Zalo DS Blog

Whack Mania: Post mortem

2013-03-11T01:12:00.001+01:00

Ok, so it's been a month already. Time enough to look backwards and see what went good about this project and what went wrong. I have decided to post these conclusions here so they can become useful for someone else. Usually I prefer to begin with the bad things and finish with the good ones (that leaves a better impression)

Bad things! (things that should be improved in the future)

- First of all, and one that disappoints me a lot, working at home. I already work 8 hours and a half daily as a programmer so adding a couple more to do the same thing plus weekends can be exhausted. It totally affected my mood, made me angry easily and obviously making me less productive on an average. I haven't fully enjoyed my holidays (not summer, neither christmas) and I have suffered insomnia several days

- Working without a release date. Even thought the game was planned to be finished in two months, I have gone way too far from that date. Nobody really seemed to be worried about it, and some people were even asking me to constantly add new stuff to the game. I haven't feel capable of just stop adding content

- Listening to everyone without filter. I should have been more strong about my decisions. I listened too much to the feedback people were giving to me… first the game was too hard, then too easy, then boring, then the camera was not accurate, then we needed to add more stuff to the levels… I think there is a difference between lack of modesty and self confidence. It's not people's fault for giving me suggestions but mine. Trying yo make a game that everyone in this world like it's very hard (or probably impossible)

- Adding way too more content. Sadly for some reason the game doesn't seen to be working pretty good. Well, to be honest… it is doing terrible. I cannot publish here the number of downloads but for me it is very disappointing. Obviously I was expecting much more. Flurry is integrated in the app so I can see which levels are people playing. A big amount of the levels are still NEVER played. So, were those 40 levels necessary? Definitelly not. Besides not all the levels have the same quality. I spent a lot of time designing some of them and just a few minutes in some others. Sadly most of these levels are at the beginning so people give up before reaching the funny stuff

- Taking it personal. I don't think it will be very proffesional to post more info about this here... so just in case you are interested to know what I am talking about, ask me in private :P It is has a lot to do with that lack of confidence I was talking about before

Good things!! (The big amount of knowledge acquired!!)

- Whack Mania has been developed using my own engine. I am pretty happy with how the engine has evolved and the way everything on it is coded. More than happy, I feel proud about it. It has been released both in IOS and Android and there is also a Windows version fully working. Versions for Nintendo DS and Wii should work also, although I haven't had the time to code some new parts on the core of the engine for them. I feel that I am perfectly capable of porting it to any platform suppoting c++ as native language. More than that I have implemented: leaderboards, advertising, in app purchases, analytics and push notifications in a clean way (they are not implemented in Windows but the game still works, and even thought the behaviour in Android and IOS is very different they share a common base that makes it very expandable). I really have that feeling that I have learn from other's mistakes.

- So many changes in my 3dsmax exporter. I am pretty happy with it. I have fixed it a lot of times. I have removed a lot of trash, kept it simple and right now there are no issues that I am concerned of. No scale issues, nothing bad orientated, animations perfectly working… and if you have already worked with maxsdk you should already know what I am talking about

- Back to maxscript! I have coded a few useful scripts for 3d max that have allowed me to use sprite atlas and define menus. Some of this stuff could have been very valuable on previous games I have been working on

- Several fights with Xcode and Eclipse had made me understand these IDEs much better (Eclipse stinks no matter what). I know how to handle libs, projects and everything. That makes me very capable of for example coding plugins for Unity3d in a safe way

- Some interesting components that I'll be using in future releases: camera safe viewport, HUD reorganization independent of screen resolution (and for the first time also, I have added a new behaviour that leaves some space for adds automatically so I didn't need to make any changes at all to support the same screen distribution both on full screen and when showing adds), input management (multiplatform), finger tracking (that I used for the movement of the planet), pathfinding (in a spherical mesh… not very different than a classic A* anyway), music synchronisation with game events (did you noticed?) . I have already coded these new behaviours into my new project in Unity without too much trouble, it is also good to realize that what really matters is the idea and not the code itself

- Interesting graphical effects (using fixed pipeline on OpenGL ES 1.1). The game doesn't use a single shader at all… I still consider that adding shaders to a cross platform engine it is a hard task and it will make things more complicated to maintain so I have refused to do it. I have used instead texture combiners for the first time that it is still fixed pipeline but when used correctly the final result can be amazing. The way for example the Wii works is pretty similar to this. I think the game performance has benefit form this a lot also.

- Implementation of my first asset manager, that can regenerate all assets form scratch just running a simple script (not true, because I coded it in C++, because it is much better :P). After too much suffering with others implementation It's hard to say how happy I am to successfully had implemented this in just one week. It is once again simple and very expandable to support new platforms or new type of assets (or even different configurations for the same platform)

And, that's it! I was right… it is good to finish with the positive stuff. As a general conclusion I'd say Whack Mania has been a total failure as a project (if you give something for free and people doesn't like it, there is no other way to consider it) and that makes me feel very disappointed, but on the other hand it has given me a great opportunity to improve my coding skills, which is something that I as a programmer need.

