Freelance deliveries can be judged on clarity, pacing, format and client requirements. Submit a recording you are allowed to use and receive one processed file for your own listening and client-side checks.
First accepted beta recording free · MP3/WAV up to 50 MB · Private delivery · Results vary.
Rigid phrasing, codec glare, phase texture and inherited ID3, LAME or Lavf strings can all invite closer review. These are different kinds of signals: an encoder tag is not an acoustic fingerprint, and neither proves a provider watermark is absent.
Our six-stage DSP workflow can adjust spectral and temporal qualities, with a separate container review where appropriate. Human QA checks the result before private delivery. We cannot promise a detector score or identify a file’s origin from a string scan.
Start with the full acoustic and phase-dispersion profile for audible texture concerns, or request a container-only review if you want to keep the sound unchanged. Specify neutral or warm tone and your preferred MP3/WAV output in the upload notes.
Processing delivers one audio file for your own review. These points help you decide whether it is suitable for the project.
The Foxwood_01 scores are point-in-time observations of one source and processed file, not a prediction for yours. ElevenLabs’ current Audio Detector can check embedded watermarks separately from its legacy speech classifier.
Source: ElevenLabs detector documentation ↗Keep your original, describe your intended use in the upload notes and compare the finished audio by ear. You can also explore Hear the Foxwood_01 before and after or How private delivery works.