Question · Typing and writing speed
How do you test your voice typing speed?
Dictate a fixed passage, then fix every error until the text matches the original. Divide the words in the final text by the minutes, counting the time spent fixing errors. Report the uncorrected errors beside the speed, and run the same passage on the keyboard so the comparison is fair.
Updated October 16, 20264 min read7 cited facts, each traced to its source
Dictate one fixed passage, fix every error until the text matches the original, then divide the words in the final text by the minutes you spent, including the fixing. Run the same passage on your keyboard with the same counting rule. Without both steps the result is not comparable, because speech and typing differ most in the errors they leave behind.
We found no standard test for voice typing speed in our search. Vendor tools each count words in their own way, and the method below follows the 2018 study and a patent that describes a dictation test.
A fair test uses one passage and one counting rule
The 2018 study that compared speech and typing on an iPhone reported speed for the finished text, and it counted the time spent on corrections 1. That is the rule to copy. Choose one passage of about 150 words that you can read comfortably, and use it for both methods. If you use different passages, a difference in speed may come from the text rather than from the method.
Count the words the same way for both methods. An accessibility study counted five characters as one word when it reported words per minute 2. That convention is useful for typing, but for a voice test the simplest rule is the number of words in the passage you were asked to produce.
Steps: time the dictation, then fix it until it matches
- Put the passage beside a blank document in the app you will use, on screen or on paper.
- Start the timer when you press the start control. In a patent's sample dictation test, the push-to-talk button starts the clock 3. Choose one start rule and use it every time.
- Dictate the passage at your normal pace, then stop speaking.
- Fix every error until the text matches the passage, then stop the timer.
- Calculate the words per minute: the words in the corrected text, divided by the minutes, including the fixing time.
- Repeat on the keyboard with the same passage and the same rules, then compare the two results.
Record each result in the same shape, so you can compare the two methods side by side.
| Measure | Typing | Speech |
|---|---|---|
| Words in the corrected text | ||
| Minutes, including fixes | ||
| Words per minute | ||
| Uncorrected errors left |
Errors that inflate a voice score
A common mistake is counting words the tool produced rather than words you spoke. The patent notes that a dictation system may accept macros to enhance effective words per minute 4. A short phrase that expands into a longer clause makes the score look faster than the speaking was. Switch off text replacement and shortcuts for the test, and count the passage's words, not the output's.
The second mistake is stopping once the text looks right. The 2018 study found that speech entered text with fewer errors but left slightly more of them in the final version: 1.30 percent uncorrected errors against 0.79 percent for typing 5. A fast result with errors left in is not an equal result, so check the final text against the original before you stop the timer.
Report the errors beside the speed
The patent allows a dictation score to be adjusted for errors, or to be reported as a separate metric 6. Report both numbers: the raw words per minute, and the uncorrected errors in the final text. Then a result of 100 words per minute with three mistakes is not confused with a result of 100 words per minute with none.
For the keyboard side, a short typing test reports speed and accuracy in one place. The 10FastFingers test runs for one minute and reports words per minute, accuracy and the number of wrong keystrokes 7. Use it only as a rough check. A one-minute test measures a different task from a passage you dictate, so it cannot stand in for the comparison.
Limits of a home test
A home test measures one speaker, one microphone, one room and one recognizer. Your result will change with the engine, the microphone distance and the background noise, so run the test in the room where you would really dictate. Treat your result as your own baseline, not as a ranking of the app you used.
For a quick check without installing anything, free dictation in your browser lets you run the same steps. Yap, which we make, is an on-device dictation app for Mac and Windows, and the same steps apply to it. You can download it here and test it on your own computer.
For the bigger picture, read can you type faster than you talk? and how many words per minute you can dictate. The guide on speaking vs typing speed walks through the studies behind these steps, and the entry on words per minute explains the counting rule.
Evidence
Every number on this page, with where it comes from. Open a fact to see the exact words in its source.
- 1.
- 2.
- 3.
- 4.
- 5.Speech produced fewer errors during entry than the phone keyboard (5.30% vs 11.22% corrected error rate) but left slightly more errors in the final text (1.30% vs 0.79% uncorrected error rate).
Sherry Ruan et al., Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT), 2018 · Original
- 6.
- 7.
Go deeper
Related questions
- Can you type faster than you talk?
- How many words per minute can you dictate?
- At what age should kids learn touch typing?
- Can AI help you type faster?
- Can typing patterns detect Parkinson's disease?
- Can you really read 900 words per minute?
- Can you touch type with one hand?
- Can you type fast without touch typing?