Voice Editing "Voice Recognition" Support Page
Operation test of the voice-recognition function
You can confirm if the voice-recognition function operates properly by following the operation procedures below:
With Ver. 1.0 Premium Edition/Ver. 2.0 Premium Edition, the voice-recognition samples are installed at the same time.
Procedure for Voice Recognition of stereo-file recording using an IC Recorder (RR-US500/065/050)
When using the voice-recognition function with Voice Editing Ver. 1.0 Premium Edition/Professional Edition/Ver. 2.0 Premium Edition, set the IC recorder (RR-US500/065/050) to "Dictation." Voice recognition does not function properly with voice files recorded with the "Meeting" setting selected (especially when stereo mode is selected). Users with a version earlier than Ver. 1.04 Professional Edition should update to Ver. 1.04.
How to use the voice-recognition function (For details, refer to the "Voice Recognition" page in the supplied PDF-format instruction manual.)
Before converting what is spoken to text, the person whose voice you wish the software to recognize must do the necessary training of the software to register his or her voice.
The Ver. 1.0 Premium Edition and Ver. 2.0 Premium Edition allow you to choose your own text for training. You can improve the voice-recognition accuracy for a particular kind of material that you wish to convert to text if you train the program using that material.
You can immediately convert what is spoken into the IC recorder (RR-US490/470/500/065/050)*1 or a commercially available microphone into text.
*1 Use the "Dictation Cord" supplied with the IC recorder. Depending on your computer, you may be able to improve the sensitivity by swapping the Straight mini plug and the L-shaped mini plug connections.
- Click here for an animated image of the procedure (in the case of Ver. 1.0 Professional Edition). 326 KB (334,336 bytes)
You can convert a previously recorded voice file to text.
- Click here for an animated image of the procedure (in the case of Ver. 1.0 Professional Edition) 744 KB (762,368 bytes)
Use the same procedure with Ver. 1.0 Premium Edition.
Ver. 2.0 Premium Edition is equipped with the dictation functions that allow you to convert recorded voice files to text files (dictation key setting, dedicated dictation screen, Dictation A Recitation mode, Dictation B File Replay mode, and Voice Command mode). For details, refer to the PDF-format instruction manual supplied with Voice Editing Ver. 2.0 or the Product Guide page (link to a newly created Web page). Apart from the supplied voice-recognition engine, link-up with Dragon Naturally Speaking (voice-input software available from Nuance Communications, Inc.) is also supported. Data of a voice file (mono HQ mode) recorded with an IC recorder can be converted to text data by using Dragon Naturally Speaking, with easy operations.
For details on Dragon Naturally Speaking, visit the website of Nuance communications, Inc.
If the video of the procedure does not start playing, choose [Save Target As] from the right-click menu and download the file first, then play it.
Interpretation Function (Ver. 1.0 Premium Edition/Ver. 2.0 Premium Edition only)
The Interpretation function works by combining voice recognition, translation, and text-to-speech functions. Words spoken into the IC recorder or microphone are immediately converted to text through voice recognition, and that text is translated into the selected foreign language (Japanese, French, German, Italian, Spanish, Chinese, Korean, or Russian), using the translation engine from Cross Language Inc. The translated text is then read aloud using the text-to-speech function.
Translation (interpretation) results are merely one example of translation (interpretation), and accuracy cannot be guaranteed. We cannot be liable for the Translation (interpretation) results in any way. Additionally, publishing, selling, or distributing of unedited translation results may violate copyright law.
Depending upon the content of the source text, it may not be possible to obtain the desired results in translation (interpretation).
Voice Recognition Engines that Can be Used in Voice Editing
You can use any Microsoft-SAPI5-compatible voice-recognition engine with Voice Editing. You can also use the voice-recognition engine bundled with Office XP/Office 2003.
- For installation of the voice-recognition engine
- For installation of the voice-recognition engine in Microsoft Word
Voice recognition is not possible with a recording from a conference in which a lot of speakers talk or if there is a lot of background noise.
Voice-recognition rates vary among individuals. However, higher recognition accuracy is generally obtained with newspaper articles or reports than with conversational sentences.
Before using the voice-recognition and interpretation functions, the person whose voice you wish the software to recognize or interpret must do the necessary training of the software to register his or her voice. The voices of people who have not performed "Training" will not be properly converted into text or interpreted. When performing conversion, please select a user name for voice recognition.