Four Voices We Helped Bring to Life in 2025
In 2025, AudioShake unlocked four voices that changed what's possible with sound—from opera legends performing beyond their lifetimes to a grandfather gaining the ability to speak to his grandchildren again. These aren't just technical achievements. They're human stories about connection, creativity, and the power of making sound as flexible as any other medium.
1. Maria Callas: Isolating an Opera Legend from 1960s Orchestra Recordings
Pablo Larraín's Oscar-nominated film Maria, starring Angelina Jolie as Maria Callas, needed to feature Callas's voice. But her iconic 1960s and 70s performances were mixed with orchestras in single tracks.
AudioShake isolated Callas's vocals from recordings like "Vissi D'Arte," giving the team clean stems to layer her voice over Jolie's performance. "Obviously, this is a movie about Maria Callas, so you want to have Maria's voice," Director Pablo Larraín told IndieWire.
2. Mike G.: Restoring a Voice Stolen by ALS
Mike G.—retired police officer, criminal justice professor, youth sports coach—spent his life communicating. Then ALS took his voice away.
Armed with only old family videos, Bridging Voice, a non-profit helping build communication tools for ALS patients, turned to AudioShake. With our multi-speaker separation technology, we could isolate Mike’s voice from background noise and voice, which gave our partner Eleven Labs clean samples to build a voice clone.
"AudioShake enabled me to once again speak to my granddaughters, and anyone else, as I once did, in my own voice," Mike shared. "It's been liberating."
3. Charlie Patton: Cleaning a 1920s Blues Recording for Modern Cinema
Indie filmmaker Alex Park wanted a 1920s Charlie Patton song for his film HORSEFLY. The wax cylinder recording was filled with hiss and distortion.
AudioShake separated vocals from guitar and removed the noise. "What we get now is a clear, unencumbered version," Park explained. The result: 1920s blues grit meets modern cinema quality.
4. Luciano Pavarotti: Reprising a Duet with Bocelli
In 1995, Andrea Bocelli and Luciano Pavarotti performed "Notte 'e Piscatore" together. Decades later, Bocelli wanted to include the duet on his album Duets—but the 1995 recording was a single mixed track, making an immersive Dolby Atmos version impossible.
AudioShake separated Bocelli's voice from Pavarotti's, even where they overlapped. This enabled Decca Records to create a spatial audio version allowing Bocelli to perform live at Teatro del Silenzio alongside the late Pavarotti's isolated vocal track for Andrea Bocelli 30: The Celebration.
AudioShake separated the mixed audio into dialogue, music, and effects—preserving the non-verbal sounds like laughter and exclamations that make the videos engaging. CrunchLabs could now localize for new languages, replace expired music licenses, and create region-specific versions. Content that was locked in English-only formats became accessible to global audiences.
Four breakthroughs that represent hundreds more happening across music, film, accessibility, and education. Four projects that show what becomes possible when you can separate mixed audio into its component parts. A film director could use archival recordings from the 1960s. A nonprofit could help ALS patients create voice clones from noisy home videos. An indie filmmaker could clean century-old recordings. A music label could create spatial audio from single-track performances. A creator could localize viral content for global audiences.
You can try AudioShake’s vocal and speech isolation technologies on our on-demand platform and APIs.