Posts: 3
Threads: 2
Joined: Jun 2025
Yesterday, 04:25 AM
(This post was last modified: Yesterday, 04:26 AM by Terry.)
I'm using an iPad Pro and the facial gestures don't also capture, and page turns can be slow. Any suggestions? I had no issue with forScore gestures.
Posts: 14,418
Threads: 302
Joined: Apr 2012
Yesterday, 04:59 AM
(This post was last modified: Yesterday, 05:07 AM by Zubersoft.)
Go to Settings->Face Gesture Settings and make sure "Use ARKit for Gesture Detection" is turned on. That will use the most battery life due to the processing required, but it will also be the most accurate. I'm guessing that forScore is using ARKit in the same way to do the facial gesture detection, so you should get the exact same result in MobileSheets. Page turns for me are instant when using that on my iPad Pro, so I'm not sure why you are saying the page turns are slow.
I suppose you could also try turning the "Use ARKit for Gesture Detection" off to use the other approach which requires less processing power to see if that works better on your device.
Mike
Posts: 1,114
Threads: 114
Joined: Dec 2015
I hijack the thread a bit (sorry Terry).
Mike, is it possible to adjust the area where the camera scans the facial gestures.
To explain what I mean: I place my iPad at different heights, depending on what I'm playing (a grand, an upright or a keyboard). And of course the reactions to the gestures differ very much for both intentional and unintentional gestures if the height differs (I'm using only 2 gestures BTW, head up and open mouth). Is there something possible to adjust so one can always use the gestures in the same way one is used too. This is probaby a keyboarder/pianist only problrm. but anyway.
Posts: 14,418
Threads: 302
Joined: Apr 2012
Yesterday, 06:39 PM
(This post was last modified: Yesterday, 06:42 PM by Zubersoft.)
No, it's not possible to adjust that through the libraries that are being used unfortunately, at least not easily. Having said that, I'm just feeding images from the camera into the library at a very fast rate, so in theory, I could always perform processing on the image before it's fed into the library, such as cropping the image to only a certain area of the video feed. I don't have a UI set up to drive any of that, or any of the code that would be needed to save that information, or code that would have to deal with the device orientation changing and all sorts of other potential problems if I'm mucking with the image that is captured.
Mike
Posts: 1,114
Threads: 114
Joined: Dec 2015
Yeah, I kind of expected that. And since it's a singular and specific keyboarder problem and as long no others speak up I don't expect you to dive into that any further. I'll cope.
