Dummy-Head Recording¶

Binaural sound capture with microphones at an artificial head's ears, preserving spatial cues shaped by its head and pinnae.

Version: v1 · 2026-09-28 · History
Domain-specific #: 9088
Domain group: Applied Sciences & Engineering
Origin domain: Engineering & Design (beyond software)
Subdomains: Binaural Recording, Audio Engineering, Acoustics → Engineering & Design (beyond software)
Aliases: Artificial-head binaural recording, Kunstkopf recording

Core Idea¶

Dummy-head recording is a specific way to make a binaural two-channel record. Microphones at an artificial head's ears capture sound after it interacts with the head and outer-ear geometry, retaining interaural timing, level, and spectral differences. The target is not merely a left-right stereo split; the artificial head shapes cues that a human listener uses to locate sound.

Dummy-head capture is one technique within binaural recording. MIT's KEMAR measurements document artificial-head paired-ear responses to sound at many positions. Human in-ear recording and software HRTF synthesis can also be binaural but lack this capture apparatus. Reproduction is usually headphone-dependent, and an average dummy head cannot guarantee individualized spatial perception.

Structural Signature¶

Sig role-phrases:

artificial head geometry — Provides head and pinna filtering rather than mere spacing between microphones. It is constitutive. Counterfactual: Two bare microphones with no head simulator can be binaural but not dummy-head capture.
left and right ear microphones — Capture separate signals at the artificial ear positions. It is constitutive. Counterfactual: A mono microphone in front of a mannequin lacks the paired-ear cue relation.
sound scene or stimulus — Provides acoustic sources whose path to each ear differs. It is constitutive. Counterfactual: An inert mannequin with no acoustic input produces no recorded binaural scene.
two-channel cue record — Retains interaural timing, level, and spectral differences. It is constitutive. Counterfactual: Combining the two channels into one erases the key comparison.
playback and fit boundary — States headphone or measurement use and limits of one head model for individual listeners. It is boundary. Counterfactual: A dummy head does not guarantee speaker playback or perfect localization for everyone.

What It Is Not¶

All binaural recording. Human in-ear and synthetic HRTF methods need not use an artificial head.
Ordinary stereo spacing. Bare paired microphones omit head/pinna filtering.
A mono mannequin microphone. Both ear channels matter for binaural comparison.
Universal spatial realism. Listener anatomy and playback conditions alter localization.
Closest near-miss. A pair of microphones worn by a real person may yield binaural recordings, but the absence of an artificial head keeps it outside this specific method.

Scope of Application¶

Binaural production. Identify artificial-head capture as one specific route to ear-cue recordings.
Acoustic research. Interpret KEMAR impulse-response measurements as reusable artificial-head evidence.
Playback analysis. Explain why separate-ear headphone presentation differs from speaker mixing.
Method comparison. Separate dummy-head capture from human-worn recording and synthesized panning.

Clarity¶

Ask where the microphones were placed and whether an artificial head and pinnae shaped the arriving sound. Two channels alone do not establish dummy-head recording. MIT's KEMAR measurements are an example of capturing ear-specific impulse responses, while later individualized playback remains a separate question.

Manages Complexity¶

The head model folds many acoustic reflections and shadow effects into a paired signal. That makes cue capture practical to describe as two channels, but the simplification hides head-model geometry, headphone response, and listener mismatch. Those omissions matter when moving from recorded cues to perceived location.

Abstract Reasoning¶

Identify the artificial head and its ear geometry.
Locate the two microphones at or near the simulated ears.
Trace the sound scene into distinct left/right cue signals.
State whether the use is headphone reproduction or acoustic measurement.
Bound localization claims by model and listener differences.

Knowledge Transfer¶

The artificial-head/paired-ear/cue relation transfers from performance capture to HRTF measurement when both use sound passing through a head simulator. One KEMAR geometry, a speaker position, or a listener's perceived externalization does not transfer unchanged to every recorder or listener.

Examples¶

Canonical¶

An artificial head with one microphone in each ear captures a room performance. The head changes each sound path before the left/right signals are stored, and headphones can present those channels separately; the construction is the method, not a promise of identical perception for every listener.

Mapped back: artificial head geometry → mannequin head and ears; left and right ear microphones → paired ear-position sensors; sound scene or stimulus → performance in a room; two-channel cue record → distinct ear signals; playback and fit boundary → headphone use and listener mismatch caveat.

Applied / In Practice¶

MIT Media Lab's 1994 KEMAR study recorded left and right impulse responses to a loudspeaker at 710 positions and published the HRTF data for research. This is a real artificial-head acoustic capture application; measured cue transfer functions are not a claim that every later listener has KEMAR's exact ears.