I also would like to show my appreciation to the other guys who worked with me in this adventure: Domingo, David, Raquel and Antonio and all the other people who helped me to improve it: Josh, Cesar, Adrián, Juanma, Josemi, Curro, my brothers and of course Lucia who has suffered it more than anyone else. Thank you very much for your help! Even thought it is a failure it is still a game that I really enjoy

I'd like to finish this with the hands on video we released a few days, so you can see the how the game works a little bit more

Thanks for reading! ;)

Whack Mania

2013-02-06T23:17:00.001+01:00

Here it comes, my next Indie video game is called Whack Mania. Is Based on the classic whack-a-mole taking advantage of the touch screen available in all devices nowadays. It's been developed using my own technology (the same used in the previous games) and will be published by Bravo Game Studios (the company I am currently working for)

Please download it Friday 8th 2013. It will be available for Android and IOS in two versions: full and free with ads. If you enjoy it please buy the full version or remove the ads in the free one. I will appreciate it ^_^

Show must go on

2011-07-19T01:03:00.005+02:00

This is my new cellphone... that is my engine :)

Some Rokoban and Jumping Jack videos

2010-11-12T09:18:00.003+01:00

Someone from nintendomax uploaded them recently on youtube. Thanks guys!

Jumping Jack in 3D

2010-10-25T00:14:00.007+02:00

Have you ever stopped to think about the first videogame you ever played? I know there are hundreds of people out there answering that it is Super Mario Bros, Tetris or any other famous one. For me it is different. It has to be one of those stored in a cassete for our old ZX Spectrum 128k (wow!) I remember myself like 22 or something like that years ago entering into my old brother's room that January 6th, sitting there and playing it... for the first time. It is hard to believe that I still spend some time playing it in some emulators :)

So.. I looked for a remake. Althought I found some games based on it there wasn't any that keeped me entertained enought so I decided to do my own one :P. It started as a very easy task but then I realize I had to model the 10 hazards (oh... it is in 3d, sorry about that ^_^UU) and animate the main character.... bla, bla

Anyway, here it is! You can find it in the download section as usual or just after this paragraph. I hope you enjoy it as much as I am doing and if you have any suggestions don't hesitate to send them
PC version
WII version

PS: oh... yes... the DS version... I am sorry to say that the poor DS is not capable of moving the game well enought beyond the 5th level, so I decided not to publish it :( I would need a new set of graphics and I am not really sure I want to do that ^_^U

Rokoban updated

2010-03-14T21:04:00.002+01:00

Althought I haven't been posting too much during this and last year I have been still working on something from time to time (not as much as I used to). Lots of improvements were added to the engine and in the meantine I was thinking that maybe I should create a new game to make some use of them. Well, creating a new game requires time so instead of that I decided to update my Rokoban using all these features. I also had the chance on one of my jobs to work pretty close to 3D max (as a tools programmer) which gave me the opportunity to learn lots of things about the program. That's why I decided to create also a new model for Roko, the main character.

I finished this version on December last year but I forgot to upload it. Well, not really. I didn't forgot I was just too lazy to do it. These are the new features added:
- New set of graphics for Wii and PC (I think the DS graphics were ok for the ds, but looked lame on my PC)
- Savegames (so now when you complete a level Rokoban will keep a record of it)
- Music and sound (finally, woo hoo!)
- Some memory leaks fixed

If you pay attention to the new character you'll see animation blending, skinning, texture animation in action :) Some other stuff I have been working on is not visible on the game. I have now a lot of debugging information that helps me a lot when I run the app in one of the platforms and I just receive a black screen (mainly asserts and memory leak detectors).

I think that's all :) Hope you enjoy it. You can find it in the downloads section as usually (there, on your right) or just here
PC
Wii
Nintendo DS

PRIXON: Time goes by... so slowly?

2009-01-10T20:10:00.003+01:00

Damn! The new year has just began and this year I realize that I have something important to celebrate. It's been 10 years since I finished Prixon! I can't believe it has already been that long...

So let me write a little bit about this game and why it is so important for me ^_^ (no boring code stuff this time). Eleven years ago I was finishing my studies and getting ready for university. I still wasn't very clear about what I wanted to do, that's the problem when you have more than one hobby. By that time I was already involved on code stuff, but only in qbasic. I had tried to do something in C++ but those annoying pointers were being a hell for my mind in that time (:D)

I wanted to create videogames since years ago. The closer attemp I did was with click'n play... but that was more like creating a game while playing (I think Microsoft is planning to create something similar for the 360 now to compete against Sony's LBA). Then, exactly 11 years ago, on Christmas my brother brought to home a demo of a new videogames creator enviromment. It was called Div Games Studio. Div for the friends :D

The demos included on the game were really impressed and the language was quite intuitive, even it included a debugger. I was amazed. And I saw a big opportunity to do something big for the first time. After months testing the language and learning it I started a project called Chula GP (the name is also a long story) that would be renamed into Prixon a few days before finishing it.

