The Super Transcriptionist: transcribing with technology
For the modern transcriptionist, technology advancements have become a double-edged sword. On one hand, transcription software, cloud file-sharing services, and transcription foot pedals have vastly improved the industry's capabilities. Not to mention, with the near ubiquity of smartphones, and the multitude of ways to both record and distribute audio, the need for transcriptionÔøΩ and for transcriptionistsÔøΩ has grown exponentially. However, the transcription process is still incredibly manual and, even with the introduction of speech recognition, remains relatively unchanged.
Despite the primary purpose of transcription (to convert sound files to readable documents), a transcriptionist's job involves much more than typing. In a 2015 interview posted to LinkedIn, Lenna K. Millar, director of Audio Transcription & Secretarial Services (ATS), outlined the complexities of her work. Proofreading, editing, and rapid knowledge acquisition are just some of the additional skills required for accurate and complete transcription, but there's more.
"We tune in to what the speakers are saying. We make judgment calls. We use our discretion," she said. Millar and her team, as well as all transcriptionists, aren't just putting words on a digital page. They're searching for a deeper meaning in the dialogue. But one primary factor can hinder their ability to find that meaningÔøΩ bad audio.
Some transcriptionists estimate that poor quality audio can increase the length of a job by 50% or more. This is in addition to the typical 3ÔøΩ ÔøΩ6 hours they already spend on one hour of audio. From noisy surroundings to bad acoustics to poor enunciation, several interferences can complicate a transcriptionist's ability to understand and capture the full dialogue. This is coupled with many clients who aren't empathetic and are displeased with the extra time it takes to produce an accurate document.
Speech recognition technology has been positioned as a savior (and sometimes, replacement) to transcriptionists for many years. Not only would it save them time but it would help them extract meaning from choppy audio and tighten the scope of their work. Sadly, up till now, speech recognition has not lived up to the promise.
According to a 2016 article from WIRED, humans are capable of transcribing speech with a 4% margin of error. Speech recognition's margin of error falls between 8ÔøΩ ÔøΩ12%, depending on a range of factors, and it too drops with poor quality audio. So even though speech recognition technology saves transcriptionists time overall, it increases the effort spent on editing and proofreading. Thus, transcriptionists are still looking for a speech recognition alternative that can improve their day-to-day experience without adding a whole load of complications.
We hope to fill that void.
We started Simon Says to make transcribing, proofing and editing as efficient, easy, and natural as possible. We're motivated by the same mission as the transcriptionistsÔøΩ to help our end-user find the deeper meaning in that dialogue.
With our website, you can upload almost any type of audio and video file and swiftly get back a transcript, with timecode sync. Those transcripts can be edited, bookmarked, annotated, and exported to Microsoft Word, Microsoft Excel and SRT subtitles. With Simon Says, we've considered the challenges and limitations of transcription in its current state and we've worked to create a site that makes the process smoother and more cost-effective.
In recent years, the tech and transcription industries have struggled to grow together. We hope Simon Says will change that narrative and pave the way forward.
To sign-up to Simon Says and receive credits to try the site for free, go to: https://www.simonsaysai.com