Data Retention Policy - Dictation Audio

When dictations come into InfraWare 360, the platform handles two versions of the audio and it stores them for various lengths of time.  The two versions are:

• Original Audio
• Compressed Audio

Original Audio is the form the audio took upon coming into the platform.  For Telephone Dictation System (TDS) jobs they are PCM WAV files with a 64k bitrate.  For handheld uploads from the IDC, they are most commonly DSS or DS2 files.  When audio is imported or retrieved from a third party dictation system, it is whatever supported audio format is provided.  Original audio is important for two reasons.  When First Draft speech recognition is performed, the engine requires original audio prior to any conversion that could result in a loss of fidelity.  The second reason is for analysis in the rare circumstance that a compressed audio file demonstrates problems that might not have been present in the original.  Original Audio is typically stored for ten days or until the job is complete.

Compressed Audio is what is used in the InfraWare 360 workflow for transcription, QA, etc.  In the first few seconds of a job's life in the platform, a compressed audio format is created that is suitable for workflow.  These files are typically much smaller in size than original audio files.  They still sound fine to the human ear, and they transfer faster to the MT computer.  The format is Windows Media Audio (WMA).  These files are used and stored for 180 days from the date of submission.

No user action is required related to audio file storage.  The platform knows what audio formats are available at each stage of the production process and uses them accordingly.