Ozeki 10
Speech Engines Speech Engines | Text to Speech engine Text to Speech engine

Speech to Text engine

You can add the Microsoft's speech engine as a connection. This makes you able to control other Ozeki 10 connections with human voice. You can always add new runtime languages to the Microsoft Speech Platform. You must create a dictionary of detectable words to use this connection, since it will only recognised the words you have included to the dictionary.

You can fill the dictionary with pre-defined word sets (e.g. colors, directions or numbers) as well.

Figure 1 - How speech to text works by using Microsoft's speech engine

Keep in mind that you can connect your microphone to either your Ozeki 10 server or browser client, which can be NOT ONLY a cell phone as you see it on Figure 1, but many different devices like laptops, PC -s or tablets. The only requirement is that the machine with a microphone must have an internet connection with a webbrowser.

Follow these steps

STEP 1: Start your Robot Controller application
STEP 2: Create a Speech to Text connection
STEP 3: Add a word set to the dictionary


  • A microphone to detect voice
  • Ozeki 10 installed on your computer
  • Microsoft Speech Platform runtime installed

STEP 1: Start your Robot Controller application

Start the service and login through the browser GUI then run your Robot Controller application (Figure 2).

Figure 2 – Start 'Ozeki Robot Controller' application

STEP 2: Create a Speech to Text connection

Now it is time to modify the connection list.

Click the 'Connect' icon on the toolbar then click the big blue 'Create new Connection' button. Here you can choose what type of connection you would like to add. Select 'Audio' (Figure 3). It contains microphone and speaker related connections, which can work alone or through speech engines.

Figure 3 – Select 'Audio' connections after clicking 'Create new Connection'

Then select 'Speech to Text' (Figure 4). Thanks to this, you will be able to choose which words to recognize with the microphone connected to your Ozeki 10 machine. It will convert the voice into text message on a selected language.

Figure 4 – Select 'Speech to Text'

Choose from two type of speech to text options, which can be a 'Browser' or 'Server' connection. Select 'Server' (Figure 5) since it connects to the Microsoft Speech Platform. You can choose from languages installed on the platform.

Figure 5 – Select 'Server'

Finally, name it, select your microphone and set Speech recognizer to English. Then set the accuracy level of the speech recognition (Figure 6). Let's say set to 80 percentage.

If you cannot choose English option, you should install the English language pack to your system.

Figure 6 – Specify connection parameters

STEP 3: Add a word set to the dictionary

There are basic wordsets integrated to Ozeki 10, like colors, directions or numbers. In this example you will see how to add a wordset containing colors. Please navigate to 'Details' next to the speech engine connection (Figure 7).

Figure 7 – Click 'Details' in the connection list

Select 'Detect words' tabpage then click on 'Create new Detectable word' (Figure 8).

Figure 8 – Detect words tab

By clicking on 'Word set' you can use our color set (Figure 9), instead of adding the colors by yourself.

Figure 9 – Adding a new wordset (1/3)

Select 'Colors' from the list of wordsets (Figure 10).

Figure 10 – Adding a new wordset (2/3)

Choose any name then click 'OK' (Figure 11).

Figure 11 – Adding a new wordset (3/3)

Now Ozeki 10 can detect colors with the help of a microphone. Do not forget to set your Speech to Text engine's state to 'Connected' (Figure 12). You have finally finished the configuration. Let’s code!

Figure 12 – Details of the Speech to Text connection

Copyright © 2000- - Ozeki Ltd | info@ozeki.hu
Home > Products > Ozeki 10 > Connections > Speech Engines > Speech to Text engine
Legal | Privacy policy | Terms of use
Page: 5832 | Login |