AI CASE STUDIES · VOICE & AUDIO

Altered.ai Case Studies: Real-World Applications of AI Voice Technology

How game developers and media production teams use Altered Studio to build and refine compelling voice performances.

These three documented examples take us from a cyberpunk action RPG to a game-concept trailer and a television post-production workflow. We examine the problem behind each production, the voice technology involved, what the original announcements actually report, and where further interpretation is warranted. The aim is not to promise automatic savings or replace performers, but to understand where AI-assisted speech can contribute to practical creative work.

Game dialogueCreative prototypingFilm & TV
Altered AI supplied red A logo
EDITORIAL SERIES · 01 / 03
The shape of a voice
Original human performanceINPUT
AI-assisted voice interpretationOUTPUT
Speech-to-speechHuman performanceCreative control
Original editorial waveform illustration; not Altered Studio interface footage or measured project audio.
Context

Understanding Altered.ai and Its Voice Technology

Altered.ai develops tools for creating and modifying spoken audio. Its media-production product, Altered Studio, brings multiple voice workflows into one environment, including speech-to-speech transformation, voice cloning, text-to-speech generation, voice cleaning, and audio editing. A separate offering, Real-Time Pro, focuses on live voice and video calls; the production examples examined here concern voice content creation rather than live calls.[4]

Speech-to-speech voice morphing begins with a human performance. Instead of replacing the words and delivery from scratch, the system processes a recorded voice to produce another vocal identity or style. That makes the technology particularly relevant when timing, energy, and intention matter: a performer can supply the rhythm of a sentence while the production team explores how the same performance might sound as another character. Altered describes models that can change timbre, accent, or speaking style, depending on the chosen workflow.[5]

Voice cloning is related but different. With appropriately authorized source recordings, a voice model can be built to resemble a particular speaker; the production can then use that model in supported generation or transformation workflows. Text-to-speech begins with written words and creates spoken audio without requiring the same recorded line. Cleaning and editing tools help prepare recordings for use or further processing. Capabilities and technical requirements vary by product and model, and output still benefits from critical listening.[6]

The three cases in this article illustrate distinct uses: broadening non-player character dialogue in The Ascent, improving the voice presentation of a prototype concept in Reclaimed, and supporting temporary dialogue in a child-actor television workflow. Each depends on creative direction and appropriate voice rights rather than AI operating as an independent author of the finished experience.

Editorial methodology: the case facts below come primarily from Altered's own dated publications. These are company-reported accounts, not independently audited impact studies. We distinguish direct reporting from educational workflow examples and our own analysis. There are no invented ROI or time-savings percentages.
CASE STUDY 01 / GAME DEVELOPMENT

How Altered Studio Helped Expand Voice Dialogue in The Ascent

In a published December 2021 announcement, Neon Giant, Sweet Justice Sound, and Altered described using a human actor's performances as the foundation for additional game dialogue.[1]

Project: The AscentDeveloper: Neon GiantSound design: Sweet Justice SoundPublished: 16 Dec 2021
800Dialogue lines produced using Altered Studio, as reported in the announcement
40Non-player characters receiving the additional dialogue performances
Human-voice inputRecordings performed by actor Sam Hughes and processed for different voices

Project background

The Ascent is Neon Giant's action-shooter role-playing game set in a cyberpunk world. Speech helps the game world feel inhabited: even a relatively brief exchange with a side character can add context and personality. Before this update, side-mission interactions included text-only dialogue. The December announcement described an effort to bring spoken performance to these exchanges, extending the voice presence beyond the core cast.

The production challenge

Voice production at character scale requires time for performance capture, casting and audio work. The published story specifically connects this need to the realities of a small development team and notes that some NPCs might otherwise remain unvoiced because of budget, time, or staffing constraints. That is a meaningful production problem, but the source does not provide a full cost ledger or a before-and-after productivity benchmark. It would therefore be misleading to convert the example into an exact savings claim.

The AI-assisted solution

Sweet Justice Sound used actor Sam Hughes’s original recordings and processed them through Altered Studio. Its speech-to-speech technology was used to produce diverse character voices from that human performance. The production did not simply type all dialogue into a generic text reader: the source explicitly credits a real actor for the recorded foundation and a sound-design team for the voice work. {ref_badge(1)}

What the implementation tells us

The verified sequence is straightforward: Sam Hughes recorded the material; Sweet Justice Sound used Altered Studio to augment his voice; the resulting dialogue was associated with supporting game characters and released as part of an update. A broader professional workflow would typically include directing performances, selecting transformations, reviewing intelligibility, and integrating audio assets. Those additional individual production steps are educational inference, not a complete documented shot-by-shot breakdown of this project.

