What I Wish My Steno Colleagues Knew About Voice Writing

[April Thompson]

Cross-posted at BFFProofreading.com

We have a guest post from Tori Pittman of Pittman Stenography. Tori was the Intersteno World Champion in Speech Capture, Speech Recognition (speed competition) for 2013 and 2015. She explains all about the unique practice of voice writing.

It seems every time I turn around, someone is asking about that mask on my face. “Are you on oxygen? What in the world? You say you’re a court reporter? I’ve never heard such a thing. Why don’t you use that funny-looking typing machine?”

Sadly, there seems to be a lot of misinformation and even disinformation about what voice writing is and whether voice writers should be called court reporters.

I'd like to clear up the confusion and share what a voice writer does, comparing it to steno reporting in input, translation, and output.

The main distinction: Input

This is obviously the biggest difference between a voice reporter and a steno reporter. A voice writer uses their voice to dictate into a speech silencer. (Many people use the term Stenomask© to describe the silencer, but that is just one specific type of mask.) This silencer is a closed microphone system that allows the voice writer to revoice/dictate what is being said in the room without the participants hearing.

In addition, the voice writer has to be cognizant of their dictation level. The certifying organization, NVRA (National Verbatim Reporters Association) calls their practitioners “the silent voice.” If voice writers aren’t silent, they’re not doing their job correctly.

A voice writer, even with all of the sound dampening, cannot dictate in their normal voice; they need to modify their speaking tone, their breathing (did you notice the mask), their sibilance, and their diction. Plus, they need to make sure their vocal cords stay hydrated and they don’t get nodes on them.

Like a steno writer, the voice writer is not only dictating what is being said, but is also dictating speaker and Q/A tokens, punctuation, and short forms/arbitraries. All this factors into . . .

The next step: Translation

The sound recording made in the mask comes in via analog sound and gets translated to digital through the sound card and a speech recognition engine (SRE) such as Dragon Naturally Speaking or other program. The SRE takes the sound provided and turns it back to its best matching phoneme.

Just like steno is written in syllables, voice is translated into phonemes, or the sound of each word part. When a voice writer begins working with the SRE, they must train the software to understand how they speak. This entails going through “basic training” in the engine, running through a dictation with different sound patterns so that the software can learn an individual’s speech/sound pattern. Once that’s complete, most trainers suggest at least one more training read to increase the SRE’s phoneme database; I use JFK’s inaugural address.

But we court reporters can individualize our SRE to make it even more accurate for our profession.

The first way we can increase its accuracy is to create what are called “custom words” – or what steno writers might call a main or job dictionary. These are the special tokens used to get very specific outputs. You can develop your own speech vocabulary to add to the SRE's already large lexicon. For example, I say “keek” for a Question token, “kak” for an Answer token, “corko” for the Court in colloquy and SPEE1, SPEE2, and so on for various speakers. Here is where you define your arbitraries, for example, “statemac” for “state your name for the record” or “obformco” for “objection to the form.” A voice writer has limitless ability to create these codes.

You can also increase translation accuracy by having the engine “analyze documents” and learn how words are said in your area. Depending on whether you’re in court or a deposition, phrases are handled differently or are used more frequently. The engine can analyze a final document, learn how words eventually are outputted, and add that information to its database to translate future questionable phoneme combinations.

The big finale: Output

Most voice writers will use their SRE in tandem with a CAT software, but it's not necessary to use CAT. Any word processing program can be used with the SRE – Word, PowerPoint, Facebook, etc. – which makes this skill a pretty good one to have for even everyday activities.

Adding a CAT software to the output from the SRE combines the powerful formatting tools of the translation system with the output from the SRE. For example, think of all the generic words in your steno dictionary that are also proper nouns. With your CAT, you can create a “conflict” (or multiple entry) so that the artificial intelligence feature will pick which choice to make. A voice writer can say “street,” and the SRE will send “street” to the CAT which then applies its AI rules and picks the correct choice.

One difference between voice and steno is where the errors occur. In steno, the longer the word (the more stroke intensive) the more opportunity for it to translate incorrectly. It's just the opposite for voice; the longer the word (more phonemes), the fewer “matches” the engine will find and the more likely it will output correctly. However, smaller words with only one sound (we're/were/where) can be much more difficult for the engine to translate correctly, even with its own embedded AI. By creating a multiple entry for the homophones – as long as they’re different parts of speech – your CAT system’s AI will correctly discern the appropriate choice and apply it.

You can also use an open microphone for remote dictation for captioning and off-site CART provision. Dragon was actually programmed for use with an open microphone headset. The ability to hear and understand a person’s speech seems to be more effective using this microphone, which stands to reason. Voice writers make their microphone work harder because we don't want those outside the microphone to be able to hear us.

Apart from the actual input methods, voice writing is the same as steno reporting. I still swear a witness. I still uphold ethical standards. I have to maintain my certifications by attending continuing education. I produce transcripts the same way whether I write in steno or voice. In the end, I’ve saved my arms and can continue to work in the profession which I love and which has enriched my life.

Tori Pittman, BA, FAPR, RDR, CRI, CVR-CM-M, RCP

[Debbie Mayer]

Tori, I finished my CEUs last year on your site stenographersworld.com.  Fantastic site for CEUs procrastinators!