More Tips & Tricks for VoiceTranslator & MouseReader
- It is possible to train the Windows speech recognition engine for the voice of a single particular speaker, for example of a news tv program or videocast presenter. In order to train the engine, you will need a clear speech source (without music or excesive noise in the background) and a text file with its transcription, the longer the better. Then use VoiceTranslator button "Train with text file" to load the text file into the training session dialogs. Remember that you should select the "Stereo Mix" (or whatever it is called in your computer) as the source for speech input.
- Speech recognition can serve as a measurement of "how good" a foreign speaker pronunciation is. If you do not understand almost anything, sometimes it is not entirelly your fault.
- Up to now it is not allowed to use scripts or automatic programs to upload/download translations to/from this web, since obviously that could interfere with its proper functioning. The dictionaries will be available for everyone and once filtered and processed you'll can download them in the form of simple text files, that you can use in tandem with the VoiceTranslator program. If we detect the use of automatic scripts we will ban your computer/organization IP.
If you have a free dictionary file
that whose definitions you want to upload, the program VoiceTranslator has the option of reading a text file with translation pairs (the first line has a text, the second line its translation, the third line another text, then its translation...and so on) and upload them to the database (Button "^"). You should check each translation before uploading it in order to avoid mistakes.
- The option of translating from one language to itself has been left enabled intentionally because in that way you can add translations between different varieties, dialects or slang of the same language, and also create a synonyms dictionary.