I was working on Prixon for a little more than a year. Not fully dedicated because that year I had to study a lot for the exams we have to do before going to university. I use a lot of stuff in this game. For the first time I got contact with tiles, mode 7, music, graphic modes, palettes... and also for the first time I got in contact with 3d max (v4, in ms-dos ); yes, the models in the game where modelated in 3D, LOL, those are the worst models I have ever done in my life. Not that I am really proud of the code today, but the thing is that I finished what I wanted. No rewards except some people telling me that they loved it, which was enought.

I think that when I finished Prixon I knew what I wanted to do with my life. I have finally become a games developer. To be honest I don't know how to fell sometimes. This is a hard profession and sometimes you wish you had a bakery or a ranch in the middle of the forest :_) I am sure I am not the only one who fells like this LOL!

When I was about to finish the game I decided to create the official comic. That was very special also. It supposed the return of Cebolleto, Carateca and Topolino. For 3 years I was dedicated to my manga style and those characters were started to be forgotten. Prixon meant the end of my manga style (or should I say Toriyama style? :D) and the return to my old characters.

Guess that I owe a lot to DIV. By the way, last summer a new version of DIV was released called Gemix Studio, fully compatible with the old one without changing any line in the code (you just need to recompile) I decided to give it a try and get a version of Prixon running under Windows. Yes, you can find it here (source code included) but, be nice :P You have to look at this game like it is, the beginning into the games programming of a university starter ;)

Happy 10th anniversary Prixon!

PD: the sound doesn't work fine.

Free Radical Closes

2008-12-19T00:29:00.003+01:00

From GamesIndustry

People who know me they know why I am posting this here. Free Radical means something to me specially since 6 months ago, when they played with me. But for some reason, these are bad news. And I am not happy for this. There were a lot of people in that studio I really wanted to work with.

It is very sad always to heard that an studio wich has made some wonderful games in the past closes its doors. As someone told me one day and I said to myself 6 months ago... "shit happens"

No resentments.

Rokoban Wii, 4th place in the Tehskeen Wii code competition

2008-12-02T22:09:00.001+01:00

http://www.tehskeen.com/forums/showthread.php?threadid=9775

All right, all right... I'll do my best for the drunken coders winter competition :P

Memory Leaks

2008-11-24T01:09:00.003+01:00

When I finished Rokoban I realized that I didn't have anything to tell me how well or bad I was working with my memory resources. So, here I'll tell you what I did to have a little bit more of control.

First of all. What's a memory leak? Well, this

void MemoryLeaker()
{
int *leak = new int[99999999];
}

is a memory leak :D

When you start learning how to make programs in C++ you soon realize that memory leaks are the more annoying part. They always teach you the basic rule: "Everything that has a new must have a delete" (or is that from the Matrix movies?). That seems easy at first, but after years of C++ development I know that no matter how hard I try, I am going to create memory leaks, and you too :P Even if you try to avoid them using smart pointers you are never safe from them.

So, is there a way to avoid them? Yeah. There are a lot of ways. You can even create a garbage collector like Java. But there is a more interesting way. It is posible not to avoid them but to let our program tell us about them. And that way you can debug your application in debug mode and finally release it without any of this checks, making it as efficient and fast as posible.

So, what is the basic idea? C++ allows you to overload new and delete. You can overload them for a class or for the whole program. And this has a lot of advantages. You can reduce memory fragmentation and you can have control of what happens any time you request some memory from the heap and when you release it. So, basically you can implement a mechanism that everytime you reserve some memory it creates an entry on a table (for example) and everytime you realease it it deletes that entry from that table. At the end of your program all the entries in your table means memory leaks.

If you really want to know more about this take a look at this link http://entland.homelinux.com/blog/2008/08/19/practical-efficient-memory-management/. I was about to write my own memory manager for my engine. But I found that Visual C implements something very interesting for you. In fact, it all began with this link http://www.highprogrammer.com/alan/windev/visualstudio.html I really recommend its reading. Here I discovered that there is already an implemented debug heap I could use instead of writing my own. It is enought for me at the moment. I can only check memory leaks on my Windows release but because I usually develop everything in Windows and then recompile it using ZEngine for the rest of the platforms I can now be sure that the common part is memory leak free. There can be still problems with all the especific parts coded for each platform but I hardly make any memory allocation there.

