To collect samples of a user, there are three training types:
During repeating, the facilitator reads out a word that the user tries to repeat.
The user reads the text from the screen. In practice the facilitator can read it out for the user to repeat as well.
During the interview the facilitator asks the user a couple of questions and record the answers.
Before you can use SSC you will likely want to adjust the training prompts for the recordings.
They are three, one for each training type, simple text files containing one line per prompt. The encoding is UTF-8.
Their location depends on your operation system:
Table 2.1. Training data files
The path for GNU/Linux uses an inline command to resolve the installation folder for you. You can also run kde4-config --install data in a shell to find out the exact path. On Microsoft Windows you have to fill in the installation folder for yourself. Usually this will be something like
C:\Program files\simon 0.3\ or similar.
texts folder will contain three subfolders containing the prompts:
To launch a training session, simply press the appropriate button for your training type in the SSC main window.
To do this you must either be connected to the server or use the offline mode.
When starting the training, you will have the option to use the power training mode.
In power training mode samples will start to be recording as soon as the associated prompt is displayed. You can then proceed to the next page simply by pressing in the wizard. When you do that the current recording will be stored, written and the next sample will start recording.
Using the power training option an averagely proficient speaker can complete the training process much quicker. However, because you don't have a break between individual prompts it might be more taxing for the speaker.
Of course the power training is not an option when letting the speaker repeat what the facilitator says as the latter will invariably be heard on the created recordings.
In this page of the wizard you can provide information about the used hardware devices (soundcard and microphone).
Please enter the model and type of the sound card in the Model and Type fields. Enter the information about the microphone in the two input boxes below.
The drop down boxes are filled with the current list of makes and models from already recorded samples. Before entering a new device please make sure that the same device, maybe named a little bit differently, doesn't already exist to avoid duplicates.
Before the recording starts, SSC will try to make sure that the microphone volume(s) are correctly configured.
To calibrate simply read the text displayed.
The calibration will monitor the current volume and tell you to either raise or lower the volume but you have to do that manually in your systems audio mixer. Once you changed the volume in any way (while the calibration is running), press the button next to the affected device. This will reset the volume status.
During calibration, try to talk normally. Don't yell but don't be overly quiet either. Take into account that you should generally use the same volume setting for all your training and for the recognition too. You might speak a little bit louder (unconsciously) when you are upset or at another time of the day so try to raise your voice a little bit to anticipate this. It is much better to have a little quieter samples than to start clipping.
In the SSC settings, both the text displayed and the levels considered correct can be changed. If you leave the text empty, the default text will be displayed. In the options you can also deactivate the calibration completely. See the configuration section for more details.
For each line in your prompts file you will get one page asking you to record a sample.
After you recorded the prompts (you may skip prompts at any time) you can upload the collected samples to the server or store them locally to send them later on.
To upload stored samples please refer to the next section.