PocketMatrix

Is it my imagination, or is the stylus on Pocket PC only accurate to within 2 pixels?

That's all the resolution I seem to get. And, looking closely at the screen, I can see little dots about twice as spaced out as the pixels, I assume these are the touch screen sensors.

This lack of accuracy makes it a bit difficult to develop games that rely on the stylus. Already, the screen area is rather small, so if I am concerned about stylus movements within a small screen area of say 32x32 pixels, I really have only 16x16 resolution to work with.

And if I am tracking stylus movement between frames, where it might only move a couple of pixels between frames, its movement will be clamped to either delta-0 or delta-2. That doesn't allow for very fine-grained stylus control, not even at the pixel level.

Are you telling me the average user can pick out each of the 240 horizontal pixels and the 320 veritcal pixels? And can then apply that in a game where the screen is most likely in a state of constant change?

Not likely. If the variance is only 2 pixels, that's pretty darn good in my book.

Out of curiosity, why would you need it to be more precise (1), can you honestly pick out each of the 76,800 pixels on the screen (2) and hit the pixel you want with the stylus in a game where speed matters (3)?

I'm not trying to flame, but it seems to me like you're being outrageously critical of something that works incredibly well.

-John

You're right and you're wrong.

It's precise enough for picking. For example, if I tap the stylus on an object, I can easily figure out which object is intended to be picked.

But it's not very precise if you consider movement. Let me explain.

Let's say I want to move the object I picked under the stylus. As I move the stylus, each frame, I relocate the object under the stylus. Dragging. OK, so 2 pixels is precise enough.

Now let's say that I want to let go of the object, but have it KEEP MOVING in the last direction and speed I was moving it in. At this point, you see the problem. Relative to the last frame, there's only a few velocities the object could be travelling:

(0, 0) stationary
(0, 2) moving down
(-2, 0) moving left
(2, 2) moving diagonal a bit faster
(4, 0) moving right even faster
(4, 2) moving ESE even faster
(4, 4) moving SE even faster

See what I mean? In theory, the object should continue to move in the direction you were dragging it. In practice, I find it "snaps" to the horizontal/vertical or diagonal, because of this imprecision.

BTW, I thought of doing a weighted average of the velocity over the last single frame, the last two frames, the last three frames, etc. Maybe this would alleviate my precision problems, but I haven't tried it yet.