I am finally doing what I found in this url http://www.cplusplus.com/forum/general/3526/:

This should go into every .cpp , after all header inclusions

#ifdef _WIN32
#ifdef _DEBUG
#include
#undef THIS_FILE
static char THIS_FILE[] = __FILE__;
#define new new( _NORMAL_BLOCK, __FILE__, __LINE__)
#define malloc(s) _malloc_dbg(s, _NORMAL_BLOCK, __FILE__, __LINE__)
#endif
#endif

And then at the beggining of my app:

// Get current flag
int tmpFlag = _CrtSetDbgFlag( _CRTDBG_REPORT_FLAG );
// Turn on leak-checking bit.
tmpFlag |= _CRTDBG_LEAK_CHECK_DF;
// Set flag to the new value.
_CrtSetDbgFlag( tmpFlag );

If you do this as anilpanicker is showing then at the end of your program you will receive a very usefull output telling you where exactly you reserved some memory that was never released :D Using this I fixed a small memory leak in Rokoban, very very small, but still it was there. Now I can sleep much better.

There is still a lot of work to be done in my engine. I wish I had all the time in the world to code it but unfortunatelly I have other things to do. Anyway, knowing all those things is always really useful, you never know when you can need them. In the company when I am working at the moment they have their own memory manager implemented. More than that. They have their own asserts. I have learnt from them that if your program can't load a resource the best thing that it can do is telling you and stop. Only in debug mode, of course. This is much better ans easier that trying to make your game work without a graphic loaded of something like that. In your release you should have all your resources so it doesn't make sense to make it work without them :)

Hope any of what I have written will be useful. And take a look at all the links. They were really usefull for me.

Downloads section updated

2008-11-22T21:57:00.002+01:00

Hi... I have just moved the downloads to a new server... hopefully this will be working better. The only change is Rokoban for the DS wich is now a single file instead of nds+resources. I haven´t had enought time to test it but I can't find the old version so I decided to upload this one

Enjoy

Getting into the Wii

2008-11-16T22:19:00.002+01:00

Helloooo, everyone!!

I have finished and I am still alive... that is good!! What am I talking about? Yeah, good question!

I have spent the last three weeks porting ZEngine to the Wii :) I was in a hurry somehow because I wanted to finish something for the Tehskeen competition, so it has been a nightmare. Three weeks programming 12 hours a day, including weekends. First at job then at home... never again!!!!

What is the tehskeen competition you may be wondering. It is the first Wii homebrew competition. Take a look here and to take a look at the entries, this link You can take a look at all the games and vote Rokoban for winner :P

The Wii scene is amazing. I think that DS has stopped and lots of people are moving to the Wii, so now it seems to be the more active. The unnoficial SDK (devkitpro with libogc) allows you to do a lot of things and people are working everyday to release new versions lots of libraries, so I think it is a good a idea to move.

To be honest, I enjoy more developing for the DS. I think there is still a lack in the Wii documentation and somedays trying to do something can easily become hard. You need to look for samples, ask in forums, irc... etc

Cool! I am free! I am going to stop developing for a couple of months, I really need it. But in the meantime there is going to be a lot of time to post here all the features that I have been making to ZEngine (and belive me, there are a lot of them)

You can download Rokoban in the tehskeen competition website, I think I can't still upload it here. But there is a new version of Gea VS CoDOS (I hate it!!!) for the Wii in the downloads section or here

Have fun!

Texture Converter

2008-09-16T22:10:00.003+02:00

Time for a new update :D Hey, I have been in the UK for 3 and a half months so my english is suppose to be better now (right? XD). Well, today I am going to post about something I did a couple of moths ago, just when I released Rokoban.

I don't know if you have realized but in Rokoban there is no use of channel alpha at all, I mean, by that time I wasn't able to load textures with transparent colors. Not even a simple bit for transparency. That's why all the texts in the game appear over a black background. Ok, if you didn't realize then I have showed every artist in this world that channel alpha is not required anymore! ... not even me can believe that

And why didn`t I give a little support for alpha channel? Because my textures were in bmp format. Bmps doesn't suppot alpha channel. Well, I am not sure at all about this... I think there are some programs that allows you to save 32 bits bmps. The problem is that you don't need your textures to be that large. 32 bits bmps has 8 bits for each channel (Red, green, blue, alpha). The DS doesn't suppor that format so you have to conver it in execution time and that is a waste of time (for the conversion) and size (your textures are taking more space that they really need)

So, basically, what are the texture formats supported by the DS? Those here:
GL_RGB32_A3 = 1, /*!< 32 color palette, 3 bits of alpha */
GL_RGB4 = 2, /*!< 4 color palette */
GL_RGB16 = 3, /*!< 16 color palette */
GL_RGB256 = 4, /*!< 256 color palette */
GL_COMPRESSED = 5, /*!< compressed texture */
GL_RGB8_A5 = 6, /*!< 8 color palette, 5 bits of alpha */
GL_RGBA = 7, /*!< 15 bit direct color, 1 bit of alpha */
GL_RGB = 8 /*!< 15 bit direct color, manually sets alpha bit to 1 */
Maybe for those of you with more experience in console development those formats are familiar, but for me, I have never used something like that. Which program allows you to save into one of these formats? Well, the last two ones (with no palette) can be more or less a 16 bits BMP and you can do some tricks like reserving one of the colors for alpha, and make that conversion when reading. That seems to be what most of the people are doing.