Documented results

Altered's announcement identifies 800 lines for 40 NPCs and says previously subtitled side-mission dialogue would receive spoken performances. The story also includes positive comments from studio and audio collaborators about expanding production scope. These statements support a conclusion about increased dialogue coverage, but they do not establish an externally verified improvement in user satisfaction, sales, retention, or profit.

Key lesson for game developers

Speech-to-speech can extend the reach of a controlled human performance when a project needs varied supporting voices. Human direction, actor rights, and sound-design review remain central to the process.

Read original case study

CASE STUDY 02 / CREATIVE PROTOTYPING

Bossa Games Uses Altered Studio for the Reclaimed Concept Trailer

Early-stage creative work still needs persuasive audio. A short Bossa Games account explains how Altered Studio helped move provisional voice tracks toward a more complete presentation.[2]

Project: ReclaimedBossa PresentsConcept-game trailerPublished: 21 Dec 2021
RECLAIMED · CONCEPT STAGE

From scratch voiceover to trailer-ready performance

Exploratory voiceover
AI-assisted narrated performance
Concept narrativeCharacter dialogue
Original conceptual illustration; no Bossa-owned trailer imagery has been reproduced.

What the source identifies

  • Bossa Presents was used to share ideas from internal game jams with players.
  • The example concerned the Reclaimed concept trailer.
  • Scratch narration and character voiceover were transformed into more professional-sounding performances.

Project background

Bossa Games described Bossa Presents as a way to put ideas created during internal game jams in front of players and help decide which concepts should advance. Reclaimed was presented as a survival-builder concept with a post-apocalyptic, solar-punk setting. The trailer had to communicate both the concept and its mood within a short piece of media.

The creative challenge

Exploratory game pitches often begin with scratch voiceover: practical first-pass audio that explains a concept without the polish expected from a finished production. According to Bossa lead game designer Mike James, communicating the idea within a short timeframe was difficult. The published account focuses on the quality and rapid iteration of narration and character dialogue, not on a financial or audience-conversion result.

How Altered Studio was used

Bossa reported turning informative scratch voiceover into professional-sounding performances for both trailer narration and character dialogue. The tool therefore served as a bridge between an initial creative pitch and the level of vocal presentation the team wanted viewers to experience. The source does not disclose exact voice presets, a full technical chain, an asset count, or production hours; those details should not be presented as documented.

Outcome and lesson

Bossa's published testimonial describes Altered as an additional tool for rapid, high-quality ideation. That is a qualitative, attributed report. A reasonable lesson is that voice transformation can make temporary creative material easier to assess and present; the case does not show measured schedule savings or prove that AI narration replaced any complete final casting process.

Why this matters

Prototyping is a legitimate use case in its own right. A concept pitch can benefit from clearer, more expressive temporary sound without making unsupported claims about finished production outcomes.

Read original Bossa Games article

CASE STUDY 03 / FILM & TELEVISION

Supporting Film and Television Production with AI Voice Technology

An Altered case study from April 2023 describes temporary voiceover for an unnamed television series featuring a child actor. The project identity and numerical savings were not disclosed.[3]

Unnamed television seriesCustom Voice ServicesTemporary voiceover & post-productionPublished: 13 Apr 2023

Background and operational constraints

Film and television production involving children has particular scheduling and safeguarding responsibilities. Actors may have limited working hours and school commitments, while dialogue sometimes needs to be revised after filming. Another challenge is continuity: a young performer's voice can change over time, even while a project or series is still being edited. The Altered article frames custom voices as a support tool for these conditions, not as an unrestricted substitute for a child actor.

The approach described by Altered

According to the published case study, the studio used existing voice recordings after receiving consent from the child actor's parents and legal representatives. The recordings were sent to Altered and used to build a custom voice. The team refined the sound for the character, then used an adult actor to perform placeholder temporary dialogue that could be transformed with the child-matching voice model and incorporated into editing and post-production. The article says the child returned for final edits.{ref_badge(3)}

Implementation process reported in the article

  1. Consent and sample preparation: existing recordings gathered with parent and legal-representative authorization.
  2. Custom voice creation: the service used the authorized recordings to develop a matching voice model.
  3. Creative refinement: the studio and provider adjusted the custom voice for the required character performance.
  4. Temporary performance: an adult performer supplied placeholder lines for the child's character.
  5. Post-production use: the transformed lines supported editing, with the child actor brought back for final adjustments.

