How this free, bundled app works and how it compares to other voice recognition software
By Enrique Serrano.
Writing faster is a productive person’s dream. The quicker you can write down your thoughts on your screen, the sooner you can complete your work. This is the main goal of any speech recognition software. Unfortunately, most professional speech-to-text alternatives are expensive and don’t offer free trials. However, if you are a Windows 7 user, then you have a free, quality voice recognition app available in your computer: enter Windows 7 Speech Recognition, available inside the Ease of Access tools menu.
Basic configuration of Windows Speech Recognition
One of the advantages of Speech Recognition in Windows is that its basic setup is extremely easy to complete. Go to Control Panel » Ease of Access » Speech Recognition, click Set up microphone, select the right kind of sound setting, read the test phrase aloud, and you’re ready.
After the basic microphone setup, you need to configure the actual voice recognition features. First, click Start Speech Recognition to initialize the program, and then, select the option Train your computer to better understand you. Reading the additional training set offered is advised – and it takes less than ten minutes. The more you train your computer, the better it will understand you.
Take note that the training is performed for the user currently logged in, with the chosen microphone, at a fixed distance. If someone else wants to use the voice recognition app on the same computer, he or she would need to close the current Windows session, log in with his or her own account, and re-calibrate the system again.
Dictating in Windows
Using Speech Recognition to dictate in Windows is intuitive. After starting speech recognition, right click the app and select Options » Enable dictation scratchpad. The dictated texts will initially appear in a small window where they can be corrected. Once your dictated text is ready, you would just need to say "Insert" to embed it on the final text.
Even when the accuracy of the software is acceptable, it improves over time, and correcting your dictation is an essential step – especially in the beginning. When you are still dictating inside the scratchpad, you would need to say "Correct [word to correct]" to open the correction menu. Replacing the word by the right one from the list is a simple as reading its number aloud, and then saying "OK." If the right word doesn’t appear on the list, you can try speaking the desired word again.
Correcting a word that appears several times on your dictation is easy, as you would be prompted to pick the right repetition of such word: just speak the number of the chosen occurrence of the word and then say "OK."
If you are experiencing problems trying to get the software to recognize a specific word (like a name or a technical term,) you can also say "Spell It" on the correction menu. After spelling unrecognized words, these new words can be added to your personal dictionary. Spelling is quite easy, and if you are familiar with the NATO phonetic alphabet, rest assured that you will be perfectly understood even when speaking fast on noisy environments.
Multilingual features in Enterprise and Ultimate versions
If you want to dictate in different languages, Windows Speech Recognition will do the trick. You just need to open Control Panel and select Region and Language » Keyboards and Languages, and set the desired display language, as this is the language used for speech-to-text functions (or install the desired language pack by clicking Install / uninstall languages.) However, you can only install and switch language packs if you either have Windows 7 Ultimate or Windows 7 Enterprise Edition – if you are a Windows 7 Home Premium user, you will have to stick to the original language of your operating system.
Once a new display language is chosen, you will need to re-open your user session for the changes to take effect. You also need to make sure to train your computer if you use Speech Recognition in a new language for the very first time. Once trained, the results seem to be acceptable: we have tested Windows Speech Recognition in both, English and Spanish, and the accuracy of the results was satisfactory in both languages.
Voice commands beyond dictation
While speech-to-text is one of the strongest points of Windows Speech Recognition, it is actually more than a dictation app. It also allows you to control your computer with your voice. The speech recognition features were designed as an ease-of-access tool, and so, they are capable of opening programs, managing windows, traversing documents, positioning your mouse cursor or expanding menu elements.
To launch applications without navigating a crowded desktop full of icons – just say "Open [application name]" and that’s all. Whisper "Copy" or "Paste" to your computer to avoid releasing your mouse, as there wouldn’t be any need to press key combinations. You can even say "Undo" to undo your last action without moving your hands. However, you may find it is faster to simply use your mouse and familiar keyboard shortcuts.
If you want to experiment with the different voice commands available in Windows Speech Recognition, you can check (and even print) the complete reference card provided with Windows, inside Ease of Access » Speech Recognition » Open the Speech Reference Card.
Windows Speech Recognition vs. Dragon NaturallySpeaking
It makes sense to ask how Windows Speech Recognition compares to other professional software for speech-to-text conversion. The most serious contender is Nuance’s Dragon NaturallySpeaking.
Even though Windows Speech Recognition is free for Windows 7 users, it isn’t an installation headache, and it works quite well, Dragon still beats Windows in several different areas:
- Windows seems to be more sensitive to noise, and the accuracy of Dragon seems to be higher.
- Trying to transcribe pre-recorded voice from a digital recorder didn’t work well in Windows.
- Only a couple of sentences fit into Windows scratchpad (beyond that point, the dictated words would be discarded) which forces you to say "Insert" every couple of lines even when no mistakes were found, and blocking you from correcting previously inserted texts. This makes dictating long chunks in Windows more complicated than in Dragon.
If you are a lawyer, there’s a more complete – and expensive – Legal Version of Dragon with specific advantages for attorneys that are lightyears ahead of what Windows Speech Recognition offers such as a complete pre-loaded set of legal terms, a solid custom commands system and automatic transcription from digital recorders.
At no extra cost for a Windows 7 user, Windows Speech recognition is a useful application to give your hands a short break, to solve an inspiration block by dictating instead of typing, or to try some dictation software to see if it fits nicely in your practice and work environment.
However, even when Windows Speech Recognition is a good start (and we are all looking forward to its evolved version in the upcoming Windows 8,) Dragon NaturallySpeaking is still a more professional alternative. If you like Windows Speech Recognition and find it useful, Dragon NaturallySpeaking is likely to take you to the next level in time-saving speech-to-text processes.
Have you ever used Windows Speech Recognition? Share your experience and your tips with us in the comments.
Also in Lawyer Tech Review: