| Many businesses need to convert recorded voice | | | | long time and a lot of work. Most wound up |
| to text and have long been looking for ways to | | | | continuing to outsource their medical transcription |
| do it quickly and inexpensively. Transcribing | | | | work. |
| medical dictation is a prime example. | | | | Of course, there are many other types of |
| Some years ago, when voice recognition | | | | situations where transcription is needed. Examples |
| software became commercially available, most | | | | include recordings of seminars, teleconferences, |
| people expected that the solution had finally | | | | interviews and classes that need to be converted |
| arrived. Businesses looked forward to cutting | | | | to text. |
| down on transcription costs and everyone who | | | | In natural speech, people tend to use lots of |
| hated typing looked forward to getting rid of their | | | | "aahs" and "umms" as well as unnecessary |
| keyboard. | | | | phrases like "you know". Current voice recognition |
| Unfortunately, the reality turned out to be rather | | | | technology is just not capable of filtering out such |
| different. Voice-to-text technology has been a big | | | | irrelevant sounds or words. |
| let down so far. | | | | In addition, people also string together several |
| The fact is, voice recognition software is easily | | | | sentences using "ands". The software can't break |
| thrown off track by many different factors. If | | | | up such speech into meaningful sentences. Nor |
| you don't speak clearly and distinctly, it may not | | | | can it break up speech into meaningful paragraph |
| give you the right output. If you try using it in a | | | | units the way a transcriptionist can. |
| noisy place, it will fail more often than not. If you | | | | And if the recording is filled with background noise, |
| have an accent, it may not understand you. Even | | | | or if more than one person is talking at the same |
| if you have a bad cold, you'll find that the | | | | time, the software will not function reliably and |
| software may give incorrect results! | | | | consistently. |
| In other words, voice recognition software works | | | | Maybe sometime in the future someone will |
| reasonably well under ideal, laboratory conditions, | | | | invent voice recognition technology that can |
| but not in a typical home or business setting! | | | | handle all the above issues. Till then businesses will |
| Healthcare professionals who attempted to use | | | | need to use transcription services, particularly for |
| voice recognition technologies to eliminate | | | | work like medical transcription, where accuracy is |
| transcription services found that they need to | | | | critical. |
| "train" the software to function well. That takes a | | | | |