Please could someone help me? My computer isn’t powerful enough to make a short movie of my game. I tried to use xvidcap and recordMyDesktop but I get a very bad result. Could someone record a small video of TUER please?
what you can do (and what I do when I need to do such) is reduce the speed of your game (speed of player, bullets, etc…) than take the video at this low speed. once done open video with virtualdub (set video to directstreamcopy) and export the video with a higher fps.
I don’t know virtualdub, I will try this solution. Thank you very much.
TUER is already able to produce one screenshot per frame at quite good speed on fast machines but I have no tool to regroup this TGA files in a single MPEG file.
I use camstudio, it let you specify a region on your desktop and than save it as an avi/divx video, but it is window and i guess you do not have it on your pc ;), but sure you will find a same software for linux. I only use virtualdub to increase the final video fps, severals other sofwtare can do that, but virtualdub is so simple!
[quote]TUER is already able to produce one screenshot per frame at quite good speed on fast machines but I have no tool to regroup this TGA files in a single MPEG file.
[/quote]
you should not spare too much time on that , there is already the “printscreen” key, so the benefit of such feature is very low.
You’re right but I didn’t spend too much time in this, JOGL has a class called Screenshot that is very useful. In Java 1.7, I hope that there will be a standard way of encoding MPEG without any need of external librairies. For me and for any user, I think that it would be fine to take some video captures directly from the game, without any other software.
Lara’s model has been removed from the zipped archive because of a problem with the license. You can get this model here:
http://www.jeuxvideopc.com/downloads/238-quake-3-arena/834-models/19944-lara-croft.php
My game works under Vista ;D
AMD X2 5200+ @ 2.7 Ghz (dual core)
ATI HD4850 - 512 GDDR3
4 Go DDR2
Vista business 64 bits SP1
http://tuer.tuxfamily.org/screenshots/screenshot_ATI_HD4850_1440_990.png
The bug of the “white wall” appears only with some graphics cards, it seems to come from the driver even though I have planned to use some workarounds.
At first, TUER is not compatible with OpenJDK 1.6 because some bugs in Java Webstart’s implementation.
On the other hand, the scenegraph is almost working, I need to perform some deeper tests, I’m going to update the source code tomorrow morning (french time).
The network of cells and portals seems correct, I’m going to perform some deeper tests to be absolutely sure of what I’m doing. The optimization allows to convert 202 cells into 182 cells that are not overlapping:
http://tuer.tuxfamily.org/screenshots/cellsmap.png
The bug fix allowed to eliminate 2 malformed cells. Now, I’m writing some tests and I’m filling the cells with the whole data.
Can’t play it because it only works under 1.6 and I am using a 32bit Mac.
Java for Mac OS X 10.5, Update 1 is available via Software Update. This update adds Java SE 6 version 1.6.0_05 to 64-bit Intel Macs. For 32-bit Macs, watch this:
http://landonf.bikemonkey.org/static/soylatte/
TUER could work with Java 1.5 but with some bugs coming from Java, I already spoke about this. In the worst case, as the game is open source, you can recompile it with the JDK 1.5 and launch it. Use the default target of the ant script to do it and use the target “run-game” to launch the game.
I will have to precise this point. My teacher had found a way of using JDK 1.6 under older version of Mac OS X (through ADC) but she had a 64-bit Intel Mac. Nevertheless, Apple removed this beta version from ADC in October 2007. Apple insists to develop its own JVM and it complicates many things. I’m sorry, I do my best.
Thanks dear
I’m currently implementing the fix for the bug that I found thanks to darkprophet. I need to put a decorator to replace the vertex sets that are too big. This decorator will contain several vertex sets but will behave like a single vertex set except on one aspect: when I ask it to draw itself, it will ask its sub vertex sets to draw themselves and each vertex set will have a size that respects the OpenGL constant GL_MAX_ELEMENTS_VERTICES. Then, the performance might be better.
I will try …
Hi!
I’m going to finish the bug fix concerning the too big VBOs today. I have a bad news. It doesn’t change the frame rate noticeably on my machine, I only have sometimes 16 FPS instead of 15 and the game is smoother. However, I think it will fix other bugs, the kind of white void on some graphics cards, it will improve the performance on small on-board graphics. I notice that GL_MAX_ELEMENTS_VERTICES has been removed of OpenGL-ES since the version 1.1 because “It seems like as a performance hint it was just not deemed necessary and therefore removed” according to Georg Kolling. Maybe some graphics cards handle reliably VBOs that don’t respect this implementation-specific threshold. Now I need to complete this bug fix.
I have not tried to split the quads into triangles for the moment but this might help.
On the other hand, I’m still more optimistic about the cells-and-portals algorithm because the frame rate is twice higher even though I send to the graphics cards 2 thirds of useless data.
Finally, I will update the JAR file and the source code tomorrow morning.
The source code and the Java archive have been updated. Please could someone test it? I would like that people who had some problems test it again, especially people with an ATI Radeon HD under Vista, people with an Intel on-board graphics and people with an ATI Radeon XPress 200M on-board graphics. Thank you very much.
Reminder: the bug fix should unfortunately improve the performance on low end machines only.
Finally, I go on working on the scene graph. When I play with Jake2, I realize that TUER is very slow :’(
Hi!
The Java archive and the JNLP file have been updated in emergency because of the issue #356. If someone has already installed JOGL 1.1.1, TUER crashed. Then, I put my fix. Please, could some people test if it still works?
Only 6 cells on a total of 177 cells are still malformed 
I hope I will fix the problem as soon as possible.
For people that can read French, I have improved my tutorial, it is not finished yet.
The source code has been updated. I’m putting some small things to allow online gaming. I’m going to use RMI temporarily. If it is really too slow, then I will use something else. The scenegraph is still not working but I’m sure that I’m going to find the root cause of the bug in a few days.
The recent video of TUER is here:
http://video.google.fr/videoplay?docid=207500017344395219&hl=fr
EXCELLENT NEWS:
[java] network integrity preserved
BUILD SUCCESSFUL
Total time: 9 seconds
I authorize you to tell me that I am very… stupid, absent-minded…
I had forgotten the bold part of the code in the method below:
public void addTopWalls(Vector topWalls){
this.topWalls.addAll(topWalls);
}
I’m so happy ;D ;D ;D ;D ;D ;D