What was reported — and what was not

The article says the studio reduced the child's workload, supported continuity, minimized some scheduling conflicts, and produced a voice that suited the character. These are provider-reported qualitative outcomes. The article does not identify the series, release independently checked production records, specify hours saved, or publish a standardized listener-quality evaluation. That limitation matters when considering how far the example can be generalized.

Responsible voice use

The stated parent and legal-representative consent is a defining feature of this example. Productions should also document the scope and duration of authorization, determine how likeness rights are handled, safeguard source recordings, and maintain human review of both creative quality and welfare. The legal requirements will depend on jurisdiction, contracts, and the actor's circumstances.

Key lesson for production teams

A consent-based custom voice can support temporary post-production audio while retaining the real performer's central role. Child welfare, rights management, and editorial review are essential parts of the workflow.

Read original film and television case study

Side-by-side view

Comparing Altered.ai’s Real-World Applications

The three examples share voice technology but involve different production goals and levels of public reporting.

FieldThe AscentReclaimedFilm & TV
IndustryVideo game productionEarly game concept productionFilm and television
ObjectiveExpand voiced NPC dialogueImprove narration and character performance in a concept trailerSupport temporary child-character voiceover
Core methodSpeech-to-speech transformation of actor recordingsAI-assisted voice transformation of scratch VOConsent-based custom voice used with adult-driven temporary performances
Documented result800 lines for 40 NPCs; previously subtitled dialogue voicedBossa described more professional-sounding performancesProvider reported continuity, scheduling and welfare benefits
Evidence limitsNo audited cost or revenue figuresNo quantified production or engagement metricsUnnamed series; no independently audited savings figures

Comparison synthesized from Altered's published examples. Descriptions in the objective and method rows are editorial summaries, not verbatim product specifications.[1][2][3]

Technology explained

How Altered.ai’s Voice Technology Works

Different tools answer different creative questions. The choice depends on whether you start with a performance, a particular voice identity, or a written script.

Speech-to-speech morphing

A performance is recorded by a speaker and transformed into another voice while retaining aspects of the actor’s phrasing and delivery.

Seen in: The Ascent

Authorized voice cloning

A model is created from voice material with the necessary rights and consent, then used in a supported generation workflow.

Relevant to: the film/TV example

Text-to-speech

Written scripts become synthetic speech. Useful for testing narration and building drafts; not every project in this article used TTS.

Tool capability, not a claim about all cases

Audio editing & cleaning

Editing and noise-reduction tools help prepare source audio and refine generated dialogue in a production pipeline.

General product capability

Product capabilities described by Altered in its voice changer and voice cloning documentation. Available models and licensing terms should be checked before production.

Production process

From Voice Recording to AI-Assisted Production

A generalized educational workflow showing where human creative and rights checks belong.

01

Define the need

Identify the character, narration, continuity, or prototyping objective.

02

Secure source rights

Record or collect permitted voice material with appropriate consent and licenses.

03

Select a workflow

Choose transformation, a custom voice, or text-to-speech as appropriate.

04

Create the output

Run the voice process and generate candidate performances.

05

Listen and refine

Review pronunciation, emotion, likeness, clarity, and context.

06

Integrate and document

Export approved audio, preserve provenance, and document rights and approvals.

Scope of this diagramThis is a recommended conceptual production sequence, not a claim that Neon Giant, Bossa, or the unnamed television studio followed identical procedures or used exactly these six steps.
Creative and business lessons

What These AI Case Studies Teach Us

The strongest lessons concern how teams structure work, rather than promises of identical financial returns.

Creative flexibility

The same recorded delivery can be interpreted through alternative voices, giving sound teams more character directions to explore.

Dialogue at scale

The Ascent illustrates how additional speech for supporting characters can be produced from an actor-driven workflow.

Rapid prototyping

The Reclaimed example shows the value of better provisional narration when teams communicate a developing idea.

Human–AI collaboration

Actors, sound designers, editors, and creative directors continue to supply judgement and direction.