Problem with that? Well, you only have support for 1 bit of alpha (wich means, no gradient) and it is using 2 bytes for each color, and it is not bad... but when you want to load the more textures as posible into memory (and you always are going to need that if you are working whit artist that always want more and more AND MORE!!!... sorry) that is not good at all.

Some people in the homebrew are using PCX. I don't know that format very well. But I know it doesn't give you suppor for any of those weird formats. So I imagine that they are converting the textures when loading then. I was doing something similar for my textures. I was reading paletted bmps and the converting then into 16 bits bmps in execution time. For the alpha channel I was plannig to load another texture in grayscale as I have seen in some games...

Ok. All of those solutions work. But are not the best. Wouldn't it be great to have a program that allows you to save in one of those formats so that you don't have to do anything in execution time? Some people in the homebrew has released programs to do that. But It's been a while and I haven't been able to download one of them. There is also de DDS format for directX, but the pixel formats supported are not compatible either (take a loook... mmmmmm,google,tacata.... here: http://www.imageconverterplus.com/help-center/about-icp/supported-formats/dds/). And there is also a program that people from ID made for quake models but I can't remember the name... so it doesn't matter XD

So... I dediced to write my own converter XDDD Yes, if you spent more time looking for something than the time that you really need to do it then you are wasting your time. The converter I have developed is better than just one specific for the DS (My engine has to be multiplatform :P) So it supports a big variety of formats. You can select how many bits you want for each channel, and in case of palettes, you can select how many bits you need for palette index. It is designed at this momento so that alpha channel in paletted textures is not stored in the palette, but in the index. So, the palette has only 3 channels (RGB) for each color and then in the index there are n-bits for the index and n-bits for the alpha channel. That't the way DS works and it is ok for the moment.

For the first time and without meaning a precedent (I think you can't say that in english) I am going to release not only the tool, but also the source code, so that you can improve it if you want ^_^. Format specification is very simple (and it is included in a .txt) and loading textures is quite straightforward for the DS. To use the program just drag and drop the files you want to convert. I decided to use C++ and MFC this time instead of Java, which is the language in wich I usually code my tools

Download it from here Any comments, suggestions are always welcome ;)

PD: by the way, Capcom... What is the meaning of MEGAMAN 9!!!! You must be thinking we are stupid
PD2: square... WHAT THE HELL ARE YOU DOING RELEASING A PORT OF CHRONO TRIGGER FOR THE DS INSTEAD OF A REMAKE!!! DO YOU THINK WE ARE SO STUPID TO BUY THE SAME GAME TWICE????

Reasons to Love Homebrew development

2008-07-20T22:24:00.002+02:00

Sometimes I think about it... why does people make their homebrew applications? Well, one thing is for sure... it is not for money. And if you think about it a little bit, that's great!

I think that there are a lot of people out there thinking, "Man, I'd like to create a game for the DS", but they don't do it. Why not? Because their companies are not interested on doing that... then, Why don't they do it by themshelves? Well, I think there are another explanations but basically some people don't do it because they are not going to receive anything in exchange.

Well, aren't they? Really? I think that people should think about it. Why are you developing videogames? Money,? That's all? I have discovered in the last months that there are something else. Of course you need money, you have to survive. But there is also a big pleasure seeing that people are intererested in what you are doing. A bit of satisfaction seeing people playing what you have created, just enjoying it. I think it's the best reward. Because that what videogames are really for, entertainment. You are not just typying lines of code, you are creating entertainment. So, the most entertained (that't this word exists?) your game is the better it is. That's all. You can make a videogme and earn a lot of money with it but if people doesn't like it you are not really happy at all.

