Transcription involves not only skilled use of language and sharp hearing, but also an understanding of nuance and cultural factors, familiarity with slang and dialect, and the ability to do some common-sense research; all implicitly human skills.
Machine learning and AI are evolving quickly into powerful tools that can help with some aspects of transcription, but the human factor still figures prominently in the production of high quality transcription services.
Photo by Annie Spratt
Who does what?
At CastingWords, we like to think we’ve found the perfect balance between automation of rote tasks, and relying on good old fashioned human ingenuity anywhere it will make a difference.
Have you ever watched a YouTube video with automatically generated subtitles? If so, you know that machine transcription will not be replacing people any time soon for final-draft transcripts, captions, or subtitles!
Our human workforce can understand accents and made-up words (“bigly,” anyone?), knows when a simple search can verify the spelling of names, places, and specialized terms, and can apply common sense to differentiate between homophones such as base and bass.
Under our current system, all transcription and editing tasks are completed by human workers.
We automate in-between steps such as website operations, audio and video processing, timestamp and caption placement, creation and distribution of tasks for our human workers, text formatting, document creation, file conversion, billing…the list goes on and on!
Anything a computer can effectively handle is automated so that our valuable human workers can focus on the tasks that only they can do: accurate transcription and editing all sorts of recorded material into polished, customer-ready transcripts.
Photo by Chris Ried
We’re always looking toward the future, and finding new ways to offer great products and services to anyone in need of transcription services.
While artificial intelligence is nowhere near being able to turn recorded speech into a clean, accurate, well-punctutated transcript, its output is far from useless. What it lacks in fidelity, it makes up for in speed and cost-effectiveness.
We are working on ways to further harness the advantages of both approaches and develop a wider range of products that can meet a wider range of our customers’ needs. Look for an announcement soon on machine transcription and who might find it useful, as well as machine transcriptions with human editing!