The issue is that the implementation of the orientation angles in 1.0 isn't actually pitch and roll, it's pitch around the x axis and pitch around the y axis.
This was deeply improved for cupcake. Yes, we've maintained compatibility with the old API too. JBQ On Mon, Jan 26, 2009 at 1:10 PM, blindfold <seeingwithso...@gmail.com> wrote: > > I'm similarly getting more and more confused with the orientation > sensor readings. With my dev phone 1 running in landscape mode, and > the phone standing on its left side, values[1] = 0 and values[2] = > -90. So far so good. Then when I lift the right side of the phone such > that the screen rotates counterclockwise (leaning just a few degrees > forward to stay away from any +/-180 jumps etc), values[1] - values[2] > remains constant at 90, while I am only rotating around the axis > perpendicular to the screen. Apparently Android sees the rotation > around the (Z) axis perpendicular to the screen as a rotation around > *both* X and Y axis. Or if I assume that the axes form an external > coordinate system with the Z axis pointing upward out of my table top, > then I would expect a values[1] azimuth dependency, but when azimuth > is kept at 0 (or any other heading for that matter) I still see the > same values[1] - values[2] = 90 effect. I do not understand this > coupling between values[1] and values[2]. How are the X and Y axis > defined to cause this strange dependency between roll and pitch? I > expected independent axes but it looks like X, Y and Z are not > orthogonal? > > As a further experiment, I first carefully put my phone in orientation > values[] = {45,45,-45} by first putting the phone standing on its left > side and then tilting the right side to rotate the screen 45 degrees > counterclockwise as described before, now also adjusting the compass > heading by rotating around the axis pointing up from my table top to > make values[0] = 45 as well. Then I found that if I very carefully > rotated the phone around its long screen axis "backward", I can add a > significant tilt without changing the {45,45,-45} orientation vector > reading much. > > Anyone here who really understands the G1 orientation sensor's > coordinate system? > > Thanks > > On Dec 24 2008, 1:20 am, Nick <nicholasdo...@gmail.com> wrote: >> Hello, >> >> If I have a pitch of 0 or 180 I can roll between -90 and 90 as >> expected. However as the pitch approaches 90 and -90 the roll range >> decreases proportionally. Should roll be independent of pitch? Also, I >> expected roll to have a range of -180 to 180. >> >> Any help is appreciated. >> Nick > > > -- Jean-Baptiste M. "JBQ" Queru Android Engineer, Google. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en -~----------~----~----~----~------~----~------~--~---