A longitudinal corpus of sustained phonation and voluntary cough acoustics from three ISS crew members

The human voice is more than just a tool for communication; it is a sensitive barometer of the body's internal state. When we speak, air from the lungs vibrates the vocal cords, a process that depends on precise fluid balance, muscle control, and a stable environment. In space, the rules of this biological machinery change. As soon as astronauts leave Earth's gravity, fluids shift toward the head, the air inside the spacecraft is drier and contains more carbon dioxide than on the ground, and the sleep cycle is disrupted by the rapid orbit of the station. These changes can subtly alter how the voice sounds, but until now, scientists had very little data on exactly how spaceflight affects the human voice because they lacked a way to measure it systematically.

To understand these shifts, researchers needed a controlled experiment that could capture the voice before, during, and after a mission using the exact same methods. This is difficult in space, where crew time is precious and equipment is limited. The challenge was to create a record of the voice that was detailed enough to reveal physiological changes but simple enough to be done repeatedly without burdening the astronauts. The goal was not to diagnose illness, but to establish a clear baseline of how the voice behaves in orbit, creating a reference point for future studies and potential health monitoring tools.

A team of researchers recently addressed this gap by collecting a unique library of voice recordings from three crew members aboard the International Space Station. The study focused on two specific types of sounds: a voluntary cough and a sustained vowel sound, where the person holds a single note like "ah," "ee," or "oo" for a few seconds. The participants, identified only as Crew M1, M2, and M3 to protect their privacy, performed these tasks twice a day during a pre-flight quarantine period and again while in orbit, spanning from mission day four through day fourteen. In total, the team gathered 254 verified recordings, creating a longitudinal record that tracks the same individuals over time under different conditions.

The researchers designed the experiment with extreme precision to ensure the data was comparable. Every session followed the exact same sequence: three separate coughs, followed by holding the three vowel sounds. The duration of each recording was fixed by the software at approximately three seconds, removing the variable of how long a person chose to speak. This consistency allowed the team to compare the acoustic properties of the voice directly, day by day, without the results being skewed by differences in recording length. The recordings were captured using a mobile application already aboard the station and then transferred to Earth, where they were processed to extract specific measurements of pitch, stability, and clarity.

One of the most significant aspects of this work is the release of the data in two parallel forms. For every recording, the team retained both the original acquisition and a version that had been cleaned up using a deep noise suppression algorithm to remove background static. While the raw audio files themselves are not distributed to the public, the study provides derived acoustic measurements for both versions. This dual release allows other scientists to see exactly how much the cleaning process changes the measurements. The study found that while the cleaning process significantly improved the clarity of the voice by reducing background noise, it left the fundamental pitch essentially unchanged, with only a minor numerical shift observed. However, it did change specific measures of vocal clarity, meaning that anyone using this data must be careful to note which version of the recording they are analyzing.

When the researchers compared the space recordings to large collections of voice data from people on Earth, a clear pattern emerged. The absolute values of the voice measurements, such as the clarity of the sound, shifted depending on where the recording was made. The space recordings and those made on consumer devices at home showed similar patterns, which were distinct from recordings made in professional, sound-treated rooms. This suggests that the environment and the equipment used to record the voice have a larger effect on the numbers than the specific health of the speaker. Consequently, the study concludes that medical thresholds for voice health derived from quiet, professional studios cannot be directly applied to recordings made in the field or in space without adjustment.

The team also looked at the timing of the coughs, breaking them down into their individual bursts of sound. They found that when the coughs were analyzed in this detailed way, the timing matched closely with data from a massive study of over 2,600 people on Earth. This agreement suggests that the fundamental mechanics of a cough remain consistent even in space, provided the analysis method is the same. However, the study also highlighted that the number of cough bursts detected depends entirely on the recording protocol. Because the astronauts were asked to cough once every three seconds, the number of bursts was lower than in studies where people were asked to cough repeatedly in a longer recording. This finding serves as a reminder that how a task is asked for can change the results as much as the environment itself.

The researchers were careful to state that this collection of data is not intended for diagnosing individual health issues or identifying specific diseases. The three participants were healthy crew members, and the study was designed to understand the effects of the space environment on voice production, not to screen for illness. The value of this work lies in its methodological rigor and the transparency of its data. By providing a verified set of measurements and the code used to generate them, the team has created a reference point for future missions. They demonstrated that it is possible to gather high-quality, repeated voice data in space, opening the door for more sophisticated monitoring of astronaut health in the future. The study confirms that while the voice changes in orbit, these changes follow predictable patterns that can be measured and understood, provided the right tools and comparisons are used.