Five Positions, Five Different Voices: Where to Place the SEM-02 clip-on
Updated: Sep 18
Before you can call a microphone position wrong, you have to say what right sounds like. For the voice, the accepted reference is a person heard from about one meter away, on-axis, in front of them. That is the timbre a listener in the room experiences, and it is the timbre an audience unconsciously expects when the same voice comes back out of a loudspeaker.
Every body-worn position is a departure from that reference. Not a small one.
The measurements
Eddy Bøgh Brixen, AES Preprint 4284: miniature omnidirectional microphones at five points on the head and body, each compared against the same talker at one meter on-axis, averaged across ten people.
Five positions, five deviation curves.
The work is theirs. What we can add is what those curves mean when the microphone is a SEM-02.
Why 2–4 kHz decides everything
Consonants make speech intelligible. They sit above 500 Hz, concentrated between 2 and 4 kHz. Vowels carry most of the energy and little of the meaning.
High-pass a voice at 500 Hz and you strip most of its energy while losing around five percent of intelligibility. Low-pass it at 1 kHz — barely touching the energy — and intelligibility falls below forty percent.
So a placement that attenuates 2–4 kHz is not making the voice duller. It is deleting the part that carries the words.
SEM-02 for voice: what the microphone brings
Omnidirectional, 20 Hz to 20 kHz. A pressure transducer with a nominally flat response — the same class used in the source measurements, which is why the curves transfer directly. It also means no built-in presence boost. Some lavaliers bake a fixed high-frequency lift into the capsule to pre-compensate for chest placement. The SEM-02 does not, so whatever a position demands, you apply.
That cuts both ways. In a neutral position it stays out of the way. In a lossy one it faithfully reproduces the loss.
Maximum SPL of 115 dB. Loud singing at the lips reaches around 130 dB RMS, with peaks well above. For spoken word the headroom is ample; for a belting singer at cheek position it is not.
A 5.8 mm capsule, 10.4 grams. Weighed on a pocket scale: capsule, full 2.5 meter cable and Microdot termination, without the XLR dongle.
Weight decides whether the most neutral position is available at all. A microphone heavy enough to drag on fabric will not stay in a hairline and will telegraph every head turn down the cable. Ten grams stays where you put it.
The five positions clip-on, ranked
Best first, by how little each alters the spectrum. The third column is the practical part: what the SEM-02 needs from you once it is there.
Two things repeat at every position. A rise near 800 Hz from body reflections, easily cut. And a deficit above it, in the consonant band — the one that costs you, and the one most often left uncorrected.
Order of operations: cut the low-mid build-up, restore the consonant band, then judge against the one-meter reference rather than against the raw track.
A correction to common advice
Thick wardrobe is usually blamed for dull chest audio, and it does make things worse. But the deficit is there under a thin cotton shirt too, because it comes from the position rather than the fabric. Treat the mid-high lift as standard for chest placement, not as a rescue for heavy clothing.
The physical side — collars, ties, strain relief, keeping fabric rustle out of the capsule — is in our SCO01 concealer guide. Two positions from it are worth noting: the collar sits closer to the mouth and improves signal-to-noise but collapses when the talent turns away; the tie points the capsule downward, adding its own colouration.
The video below runs through concealed chest placement and ends with a sound sample recorded that way.
The ALC01 alligator clip is the visible alternative: capsule outside the collar or lapel, nothing to hide, fabric noise easier to control. Same acoustic neighbourhood, same correction. There is a separate walkthrough and sound sample for it.
The same physics, on an instrument
None of this is specific to the human voice. Directivity rises with frequency, so moving off-axis costs you the top first — true of a talker, an acoustic guitar and a saxophone bell alike. Bodies reflect. And proximity changes the spectrum, not just the level, even with an omni capsule that has no proximity effect in the cardioid sense.
Which is why a SEM-02 parked over a guitar soundhole makes the same category of error as a lavalier buried under a lapel: sitting inside a resonance, off the radiation axis of the frequencies that carry articulation. Move the microphone before you reach for the EQ.
Four rules
Define your reference before evaluating a position. One meter, on-axis.
