Speech synthesis
Artificial production of human speech
Follow Speech synthesis on Notably News to receive short updates to your email — rarely!
We include updates on ElevenLabs, 15.ai, Audio deepfake, Software Automatic Mouth, VoiceXML, Dr. Sbaitso, DECtalk, PlainTalk, Mockingboard, PSOLA, Currah, Voice browser, Phase vocoder, Inverse filter, Source–filter model, Remote infrared audible signage ... and more.
| June 2025 | ElevenLabs released Eleven v3, a new Text to Speech model that supports more than 70 languages, natural multi-speaker dialogue, and audio tags like [excited], [whispers], and [sighs]. |
| May 22 2025 |
Audio deepfake
Claim emerged that Hoya Corporation used Gayanne Potter's recordings without full permission, now using her voice as 'Iona' for ScotRail train announcements, extending beyond the original intended use.
|
| February 2025 | ElevenLabs released Scribe, a speech-to-text model that transcribes audio with character-level timestamps and speaker diarization. |
| February 2025 | ElevenLabs introduced a new platform enabling authors to create and publish AI-generated audiobooks directly on its Reader app. |
| January 30 2025 | ElevenLabs completed a $180 million Series C funding round, raising the company's valuation to $3.3 billion. The round was co-led by a16z and ICONIQ Growth, with strategic investors including Deutsche Telekom, LG Technology Ventures, and others. |
| 2024 |
Audio deepfake
OpenAI corroborated the 15-second data efficiency benchmark for audio deepfake generation originally demonstrated by the MIT researcher in 2020.
|
| 2024 |
Audio deepfake
Over 20,000 New Hampshire voters received robocalls with an AI-impersonated President Joe Biden urging them not to vote, which violated state election laws.
|
| 2024 | ElevenLabs limited access to its voice cloning feature to paid subscribers to improve user accountability and prevent potential misuse of the technology. |
| July 2024 | ElevenLabs released 'Voice Isolator', a tool that removes background noise from audio. |
| June 2024 | ElevenLabs released the ElevenLabs Reader App on iOS and Android, allowing users to listen to articles, PDFs, and ePubs with AI Voices on their phone. |
| May 2024 | ElevenLabs shared samples from their text-to-music model, though it was not yet available for public use. |
| May 2024 |
Audio deepfake
The FCC proposed a $6 million fine against Steve Kramer for spoofing a local political figure's number, and four New Hampshire counties indicted him on felony voter suppression and candidate impersonation charges.
|
| May 2024 | ElevenLabs launched a text-to-music model. |
| February 2024 |
Audio deepfake
US Federal Communications Commission banned the use of AI to fake voices in robocalls.
|
| February 2024 |
Audio deepfake
Political consultant Steve Kramer admitted to commissioning the AI Biden robocalls for $500, claiming he wanted to highlight the need for AI campaign regulations.
|
| January 2024 | During the New Hampshire Democratic primary, AI-generated robocalls mimicking Joe Biden's voice were sent to thousands of residents, discouraging voting. The New Hampshire attorney general's office investigated the incident, with audio experts confirming the calls were created using ElevenLabs technology. |
| January 22 2024 | ElevenLabs raised $80 million in Series B funding, increasing the company's valuation to $1.1 billion. The round was led by Andreessen Horowitz, Nat Friedman, Daniel Gross, and Sequoia Capital. The company also announced new products including Voice Marketplace, AI Dubbing Studio, and a mobile app. |
| November 13 2023 |
CeVIO
HARU (羽累), the fifth female AI singing voicebank by Kamitsubaki Studio for CeVIO AI, was released as a 'musical isotope' of virtual rapper/singer Harusaruhi.
|
| October 2023 |
Audio deepfake
An audio deepfake of Slovak politician Michal Šimečka falsely claimed to capture him discussing election rigging methods.
|
| October 2023 |
Audio deepfake
An audio deepfake of Labour leader Keir Starmer was released, falsely portraying him verbally abusing staffers and criticizing Liverpool during the Labour Party conference.
|
| October 2023 | ElevenLabs presented 'AI Dubbing', a tool capable of translating speech into more than 20 languages while preserving the original speaker's voice, emotions, and intonation. |
| September 2023 | ElevenLabs released the 'Projects' tool for creating long-form spoken content. |
| August 2023 | ElevenLabs expanded its voice generation capabilities to 28 languages, using an in-house AI model that automatically detects languages like Korean, Dutch, and Vietnamese, and officially exited its beta phase. |
| July 2023 | ElevenLabs announced 'Projects', a tool for creating long-form spoken content like audiobooks and dialogue segments with contextually-aware synthetic or custom voices. |
| June 2023 | ElevenLabs raised a $19 million Series A funding round at a valuation of about $100 million, with co-leadership from Andreessen Horowitz, Nat Friedman, and Daniel Gross. |
| June 2023 | ElevenLabs reached over one million registered users since its launch in January, demonstrating rapid user adoption and market interest. |
| June 20 2023 | ElevenLabs released the AI Speech Classifier, a tool designed to determine if an uploaded audio sample originates from their AI technology, which they claim is the first of its kind and accessible through an API. |
| June 2 2023 |
CeVIO
Techno-Speech published an article about the development of a new Vocoder, which included a demonstration of an unknown female talk voice.
