Skip to main content

Voice ReGen Service Overview

Voice ReGen API brings Waves’ voice enhancement engine into your product workflow, so you can clean and polish speech at scale. Built on the same award-winning audio technology used across major music and film productions, the API lets you programmatically process uploaded audio or video to reduce background noise, tame room echo, improve speech clarity, and smooth uneven levels, then return production-ready outputs through your own pipeline. For teams shipping creator tools, media platforms, or communications products, Voice ReGen API turns “bad audio” into a reliable backend capability: ingest, process asynchronously, track status, and deliver cleaner voice results automatically. You register files for processing, start processing, track progress, and download processed outputs.

The Voice ReGen service supports various input file formats including audio and media containers with audio tracks. Only the audio stream is processed. Files are registered both via direct uploads as well as via URL-based registration.

If an input can’t be processed (e.g. unsupported format / multi-channel / multi-track), you’ll see it reflected in the file status and error codes.

Supported Audio

Supported Audio Input Format

Audio CodecFile Extension
LPCM (Linear Pulse Code Modulation).wav
MPEG-1/2 Audio Layer III.mp3
LPCM.aiff, .aif
Advanced Audio Coding.aac, .m4a
Free Lossless Audio Codec.flac
Opus.opus
Vorbis.ogg

Input audio can be mono or stereo. Surround audio is not supported.

Supported Audio Input Quality

The following audio input quality is supported:

  • Sample rate: 4kHz <= Sample rate <= 96kHz
  • Bit depth: 8-bit <= Bit depth <= 32-bit

Input Audio Maximum Duration

  • Maximum audio duration: 5 hours

Audio Output

The processed audio file is provided in the same codec and file format as the input file except where noted, maintaining the original file metadata. The output audio for uncompressed PCM formats has the following characteristics:

  • Sample rate: 48k
  • Bit depth: 24bit
  • Mono

Notes

  • The audio is delivered in the original file format with the same or a compatible audio codec whenever supported. During this final conversion, some codec-specific properties may change. In particular, bit depth is preserved as a fixed property only for uncompressed PCM formats. For compressed formats such as MP3, AAC, Opus, and Vorbis/Opus, bit depth is not represented in the same way and should be considered codec-dependent rather than a guaranteed 24-bit output property.
  • Vorbis input will generate Opus as output

Supported Media Containers and Video Input

What are Media Containers?

Media containers are file formats that package together different types of multimedia data—such as video, audio, subtitles, metadata, and chapter information—into a single file without necessarily defining how the media itself is encoded. A container acts like a wrapper that organizes and synchronizes these streams so media players can play them correctly. For example, an .mp4 file may contain H.264 video and AAC audio. Containers are separate from codecs: the codec compresses and decompresses the media data, while the container stores and manages it.

Supported Media Containers

The following media containers are supported:

  • mp4
  • mov
  • webm
  • ogg
  • m4a
  • avi

The media container file may only include a single audio channel or track. In case there are multiple tracks, the processing will return an error.

The supported audio codecs include all supported input codecs mentioned before.

Input Media Containers Maximum Size

  • Maximum media container size: 10GB

Processing Media Containers Files

When processing media container files, the Voice ReGen service extracts the audio track, processes it and then replaces the audio track within the media container file. No change is applied to other tracks in the file. While the Voice ReGen service will maintain the same codec that was used in the original file, the output audio will have the audio quality described before.

User-Specified Output Format

Set options.output_type in startProcess to request an output format that differs from the input format. Supported output formats are:

  • wav
  • mp3
  • flac
  • aiff
  • ogg
  • opus
  • aac
  • m4a

Notes:

  • If options.output_type is not provided, the output uses the original file format.
  • If options.output_type is the same as the input file format then the option has no effect.
  • If options.output_type is different then the input file format then setting output_type returns an audio-only output and does not preserve video/subtitle/container metadata.

See Getting Started section to find a step-by-step walkthrough using cURL.