BlitzDGS (Blitz Dynamical Grass System)

Community Forums/Showcase/BlitzDGS (Blitz Dynamical Grass System)

Luis de la Cruz(Posted 2009) [#1]

DreamsSoftGames are working in the Beta 0.5 of BlitzDGS.

BlitzDGS (Blitz Dynamic Grass System) it's a powerful system to show a lot of grass in your games.
With another grass systems, we have speed problems in games with critical proves with around 40.000 grass covering the land.
Also problems comes with big extensions of terrain where systems becomes too slow.
BlitzDGS is thinking to shows millions of grass very fast in big terrains with slow or middle computers.

BlitzDGS precalculate exactly grass position in your land and save every grass coordinates into a file.
When BlitzDGS runs in your game, only system places every grass in land into your defined range LOD, without calculate height position, without consume more hardware process time. Also not necessary comprobe entity distance to every grass point/zone. With this system and 2 maths operations it knows in wich grass zone we are.
This result is a fast processing which leaves the cpu free for other tasks (IA enemies, sceneries, particles, shadows, effects,...)

Another important feature of BlitzDGS: with some grass systems is easy put aleatory grass on land surface, but why do you want grass into your buildings, ways or another elements? With BlitzDGS is very easy put grass only where you want land coverings.
You only need a bitmap image edited with Photoshop or similar, of same size of your heigthmap, when you paint spray grass spots onto your colormap, every point is a grass point and it generates grass in density proportion selected.
BlitzDGS places only in this zones the grass covering.

;----------------------------------
;limitations of this beta version:
; - Only for Blitz3D
; - Only for heightmap land generates by a bitmap image.

;Future features:
; -Suport for mesh terrains.
; -Wind effects in grass.
; -Different types of grass
; -Editor in real time

Download beta 0.5 demo in
Click here!

This beta is not a critical demo, only density factor 5. Place around 1,770,000 grass elements in land and more 350,000 grass points.
Please, agreed PC specs and fps. Thanks.


Nate the Great(Posted 2009) [#2]
looks fun. ill try it out when I have time.


Mike0101(Posted 2009) [#3]
I was try. Cool, realy fast and efficency. Keep up your work!


puki(Posted 2009) [#4]
Interesting.

Someone test it and report back.


Xzider(Posted 2009) [#5]
About to test this, though you should put up some screenshots.

Edit: Not bad but what's dynamic about it? Also the zip was 3.77 MB-unzipped 6.17MB after running it the folder was 80.6 MB. Can you explain this further?


plash(Posted 2009) [#6]
Edit: Not bad but what's dynamic about it? Also the zip was 3.77 MB-unzipped 6.17MB after running it the folder was 80.6 MB. Can you explain this further?
The 'zone pre-calculating' bit creates 14400 files. I decided not to bother, as it takes a long time for winblows to decide it wants to respond to my request for deletion.


Nate the Great(Posted 2009) [#7]
ok so I tested it and its interesting certainly. The only part I don't like is the fact that the terrain looks so plain and bare after the grass dissapears.. maybe a little texture tweaking? also, maybe if instead of just fading them out slowly, first one would fade out, while the other fades much slower and rotates to face the camera. I think that would give a much nicer effect. It looks like it needs a little life in it too... maybe small swaying movement?


ZJP(Posted 2009) [#8]
Hi,

90-110 Fps.
C2D E6600 @ 3.2Ghz, GeForce 9600GT

JP


Xzider(Posted 2009) [#9]
I see, that sure takes a lot of space.

FPS: 78-110

Specs:

Windows Vista 32-bit Home
NVidia Geforce 8500 GT
3GB RAM
Intel Quad Core CPU


Nate the Great(Posted 2009) [#10]
perhaps if it stored all of the info for all of the grasses in one or two files... it wouldnt be that hard to write a simple parser would it?


Luis de la Cruz(Posted 2009) [#11]
I very much appreciate all your suggestions. We are already working in the beta 0.6 to try to correct the defects found. Thank you very much again. We will report as soon as possible.


Luis de la Cruz(Posted 2009) [#12]
Ok, first a clarification on the term dynamic used in BlitzDGS because something ask about this. System is constantly loading and unloading areas of grass next to the camera in real time, allowing them to consume only the minimal hardware resources and memory, to be attached to the positions and heights precalculate field , processing time is reduced.

Now improvements in the Beta 0.6:
- Optimized the number of files: only 128 instead of 14,000 from the previous beta.
- Optimized file size of positions precalculate now about 18 MB instead of the nearly 80MB from the previous beta.
- Maintaining the processing speed in generating grass and decrease generation time of grass zone files.
- Added a first approximation the motion of the grass by the wind.
You can download here!


Stevie G(Posted 2009) [#13]
1st Demo 77-100 fps
2nd Demo 60-65 fps ( prob due to the sway effect )

G8600GTS

Personally, I don't think this is fast enough to be practical given that there are never more than 12,000 poly's visible. I can't help thinking that their must be a faster approach, although I may be wrong as I don't know what the benchmark is for this kind of thing.

The speed that the grass is created is much better in the 2nd demo. I still think that 128 files totalling 18MB per level is too large for simply adding grass.


Luis de la Cruz(Posted 2009) [#14]
New Beta 0.7 of BlitzDGS (Blitz Dinamic Grass System).

;Improvements in the Beta 0.7:
- Optimized file size of positions precalculate: now about 10% less respect last beta.

- This demo is done on a smaller field (1024x1024 blitz units). The files that store the positions occupied by grass precalculate only 1Mb in 64 small files. However, it works perfectly well in large areas, requiring larger file of course.

- For the far land does not appear simple and bare a percentage of each zone remains fixed charge on the stage.

You can download here!