Edge Rewrite
translate Source code View original
Jump to content

Draft:Speech Graphics

From Wikipedia, the free encyclopedia
  • Comment: Slightly rewording AI writing doesn't make this article any less AI generated. You need to take WP:TNT to it, blow it up and rewrite it completely using sources you find and in your own words. Sulfurboy (talk) 18:27, 7 August 2026 (UTC)
  • Comment: Please be aware that use of LLMs to generate content for Wikipedia is not allowed, see the guideline at WP:NOLLM. They are very bad at actually summarizing sources: you will need to hand-check every claim to make sure the source actually supports it, and also that it's not so closely paraphrased as to be a copyright violation. M kuhner (talk) 01:04, 7 August 2026 (UTC)
  • Comment: Have you used AI to generate any of this content? Theroadislong (talk) 16:11, 6 August 2026 (UTC)
  • Comment: In accordance with the Wikimedia Foundation's Terms of Use, I disclose that I have been paid by my employer for my contributions to this article. GregorHofer (talk) 14:51, 6 August 2026 (UTC)
  • Comment: Declaring a conflict of interest: I am the CEO and co-founder of Speech Graphics. Paid/COI disclosure is on my user page and on this draft's talk page. I am submitting through AfC rather than publishing directly.
    Strongest independent, secondary sourcing for notability:
    * Metzger et al., Nature (2023) — the company's facial-animation technology, and its CTO, were part of the study.
    * Mashable (2023) — mainstream coverage naming the company and its role in the study.
    * GamesIndustry.biz (2024) — feature discussing the company and its Rapport platform.
    *GamesBeat (2022) - coverage of company fund raise and new product
    * Alphr (2020) and DigitalNation.scot (2026) — profiles of the company.
    Funding and acquisition coverage (Silicon Canals, FinSMEs, Built In Edinburgh) supports specific factual claims rather than notability. GregorHofer (talk) 14:51, 6 August 2026 (UTC)
Reply: I have completely rewritten all sections without the use of AI while keeping the orginial strucuture. I have also added additional sources that I have verified myself including the Times and Gamesbeat. GregorHofer (talk) 16:35, 7 August 2026 (UTC)


Speech Graphics
TypePrivate
Industry
Founded2010; 16 years ago (2010) in Edinburgh, Scotland
FoundersMichael Berger, Gregor Hofer
Headquarters
Edinburgh
,
Scotland, United Kingdom
Key people
Gregor Hofer (CEO); Michael Berger (CTO)
ProductsSGX, Rapport, FaceFX
Number of employees
~60 (2025)
Websitewww.speech-graphics.com

Speech Graphics is a technology company that develops animation technology and is based in Edinburgh, Scotland. It was founded in 2010 as a spin-out from the University of Edinburgh. Speech Graphics core technology generates facial animation automatically from audio, without the use of motion capture, by analysing the voice to drive a model of the facial muscles involved in speech and expression.[1][2] [3] The company's software has been used in many video game titles including Middle-earth: Shadow of Mordor, The Last of Us Part II, Hogwarts Legacy and Monster Hunter Wilds.[4][1][3]

History

[edit]

Speech Graphics was founded in 2010 as a spin-out from the School of Informatics, University of Edinburgh by Michael Berger and Gregor Hofer [5][2] while finishing their PhD's in speech technology and machine learning. [3]

The first public use of the company's technology was the facial animation for Kanye West's 2014 music video "Black Skinhead", also in 2014 the company provided facial animation for Middle-earth: Shadow of Mordor, published by Warner Bros. Games.[6] In 2016, Gears of War 4 was the first video game to make use of SGX.[7]

The company raised a Seed round in 2018 [8] and in 2022 raised a Series A funding round led by Sands Capital to launch an enterprise-facing brand, Rapport, for conversational avatar applications.[9]

In September 2025 the company the FaceFX facial-animation software from OC3 Entertainment. FaceFx is used in games such as Halo, Gears of War, Grand Theft Auto V and Baldur's Gate 3; OC3's co-founder Jamie Redmond joined Speech Graphics as part of the transaction.[4][10] The company employs around 60 staff and is headquartered in Edinburgh, with further offices in Austin, Budapest and Singapore.[4][3]

