How Long Does Voice Writing Training Take? A Realistic Timeline
How Long Does Voice Writing Training Take? A Realistic Timeline
Three to six months. That is the figure industry sources quote for voice writing training to working speed, and it is roughly right.
What the figure hides is that months are the wrong unit. The real unit is practice hours, and the real variable is not how fast you can talk. It is how quickly you stop making the same eleven mistakes.
Ninety minutes a day, six days a week, for twenty weeks comes to about 180 hours of drilling. Somebody who does that gets there in five months. Somebody who practices forty minutes on Tuesdays and Saturdays does not get there in five months, and no course can fix that.
Here is what those twenty weeks actually look like.
Weeks 1 to 3: Learning to Talk Into a Mask
The first skill is physical, and it is stranger than people expect.
Voice writing means repeating everything you hear into a padded stenomask connected to speech recognition software. Everything. Every word the witness says, plus the speaker identification, plus every comma, period, question mark, and quotation mark, spoken out loud.
Week one is about producing a steady, level, unremarkable voice. Not loud. Not dramatic. Speech recognition punishes variation. You will also spend real time training a voice profile so the software recognizes you specifically.
Expect to feel absurd. Expect a sore throat. Expect your accuracy at 100 words per minute to be embarrassing.
Weeks 4 to 8: Punctuation Becomes Automatic
This is the phase that separates people who finish from people who quit, and it has nothing to do with speed.
You are building a reflex where hearing a pause produces the spoken word "comma" without any conscious decision. Same for a question, which requires you to hear the shape of the sentence and produce "question mark" before the next speaker starts talking.
Speaker identification is layered on top. In a deposition you tag every switch: examining attorney, witness, defending counsel objecting mid-answer. When two attorneys interrupt each other, you have to keep the tags straight while the words are still coming.
Most beginners at this stage can hit 200 words per minute for thirty seconds and then fall apart at ninety seconds. That is normal. Stamina comes later.
By week eight you should be running clean five-minute takes at 140 to 160 WPM.
Weeks 9 to 14: The Speed Ladder
Now speed becomes the project, and the correct way to climb is boring.
You do not move up in speed until you hold your accuracy target at the current speed. Moving up early feels like progress and produces months of plateau, because you are rehearsing errors at higher volume. The discipline is: hold the tier, then advance.
The three CVR target speeds give you the ladder:
| Leg | Speed | What makes it hard |
|---|---|---|
| Literary | 180 WPM | Continuous prose, no natural breaks, vocabulary you did not prep |
| Jury charge | 200 WPM | Formal legal language, long sentences, dense punctuation |
| Two-voice Q&A | 225 WPM | Fast speaker turns, overlapping speech, constant speaker IDs |
Q&A is the fastest leg and often the one people pass first, because real conversation has short bursts and natural pauses. Literary at 180 has no mercy in it at all.
Weeks 15 to 20: Chasing 95 Percent
The CVR requires 95 percent accuracy. Graders score against 17 defined error types.
At this stage you stop practicing broadly and start practicing surgically. You need to know which of the 17 error types you personally keep committing, because almost nobody makes all seventeen. A typical candidate is losing points on four or five specific things: dropped speaker IDs under crosstalk, homophones like "principal" and "principle," a comma that changes the meaning of a sentence, dropped words on the front end of a fast answer.
Five error types, drilled one at a time, is a solvable problem. "Get more accurate" is not a solvable problem. It is a wish.
This is the phase where most self-taught candidates stall, and the reason is simple: they cannot see their own error distribution. They know their score. They do not know why.
The Exam Itself
The CVR from the National Verbatim Reporters Association has two components. A written knowledge test covering legal terminology, procedure, transcript format, and professional practice. Then the three five-minute dictation legs above, at 95 percent accuracy.
You do not have to pass everything in one sitting under most testing arrangements, and legs can be attempted again. Check current NVRA policy and fees directly, because both change.
Why This Is Not Stenography
The comparison is the whole reason voice writing exists as a career entry point.
| Voice writing | Stenography | |
|---|---|---|
| Time to working speed | 3 to 6 months | 2 to 4 years |
| Reported attrition | Near 10 percent | 85 to 90 percent |
| Core skill | Controlled speech, punctuation reflex | Chorded keyboard theory, then speed |
| Startup cost | Roughly $1,100 to $2,600 | Machine alone often several thousand |
Both attrition figures come from industry reporting rather than a controlled study. Treat them as strong directional signals, not laboratory results.
The reason steno takes years is that a student must first learn an entire chorded writing theory before speed practice can even begin. A voice writer already knows how to speak English. The training is about control and punctuation, not a new symbolic language.
What the Career Pays After
Worth knowing before you commit twenty weeks. The Bureau of Labor Statistics reports a median annual wage of $72,420 for court reporters and simultaneous captioners as of May 2025, with the bottom ten percent under $39,100 and the top ten percent over $127,020. BLS projects roughly 1,700 openings annually from 2024 through 2034 and describes the outlook as little or no change in total employment.
Freelance deposition work is where the upper end concentrates, and it has no salary floor and no benefits. Honest math, both directions.
What Slows People Down
Practicing without measurement. Volume with no error breakdown produces plateaus.
Advancing speed too early. It feels productive and costs a month.
Skipping punctuation drilling because it is dull. It is the highest-scoring category on the exam.
Practicing only clean single-voice audio. Real depositions have crosstalk, accents, and a witness who mumbles.
What To Do Next
Pick a schedule you can actually hold, then measure the right thing. Not your speed. Your error distribution.
VerbatimPrep is practice software for voice writers. It scores your dictation against the same 17 NVRA error types graders use, charts accuracy by error type over weeks so you can see the four mistakes actually holding you back, and runs a speed ladder that unlocks the next WPM tier only after you hold the accuracy threshold at the current one. Start at verbatimprep.com.
Ready to start? Practice voice writing free with real courtroom audio. VerbatimPrep scores your accuracy in real time and coaches you through every mistake. No credit card. No equipment to buy.
Sources: NVRA CVR exam standards; BLS Occupational Outlook Handbook, SOC 27-3092 (May 2025); training duration and attrition figures per industry reporting. VerbatimPrep is practice software, not a school, and makes no employment or income representations.
← All posts