Speech Technology Magazine

 

To Dictate or To Record?

If you were on trial for a criminal offense, facing the possibility of life in prison, would you rather have the record of your trial created based on a recording or a court reporter sitting in the courtroom transcribing what everyone is saying?
By Robin Springer - Posted Nov 9, 2006
Page1 of 1
Bookmark and Share

If you were on trial for a criminal offense, facing the possibility of life in prison, would you rather have the record of your trial created based on a recording or a court reporter sitting in the courtroom transcribing what everyone is saying? Some jurisdictions have sought to save line items on their budgets by eliminating certified court reporters and replacing them with electronic recording systems (ERS), a tape or digital recorder with an open microphone that records the sound in the room from which a transcript is created.

Previously, court reporters created transcripts of legal proceedings by writing shorthand on paper, subsequently speaking the shorthand into a tape recorder and handing it off for someone else to listen to and generate a typewritten record.

In the 1940s, Horace Webb, a court reporter frustrated by the repetitiousness of the process, created a mask with a microphone inside. Using this Stenomask, Webb talked into a tape recorder instead of taking shorthand. At the end of the day, the recording had already been generated, and off it went to be typed. One step had been eliminated from the process.

With the 1990s came desktop dictation. Could it be possible to use the microphone attached to the computer instead of the tape recorder, further refining the system? If it worked, court reporters would be able to dictate directly into the computer using speech recognition and the typed transcript could be created on the fly. Proof for misrecognitions, make the corrections, and, voila, you've saved another step.

Court reporters including Linda Drake, former president of the National Verbatim Reporters Association, and Bettye Keyes, are ardent proponents of speech recognition. Drake first used Dragon NaturallySpeaking desktop dictation for court reporting in 1997. Working on a case that included 40 depositions of nearly 200 pages each and requiring a two- to three-day turnaround, she says of Dragon, "It missed a lot, but I had something on the screen to work with."

The technology improved with each release and users gained experience tweaking the system. Now, Keyes, author of "Voice Writing Method," dictates 180 to 200 words per minute with greater than 96 percent accuracy, even after taking into account punctuation and annotations.

ERS has been the next step in the courtroom's technology evolution. But critics of ERS say courts have gone too far because ERS is too unreliable and leaves no one accountable for the results. Horror stories include a criminal action in which part of the trial was inaudible and a legislative session that was left without a record when the ERS failed.

When a recording system fails during a trial, it is a point of error on which to base an appeal, the ramifications of which can affect a defendant's civil liberties, contribute to court congestion, and increase the costs associated with litigation, including increasing requisite taxpayer dollars to try, appeal, and retry a case.

There are fewer than 100,000 court reporters nationally, fewer than 10 percent of whom use their voices for court reporting, and fewer still who use their voices with speech recognition. However, speech recognition appears to be gaining popularity. While established court reporters may be resistant to change their processes, the younger generation is exploring the benefits of speech recognition from the outset. Court reporting schools now even include speech recognition in their curriculums.

How can we support voice writers? We can be mindful of the needs of niche markets. For example, Dragon NaturallySpeaking version 9 is the first Dragon release that allows users to accept dictation of contractions while rejecting abbreviations. While this may seem trivial to most, court reporters use contractions but not abbreviations so this feature is critical for them.

We can collaborate with companies that make complementary products for speech recognition. It wouldn't be feasible to dictate in a courtroom without an adjunct like Sylencer, from TalkTechnologies. Sylencer, successor of the Stenomask, was designed specifically for court reporters to use with speech recognition.

Technology assists us by automating tasks and streamlining workflows but at some point we need to take a step back and remember that technology is only as effective as the human being in control of it. We need to listen to the people who are controlling our technology to allow users ultimate levels of performance.



Robin Springer is the president of Computer Talk (www.comptalk.com), a consulting firm specializing in the

design and implementation of speech recognition and other hands-free technology services. She can be reached at (888) 999-9161 or contactus@comptalk.com.

 

 

Page1 of 1