ElevenLabs just got nuked by open source

Quen 3 TTS: A Game Changer for Voice Cloning?

Introduction to Quen 3 TTS

• Quen has released Quen 3 TTS, raising concerns for ElevenLabs regarding voice cloning capabilities.

• The speaker shares a personal experience of having their voice cloned using ElevenLabs last year, noting improvements in the technology.

Accessibility and Ease of Use

• Open-source Qwen by Alibaba Cloud allows users to download and run TTS models on various systems, including Raspberry Pi and smartphones.

• Users can easily clone voices by uploading audio recordings along with transcripts; this process is straightforward enough for anyone to replicate.

Demonstration of Voice Cloning

• The speaker demonstrates the ease of generating cloned voice audio by inputting text into the system.

• Despite some limitations in intonation and vocal range, short phrases can convincingly mimic a person's voice.

Implications of Voice Cloning Technology

• Short snippets generated by the model can be convincing enough to fool listeners unfamiliar with the original speaker's voice.

• The potential misuse of this technology raises concerns for individuals whose voices are integral to their online presence and revenue generation.

Personal Concerns About Voice Misuse

• The speaker expresses discomfort about unauthorized use of their cloned voice, emphasizing that they have never used a cloned version in any content.

• There is an increasing worry about how accessible and easy it has become to clone voices without consent.