voice key problem

SuperLab Support for Earlier SuperLab Versions
1

i’m using superlab 2.03 with a serial response box from psychology software tools. we’re measuring latency of response to auditory stimuli. the problem is that for each subject utterance we get multiple voice key responses - some 10 or 20. lowering the mike sensitivity reduces this, but then misses some of the subject responses. is this a problem with the response box? can the software be programmed to ignore voice key responses for a second or so after the first one?

thanks,
noam.

2

Hi Noam,

This is problem with detecting voice key in general. There is a certain threshold, or mic sensitivity. Every time the participant speaks, i.e. every time the voice drops then rises above that threshold level, it’s as if the participant pressed a key. If they are responding with a sentence instead of a single word, you end up with a lot of “keys” pressed.

SuperLab records ALL responses until the correct response is detected. The solution is to simply have your trial/event:

  • end on the correct response, with the correct response being the microphone, OR
  • have the criterion response be any response
In the past, I've also recommended that users add an extra waiting event to let the participant finish speaking. This can be a fixed duration or you can have the participant press a key to start the next trial.

Hisham.

3

voice key again

i imagined that was how the voice key works. our responses are single CVC words, so it was surprising to get 10-20 detections from a single word. the problem is not the flow of the experiment, since the event we wait for is for the experimenter to press a key (signaling correct/incorrect response). the problem is that the log file has many “false” lines in it that have to be removed manually.

we got a better result with your own company’s voice key box, but that one doesn’t have a regular key, so to use it we have to install superlab 4, which allows the computer keyboard to be used for extra keys. that removes the installation of superlab 2, so we have to reprogram the whole current experiment.

we’re experimenting with your voice key box (sv-1, i think) on another computer, we’ll see how this works. buy i’m still stumped that one word would create so many responses.

thanks,
noam.

4

I’m stumped as well. I haven’t seen the behavior that you’re describing in-house or with other customers. If you are suspecting a defective SV-1 unit, we’ll be happy to replace it for you.