ElevenLabs launched a Voice Isolator tool

ElevenLabs, the AI company best known for specializing in voice and sound generation technologies, recently launched a free-to-try AI Voice Isolator tool that allows creators to remove unwanted ambient noise from content directly uploaded to the platform, or recorded on the go using the platform controls. The audio fed to the tool is processed to detect and remove background noise, producing studio-quality speech output. This new tool follows the launch of ElevenLabs' iOS Reader app, which enables users to transform provided text, files, or imported links into an audio narration, further diversifying ElevenLabs' offerings in the voice and sound generation space.

The Voice Isolator recording option is set to suggest users a quick demo, where several prompts invite them to say three different sentences while making common noises: the first sentence should be said quietly while clapping, the second should be said louder while banging the table, and the instruction for the third one is to say it calmly while snapping one's fingers. The Isolator does an excellent job at removing these common noises, and additional testing I performed showed that the tool is also well-equipped to isolate speech from instrumental background music. The Voice Isolator FAQ warns users that the tool is not meant to isolate vocals from music, although they may achieve varied degrees of success depending on the recording's features.

ElevenLabs also clarified that each processed audio minute costs 1000 characters, with the free tier providing users with 10,000 characters; enough to process up to 10 minutes monthly. Paid plans start at $5/month for larger files, and the Voice Isolator can handle files up to 500MB and 1 hour in length. Although the Voice Isolator does not support an API or real-time streaming support yet, ElevenLabs plans to extend the tool's capabilities to fill those gaps soon.