Hi all,
I'm trying to adapt the Automated OSPAN task of the Inquisit library to a verbal response mode, and I've been only partially successful. My main concern is whether the speech recognition engine is able to recognise letters.
Here are the different steps I followed:
First, changing the response mode to voice input:
<defaults>
/ canvassize = (100%,100%)
/ canvasaspectratio = (4,3)
/ screencolor = (white)
/ txcolor = (black)
/ txbgcolor = (transparent)
/ fontstyle = ("Verdana", 4.0%, true)
/ inputdevice = voice
/ minimumversion = "4.0.0.0"
/ halign = center
/ valign = center
</defaults>
Then, changing the way to progress after each instruction screen from "left mouse button" to the vocal command "Next"
<trial instructions>
/ pretrialpause = 250
/ posttrialpause = 250
/ stimulustimes = [1=instructions]
/ validresponse = ("next")
/ responsetime = 2000
/ recorddata = false
</trial>
So far, everything is working, I can go through the 3 first instructions screens saying "Next", meaning that the voice input is correctly recognised by Inquisit.
The problems starts when I start the practice, and I need to name the letters. Then Inquisit don't seem to recognise them, nothing happens, and even the words "Blank", "Clear" and "Exit" seem not to be recognised:
<trial recall_letter>
/ ontrialbegin = [values.recallcount+=1]
/ ontrialbegin = [if(values.recallcount==1)values.recalldelay=500 else values.recalldelay=0]
/ pretrialpause = values.recalldelay
/ stimulusframes = [1=WhiteScreen, F, H, J, K, L, N, P, Q, R, S, T, Y, _, clear, exit, recalledletters, recallprompt, letterstrings]
/ validresponse = (F, H, J, K, L, N, P, Q, R, S, T, Y, blank, clear, exit)
/ monkeyresponse = ("F", "H", "J", "K", "exit")
/ ontrialend = [if(trial.recall_letter.response!="exit" && trial.recall_letter.response!="clear")
{item.RecalledLetters.item=trial.recall_letter.response; values.recalledletters=concat(values.recalledletters, trial.recall_letter.response)}]
/ ontrialend = [if(trial.recall_letter.response=="clear")
{clear(item.RecalledLetters); values.recalledletters=""}]
/ responsemessage = (F, clickF, 150)
/ responsemessage = (H, clickH, 150)
/ responsemessage = (J, clickJ, 150)
/ responsemessage = (K, clickK, 150)
/ responsemessage = (L, clickL, 150)
/ responsemessage = (N, clickN, 150)
/ responsemessage = (P, clickP, 150)
/ responsemessage = (Q, clickQ, 150)
/ responsemessage = (R, clickR, 150)
/ responsemessage = (S, clickS, 150)
/ responsemessage = (T, clickT, 150)
/ responsemessage = (Y, clickY, 150)
/ responsemessage = (clear, clickclear, 150)
/ responsemessage = (exit, clickexit, 150)
/ responsemessage = (_, click_, 150)
/ branch = [if(trial.recall_letter.response=="exit")trial.letter_feedback else trial.recall_letter]
/ trialdata = [recalledletters]
/ recorddata = true
</trial>
Any ideas how I could make Inquisit to recognise the letters when a person say them? I even tried to write them "phonetically", like I thought the speech recognition engine would recognise them, but nothing changed.
Many thanks in advance for your help,
Sylvain