Products

[edit]
  • SGX generates facial animation from voice recordings which can directly be imported into game engines and 3D animation software. It is distributed with plug-ins for Autodesk Maya and Unreal Engine.[3]
  • SG Com generates facial animation in real-time from either live or synthetic audio input; it has been used for real-time avatar voice chat, including the player-to-player lobby chat in Fortnite.[3]
  • Rapport is a cloud platform for AI-driven avatars, which was launched as a brand in 2024. It is used in enterprise applications such as scenario based roleplay simulations.[5][2]
  • FaceFX is the facial-animation software acquired from OC3 Entertainment in 2025.[4][10]

Notable projects

[edit]

Speech Graphics' technology has been used in many AAA video games, including Middle-earth: Shadow of Mordor, The Last of Us Part II, Hogwarts Legacy, Resident Evil Village, Gotham Knights, Dying Light 2 Stay Human, High on Life and Monster Hunter Wilds.[3][4] For Hogwarts Legacy, the company's SGX suite was used to produce facial animation across multiple languages,[3] and for Monster Hunter Wilds it was used to generate lip sync and expressions in Japanese, English and Chinese.[11]

Research applications

[edit]

Through a colaboration with researchers at the University of California, San Francisco and the University of California, Berkeley SG Com was used as part of a brain–computer interface study, published in the journal Nature in 2023, that enabled a severly paralysed patient to communicate through a digital avatar.[12][2] As the patient attempted to speak, neural activity was recorded and used to synthesise speech as well as avatar movements with Speech Graphics' technology .[12] The SG Com real-time software was used to convert decoded muscle movements into 3D facial animation in a game engine, similar to a video game.[13] [14]

References

[edit]
  1. 1 2 "Speech Graphics: Revolutionizing Facial Animation with Audio-Driven AI". DigitalNation.scot. 5 April 2026.
  2. 1 2 3 4 Freeman, Will (8 July 2024). "Levelling up healthcare: Understanding games' life-changing medical potential". GamesIndustry.biz.
  3. 1 2 3 4 5 6 7 8 "A Deep Dive Into Speech Graphics' Facial Animation Solutions". 80 Level. 20 March 2023. Interview with co-founder and CTO Michael Berger.
  4. 1 2 3 4 5 "Speech Graphics Acquires OC3 Entertainment". FinSMEs. 2 September 2025.
  5. 1 2 "Scotland-based Speech Graphics raises €6.22M for its facial animation technology; introduces a new brand, Rapport". Silicon Canals. 25 March 2024.
  6. ↑ "Inside Speech Graphics: creating the virtual faces of the future". Alphr. 3 June 2020.
  7. ↑ "Tech firm talks a great video game". The Times. 16 November 2016.
  8. ↑ "Speech Graphics pays more than lip service for growth". The Times. 27 February 2018.
  9. ↑ Takahashi, Dean (23 February 2022). "Speech Graphics raises $7M for lip-synced facial animation". GamesBeat.
  10. 1 2 "Speech Graphics Acquires OC3 Entertainment's Game Animation Tech". Built In Edinburgh. 2 September 2025.
  11. ↑ "Speech Graphics releases a case study on the use of "SGX" in Monster Hunter Wilds". CGWORLD.jp (in Japanese). 16 July 2026.
  12. 1 2 Metzger, Sean L.; Littlejohn, Kaylo T.; et al. (2023). "A high-performance neuroprosthesis for speech decoding and avatar control". Nature. 620 (7976): 1037–1046. doi:10.1038/s41586-023-06443-4. PMC 10826467. PMID 37612505.
  13. ↑ "AI-Powered Video Game Tech Helps Paralysed Woman To Speak Again". BeyondGames.biz. 1 September 2023.
  14. ↑ Dobriyanova, Teodosia (24 August 2023). "Video game technology helped a woman regain her speech". Mashable.
[edit]

Category:Companies based in Edinburgh Category:Software companies of Scotland Category:Animation companies Category:Companies established in 2010 Category:University of Edinburgh

Retrieved from "https://en.wikipedia.org/w/index.php?title=Draft:Speech_Graphics&oldid=1368211458"