View Issue Details
| ID | Project | Category | Date Submitted | Last Update | |
|---|---|---|---|---|---|
| 0003679 | AI War 1 / Classic | Graphical Bug | Jul 20, 2011 10:15 am | Aug 1, 2011 11:28 am | |
| Reporter | annikk.exe | Assigned To | Chris_McElligottPark | ||
| Status | considering | Resolution | open | ||
| Product Version | 5.011 | ||||
| Summary | 0003679: Parallax Scrolling Bug | ||||
| Description | DESCRIPTION When in System view, if you zoom all the way out and pan the camera around, the background elements such as the planet move incorrectly; they appear as if they are in front of the camera. TO REPRODUCE THE BUG Simply look at the view of any system, zoom all the way out to maximum, then pan the camera around and note how the planet in the background moves compared with any ships or other objects in the system. ADDITIONAL I have worked with parallax before and am doing my best to remember how the maths for it works. I assume this behaviour comes about because the background elements such as stars and the planet are being assigned a parallax factor that is less than 1. This causes them to move as if they were "in front of" the flat playing field plane where ships, missiles, metal chunks etc all exist. As far as I can remember, layers with a parallax factor less than 1 appear in the foreground, a factor greater than 1 causes layers to appear in the background, and exactly 1 appears on the aforementioned flat playing field plane. A distant object therefore might have a parallax factor of 1.01, and a very very distant object might have a factor of 1.001. Also, parallax in general is not a good approach if you want to include zooming in and out. This is because Parallax does not move objects in the background based on zoom level. A better approach is to calculate the rendered position of background elements having given them an actual z-depth. I am able to provide conceptual diagrams and notes on the maths for implementing such a system if you would find that useful - I have previously implemented a 3D starfield engine for Eufloria and my work is free to use, copy and modify. See here for more details: http://www.dyson-game.com/smf/index.php?topic=1227.0 Please contact me if I can be of any assistance with this!! I am eager to help for free in any way I can. [email protected] | ||||
| Tags | No tags attached. | ||||
| Internal Weight | |||||
|
|
Interesting, we've been using the same logic for two years and three months, sold over 120k copies including expansions, and you're the first person to ever bring this up to my recollection. :) I appreciate the detailed post and the the offer of code. I'll have to look at this at some point, but right now we aren't really focused on things like this for AI War as time is very tight for us. Given that nobody else has mentioned it, this seems like an incredibly niche one. Thanks very much, though! |
|
|
* Some starfield changes in the next version (5.013): ** Slightly improved the draw efficiency of the starfields by reusing a single mersenne twister rather than initializing a new one every frame. ** Changed it so that there is only now one layer of starfield, rather than three. This makes for a more cohesive style with most of the starfields, as well as making the sky not QUITE so incredly dense with stars. This also removes the parallax effect from the stars themselves, which made the foreground stars really look more like dust than stars. *** This also is a small boost to performance, and a slight reduction in VRAM usage per frame. On most computers it won't make a substantial difference, though. ** Added in a parallax effect to the nebulae, making it so that they now have a slight motion and will wrap around the screen view if players scroll enough. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| Jul 20, 2011 10:15 am | annikk.exe | New Issue | |
| Jul 20, 2011 10:22 am | Chris_McElligottPark | Note Added: 0012750 | |
| Jul 20, 2011 10:22 am | Chris_McElligottPark | Assigned To | => Chris_McElligottPark |
| Jul 20, 2011 10:22 am | Chris_McElligottPark | Status | new => considering |
| Aug 1, 2011 11:28 am | Chris_McElligottPark | Note Added: 0012785 | |
| Apr 14, 2014 9:29 am | Chris_McElligottPark | Category | Bug - Graphical Issue => Graphical Bug |
