-
Voice Insert Tutorial No.1
Demonstrates urDictation mode for VB 6.0
Remarks
Please understand the example in four parts :
1. Getting the computer info. (See Form Load Event )
2. Setting the computer info. (See Code for all Combo Box Clicks)
3. Actual Start of the Recognition Activity. (See Start, Stop, Pause and Resume Button code)
4. After a word is recognized. (See all Events)
-
Voice Insert Tutorial No.2
Demonstrates urCommands/urPasswords mode for VB 6.0
Remarks
Please understand the example in four parts :
1. Getting the computer info. (See Form Load Event )
2. Setting the computer info. (See Code for all Combo Box Clicks)
3. Actual Start of the Recognition Activity. (See Start, Stop, Pause and Resume Button code)
4. After a command/password is recognized. (See all Events)
-
Voice Insert Tutorial No.3
Demonstrates urLimitedDictation mode for VB 6.0
Remarks
Please understand the example in four parts :
1. Getting the computer info.(See Form Load Event )
2. Setting the computer info.(See Code for all Combo Box Clicks)
3. Actual Start of the Recognition Activity. (See Start, Stop, Pause and Resume Button code) (How VoiceInsert takes the XML file. Also see Pro.XML or First. XML in path for sample scripts.)
4. After a word is recognized.(See all Events) -
Wave To Text Tutorial No.4
Converts wave files test.wav and test1.wav. Also displays the starting time and ending time of a word within a wave. How to calculate ? Divide the "startwaveposition" by the frequency of the wave recording file and you will get the starting time of the recognized word. Similarly divide "endwaveposition" by the frequency of the wave recording and you will get the ending time of the word. The Audio Size information of every word and the Audio Time presented is very important in determining and differentiating two files or even patches within a wave recording. -
WaveDiffPro Example
Before you execute this example, please make sure you have set WaveDiffPro v1.2 on your machine.
The sample aims at identification of a wave prompt in a wave file. As you see, all information about every word is returned to us from a pre-recorded wave file. That is, every word has a starting and ending location we require that for playback or storage purpose. Yes you can store critical word information by storing the length of the word found by the following formulae : EndWav-StartWav. But more, you can also store a word info by applying FFT over the audio patch given as required by WaveDiffPro DLL. The use of the WaveDiffPro DLL and Voice Insert DLL is presented in the example.
Please select vip.wav which says four words "Hello, Today is Saturday!". As you see clicking 'Add single prompts' processes the wave file using SAPI 5.1 and gives us StartWav and EndWav Info. Please input a identification title as require by WaveDiff for every prompt. Lets say HELLO = HELLO1, TODAY= HELLO12, IS = HELLO13 and so on. Now click 'Add all to Voca.fft' to add a FFT info of all four prompts to file Voca.FFT. Before you do this make sure you have deleted a earlier file using "Delete FFT file" button.
In the second step we have to identify the stored prompts from a new recording. Select hpvip.wav and click 'Process new recording'. This again at the start finds for us the required StartWav and EndWav values. Now you can compare the stored prompts for a close match with this processing. Click 'Compare prompts with new recording' to get a result. The best score is the one that is less. That is zero is the best match. The higher the score, lower the match.
Thus you see Voice-insert and WaveDiff can be used in Multithreaded Environments to Identify Prompts of any Frequency, even non-speech range as some times require in telephony, or security projects
