Here is a guide for using Watson to generate captions over IBM Watson Media. This file can be used in any text editor to view the transcription, or uploaded to services that recognize the WebVTT format. This transcription is then associated with the video file, and can be used over the IBM’s video streaming and enterprise video streaming platforms or downloaded as a WebVTT text file. So a 20 minute video would take roughly 20 minutes to generate the transcription. Once the language is set, it takes roughly the length of the asset to create a transcription of the audio. Setting the language can be done manually or setup to automatically transcribe content uploaded in the future. Content owners simply need to upload their video content to the service and select a language for the transcription process to start. Furthermore, this is included into the service for these offerings, at no additional cost. With IBM’s video streaming and enterprise video streaming offerings, this is done through using the IBM Watson technology. For example, there is a big difference between someone saying “we have a mail problem” versus “we have a male problem” even though both sound the same.Ĭonsequently, the success of a transcription service will depend on the strength of the AI behind it. It then needs to not only decipher what they were saying but the meaning behind it as well. This is a complex process that needs to separate dialogue from other noises to focus on transcribing the spoken language. Variables that lower automated transcription accuracyĪutomatic speech to text relies heavily on machine learning.

Considerations when using automated audio to text.Benefits of automated video transcription.Note: some of these topics were covered in our Simplify Your Corporate Video Strategy webinar, with the archived version available for immediate viewing.

Using this technology, content owners can start generating transcripts through simply uploading a file. This takes AI (artificial intelligence) and uses it to transcribe speech through combining information about grammar and language structure.

For those looking for a solution to scale or speed up video transcription, a solution is automated audio to text. Transcribing audio can be a slow process.

Video transcriptions