Oct 9, 2026, 6:08:52 AM • 8 min

Synthetic Voice in Award-Winning Films: What 5 Oscar Productions Teach Us

•••

At first glance, The Brutalist, Emilia Pérez, Better Man, and Bohemian Rhapsody have almost nothing in common. But behind their Oscar runs lies the exact same post-production secret: synthetic voice technology used to fix a performance on screen.

AI voice in award-winning films didn’t begin with the current AI boom. Back in 2018, Bohemian Rhapsody was already piecing performances together voice by voice. By the time The Brutalist arrived, the work could come down to individual sounds.

These 5 cases show how synthetic voice in Oscar films got from one to the other, and what production teams can learn along the way.

Key Takeaways

  • Solve the voice problem you actually have. Bohemian Rhapsody shows how multiple vocal sources can be combined when no single recording works for the scene.
  • Don't assume a lost voice means replacing the actor. Top Gun: Maverick shows how voice recreation can be explored when illness changes what an actor can physically record, while the actor remains central to the character.
  • Go as small as the problem allows. The Brutalist shows that the intervention can come down to individual vowels and consonants rather than rebuilding an entire performance.
  • Extend the performance when the scene demands more. Emilia Pérez shows how voice conversion can help bridge the gap between an actor's dramatic performance and the vocal demands of a musical number.
  • Plan for voice continuity across the whole production. Better Man shows the challenge of keeping one recognizable voice consistent across archival recordings, new material, speech, and singing.

Case 1: Bohemian Rhapsody (2018) – The Pre-AI Benchmark

“Open heart surgery-style editing” – that's how sound editor Nina Hartstone described some of the vocal work behind Bohemian Rhapsody. And that was a literal description of what they had to do.

The team had to match Rami Malek's lip-sync to Freddie Mercury, except key Mercury tracks were missing or unusable. The solution was to build performances from three vocal sources:

  1. Mercury's original master tapes
  2. Marc Martel's recordings
  3. Malek's own vocals

For scenes such as “Happy Birthday” and “Love of My Life,” all three voices had to become one convincing performance. The sound team took home the 2019 Oscars for both Editing and Mixing for pulling off the vocal trick.

The production lesson: Synthetic voice in Oscar films existed before today's AI systems. It was simply slower, more manual, and far more dependent on usable archival material. Sometimes it came down to a single consonant and whether it worked with the close-up.

Case 2: Top Gun: Maverick (2022) – Rebuilding a Voice After Illness

Throat cancer, diagnosed in 2014, and the tracheostomy that followed took most of Val Kilmer's voice. When he partnered with Sonantic in 2021, they turned to recordings of his earlier voice as their source material. The team reportedly generated more than 40 voice models before finding one that captured the qualities Kilmer wanted.

Kilmer's work with Sonantic is confirmed. Whether the resulting voice appears in Maverick is not: Paramount disputed reports that it was used. That same film took home Best Sound at the 2023 Academy Awards.

The production lesson: AI voice in award-winning films isn’t always there to make a voice younger, cleaner, or suspiciously well-rested. Sometimes the actor is still the only person who should play the part, but illness has changed what their voice can do. 

Case 3: The Brutalist (2024) – Down to a Single Sound

Adrien Brody and Felicity Jones spent months working on their Hungarian. They got close, but Dávid Jancsó, the film's editor and a native Hungarian speaker, wanted closer.

By the time we entered the process, this wasn't about entire lines or performances. It was about individual sounds. We've written about Respeecher Oscar films before, but some myths about voice cloning film controversy have excellent staying power. So, one more time:

  • no, AI didn't replace Brody or Jones.
  • no, it didn't rewrite their English dialogue.
  • yes, their original delivery and emotion remained.
  • what changed were specific Hungarian vowels and consonants, based on native pronunciation.

A native Hungarian speaker provided the pronunciation reference, and Respeecher worked on the sounds that needed it. On Oscar night, The Brutalist had three wins to its name: Best Actor, Cinematography, and Original Score.