|
| March 2023 |
Audio deepfake
US Federal Trade Commission issued a warning to consumers about AI being used to fake family members' voices in distress, requesting money.
|
| January 2023 | The company publicly released its beta platform for AI voice technology. |
| January 2023 | ElevenLabs secured a $2 million pre-seed funding round led by Credo Ventures and Concept Ventures, highlighting their potential in AI voice intelligence. |
| January 2023 | ElevenLabs launched, introducing an AI voice generation platform with high-quality voice output and fast generation times. |
| January 25 2023 |
CeVIO
COKO (狐子), the fourth female AI singing voicebank by Kamitsubaki Studio for CeVIO AI, was released as a 'musical isotope' of virtual singer KOKO.
|
| 2022 | ElevenLabs was co-founded by Piotr Dąbkowski and Mati Staniszewski, two Polish entrepreneurs with backgrounds in machine learning and deployment strategy. |
| December 21 2022 |
CeVIO
CeVIO releases POPY (AI), a female vocal based on Kasumi Toyama from Poppin'Party, with voice data recorded from previous Poppin'Party songs and voiced by Aimi.
|
| December 21 2022 |
CeVIO
CeVIO releases ROSE (AI), a female vocal based on Yukina Minato from Roselia, with voice data recorded from previous Roselia songs and voiced by Aina Aiba.
|
| December 2 2022 |
CeVIO
Futaba Minato, a female AI vocal voicebank by Gasoline Alley for CeVIO AI, is released. Her voice is provided by voice actress Sachika Misawa, and she is designed as a youth girl's singing voicebank.
|
| October 25 2022 |
CeVIO
RIME (裏命), the third female AI singing voicebank by Kamitsubaki Studio for CeVIO AI, was released as a 'musical isotope' of virtual singer RIM.
|
| April 29 2022 |
CeVIO
SEKAI (星界), the second female AI singing voicebank by Kamitsubaki Studio for CeVIO AI, was released as a 'musical isotope' of virtual singer Isekaijoucho.
|
| February 25 2022 |
CeVIO
CeVIO announced the Kizuna AI vocal product (#kzn), a female AI vocal for CeVIO AI and VoiSona based on the VTuber Kizuna AI, capable of singing.
|
| January 2022 |
WordQ+SpeakQ
WordQ Chrome browser version was scheduled for release, expanding the software's accessibility to Chrome users.
|
| 2021 |
Audio deepfake
Actress Gayanne Potter provided recording work for Hoya Corporation's ReadSpeak, understanding it would be used for accessibility and e-learning software.
|
| July 7 2021 |
CeVIO
KAFU (可不), the first female AI singing voicebank by Kamitsubaki Studio for CeVIO AI, was released as a 'musical isotope' of singer and VTuber KAF.
|
| 2020 |
Audio deepfake
AI voice impersonation technique was used to convince a branch manager to transfer $35 million by mimicking a company director's voice.
|
| March 2020 |
Audio deepfake
A Massachusetts Institute of Technology researcher demonstrated data-efficient audio deepfake generation through 15.ai, a web application capable of generating high-quality speech using only 15 seconds of training data. The system introduced a unified multi-speaker model with speaker embeddings, allowing simultaneous training of multiple voices and learning shared patterns across different emotional contexts.
|
| 2019 |
Audio deepfake
Scammers used AI to impersonate the voice of a German energy company CEO, directing the UK subsidiary CEO to transfer funds in a sophisticated voice fraud scheme.
|
| January 2019 |
WaveNet
DeepMind published the 'Unsupervised speech representation learning using WaveNet autoencoders' paper, enhancing the automatic recognition and discrimination of dynamic and static voice features for more reliable voice swapping.
|
| 2018 |
CeVIO
Techno-Speech published an article showcasing samples of their upcoming AI technology, including a Chinese Female Vocal for CeVIO Creative Studio, alongside Japanese (Sato Sasara) and English (IA) voice samples.
|
| September 2018 |
WaveNet
DeepMind published the 'Sample Efficient Adaptive Text-to-Speech' paper, showing WaveNet could reduce voice sampling requirements to just a few minutes of audio data while maintaining high-quality results.
|
| June 2018 |
WaveNet
DeepMind published a paper demonstrating WaveNet's ability to perform audio and voice 'content swapping', allowing conversion of a speaker's voice while maintaining the original speech content.
|
|
We are only showing the most recent entries for this topic. |
|
This contents of the box above is based on material from the Wikipedia articles Audio deepfake, WaveNet, ElevenLabs, CeVIO & WordQ+SpeakQ, which are released under the Creative Commons Attribution-ShareAlike 4.0 International License.