Yes; it is conventionally described as a voiced bilabial nasal.

Are /f/ and /v/ bilabial?

No; their usual classification is labiodental.

Does bilabial determine voicing?

No; place and voicing are separate descriptive dimensions.

AI Glossary / Natural language and speech / Advanced

Bilabial

Definition

Bilabial describes a speech sound whose primary constriction is made using both lips. It is a place-of-articulation category: it identifies where the relevant restriction occurs. Familiar English examples are /p/, /b/ and /m/. These sounds share a location of articulation while differing in other properties, so bilabial does not identify a single sound or specify its complete phonetic description.

What it means

The term distinguishes two-lip articulation from other ways of shaping the vocal tract. For example, /f/ and /v/ involve the lower lip and upper teeth and are labiodental. They involve a lip, but they are not bilabial in this classification.
Place, manner and voicing answer different questions. Place concerns where constriction occurs. Manner concerns how airflow is restricted or directed. Voicing concerns vocal-fold activity. A useful sound description combines these dimensions rather than assuming that one label supplies all of them. In speech technology, these distinctions help organize sound inventories and interpret recognition errors.

How it works

For the familiar bilabial plosives /p/ and /b/, the lips form an oral closure. Their conventional broad descriptions distinguish a voiceless plosive from a voiced plosive. Actual timing and realization depend on language and context, so the labels should not be mistaken for an identical acoustic pattern in every recording.
The bilabial nasal /m/ also uses a lip closure, but air can pass through the nasal route. Sharing the lip position therefore does not make /m/ the same kind of sound as /p/ or /b/.
Bilabial is broader than these three English examples. The International Phonetic Alphabet chart also distinguishes bilabial sounds with other manners of articulation. Not every bilabial sound requires the same complete closure as a plosive. The defining feature is the involvement of both lips in the primary constriction, not membership in a three-letter spelling list.

Three practical examples

A pronunciation lesson contrasts the initial sounds of “pat,” “bat” and “mat.” All three begin with bilabial consonants, while the contrasts illustrate why place alone is insufficient to distinguish them.
A speech-recognition evaluation groups substitutions by articulatory features. Confusing /p/ with /b/ preserves the bilabial place category, but still produces an incorrect sound label.
A speech-dataset team documents the phonetic inventory used in its annotations. Separating bilabial and labiodental categories prevents all sounds involving a lip from being merged into one misleading category.

Think of it this way

Think of describing an event by its location and its type. Two events can happen at the same location while being very different events. Bilabial supplies the location component of a sound description. Additional information is needed to specify what the speech organs and airflow are doing there.

Worked example

Consider six invented reference sound labels: /p/, /b/, /m/, /f/, /v/ and /t/. The corresponding predictions are /b/, /p/, /m/, /v/, /f/ and /t/. Treat these as broad sound categories for a simple evaluation exercise, not measurements from an actual recognizer.
Only /m/ and /t/ are predicted exactly. Exact-label accuracy is therefore 2/6, or approximately 33.3%. The substitutions /p/ to /b/ and /b/ to /p/ are errors even though both members of each pair are bilabial.
Now map every reference and prediction to a place category. The first three reference labels are bilabial, and their predictions are also bilabial. The next two are labiodental, as are their predictions. The final /t/ is alveolar in the broad classification used here, and its prediction matches. Place-category accuracy is consequently 6/6, or 100%.
The calculations answer different questions. The system gets the articulation location right for every item in this tiny example while getting only two exact sound labels right. Reporting the place score as overall recognition accuracy would substantially misrepresent the result.
The example also shows why bilabial should not be treated as a spelling feature. The labels are phonetic categories. A written letter can represent different sounds depending on the language and word, and some written letters are not pronounced. Sound classification should follow the transcription convention and actual speech being described.

Illustrative case study

In a fictional review of a speech model, a dashboard displays “100% recognition” for the worked predictions. A reviewer checks the labels and finds four substitutions. The calculation behind the dashboard had first converted each sound into a broad place category, then compared those categories.
The team corrects the display to distinguish exact sound-label accuracy from place-category accuracy. The latter remains useful: it suggests that the model's errors in this example occur within the broad place groups. It does not establish the acoustic cause of those errors or prove that the model has learned human-like articulation.
The team retains both views and expands its evaluation beyond the six invented items. It documents the transcription scheme and examines errors across contexts. A feature-level summary becomes a diagnostic aid rather than a replacement for the task's actual recognition measure.

Common misconceptions and mistakes

Bilabial does not mean every sound produced with rounded lips. Lip rounding can accompany sounds whose primary articulation is elsewhere. Similarly, labiodental and bilabial are different categories despite both involving the lips.
Do not infer the exact sound from a visible lip closure alone. /p/, /b/ and /m/ share important aspects of their lip configuration, while differing in properties that the closure itself does not fully specify. Also, the three common English examples do not exhaust the bilabial possibilities described in the IPA.

Business use

Phonetic categories help teams document speech data, inspect errors and communicate about pronunciation features. Use them at the level appropriate to the task. A diagnostic improvement in bilabial classification is useful evidence about that category, but it should not be presented as an equivalent improvement in word recognition, transcription accuracy or overall communication quality.

Check your understanding

Answer: Which three reference sounds in the example are bilabial? /p/, /b/ and /m/.

Why is /p/ predicted as /b/ still an error?

Answer: The exact sound category changed.

What does the 100% score measure?

Answer: Agreement of broad place categories, not exact sound recognition.

Related terms

Place of articulation
Manner of articulation
Phoneme
Phone
voicing
labiodental
plosive
nasal consonant
International Phonetic Alphabet

Learn next

Compare /p/, /b/ and /m/ across place, manner and voicing, then inspect how transcription conventions represent phonetic detail.

Sources

Jurafsky and Martin, Speech and Language Processing, chapter 15, section 15.2: articulatory phonetics and bilabial versus labiodental sounds. https://web.stanford.edu/~jurafsky/slp3/15.pdf International Phonetic Association, IPA chart, revised 2020: consonant places and manners. https://www.internationalphoneticassociation.org/IPAcharts/common_files/pdfs/pdfs_IPA_charts_archive/IPA_unitipa_2020.pdf The six-label evaluation and fictional dashboard case are original teaching illustrations.

Previous: Bigram Next: Bilateral filtering