Responsible deployment

Consent, likeness rights, safeguarding, and an accountable approval process are prerequisites for credible use.

Industry context

Where AI Voice Technology Can Be Applied

Only gaming and television are customer examples detailed on this page. The other entries are illustrative uses of the underlying tool capabilities, not verified client success stories.

Video game production

Supporting character dialogue, prototypes, and narrative variations.

Documented case

Film and television

Temporary dialogue and voice continuity under appropriate authorization.

Documented case

Animation

Exploring character voices before final casting and production.

Potential application

Podcast production

Testing segment narration and selected character voices.

Potential application

Digital advertising

Producing narration variants with approved voices and claims.

Potential application

Audiobooks

Prototyping readings or supporting authorized voice production.

Potential application

eLearning

Drafting instructional narration and localized script variants.

Potential application

Content creation

Experimenting with vocal identities for original media projects.

Potential application
Synthesis

Key Findings from Altered.ai’s Case Studies

The clearest measurable claim is dialogue output

The Ascent announcement identifies 800 lines and 40 NPCs. Those are production-output figures, not a measured return on investment.

A recorded human performance can remain central

The Ascent account explicitly credits Sam Hughes; the film/TV example describes an adult performer providing placeholder lines.

Prototyping benefits can be qualitative

Bossa described stronger narration and dialogue presentation, but did not publish measured increases in engagement or adoption.

Rights and safeguarding shape implementation

The child-actor case states that parents and legal representatives authorized the voice samples used for a custom voice.

Independent benchmarking is still missing

None of the three original articles provides a controlled study proving a general speed, cost, or quality improvement across all projects.

Expert perspective

What Altered.ai’s Case Studies Mean for the Future of Voice Production

Editorial analysis, not a claim of future guaranteed results: these examples show AI voice systems becoming useful at different points in the production lifecycle. In games, the opportunity is not merely to generate more speech but to make supporting character dialogue viable within limited creative resources. In concept development, the opportunity is to judge an idea with a closer approximation of its eventual audio presentation. In television post-production, a custom voice may give editors temporary continuity while reducing the need for repeated recording sessions by a child performer.

The common thread is that voice generation moves from being an isolated software feature into a supervised production workflow. A voice model by itself is not a performance brief, a casting decision, or a finished audio mix. It requires an appropriate recording, clear direction, choices about tone and character, technical checks, and final editorial approval. The examples with Sam Hughes and the adult actor underscore the productive role of human performance rather than demonstrating its disappearance.

There are plausible efficiency advantages: teams may be able to audition variations, fill temporary gaps, or experiment with dialogue without recreating a full recording process each time. Yet the source material does not quantify how many hours were saved or whether synthetic voices consistently outperform traditional methods. Different languages, emotional scenes, voice models, and recording environments can produce different results. Creative teams should conduct listening tests on their actual material instead of extrapolating universal gains from a single case.

The ethical dimensions are equally material. Voice likeness is a personal and potentially commercially valuable attribute. Consent should specify what recordings may be used, which outputs may be created, who approves them, how long they may be retained, and whether they may be distributed or repurposed. In the case of minors, stronger safeguards are appropriate. The likely direction of responsible adoption is therefore neither all-human nor fully automated; it is a carefully governed collaboration where AI expands options but people remain accountable for quality, rights, and context.

Necessary caveats

Challenges of AI Voice Production

AI can extend a creative workflow, but a production-ready system must meet technical, legal, and human standards.

Source recording quality

Noise, reverberation, and performance inconsistencies can compromise synthesized output.

Likeness is imperfect

A recognizable timbre does not automatically reproduce every subtlety of a real voice.

Pronunciation and accents

Names, technical vocabulary, code-switching, and unusual phrasing need review.

Emotion and naturalness

Dramatic intent and subtle expression may require multiple iterations and human direction.

Copyright and contracts

Rights to scripts, performances, training audio, and output usage must be verified.

Consent and safeguarding

Likeness usage needs permission, with added protections for minors and vulnerable people.

Identity misuse risk

Synthetic speech can impersonate speakers; provenance and approval controls matter.

License and deployment limits

Commercial permissions and model availability can vary by service and subscription.

Quality assurance

Editors need repeatable listening criteria and a final human sign-off before publication.

For current conditions, consult Altered's published ethics principles and terms and conditions, and obtain legal advice when contractual voice rights are material.

