millisec test timings and screen refresh rate

[Mark Talluto]

On Apr 19, 2007, at 5:19 PM, Phil Davis wrote:

> Some tests require the Rev app to track the millisecs elapsed from  
> when an image is displayed on screen to when a user-initiated event  
> occurs, like a keypress. This means RR needs to know as precisely  
> as possible *when* the image became visible on screen.

It looks like you are after the time between showing the image and  
the user hitting a button.  Are there other factors like the user  
hitting the button before the image is shown and tracking those  
activities as well?

I would then start with showing the image, taking the milliseconds on  
the following line.  Once the user hits the button on a mouseDown,  
you could take the milliseconds again and do the subtraction.

The exact time it takes for the picture to show after you issue the  
command set the vis of img x to true might be statistically  
insignificant.

The following script showed a 5 megapixel image in 30 milliseconds  
regularly on my laptop.  I did the test on an image that was 78x78  
pixels and got an average of 8 milliseconds.

on mouseUp
     put the milliseconds into temp
     set the vis of img 2 to true
     put the milliseconds - temp

     wait 2 seconds
     set the vis of img 2 to false
end mouseUp

I think the big thing is to have the image in memory before starting  
the test as loading and showing all in one swoop would produce the  
most slow down.

The refresh rates on newer LCD monitors is really getting close to  
what CRTs can do.  These are just some of my thoughts on the subject.


Mark Talluto
-- 
CANELA Software
http://www.canelasoftware.com