Thanks, I've made much more progress. I'm making Rotation the primary if it
exists and smoothing the Mag-accel via exponential smoothing of the
vectors.
I have seen in some examples where gravity sensor is used as if it were a
backup to the accelerometer. However, I have found no device that h
Hi Nathan, I had posted this earlier in the thread --
https://github.com/ratana/rotation-vector-compass
This is my project where I attempt to get the canonical bearing with
respect to where the user is pointing, which many of the above can get
wrong. It has a few implementations of filters / smo
I have found these examples of using the rotation sensor. Are they correct?
https://github.com/kplatfoot/android-rotation-sensor-sample/blob/master/src/com/kviation/android/sample/orientation/Orientation.java
https://android.googlesource.com/platform/development/+/master/samples/ApiDemos/src/com/e
I have collected more data on the topic that you might be interested in.
Trying to give back.
In five days of Analytics data, I have found
85 devices returning no orientation sensor.
Of these, most do not have a magnetic sensor either, such as the Samsung
SCH-S738C Galaxy Centura. So some peopl
On Tuesday, July 15, 2014 7:53:52 PM UTC-7, Adam Ratana wrote:
>
> Nathan, I would suggest just using a simple low pass filter (in this case
> an EWMA - http://en.m.wikipedia.org/wiki/Exponential_smoothing) and
> tuning the weighting to taste. The more samples usually the smoother you'll
> get
Nathan, I would suggest just using a simple low pass filter (in this case
an EWMA - http://en.m.wikipedia.org/wiki/Exponential_smoothing) and tuning
the weighting to taste. The more samples usually the smoother you'll get. I
use the GAME rate, but I believe you can specify a rate past a certain api
For those of you who have smoothed the jitter.
On Wed, Jul 9, 2014 at 11:55 PM, gjs wrote:
>
> One way to smooth the jitter is to compute an average value from a small
> array of recently stored recent values.
>
For those who have filtered to get a smooth value, what is a good buffer
size and
Hi,
Some devices lack the requisite magnetometer so Orientation / compass does
not work at all, eg: Nikon Coolpix S800c & S810c.
One way to smooth the jitter is to compute an average value from a small
array of recently stored recent values.
Regards
On Thursday, July 10, 2014 3:53:45 AM UTC+
On Monday, March 17, 2014 4:40:21 PM UTC-7, Nathan wrote:
>
> Last time I worked on a compass reading (which was years ago), I did
> something like this.
>
> A:
> http://www.javacodegeeks.com/2013/09/android-compass-code-example.html
>
> Of course, this example has a locked portrait orientation
Hi guys -
Yes, the orientation sensor is definitely broken in some devices. Some
firmware updates by carriers have also broken it in some cases, in the
past, though that's rarer these dayes. I have an app called Sun Surveyor
(+Lite) which used to use the Orientation sensor by default for it
On Thursday, March 20, 2014 4:44:23 AM UTC-7, gjs wrote:
>
> Hi,
>
> Risk - that B won't work as well A and (many?) users will complain. Reward
> - you won't be bothered by the deprecation warnings.
>
Put that way, it's a no brainer. Deprecation warnings only scratch at the
perfectionist in me
Hi,
Risk - that B won't work as well A and (many?) users will complain. Reward
- you won't be bothered by the deprecation warnings.
A little while back I also implemented & tested method B for the same
reasons & I found that method B had more jitter (than method A) & had to be
smoothed somewha
Putting it more concisely. . .
If I have some legacy code implementing a compass reading using the
deprecated Orientation_Sensor, in use by thousands already.
What is the risk/reward of implementing something like the Implementation
here:
http://sunil-android.blogspot.com/2013/02/create-our-and
13 matches
Mail list logo
