| ed by voice recognition software? | | | | takes to transcribe. He asked if I would be |
| This article indicates a number of reasons why | | | | prepared to use a voice recognition program and |
| transcription will not be replaced by voice | | | | re-speak his files for him at a rate that reflected |
| recognition software, at least for the foreseeable | | | | the 2.5 hours per hour. I agreed to purchase the |
| future. Well, maybe as a professional | | | | software as I’d been toying with the idea |
| transcriptionist I could be considered biased, but | | | | of doing so anyway, and do a trial tape for him. I |
| there are issues with voice recognition software | | | | said I wasn’t prepared to agree to 2.5 |
| that I don't think will be solved in the next few | | | | hours per hour until I’d proved it to |
| years. | | | | myself, and he accepted this. |
| Well maybe as a professional transcriptionist I | | | | Funnily enough, several months on, I’m still |
| could be considered biased, but this article | | | | waiting for that trial tape. Having experimented |
| indicates a number of reasons why transcription | | | | with the software I find that re-speaking actually |
| will not be replaced by voice recognition software, | | | | takes me
guess how long — once the |
| at least for the foreseeable future. | | | | checking has all been done, about four hours per |
| What is voice recognition software? | | | | hour of recording! |
| Voice recognition software (the best known | | | | I do use the software now and then, because on |
| example is probably Dragon Naturally Speaking by | | | | a busy week it saves the fingers and wrists, but |
| Nuance) is software that allows the operator to | | | | what it doesn’t save is time! When |
| speak in the computer in order to give the | | | | I’m transcribing I proofread as I type. |
| computer commands (e.g. to save a file) or to | | | | Then just a quick skim through is required |
| dictate speech into the computer. Over the past | | | | afterwards to make sure no errors have sneaked |
| ten years it has been hailed a number of times as | | | | through. Using the software to dictate one can |
| the answer to fast and accurate dictation | | | | probably do this, and there may be a little time |
| — but we’re still waiting. The latest | | | | saved, but one cannot do this while re-speaking |
| version of Dragon claims to be 99% accurate and | | | | because one is concentrating on listening to the |
| three times faster than typing, so I recently | | | | words being said and repeating them a few |
| purchased a copy to see. | | | | words after, while trying to continuously play the |
| Dictation only? | | | | recording. |
| It is important to remember, especially if you are | | | | Also because the speech recognition software is |
| producing interviews rather than one-person | | | | not intelligent it often doesn’t know which |
| dictation, that the software will only recognise the | | | | homonyms to use. This is something that has |
| one voice it’s trained in. If you try to | | | | been vastly improved over the years, and in |
| record an interview into it, it will pick up your | | | | simple sentences it is often capable of working |
| voice but not the interviewee’s. It is | | | | out whether to, two or too is required; or |
| possible to re-speak an interview — listen | | | | whether it’s here or hear. However in a |
| to the interview recording and speak the words | | | | longer and more complex sentence it frequently |
| of both interviewer and interviewee a few words | | | | struggles with this and very careful proofreading |
| after they’re said. However, that’s | | | | is required afterwards. |
| an art in itself, and even a trained court | | | | Also, when you’re dictating you |
| stenographer would struggle to listen to an | | | | don’t really want to also have to use the |
| interview straight through, with no pausing, and do | | | | keyboard, so the software allows you to include |
| this, so how much time does it really save? | | | | commands for making words bold, underling them |
| Why is this software not about to take over the | | | | etc. However, you need to leave a small pause |
| world? | | | | between the command and general dictation or |
| A recent seminar given by a seller of digital | | | | the software just things the command is part of |
| dictation systems and associated voice recognition | | | | the words being dictated. That also slows things |
| the presenter indicated that voice recognition | | | | up, as compared to professional transcribe using |
| would probably only work well for three in ten | | | | keyboard shortcut keys for these commands. |
| people. This is because the software has to be | | | | And re-speaking an interview is even worse |
| first trained to recognise your voice (a process | | | | because every time there’s a change of |
| which really does only take a few minutes) and | | | | speaker that needs to be indicated either with the |
| then continuously corrected and trained to | | | | name or initials and a tab, or the name and a new |
| recognise words that it’s either not familiar | | | | paragraph, so there are a number of commands |
| with or simply doesn’t recognise from | | | | involved around that. |
| your speech patterns, which is an ongoing process | | | | To save time the latest version of Naturally |
| and does take time. | | | | Speaking will actually try to punctuate for you
|
| I used this software first about five years ago | | | | so you dictate and it puts the commas and full |
| and I have to say it’s come on in leaps | | | | stops in the right places. That’s the theory |
| and bounds since then, but even the latest | | | | anyway. Suffice it to say that I wouldn’t |
| version can be a struggle! I purchased the latest | | | | recommend it! |
| version on the advice of a potential client who | | | | Conclusion |
| claimed that even re-speaking his interviews | | | | All in all I’d say Dragon 9, the latest |
| (listening to a recording and then speaking the | | | | version, is a very useful tool, but it will be some |
| words into the software as they were said) he | | | | time before it’s really ready to replace |
| could get an hour of recording down in 2.5 hours | | | | transcriptionists. |
| rather than the four hours per hour it usually | | | | |