In a fascinating blend of biology and technology, Benn Jordan—a musician and science hobbyist—has shown that birds can act as living data storage devices. By converting a simple PNG sketch of a bird into an audio waveform and embedding it in a starling's song memory, Jordan not only retrieved the image from the bird's vocalization but also calculated potential data transfer rates of up to 2 MB/s. The experiment, shared in a detailed YouTube video, pushes the boundaries of what's possible in bio-acoustic data handling, though it raises as many questions about practicality as it does excitement.

Article illustration 1

Why Birds? The Science of Avian Memory

Songbirds, like the starling Jordan used, possess exceptional vocal capabilities due to their unique anatomy. Unlike parrots, which are often celebrated for mimicry, songbirds have a specialized organ called the syrinx at the junction of their trachea and bronchi. This structure allows for independent control of pitch and speed through muscle groups, enabling complex vocal gymnastics—including phase and wave interference effects. Jordan chose a young starling that had been rescued after being abandoned near a train track. Raised by humans, this 'blank canvas' bird was already adept at imitating unnatural sounds, such as camera shutters and reverberated speech, making it ideal for the experiment.

Encoding and Retrieval: From PNG to Birdsong

Jordan began by creating a basic line-drawing PNG of a bird, then transformed it into an audio waveform using a spectral synthesizer. This process essentially allowed him to 'play' the image as sound. During a recording session, he repeatedly exposed the starling to this audio.

"This little bird successfully learned and emulated the sound in the exact same frequency range that he heard it, effectively transferring about 176 kilobytes of uncompressed information," Jordan explained in the video.

Initially, Jordan didn't realize the bird had replicated the sound. It was only after sifting through gigabytes of session audio that he spotted a familiar waveform—one that matched his original image spectrogram but appeared spontaneously, confirming the bird had reproduced it independently.

Article illustration 3

Data Transfer Claims and Caveats

Jordan's back-of-the-envelope math suggested that if this were a compressed audio transfer protocol, it could achieve speeds near 2 MB/s. However, he emphasized this is highly theoretical. Key limitations include:
- Lossy Reproduction: The retrieval wasn't a perfect PNG; it was a spectrogram-based approximation, losing some fidelity.
- Biological Constraints: Birds aren't programmable devices—training is time-intensive, and 'data' can't be easily overwritten.
- Scalability: Real-world applications would face challenges like environmental noise and the bird's limited memory capacity.
As one commenter on Tom's Hardware noted, "Lower bit rates are more robust," hinting that practical implementations would require error correction and compression akin to traditional systems.

Implications for Bio-Inspired Tech

While storing data in birds isn't a viable solution today, this experiment underscores broader possibilities in bio-acoustics and unconventional computing. Concepts like using natural systems for edge data storage or developing new audio-based transfer protocols could emerge from such research. Jordan's work also echoes historical methods like slow-scan TV, but with a biological twist that could inspire innovations in resilient, low-power data handling. Ultimately, it reminds us that nature often holds unexpected blueprints for technological advancement—even if the next step isn't an ostrich running Doom.

Source: Based on reporting by Tom's Hardware and Benn Jordan's original video.