![]() |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Did you add sound support for the n900?
Because in my game i can not heard anything and sound support still is on your todo list |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Brian sorry i made a mistake,
but sound event works on Microemu-demo :) by the way thx again |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
@DavyP: There are 2 things you can try to increase performance. Search in Qt documentation for their effect.
First one is Fremantle specific, In your MainWindow constructor (in the same #ifdef..#endif block where WA_Maemo5AutoOrientation lives, place: Code:
setAttribute(Qt::WA_Maemo5NonComposited); Second tweak should work in both Fremantle and Harmattan, in your MainWindow constructor place: Code:
setAttribute(Qt::WA_OpaquePaintEvent, true); BTW I found a bug in your build using WA_Maemo5AutoOrientation: If you enable "Enable portrait mode" in MIDlet Settings, then window appears rotated 90 degrees (shows landscape when in portrait and vice versa) Hope the above will help you a bit, thanks a lot for your work, keep it that way :) |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Quote:
messages. I hope the next steps to play custom samples will be straightforward. If I also find a nice way to play tones of a particular frequency and length, then I hope I will have most sound playback support covered. Unfortunately, MeeGo does not play ball (as the audio-playback does not seem to work on the N9). Davy |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Quote:
About the bug: The "Enable portrait mode" feature is only meant for devices on which phoneME always starts in landscape mode. So if auto-rotation works, you should not enable this feature. With this feature enabled, I actually rotate the midlet 90 degrees. That is why you are getting the landscape version in portrait mode and vice versa if auto-rotation works. Perhaps I should call the feature "Rotate midlet 90 degrees". That description would be more accurate. Davy |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
I added the above two tricks and used both JBenchmark and
JBenchmark2 to measure the performance improvement, and I get overall scores that are up to 8% higher than without the two tricks. Not world shocking, but a nice improvement nonetheless. Davy |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Quote:
EDIT: non-compositing should have bigger effect in fullscreen |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Quote:
for every two lines I add, I am willing to add even more lines like that :-) BTW, I ran the benchmark in full screen. Perhaps to low improvement is due to the fact I emit a Qt4 signal to trigger for every midlet window update, and perhaps Qt4 is smart enough to coalesce repaint events. Also note that the JBenchmarks probably measure more than just the rendering speed. Davy |
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Quote:
|
Re: [TESTING] PhoneME Advanced (Java Mobile) prototype release
Quote:
clipping ranges, but they are usually much bigger than the bounding box of the pixels that have changed that it usually does not make much of a difference when painting the full midlet window or just the area within the clipping ranges. Especially midlets that paint their own UI on a Canvas or GameCanvas (like Opera Mini or J2ME Polish/LWUIT midlets), this definitely does not make much of a difference. This was my experience on Android, I have not tested this on the N900. So, for the time being, I think the rendering speed is good enough so I would rather focus on some more interesting features. Davy |
All times are GMT. The time now is 07:13. |
vBulletin® Version 3.8.8