EasyCE 1.6 release

[Phantom]

[LLcooL]

I see you decided not to include the palette device (HP525) in this release. Will you include it in future releases?<br><br>LLcooL Last modification: LLcooL - 12/18/01 at 08:59:19

[Zensunni]

I don't know if it was intended, but you are still printing that you detected a iPAQ3800 during each updatedisplay();

[Phantom]

Ehm oops, yep, there's a debugging message in the 1.6 release. Well I'll leave it as an excercise to the coders to get rid of it. <br><br>About the palette code: It's a bit risky to include it; I can't test it... Anyway, if I find some time tomorrow I'll add it. That will be a 'beta feature', I'll wait for your comments before I call it 'stable'.Last modification: Phantom - 12/18/01 at 11:56:48

[Phantom]

New bug. Sigh.

[BadBazza]

Phantom,<br><br>Is that in the palette code or some where else?<br><br>Cheers<br>Bad

[BigCanoe]

Can this be run in the PocketPC emulator?

[Phantom]

The 'new bug' refers to a bug in the message board system of PocketMatrix, not to a bug in EasyCE. Sorry for the confusion.<br><br>About using EasyCE on the emulator: I have no idea, I never tried it. When I don't want to test on a PocketPC, I test on the PC (native) by compiling the project using Visual C++. This gives me all the advantages of native debugging, and it works great. This fails only when I encounter data alignment errors, but I suspect that the emulator doesn't crash on those either, so I don't see any reason at all to use the emulator.

[LLcooL]

New bug, will be else where as the palette code as not been released.<br><br>To test the palette code, Lemmings could be modified and release as a binary only to HP525 beta testers, the a demo beta version of the Nutcracker.

[Dan East]

Here's a suggestion for the next release. I'm doing some experimenting, and noticed display problems due to my drawing the iPaq screen from the top down (portrait-wise). The iPaq does it's hardware drawing from left to right. I checked the EasyCE source and see you do the same thing I've been doing - always drawing relative to your backbuffer from top to bottom. Of course on the Casio that coincides with the hardware layout, but on the iPaq you really should draw exactly as the GAPI properties dictate, and jump around in your backbuffer appropriately. You can really see the effect of the "scan lines" during smooth scrolling if you are drawing perpendicular to them.<br><br>Dan East

[Phantom]

LLCool,<br><br>I just looked through your 256 color stuff, and I'm afraid this approach is not going to work very well. First of all, you included a huge table (65536 elements). This is quite logical, since calculating this table will take about 1-2 secs on a PPC, wich is not nice, especially not if this would happen a couple of times at runtime. However, if you would load a tga file and you wanted to change the palette, this 'color cube' would have to be regenerated, wich would thus take 2 secs. You omitted this code, but adding it doesn't seem to make sense, because of the expected delays.<br><br>Right now, I've taken a different approach: I took your code to initialize the palette, and changed it so that it now essentially makes a 2-3-2 hicolor palette. Then, when updating the display, I turn each 16bit color (5-6-5) into an 8-bit color (2-3-2). This color is then plotted. Given the terrible color resolution, this would not look very good, but it would work, and it should work under all circumstances with unchanged code (outside EasyCE). So for the moment, I'll keep this.<br><br>There are two ways to improve the image quality: The first one is dithering; I could use the extra bits to address a dithering table. This would probably lift the image quality to an acceptable level.<br><br>The other way is less attractive: Simulate 3-4-3 or so using the 2-3-2 palette by rapidly changing the color value of each pixel. I did this to simulate 32 greyscales on 2 bit devices, but I don't know how good it would look on color devices. Biggest problem with this is that it requires calling 'update' as often as possible, but at least 50 times a second. That makes this approach unusable for The Nutcracker.<br><br>I'll upload my changes tomorrow; if you can verify that they at least work, then 525 owners can finally play all those cool games that work with EasyCE. It won't be pretty for the moment, but it should get better over time.<br><br>- Jacco.

[GeoffSulcer]

[Phantom]

[Peter Balogh]

[Zensunni]