Mapped back: artificial head geometry → KEMAR head/ear simulator; left and right ear microphones → recorded ear response channels; sound scene or stimulus → loudspeaker at documented positions; two-channel cue record → paired impulse-response/HRTF data; playback and fit boundary → research use, not universal individual fit.

Structural Tensions¶

T1 — Standardized Capture versus Individual Hearing. A reusable head model supports comparable recordings while each listener's anatomy may localize differently.

Diagnostic: Whose head-related cues are being preserved?

T2 — Spatial Impression versus Playback Dependence. Ear-channel cues can be compelling under headphones but conventional speakers mix channels acoustically.

Diagnostic: Is the intended reproduction or analysis context stated?

Structural–Framed Character¶

The approved DAG parent is Binaural Recording: two ear-related channels retain spatial localization cues. Dummy-head recording adds an artificial human head with microphones near its ears; perceived externalization varies by listener.

Evaluative weight: Spatial fidelity is listener- and setup-dependent. Human-practice-bound: Moderate, because head geometry and microphone placement are designed while acoustics governs filtering. Institutional origin: Audio research and production use different head simulators, not one universal shape. Vocabulary travels: Performance capture and HRTF measurement may fit; one simulator's response does not transfer unchanged. Import versus recognize: Recognize the child by artificial-head capture; bare stereo spacing imports no head-shaped cues.

Its character: An acoustic recording subtype with portable paired-ear capture and an artificial-head carrier.

Structural Core vs. Domain Accent¶

Skeletal core. Two ear-related signals preserve spatial acoustic differences.

Domain-bound accent. An artificial head shapes the sound field before microphones capture left and right channels.

Why not prime. Binaural recording is broader; human-worn or wholly synthetic routes lack this dummy-head mechanism.

This entry is a kind of Binaural recording.

Strict parent — binaural recording. Ear-position channels and spatial cues satisfy the broader binaural record; artificial-head capture makes this a narrower method.
Related — representation. The recorded channels represent selected acoustic spatial relations, but the nearest parent is the existing binaural-recording method.

Relationships to Other Abstractions¶

Current abstraction Dummy-Head Recording Domain-specific

Parents (1) — more general patterns this builds on

Dummy-Head Recording is a kind of Binaural recording Domain-specific

Dummy-head recording is binaural capture narrowed to an artificial head with microphones at its ears and preserved left/right spatial cues.

Hierarchy path (1) — routes to 1 parentless root

Dummy-Head Recording → Binaural recording → Encoding And Decoding → Transformation → Function (Mapping)

Neighborhood in Abstraction Space¶

Dummy-Head Recording sits in a moderately populated region (58^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Musical & Poetic Form (11 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With¶

Human in-ear recording. Tell: Is the head a physical person or an artificial simulator?
Conventional stereo. Tell: Did head and pinnae shape the paired signals?
Binaural synthesis. Tell: Were ear cues recorded through a dummy head or generated in software?
Headphone playback. Tell: Is this a use condition rather than the capture method itself?

References¶

MIT Media Lab, KEMAR dummy-head HRTF measurements: https://sound.media.mit.edu/resources/KEMAR.html
Frozen Wikipedia discovery revision: https://en.wikipedia.org/wiki/Binaural_recording (revision 1367201988).
Preserved source candidate: https://www.researchgate.net/publication/233582452
Preserved source candidate: https://web.archive.org/web/20190129001035/https://www.researchgate.net/publication/233582452_Binaural_Recording_Technology_A_Historical_Review_and_Possible_Future_Developments
Preserved source candidate: https://www.sciencedirect.com/science/article/abs/pii/S0003682X20306782#:~:text=Intelligibility%20scores%20from%20trained%20subjects,-2,%20Reverb-2
Preserved source candidate: https://hearingreview.com/hearing-loss/patient-care/hearing-fittings/a-brief-history-of-binaural-hearing-aid-fittings
Preserved source candidate: https://books.google.com/books?id=piQEAAAAMBAJ&q=street+hassle+binaural&pg=PP1
Preserved source candidate: https://www.soundonsound.com/techniques/binaural-panning-logic-pro
Preserved source candidate: http://ent.163.com/15/1125/07/B98I0FGK00031H0O.html
Preserved source candidate: https://www.tonyawards.com/news/gareth-fry-and-pete-malkin-to-be-honored-with-2017-special-tony-awards/

The frozen Wikipedia revision is discovery provenance. The retained source set was reviewed for identity, formal or operational relation, and scope. The encyclopedia's structural synthesis is bounded to those claims; a thin authority surface is recorded as a nonblocking source-strengthening repair rather than concealed.