that is missing at the outputs of the pipes.
• One omnidirectional microphone (AKG 417, not used in Amsterdam, audio 8) attached to the mouthpiece (outside) close to the mouth, to capture the voice and air sounds as a more broadband alternative to the microphone inside the windchest.
Audio 7 — Trio improvising in residency — Excerpt of a free improvisation between Sheng (Wu Wei) and two electronic music performers (André Serre-Milan and Alexis Baskind) during a residency in GMEM/Marseille in May 2021. The sheng and the electronics by André Serre-Milan provide most of the sound material, which is spatialized in a performative way by Alexis Baskind. The sound projection has been done there on 13 channels consisting of: an outer ring of 8 loudspeakers regularly spaced projecting inwards (octophony), a top lautspeaker (“voice of god”), and an inner ring of 4 speakers on the floor projecting outwards.
This complex setup has been down-mixed to two-channels for headphone presentation (binaural processing) in post-production.
[See the audio “Trio improvising in residency (2021) ” at the top of the page]
Audio 8 — Excerpt of a duo concert — This is an excerpt of the very first public concert for the project “a forest of pipes”, given at Eye Museum in Amsterdam in May 2021 during the Holland Festival. The music presented here is a mixture of traditional music rearranged by Wu Wei, original music written by Wu Wei, and improvisation. The sheng is in duo with electronic music (performed by Alexis Baskind), which here consists of improvisations based on pre-recorded material projected on a multichannel setup.
The concert venue was a movie theatre with a 7.1 sound installation. This setup was used for the live sound projection, and again, down-mixed in post-production for this two-channel headphone presentation.
[See the audio “Excerpt of a duo concert (2021) ” at the top of the page]