The production lesson: If you're looking at AI voice post-production Oscar workflows, The Brutalist shows just how small the intervention can be. Sometimes you don't need another take, just one vowel change. Speech-to-speech lets the team work on that sound without rebuilding everything around it.

Case 4: Emilia Pérez (2024) – When the Score Demands More

Hollywood loves a good musical. But asking dramatic actors to hit operatic high notes is a quick way to stall a production. On Jacques Audiard’s Emilia Pérez, lead actor Karla Sofía Gascón delivered a standout performance, but key musical passages stretched beyond her vocal limits.

Instead of going full old-school Hollywood and replacing the actor's vocals with a ghost singer, re-recording mixer Cyril Holtz used Respeecher and opted for a smart AI singing voice conversion workflow:

  • keep Gascón's performance
  • bring in a professional singer, where the range calls for it
  • work on the notes and sounds that need help
  • keep the result sounding like one character

And what started as a supporting role for the technology got bigger. Speaking at Cannes, Holtz said the AI approach ended up handling more of the vocal work than the team had originally planned. Come Oscar season, Emilia Pérez had 13 nominations and two wins: Supporting Actress and Original Song.

The production lesson: The score doesn't care about the limits of human vocal cords. Synthetic voice actor performance can give filmmakers more room to follow the creative idea when the voice itself can't quite follow.

Case 5: Better Man (2024) – Making a CGI Ape Sound Like a Pop Icon

Putting a CGI chimp on screen and telling the audience "that's Robbie Williams" only works if the voice is 100% indisputable.

On Better Man, the audio team had decades of Williams' vocal history to work with. The problem was getting it all to live in the same film: older compressed recordings alongside pristine new studio takes, across both speech and singing. Every laugh, sob, scream and signature rasp still had to sound like the same person.

Respeecher spent more than two years refining the voice with the production team, with Williams recording additional material when needed. Even his characteristic vocal fry became part of the job.

Respeecher powered voice work on The Brutalist and Emilia Pérez, two Oscar-winning productions from the 2024–2025 season. Explore Respeecher Oscar films and voice solutions for complex post-production challenges.


What 5 Oscar Productions Reveal About Synthetic Voice

Film

Oscar recognition

Voice approach

Production lesson

Bohemian Rhapsody

Won Sound Editing & Sound Mixing

Composite voice, manual editing

Synthetic/composite performance predates modern AI

Top Gun: Maverick

Won Best Sound

Sonantic voice recreation project; final-film use disputed

A lost speaking voice can be recreated with the performer's participation

The Brutalist

Won Best Actor, Cinematography & Original Score

Respeecher pronunciation refinement

Specific phonemes can be corrected without replacing the performance

Emilia Pérez

Won Supporting Actress & Original Song

Respeecher singing refinement

Voice technology can extend a vocal performance

Better Man

Nominated for Visual Effects

Respeecher voice conversion

Voice continuity can span eras, recordings and performance conditions

Line them up and the same shape appears every time:

  1. The actor gives the performance.
  2. The technology handles a specific technical limit, a lost voice, a stubborn vowel, a register out of reach, a career that spans decades.
  3. None of these are cases of a machine inventing a performance from nothing.

It's also why consent and provenance sit at the center of this work rather than off to the side. Every one of these projects rested on documented permission and a real human performance underneath.

When a vendor can't explain how they handle that, they're not ready for a cinema or broadcast credit.

 

The Controversy: Where Does Performance End and AI Begin?

If individual sounds in an actor's dialogue are adjusted, is the finished performance still entirely theirs? What about a singing voice blended with another vocalist? Editors have reached for pitch correction and CGI for years, and the Bohemian Rhapsody edit won an Oscar doing it by hand. Is AI doing something new here, or just doing it faster?

No one agrees yet.

Dialect coach and NYU Tisch associate arts professor Jane Guyer Fujita told NPR after Adrien Brody's Oscar win that she saw the Bohemian Rhapsody comparison as relevant and didn't consider the use in The Brutalist disqualifying. Her distinction was centered on whether technology displaced human work and whether the actor's contribution remained fundamental.