Why do I love homebrew development for the DS? Well, first because I do what I really am interested on, and do it my way. Second, because I love this machine, really! And third and most importan: People are interested. They care about it. I released Rokoban just two weeks ago, a Sunday night. Next day, on Monday, I googled for it. How many url's were talking about the project??? It was awesome. I was completely amazed. But above all, I was really really happy. And so is Roko

Thanks, everyone

PD: anyway... don't spend your life developing all the day ;) Just enjoy it

Rokoban, a Sokoban version, in 3D

2008-07-07T00:00:00.005+02:00

I remeber when was the first time I played Sokoban. It was on my Gameboy. I was only 12 and it was summer. That summer I bought one of those 16-in-1 cartdridges. There were a lot of awesome games there but there was one that I remember specially. I didn't even knew his name until some years later. It was an awesome game that made me spent a lot of time thinking and having fun. There was a guy with a hat and some boxes that you had to push to some cells that were marked with points. I didn't need some instructions or something like that to play it, the game was in Japanesse. And I think that is one of the wonderful things of this game. It was intuitive, fun and gave you that felling of satisfaction when you finished a level you have been trying for days. It was Sokoban.

A lot of ports of this game have been released since then, for all platforms, including the DS. But anyway I wanted to make my own version :D I have spent a lof of time on it. Not because it is very hard to develop (I coded the logic of the game in just an afternoom and in 3 weeks I had my first playable demo) but because at the time of developing this game I have also been improving ZEngine. The version that I've made is on 3D and I thinks is the first one like that released on the DS. You can also load set of levels from your card and I think this is also new too. And the main character of the game is Roko, one of the character that I draw for my weekly comic strip tinomi. I had the model done years ago, it was going to be on 1010 Locos por el tenis, and finally I have been able to place it on a game

Ok, that's all from me, at the moment. Here I leave you the versions for Windows and Nintendo DS. There is no version for Symbian. You now what? I hate Symbian, it is horribleeeeeeee. Yes that's the truth, you don't program for Symbian, you fight! XDDDD And I am tired. I don't know if I'll continue supporting Symbian versions or not. Who knows. Hope you enjoy the game. For any suggestion of bug report, you know where to find me :P

Download Windows Version
Download Nintendo DS Version

Gea vs Codos working on R4

2008-06-20T14:36:00.001+02:00

Oeee, ooeeeeeee, finally!!!

You can download it from the downloads section.
Thanks to Blatbaster for testing it ;)

Painting in 2D using 3D (III)

2008-06-18T12:59:00.004+02:00

It's been a while. Again. LOL XDDDD

Ok, let't finish with this. When I wrote the previous blog I was still doing something bad. And I am going to expose it here so that nobody else has the same problem.

The problem is that in libnds glOrtho is defined like this

void glOrtho(float left, float right, float bottom, float top, float zNear, float zFar) {
glOrthof32(floattof32(left), floattof32(right), floattof32(bottom), floattof32(top), floattof32(zNear), floattof32(zFar));
}

if we take a look at floattof32

#define floattof32(n) ((int32)((n) * (1 << 12)))

Then we see basically that the params that glOrthof32 receive are 24:12 fixed points. Now I am going to show you what I was doing wrong. I called glOrthof32 like this

glOrthof32(0 << 12, 256 << 12, 0 << 12, 192 << 12, -1 << 12, 1 << 12)

Ok, the problem here is that as I say in a previous post, vertex coordinates in DS must be between [-8, 7] range. So, it was going to be impossible to draw on my current perspective (wich extends from 0 to 256 and 0 to 192). I was only going to be able to draw in a very low part of the screen. So... what did I do? I used a glScale for my vertex... and... well, I think it is better if I tell you how to do it right, because it is possible to do it like that, but it is more complicated.

So. I was curious about how other people did this. It couldn't be that complicated. And I see a portion of code when someone did something like this...

glOrthof32(0, 256, 0, 192, -1, 1)

Yes, as simple as that. Now you must be thinking that I am stupid or something similar XDDD. The problem is that I has a Fixed point class that manages all my conversion. So I don't really called glOrtho with the params displaced 12. My class did. Id you take a look at the previous call and you think about its meaning, really what it os doing in floats world is this:

glOrthof32(0 / 4096, 256 / 4096, 0 / 4096, 192 / 4096, -1 / 4096, 1 / 4096)

which is

glOrthof32(0, 0.0625, 0, 0.046875, -0.000244140625, 0.000244140625)

This projection is perfect for 2D painting!!
Ok, so the conclussion basically is... I was trying to scale each vertex when I could have just scaled the perspective!

Thank you very much.

By the way. I am going to write this week. I have something to upload ^_^ I have almost finished a new simple game.

Painting in 2D using 3D (II)

2008-02-24T15:57:00.002+01:00

Ok, some people who read the previous post told me that they dindn't understand anything of what I was saying :D So I'll try to go explain everything a little bit more better this time. Let's just focus on the first issue I was talking about.

When you are painting on the screen and you are using a 3D Engine (like opengl) there is something called the pipeline. You just don't draw 2d vertices on the screen as usual, but in 3d. You are painting on 3d wether you like it or not. You can check this pipeline on the internet (here, for example http://june.ncviz.com/openglpipeline.html) Basically what goes on with a 3d vertex that you draw is that it is being transformed throught the pipeline until it is finally painted in 2d into the screen.