Questions and answers

Frequently Asked Questions About Altered.ai Case Studies

What is Altered.ai used for?

Altered.ai offers speech-to-speech voice transformation, voice cloning, text-to-speech and voice editing tools for media workflows. It also markets a separate product for real-time voice calls.

How was Altered Studio used in The Ascent?

Sweet Justice Sound processed recordings by actor Sam Hughes using Altered Studio, resulting in 800 additional lines for 40 non-player characters, according to Altered’s December 2021 announcement.

Can one human performance support multiple characters?

Yes, as described for The Ascent. Voice transformation can create different vocal presentations from recorded performances, though quality and suitability require creative review.

How did Bossa Games use Altered Studio for Reclaimed?

Bossa said it transformed scratch voiceover into more professional-sounding narration and character dialogue for a concept-game trailer. It did not publish numerical time or cost results.

Did Altered name the television show in its child-actor example?

No. The April 2023 article describes an unnamed television series and a temporary voiceover workflow using a consent-based custom voice and an adult performer.

How are voice morphing and voice cloning different?

Speech-to-speech morphing changes a recorded performance into a target voice or style. Voice cloning creates a voice model resembling a particular speaker from authorized reference audio; a cloned model may then be used in supported workflows.

Is permission needed to clone another person’s voice?

You should obtain explicit, appropriate authorization and verify the relevant rights and license terms. Altered’s terms require rights to the source voice and consent from the speaker when the user is not that speaker.

What are the limitations of AI-assisted voice production?

Source audio quality, emotional consistency, pronunciation, voice similarity, licensing, and identity protection may require careful management. A human editor should approve the final performance.

Primary-source record

Sources and Further Reading

Direct links to original reporting and product documentation. Publication dates are included only where the official page visibly states one.

  1. [1] Neon Giant teams up with Altered for voice in action RPG game The AscentAltered · 16 December 2021https://www.altered.ai/blog/neon-giant-teams-up-with-altered-for-voice-in-action-rpg-game-the-ascent/
  2. [2] Bossa Presents uses Altered Studio to voice the protohype trailer for ReclaimedAltered · 21 December 2021https://www.altered.ai/blog/bossa-presents-uses-altered-studio-to-voice-the-protohype-video-trailer-for-concept-game-reclaimed/
  3. [3] How film and television studios use Custom Voice Services to support child actorsAltered · 13 April 2023https://www.altered.ai/blog/revolutionising-the-industry-how-film-and-television-studios-use-altered-studios-custom-voice-services-to-support-child-actors/
  4. [4] Altered: Voice content creation platform and product overviewAltered · date not shownhttps://www.altered.ai/
  5. [5] Voice Changer for Media ProductionAltered · date not shownhttps://www.altered.ai/voice-changer/
  6. [6] Local Voice CloningAltered · date not shownhttps://www.altered.ai/voice-cloning/
  7. [7] Altered Studio ShowcaseAltered · date not shownhttps://www.altered.ai/showcase/
  8. [8] Our Ethics CodeAltered · 23 February 2021https://www.altered.ai/blog/altereds-ethics-code/
  9. [9] Terms and ConditionsAltered · page date not verifiedhttps://www.altered.ai/terms-and-conditions.html

Research reviewed 12 November 2026. Publication dates describe the original articles, not the date on which Altered's current product capabilities were independently certified.

Closing perspective

Conclusion: The Real-World Impact of Altered.ai

Altered.ai’s published examples show three concrete ways that voice technology can support creative work. In The Ascent, a recorded actor performance was transformed into additional dialogue across 40 supporting characters. In Bossa Games’ Reclaimed trailer, scratch narration and dialogue became stronger material for presenting an early creative concept. And in a television example involving a child actor, a consent-based custom voice supported temporary post-production performances and the company reported improved workflow flexibility.

These cases are informative precisely because they are not interchangeable. Game dialogue expansion, prototype communication, and temporary voice continuity have different needs and different measures of success. The available sources document one specific dialogue-output count and several qualitative observations; they do not establish universal savings or benchmark synthetic speech against every conventional alternative. The broader opportunity lies in combining accessible voice tools with real acting, sound design, editorial judgement, and well-defined permissions. As AI voice production matures, its most useful applications will be those that preserve the dignity and agency of the speakers whose performances make the technology possible.

Explore Altered.ai’s official website