The Academy's position

The Academy has since given the industry a similarly non-binary answer. Under rules introduced for the 98th Oscars, generative AI and other digital tools "neither help nor harm" a film's chances of nomination. Instead, Academy branches are instructed to consider the extent to which human creative authorship sits at the center of the achievement.

That makes transparency more useful than drawing a simple AI/no-AI boundary.

Better to be specific. A production that explains what was changed, who approved it, what source material was used, and what remained the actor's own gives audiences and collaborators something concrete to assess. “We used AI,” on its own, tells them very little.

Final thoughts

Six years separate Bohemian Rhapsody’s hand-mixed vocals from the vowel-by-vowel work on The Brutalist. The process changed, some of the manual work disappeared, but the reason for doing any of it never did. Both teams wanted the same thing: a technical fix that didn’t cost them the performance.

And the performance still starts with a human. Synthetic voice can extend or refine it, but productions also need to be clear about where that technology enters the process.

Studios choose Respeecher for Emmy-winning technology, a public ethics policy their union and legal teams can review, and a track record of credited work on Oscar-nominated films.

Learn more →


image1

FAQ

Examples of AI voice in Oscar-winning productions include The Brutalist, Emilia Pérez and Better Man, all of which used Respeecher in their voice workflows.

Bohemian Rhapsody is an important earlier comparison because it created a composite vocal performance without modern AI.

In Top Gun: Maverick, Val Kilmer worked with Sonantic on an AI recreation of his voice, but Paramount disputed reports that the technology was used in the finished film.




The sound team combined original Freddie Mercury recordings with performances by Marc Martel and Rami Malek.

Editors then painstakingly blended the material so the finished vocals appeared to come naturally from Malek's on-screen performance.


 

The public record is conflicting. Kilmer confirmed working with Sonantic to create an AI-powered version of his speaking voice, but Paramount disputed reports that AI generated his voice in Top Gun: Maverick. It's more accurate to acknowledge both accounts than present the film use as settled fact.

 

In The Brutalist, Respeecher worked down to individual Hungarian vowels and consonants in Adrien Brody and Felicity Jones's performances.

In Emilia Pérez, the technology was used to fine-tune specific elements of Karla Sofía Gascón's singing.



 

The Academy does not automatically view the use of AI as either a benefit or a disadvantage. Its rules state that generative AI and other digital tools do not affect a film’s chances of being nominated, with the focus instead placed on the extent of human creative authorship.

A production can work at the level of specific sounds, pronunciation, pitch or selected lines while retaining the actor's timing, emotion and broader delivery. The Brutalist is a particularly clear example: the intervention targeted Hungarian pronunciation rather than rebuilding the performance.

 

A composite voice can be assembled manually from several existing vocal recordings or performers, as happened in Bohemian Rhapsody. Voice cloning in Hollywood productions builds a model of a particular approved voice from recordings so new or converted speech can retain that voice's characteristics.

Glossary

Synthetic Voice

Speech created or modified with digital voice technology, either by generating new speech or transforming an existing performance.

Composite Voice

A single finished voice assembled from recordings or performances by multiple vocal sources.

Voice De-Aging

A technique that makes a performer's voice sound younger while preserving the underlying performance.

Speech-to-Speech (STS)

A process that converts a recorded performance into a target voice while preserving elements of the original delivery, such as timing, phrasing and emotion.

Re-Recording Mixer

The person who combines dialogue, music and sound effects into the film's final soundtrack.

AI Disclosure

A clear explanation of where and how AI was used, including what was changed, whose material was involved and how its use was authorized.

Voice Model

A model built from recordings of a specific voice that can be used, with appropriate permission, for voice conversion or speech generation.
Previous Article
A Producer’s Checklist for Recording Clean AI Voice Training Data
Clients:
Lucasfilm
Blumhouse productions
AloeBlacc
Calm
Deezer
Sony Interactive Entertainment
Edward Jones
Ylen
Iliad
Warner music France
Religion of sports
Digital domain
CMG Worldwide
Doyle Dane Bernbach
droga5
Sim Graphics
Veritone

Recommended Articles