A revolutionary advance
in voice input less affected
by noise clipped to a mask,
mask voice clip is an input device
that isolates the wearer’s voice Even in noisy environments or where several people are speaking at once,
highly accurate real-time isolation is possible by detecting subtle vibrations on the mask.

Winner, Minister of Digital Agency Award, 2025 CEATEC Award Program

Description

Partners interested in mask voice clip applications that solve issues with voice input and communication in noisy environments
Products used

Mask voice clip, a voice input device clipped to masks
*Currently under development and not commercially available.

MATCHING
POINT

・Those considering voice input deployment where masks are worn (in medical settings, manufacturing, or construction)
・Those with ideas for a voice input device clipped to masks

Mask voice clip is a voice input device clipped to masks for highly accurate real-time voice isolation even in noisy environments.
Murata is looking for parties interested in this technology, those with ideas for potential applications, and those who might consider voice input deployment in medical settings, manufacturing, or construction. We look forward to hearing from you.

Mask voice clip, a voice input device clipped to masks: Winner, Minister of Digital Agency Award, 2025 CEATEC Award Program

Click here for Details

Contents

Background of Development
Features of mask voice clip
Usage scenarios for mask voice clip
Message from the developers

Background of Development

Clips on a mask

Voice input is becoming the most intuitive way to interact with the digital world, as generative AI use continues to surge. Yet in the physical world, adoption of voice input has not spread to medical, manufacturing, or construction sites that are often noisy and require people to wear masks.
One factor for this may be that some things present challenges for conventional voice input technology. Ambient noise or other people's voices may cause misrecognition, and speech through masks may be difficult to pick up.
Mask voice clip was developed as a solution. This voice input device clipped to masks provides highly accurate voice isolation even where there is background noise or audible disturbances.
It picks up minute vibrations on the surface of the mask worn by a speaker, enabling accurate input of their voice in real time undisturbed by ambient noise or other speakers’ voices.
This next-gen technology was recognized with a Minister of Digital Agency Award at CEATEC 2025 and has drawn attention from many fields.

Features of mask voice clip

Feature01Physically isolates and reliably picks up the speaker's voice, even in noisy environments

Mask voice clip incorporates a Murata Picoleaf^TM piezoelectric film sensor that detects minute vibrations on the mask.
Most headset microphones detect vibrations in air, which makes them susceptible to picking up ambient sounds.
In contact microphones such as throat microphones, high-frequency components tend to be attenuated as they pass through the body, which may affect identification of consonants. They may also be uncomfortable and prone to generating noise from rubbing against the neck. In contrast, mask voice clip works by measuring minute mask vibrations as the user speaks. External sound is physically shut out and does not affect the highly accurate sensing of a speaker’s voice.
This makes it a versatile, promising approach for noisy environments, places with multiple speakers, or situations requiring privacy.

Voice input mechanism

CER comparison: typical lavalier mic versus
mask voice clip¹

Feature02Convenient clip-on device that also supports mask hygiene

With its clip structure, mask voice clip is easy to attach or remove from masks. This eliminates the need for special equipment or complicated fasteners, enabling easy attachment and removal before and after work or when switching users. The device itself can be reused by people using disposable masks, supporting mask hygiene.

Feature03Low-latency, energy-efficient voice input without software processing

Mask voice clip hardware design enables low-noise voice input. With no need for post-processing such as AI noise reduction, it does not rely on a GPU or other sources of high-performance computing, which minimizes audio input latency. This makes it an ideal voice interface for manufacturing or medical settings or similar applications requiring immediate operation and reliability.

Usage scenarios for mask voice clip

Usage scenario01Giving site-specific instructions or taking voice memos

Because users can record themselves accurately without recordings being affected by ambient noise, mask voice clip enables them to instruct others or record memos without worrying about disturbances around them, especially at noisy construction sites or production lines.

Usage scenario02Hands-free communication or voice commands

Mask voice clip is convenient where users need to use both hands for other work. Hands-free operation while using factory or medical equipment is one example of this.

Inquiries

Working toward commercialization, Murata is involved in the development and testing of the mask voice clip voice input device clipped to masks.
We are looking for partners who are considering voice input deployment where masks are worn, or who have ideas for potential applications.
If you are interested, we welcome inquiries through the contact form.
*This product is under development.Therefore, the product specifications are subject to change without advance notice.

Murata Manufacturing Co., Ltd. developers

Message from the developers

Murata Manufacturing Co., Ltd.
Strategic Technology & Innovation Department
Corporate Technology & Business Development Unit

Masanori Morita Miku Nonogaki

Mask Voice Clip voice input technology was developed through joint research with Rekimoto Lab, University of Tokyo Interfaculty Initiative in Information Studies.¹
At the 2025 CEATEC awards, jury members praised this work for balancing sound source selectivity with noise immunity through the structure adopted, noting how it is a very useful next-generation voice interface for noisy environments where masks are required, such as in manufacturing or medical settings.²
We view the award as recognition of our approach. We have treated speech and voice recognition for mask wearers not as problems but as prerequisites for creating new value. We believe the technology has the potential to expand communication itself through future applications including subvocal recognition and new voice input solutions. Knowing that the technology will show its true value only when it helps people in the field as a practical solution, we are working toward commercialization and business development, envisioning widespread adoption.

We welcome inquiries by those interested in the technology. We look forward to meeting partners to join us in building the future of next-generation voice interfaces.

References

（1）Hirotaka Hiraki and Jun Rekimoto. 2025. MaskClip: Detachable Clip-on Piezoelectric Sensing of Mask Surface Vibrations for Real-time Noise-Robust Speech Input. In Proceedings of the Augmented Humans International Conference 2025（ AHs ’25 ）. Association for Computing Machinery, New York, NY, USA, 333-344.
https://doi.org/10.1145/3745900.3746088
（2）2025 CEATEC awards included the Minister for Internal Affairs and Communications Award, Minister of Economy, Trade and Industry Award, Minister of Digital Agency Award, and category awards such as the new Mobility Award.
https://www.jeita.or.jp/japanese/topics/2025/1007award.pd