Thanks for the replies.
I attach a simple example here. It has 4096 points and I try to catch every peaks with the cursor. You may limit the x axis to a small range (say, 1-300), and you can get the first peak. But for the whole range (0-4095), I can not get the first peak point. I attach a picture to show the closest point I can catch.
I am also thinking about that it should be possible to find that peak point with the cursor if that point can be displayed. Maybe the mouse is the problem? Any way to change the configuration of the mouse for higher resolution (pixel by pixel)? I looked through the mouse setup in control panel but find nothing related to this.
Well done ! You are right, I can't catch the peak highest points using the mouse ! That's something new for me.
However, you can still reach the points using the cursor control button (which was offscreen on your vi window).
But this doesn't explain why some points can't be reached. Seems that there is a flaw (not to say a bug..) in the cursor positionning algorithm.
Yes, zoom in can help the cursor to reach every point.
Just curious whether there is anything limit the resolution of the cursor?
"...whether there is anything limit the resolution of the cursor?"
I would say the resolution of the display.
Lets say you had 5000 points plotted and the graph was 500 pixels wide. Then every pixel would have to represent 10 data points. In that case, which of the 10 data points should the cursor snap to?
Just my thoughts...
The problem is that that peak is displayed in the full range graph. There should be a pixel corresponding to this point, right? Why can't we reach this pixel?
My guess is that LabVIEW does some compression for the image when the resolution is not enough to show all the data points. Some algorithm will skip some data points and still make the whole image look roughly the same as the non-compressed one. It causes the trouble to match every point on the image.
If this is the reason, I am afraid there is no other solution except "zoom in". 😞
I like your guess!
It is sorta like pointing at a car lot and saying "I want that car." Which car?
It is consitent with my (old and failing) memory of something I think Greg McKaskle wrote years ago. I seem to remember him saying that in situation were there was more data than pixels, LV would attempt to render the details as best it could. An example would be if (in my previous example) there was a spike in the reading. This would be represented by a verticle line, all at the same horizontal pixel location.