WatsonSpeechtoTextTool from M2ComSys is implemented with IBM Watson Speech to Text Conversion is an essential tool for Transcribers to improve their productivity. They can increase their productivity at-least 50%
How To Use this Voice Recognition Tool
This tool uses IBM Watson Engine for Voice Recognition. Obtain a license key from IBM Watson web site.
Note: Please enroll for a paid plan since it requires a paid plan to use acoustic and language train option. Follow below given URL to create an account
Register with your watson account details in this tool.
Each dictator has to reserve his identity in Watson by means of profile. This tool provides UI to create your own profile.
Once the profile is created, train your Watson profile with this tool.
For Better text quality, IBM Watson demands two types of training – Language and Acoustic. Language Training needs to be completed first. Upload as many available text (speech converted) files of the dictator for language training. After successful completion of language training, we can move forward to acoustic training. Acoustic training requires a minimum of 10 minutes of audio and maximum 50 hours of audio. This is a time consuming process.
After successful training completion, one can directly use it for transcription (Speech to Text conversion).This will give you the out of the box accuracy of IBM engine.
On VR Completion, the transcript text from watson can be download as document from this tool and can be editted using the provided text editor. The inbuilt media player in the editor UI makes editing easy. Editor UI displays text in different color code based on the IBM Watson returned confidence factor. A legend is provided in the tool to show the IBM Watson returned confidence factor. It also provides a legend showing the short cut keys to control the audio playback.
You can finalize the transcript after editing and RTF file is returned.