Speech input - Amazon Transcribe

Speech input

Amazon Transcribe Medical can transcribe speech as either an audio file or a real-time stream. Your input audio must use the encodings and formats described in the following sections.

Containers and formats for batch transcription

When you transcribe an audio file using the StartMedicalTranscriptionJob API or the Amazon Transcribe Medical console, make sure that the file is:

  • In FLAC, MP3, MP4, Ogg, WebM, AMR, or WAV file format

  • Less than 4 hours in length and less than 2 GB in size

  • Encoded at a sample rate of 16,000 Hz or higher

Note

For AMR, Amazon Transcribe Medical supports both Adaptive Multi-Rate Wideband (AMR-WB) and Adaptive Multi-Rate Narrowband (AMR-NB) codecs.

For the Ogg and WebM file formats, Amazon Transcribe Medical supports the Opus codec.

For best results:

  • Use a lossless format. You can choose either FLAC, or WAV with PCM 16-bit encoding.

Audio containers and formats for streaming transcription

When you transcribe a real-time stream using the StartMedicalStreamTranscription API or a WebSocket request, make sure that your stream is encoded in:

  • PCM 16-bit signed little endian

  • FLAC

  • OPUS encoded audio in the Ogg container

Your stream must use a sample rate of 16,000 Hz or higher.

For best results:

  • Use a lossless format, such as FLAC or PCM encoding.

For more information on using a WebSocket request to transcribe your streaming audio, see Establish a bi-directional connection using the WebSocket protocol.