Researchers at Beijing Institute of Technology have developed a microphone that does not detect vibrations in the air like traditional microphones, but instead captures light reflected from subtle surface movements. The new device, called a "visual microphone," relies on a surface's response to sound waves, using these tiny vibrations to reconstruct audible information.
This advancement opens up the possibility of listening in environments that are difficult to achieve with traditional microphones, such as communicating through glass windows or monitoring sounds in an isolated space without directly transmitting the sound.
"Our method simplifies the process of using light to capture sound, reduces costs, and can also be applied to scenarios where traditional microphones cannot function, such as talking through a glass window," explained Yao Xuri, the lead researcher on the project. "As long as light can pass through, sound transmission is not necessary."
Previous attempts to use light to capture sound have relied on complex and expensive equipment, such as lasers or high-speed cameras. The team took a different approach. Their system uses a technology called single-pixel imaging, which eliminates the need for camera sensors with millions of pixels. Instead, it utilizes structured light patterns projected by a single light detector and spatial light modulator.
Essentially, the technology works by casting manipulated light onto a target and capturing tiny changes in the reflected brightness as the target vibrates in response to nearby sounds. These tiny intensity changes are detected and converted back into sound signals through computational algorithms. This approach not only reduces cost and complexity, but also makes the technology more readily available.
Professor Yao said: "By combining single-pixel imaging with Fourier transform-based positioning methods, we are able to achieve high-quality sound detection at lower cost using simpler equipment. Our system is able to perform sound detection using everyday objects such as paper cards and leaves under natural lighting conditions, and does not require vibrating surfaces to reflect light in a specific way."
The researchers reconstructed the audio signal by imaging the vibrations of a paper card (ac). They applied signal processing filters to enhance the high frequency content (df) of the signal.
To demonstrate the system's capabilities, the researchers tested common materials such as paper cards and leaves and placed them about half a meter away from audio sources, such as speakers playing spoken digits and musical clips. The visual microphone successfully recreated clear and intelligible audio, especially when using paper cards. Low-frequency sounds were captured with great accuracy, while high-frequency sounds suffered from some distortion—a limitation the team partially alleviated using signal processing techniques.
The setup is also data efficient, producing a modest data stream of around 4 MB per second, making it suitable for long-term or continuous recording, and suitable for storage or internet transfer.
Researchers envision applications for the technology in a variety of industries, including communications through solid barriers, remote environmental monitoring, non-invasive medical observation, and advanced industrial diagnostics.
Professor Yao pointed out: "Currently, this technology is still in the laboratory stage and can be used in special scenarios where traditional microphones cannot work." The team is currently working to improve sensitivity and accuracy, and to develop a more portable version of the system and expand its detection range to achieve longer-distance sound detection.
Looking to the future, researchers see potential beyond the lab—from detecting heartbeats and pulses at a distance to aiding in search-and-rescue operations where microphones cannot be deployed directly. For now, the system represents a new way of "listening" with light, opening up the possibility of communicating and monitoring in environments beyond the reach of traditional microphones.