Millisecond Forums

Could you

(1) Provide the Inquisit code that gave you those measurements / negative coordinates?
(2) Could you elaborate what exactly " need to correct for the various monitor sizes used" means in your case? What are you trying to correct for? The image size in pixels or millimeters? Its position?
(3) Are you making use of /canvassize, /canvasaspectratio and/or /canvasposition in your script?
(4) What are the return values of display.height, display.width, display.canvasheight, and display.canvaswidth?
(5) Are multiple monitors attached to the system? If so, what are their respective dimensions and which is the primary monitor?

Thanks.

1) I have added the script file. This is a change blindness experiment from the library that used to collect typed responses.  We have modified it so that the participant clicks on the part of the image that has changed.
2) To date, we have been running the study on a standard computer on the lab (monitor of 1920 x 1080).  We defined a set of correct regions (x,y,radius) by clicking on the part that was changed and then we can compare the participants response against these predefined regions.  This has worked fine.  The image itself is displayed at 70% monitor height (with width adjusting proportionally).  We assumed that when we try on different resolutions that it would be a simple rescaling (a multiplier) of the correct coordinates, but it soon became apparent that something was wrong.  When we clicked on the 4 corners of the monitor to see what coordinates were returned we were very surprised that the upper left corner returned an x-value of -240 instead of 0. 
3) As we imported it from someone else, we didn't think to check (no idea about this).  I now see that the script says "/canvasaspectratio = (4,3)" so I am guessing this is the origin of all our woes!

Jamie

/ canvasaspectratio = (4,3) restricts the screen area Inquisit draws on to a 4:3 screen region, i.e., that area is smaller than the total (horizontal) surface your widescreen monitor (16:9 ratio) provides. The idea behind /canvasaspectratio is to facilitate consistent results in terms of (relative) stimulus sizing and spacing across varying display ratios. Now, when you click *outside* of the designated 4:3 area, the returned coordinates will (and should be) negative.

To illustrate, run

<defaults>
/ canvasaspectratio = (4,3)
/ canvasposition = (50%, 50%)
/ screencolor = white
</defaults>

<values>
/ x = -1
/ y = -1
</values>

<block myblock>
/ trials = [1-4=mytrial]
</block>

<trial mytrial>
/ ontrialend = [values.x = trial.mytrial.responsex; values.y = trial.mytrial.responsey]
/ stimulusframes = [1=myshape]
/ inputdevice = mouse
/ validresponse = (lbuttondown)
</trial>

<shape myshape>
/ shape = rectangle
/ color = black
/ size = (100%, 100%)
</shape>

<data>
/ columns = [trialnum trialcode values.x values.y]
/ separatefiles = true
</data>

The black area is the designated 4:3 section. If you click the 4 "corners" of the screen in the *white* area, you'll see negative x-coordinates.

Hope this clarifies.