You can't avoid these transformations, all right? Let's focus on the first two steps of this pipeline. You define a vertex and it is first multiplied by the modelview matrix and then by the projection matrix (sometimes modelview martrix is divided into camera matrix and object transformation) and after that we don't care now (if you are interested read it in the previous link :D)

The first matrix transforms the vertex coordinates into view coordinates (still on 3d) and the second (which we are specially interested on) projects them on 3d according to the perspective we have defined. You usually define a 3d perspective with a frustrum (a truncated square pyramid). With thise kind of perspective closest objects are seen bigger than further ones

This is the way you paint on 3d. But there is also a special perspective known as orthographic in which the size of the objects doesn't depend on the distance to the camera. It doesn' matter how far they are you'll see them always the same. This kind of perspective is defined on OpenGl using glOrtho(). http://www.cs.utk.edu/~vose/c-stuff/opengl/glOrtho.html

void glOrtho( GLdouble left,
GLdouble right,
GLdouble bottom,
GLdouble top,
GLdouble zNear,
GLdouble zFar )

Using glOrtho(0, SCREEN_WIDHT, 0, SCREEN_HEIGHT, 1, 0) we can say that you have created a full screen canvas. Now if you paint a vertex on (x, y, z) the point will be painted on the screen on (x, y) and z will define wich point goes behind the others. Easy!!

Mario Kart, April 11 in Europe!!!!

2008-02-20T23:48:00.001+01:00

Sorry, this is a little bit off-topic... but I have to post IT!!!!!!!!!!!!!!!!!!!

KIAAAAAAAAAAHHHH!!!! IAAAAAAAAAAAAAAAAHHH!!! AHHHHHHHHHHHHHHHHHHHHHHHHHHH!!!

http://www.cubed3.com/screens/1221

Sorry, I am still working on the new post XDD

Painting in 2D using 3D (I)

2008-02-11T02:12:00.000+01:00

Hi.

Well, now it is when I realize how annoying is to have a blog like this XDDD, why can't it just auto-update by itself... do I have to do everything :D? For those of you who think that I have abondoned DS homebrew I have something to say about it. I have been developing these moths, but I haven't updated the blog. And I think it is time to solve this. So I am going to explain deeply (well, more deeply than in previous posts) what I have been doing.

Let's begin with 2D stuff. For those of you not familiarized with the DS hardware you can think that it is stupid to write about painting in 2D using 3D. If you are familiarized with opengl then you know how to do that, and it is very simple:
1.- Call glOrtho and create a 2d View
2.- Paint a square with the texture you want to draw

As simple as this. And what happen with the DS then? Does this work? Well... hmm.. no. Why not? There are a few problems here.
1.- If you take a look at libnds implementation of glOrtho you can see the next line: MATRIX_MULT4x4 = divf32(inttof32(2), top - bottom); ; there is a division here. If you want to create a Canvas whose height has to be SCREEN_HEIGHT = 192 then you have a problem, because 2 / 192 = 0,010416666666666666666666... yes, one of those horrible numbers with infinite decimals. It doesn't matter if you use fixed point, it won't solve the problem

2.- Vertices format in DS is 4:12 so... What happen if you want to draw your texture
in (50, 50) for example? (remember that 4:12 range is -8 to 7)

Looking just to this two problems made me think if it was really possible to draw in 2D this way. What most games do is using the 2D capabilities of the DS in another layer and painting the sprites there... but this solution wasn't good for a multiplatform engine; and there is also the sprites size limitations this hardware have.

So, this was the problem. How can you solve this? See you on next post :D

DLDI version

2007-10-16T00:46:00.000+02:00

Ok, I posted it in gbadev.org and now there are a lot of webs talking about it, but not mine :P
Last friday I released a new version of Gea vs coDOS with DLDI support hopping that it will fix that weird bug in R4. But it is still the same. Yoshi's screen appears and then everything freezes. I am still investigating.

Anyway, here is the new version. I also fixed a memory leak in the menu. So it is better if you update your version anyway.

Campus Party 2007, Gea vs Codos and ZEngine

2007-09-20T01:24:00.000+02:00

Howdy, folks ^_^!!! (I had to say that)
Well I have a lot of things to write about this time, and I hope I won't forget anything. I hope you have had a good summer, at less as good as mine, hahaha... well, let's begin.

