Resample and encode audio

POST /v1/audio/upsamplings

Unreleased. Accepts single-track WAV PCM16/24/32/float32, FLAC, MP3, M4A/AAC, ADTS AAC, Ogg/Vorbis or Ogg/Opus. Limits: 1–8 channels, 32 MiB, 120 seconds or stricter gateway duration. Higher rates use AudioVAE at 48 kHz then downsample as needed; lower rates downsample directly. Default WAV delivery preserves original bytes only for equal-rate WAV without encoding overrides. Preserve speaker positions and synchronized channels, remapping storage order between containers as required. Any channel failure fails the whole request. Audio upsampling is included as a free extra and does not use your EdgeSpeak token allowance; authentication, license checks and file limits still apply. Separate transcription or speech-generation calls remain metered. FFmpeg codecs are required for compressed formats. Unknown/repeated fields are rejected.

Guide and examplesTranscription and speakersAll endpoints

Request body

multipart/form-dataRequired

filefileRequiredInput audio file; detected from content, not filename or MIME.
sample_rateintegerTarget Hz. Opus requires 48000; MP3/AAC/M4A accept supported discrete rates only.1–48000Default 48000
response_formatstringOutput container/codec: ogg=Vorbis, opus=Ogg/Opus, m4a=MP4/AAC, aac=ADTS/AAC. MP3 supports mono/stereo only.wav | flac | mp3 | m4a | aac | ogg | opusDefault "wav"
bitrateintegerTotal target bit/s across all channels. Lossy formats only. Default per-channel target: 96000 for MP3/AAC/Vorbis, 64000 for Opus, capped at 512000 and the supported rate range. Not an exact file-size guarantee.8000–512000
pcm_formatstringWAV sample representation; FLAC supports s16/s24 only. Rejected for lossy formats. Default preserves source precision where possible; compressed/model output defaults to WAV f32 or FLAC s24.s16 | s24 | s32 | f32

Responses

200Binary audio in response_format at the requested sample rate. Raw AAC may retain codec padding; use WAV/FLAC for exact timelines.
400Invalid audio, encoding combination or channel layout; invalid parameters include error.param.
401Authentication required.
403License or authorization required.
413Upload or audio duration limit exceeded.
503Runtime or codec unavailable, or encoding failed.