Choose the position first, then confirm the microphone suits it.
A flat microphone gives an honest capture, so the compensation is your job. Decide where in the chain before you record.
Convenient, invisible and neutral are three different criteria. No position wins all three.
Common questions
Where should you place a microphone on the body for the most natural voice?
The hairline. Of the five positions clip-on measured — hairline, cheek, over the ear, chest and neck — the hairline alters the speech spectrum least when compared against a talker heard from one meter away, on-axis. It is also the hardest to hide on camera, which is why chest placement remains the broadcast default despite sounding worse.
Why does a chest-mounted lavalier microphone sound muffled?
Because chest placement loses a substantial amount of energy between 2 and 4 kHz, where the consonants that carry intelligibility live, and adds a rise near 800 Hz from chest reflections. The result is a voice that is boxy underneath and short of articulation on top. Thick clothing makes it worse but does not cause it — the deficit is there under a thin shirt too.
How much EQ does a chest-placed microphone need?
Roughly 5 to 10 dB of boost around 3 to 4 kHz to return a flat-response microphone to neutral, plus a cut to the build-up near 800 Hz. Microphones sold specifically as broadcast lavaliers often build that lift into the capsule. The SEM-02 does not, so the correction is yours to apply.
Can the Seruniaudio SEM-02 be used for voice?
Yes. It is an omnidirectional clip-on microphone with a 5.8 mm capsule weighing 10.4 grams, small enough to conceal in a hairline or a costume seam, and it terminates in Microdot so it adapts to both phantom-powered inputs and wireless bodypacks. Its response is flat, with no built-in presence lift, so any position-dependent correction has to be applied by the user.
What is the maximum SPL of the SEM-02, and how does it affect placement?
115 dB. Loud singing measured at the lips reaches around 130 dB RMS with peaks well above that, so the SEM-02 should not sit at the corner of a loud vocalist's mouth. For spoken word at conversational level, the headroom is ample at any of the five positions.
How much does the SEM-02 weigh?
10.4 grams for the microphone, cable and Microdot connector, measured on a 0.01 gram pocket scale without the XLR dongle. The 300 gram figure on the specification sheet is the boxed shipping weight.
Is the neck a good position for a lavalier microphone?
No. Of the five measured positions it shows the heaviest loss of consonant frequencies, and equalisation cannot fully recover it. It persists in broadcast because it is often the only place a microphone can go on a T-shirt or under a coat, which makes it a wardrobe decision rather than an acoustic one.
Related reading
What is a lavalier microphone? — start here if the terminology is new.
Concealing the SEM-02 with the SCO01 — this article is the why; that one is the how.
Omni or cardioid? — everything above assumes an omnidirectional capsule.
Watch
How to mic a voice with the SEM-02 — concealer technique, with sound sample.
How to mic a voice with the SEM-02 — alligator clip technique, with sound sample.
Sources
The placement curves, the one-meter reference condition, the ten-talker averaging, the rise near 800 Hz, the 5 to 10 dB chest compensation figure and the sound pressure levels of loud singing at the lips all come from Brixen's AES work, listed first below.
Brixen, Eddy B. — Spectral degradation of speech captured by miniature microphones mounted on persons' heads and chests. AES Convention 100, Copenhagen. Preprint 4284.
Brixen, Eddy B. — Near field registration of the human voice: Spectral changes due to positions. AES Convention 104, Amsterdam. Preprint 4728.
French, N. R. & Steinberg, J. C. — Factors governing the intelligibility of speech sounds. Journal of the Acoustical Society of America, Vol. 19, No. 1, 1947.
Chu, W. T. & Warnock, A. A. C. — Detailed Directivity of Sound Fields Around Human Talkers.
Preprints 4284 and 4728 are available through the AES E-Library. A summary of the same material, written by Brixen, is published as Facts about Speech Intelligibility in DPA Microphones' Mic University.
SEM-02 figures come from the SEM-02 product page, except the weight, measured in-house on a 0.01 g pocket scale — capsule, 2.5 meter cable and Microdot connector, XLR dongle excluded.
Comments