Demo of Digital Compass Used in an iPhone App

[DELLsFan]

Oh sweet ANOTHER useless article about the very small upgrade to the iphone. Can we post another 50 threads of "hey look how much faster you can plug in the charger cable then the regular 3g" macrumors?

For real! I love my iPhone ... but damn - I don't think the new 13" MBP has got this much buzz!

 

[-wiseguy-]

this hurts my head haha

 

[iNassar]

seems a nice app for a good price ... perfect for showing the new iphone´s performance

 

[smileyborg]

No, no... I created an account just to reply...

First of all, you are mixing applications (Coverflow) with hardware elements (Accelerometer, Magnetometer).

You're right that the accelerometer covers two dimensions but the magnetometer covers just one. They complement each other to make three.

The accelerometer measures the phone's orientation with respect to gravity. So it can measure pitch and roll, because as you pitch and roll the phone one end is going to be higher than the other end and the accelerometer can detect the slope.

When you yaw the phone its orientation does NOT change with respect to gravity. It remains flat all around. So for 12 months Apple has told developers, "we can give you the location, pitch and roll of the phone, but not the yaw."

The magnetometer fills in the yaw by using the magnetic field of the Earth rather than the gravitational field. Now developers have everything about position and orientation. (Except a more precise GPS antenna...)

This app uses both the accelerometer and the magnetometer to show how pitch, yaw and roll all work in unison, and I think it looks great.

That's actually not quite correct...the magnetometer works in 3 axes...in other words, it locates magnetic north as an absolute vector in 3D...so saying you point your iPhone's screen upwards with the headphone jack pointing directly to the magnetic north pole of the Earth. Then ANY movement from this position will be measured by the magnetometer -- not the movement itself, but the resulting end position. The "Compass" app is only a 2D compass, but it could theoretically be a 3D one with a 3D arrow that points to the Earth's magnetic pole (sometimes this would be up on the screen, sometimes to the left, etc...but also sometimes IN TO and OUT OF the screen.)

The accelerometer is actually capable of detecting acceleration on 3 axes as well, the limitation is that it is designed to be sensitive to acceleration on a scale relative to the magnitude of gravity. So it DOES sense a yaw-motion as we have discussed, but that "yaw" is so negligible in magnitude compared to the pull of Earth's gravity that it doesn't do very well with the interpretation of the motion. To illustrate further, imagine holding your iPhone flat in front of you with the screen upwards. A tilt left or right, or a tilt up or down changes the phone's orientation relative to gravity...making a easily measurable change for the accelerometer. But if you just spin the phone keeping the screen up ("yaw"), the phone has not changed in any way relative to gravity (ie the phone remains in the same z-plane). This motion is not easily detected by the accelerometer, as the acceleration you create by spinning it is far less than the pull of gravity on the phone. But now, imagine the phone being held in front of you with the screen facing you, as you hold it when steering a car in a driving game app. Now, that same "yaw" motion (turning the iPhone left and right) DOES change the phone's position relative to gravity, and a tilt forwards or backwards also does...so these motions are registered by the accelerometer. However, now you have lost a different axis -- the twisting of the device (moving either the headphone port or dock port closer or farther).

 

[randfee2]

i really wonder how accurate this magnetometer is, wanna see an app with a direct readout of 3 axis fieldstrength, can any of the developers just do that for free... I guess it'd be a 5 minute job with the magnetometers api.

 

[oticon6]

I honestly don't see the compass being used here in a classical way, since when the roller coaster does a 360 and the viewing angle would be fixed to one geographic direction (which would be a fixed one), the view should spin 360° as well, always looking the same way.

Sorry if someone has already mentioned this... but...

That would make no sense. If you're sitting in a rollercoaster, and it turns a corner, your head turns with it. If your head is tilted 90 degrees to the left of when you started, you're looking out the left side of the cart.

 

[iPhoneNYC]

I think the compass will be very useful when someone figures out what to do with it. I don't hink this is the answer yet..