It all began on Campus Party in Valencia, here in Spain. The previous days to that party I was working hard on my engine so I can develop something for the compo that every year takes place there. Basically, you arrive there and when the competition begins they announces some stuff that may be in your game. Then, you have 72 hours (3 days) to develop whatever you want. This year there were 9 conditions and you had to choose 3 of them. I don't remember the 9, but we chose these:
- Super powers
- Climatic change
- One Button Game
And there they were: "Herederos de Turbo Yoshi" (I am not going to translate this XDDD). We were:
Andrés: working in main characters an tiles
David: working in enemies, skybox and animations
Jorje: main Gea vs Codos programmer
Zalo: ZEngine developer
Sadly for us ZEngine is still in its very early version... so we haven't something like fonts, music, 2d... It was a pity. But I wanted to work with ZEngine anyway and I would had done it even if I had to compete alone :D. We couldn't finish symbian and nds version for the constest and only pc version was available...

Even worse... the stress was... how can I say it.. the last day of the competition I could have killed someone!!! That was really amaziiiiiiiiiiiiiiiiiiingg!!! YEAHHHHH!!! It has been a long time since I worked with someone in my spare time (not at work) and I have to admit that it was very fun (well, now it seems to me very fun, but I really could have killed someone). You can read a blog we wrote during Gea vs Codos developing (that was a contest obligation) and see our state in the last hours, hahhaha... it is here http://herederosdeturboyoshi.blogspot.com/ (it is on spanih... sorry ^_^U)

Well, Campus Party was very fun, but that was supposed to be part of my holidays and I got back even more tired than I left. So after two weeks I traveled to Idaho (in the USA) to visit a friend. I were in USA for 2 weeks and I didn't take my laptot with me (I have bought a laptot!!) because I really needed not to see a computer in days.

After that, I was still tired and still have some days for vacation so I went to the beach in Murcia for a week 8-D. And then... it was time!! Gea vs Codos had to be ported to Symbian and DS whatever it costed!!! And that's what I have been doing these weeks. The game is already working. I have been optimizing ZEngine (That's how I have called my multiplatform 3d engine) also so that it can be more intuitive.

I wanna write about ZEngine, and how it works. I think that my next post will talk about it. Linux platform will be the next supported (it is very similar to Windows, so I don't think it will take me much time).

Anyway, here you have it. Gea vs Codos is already finished. There are a lot of bugs that have to be fixed, but that is not my work :P I am already working on something else (somethign I was already working before Gea vs Codods) you'll see it soon ^_^.

In Gea vs Codos you take control of GEA a super heroin that has to capture the photons trapped in the atmosphere by the greenhouse effect. Co2 particles won't make the things easy for you. So, what are you waiting for? Gea needs your help!!!!
Whatever troubleshotting you have, please let my know. We'll try to fix it. And, now, it's time to save the world

Download Gea vs CoDOS here

Windows port

2007-06-25T00:26:00.001+02:00

It has only taken me a few hours to port my engine for Windows :P. Here it is if you want to try it. You can change between cameras with backspace

Windows port

I am going to upload the source code of all this stuff to sourceforge or something similar soon, so that it can be available for anyone.

3D More and More!!

2007-06-18T00:29:00.000+02:00

Here I am again. I am starting to realize how I hate writing into my blogs XDDD. Well, for some of you that may be wandering what I have been doing I am going to post it as brief as I can.

My engine did't work in some supercards
And I say didn't because now it works perfectly (well, I hope so). The day after I posted the previous entry in my blog, Olaia (one of the many girls we have at work :D) told me that the app I upload didn't work in her DS. First I thought it was because it was pink but after searching a little in some forums I realized that there was a problem with the SC Lite Rumble (or SC Rumble Lite, I never remember the name) that was also pink. It took me two weeks to realize what was goin on, but then I found a post saying that this kind of cards doesn't have an extra memory space where I used to store my application resources. Luckily there is another way to store resources using the bin2s application (whose developer mustn't be very proud of his application, I think, because there is almost no documentatin about how to use it >:/ )
Also there was a problem with the latest devkitARM version (r20) that makes any of your application not to work on that Rumble cards, so I am using r19 instead.

Problem with vertex boundaries
Yes, I wrote about this in the previous post. Vertices in DS must be in 4:12 fixed point format. That means any coordinate of your vertices must be between [-8, 7]. If you have a mesh in your 3d studio max and reescale it so that it fits into a big cube of that dimensions (16, 16, 16 and centered) you basically are doing nothing, because reescaling in max affects meshes's matrices but not vertices. So I changed my exporter to export normalized matrices and reescale vertices (well, in fact it was a little more complicated than that ^_^ but basically that's the idea).
With these changes in my exporter I also fixed a problem with normals (yeah, use always normalized matrices, my friend; you will never regret it XD)

Keyboard implementation
Sorry Ds but this is your fault. Keys interrupts in DS only work with key-pressed but not with key-released!! That made me change the class I implemented for the keyboard. I have also improve it so that it can tell me some screen touch coordinates info. And, of course all these things work in Symbian too :P (well, my n-gage doesn't have a touch screen... so there was nothing to do with that ^^U)

And... something new to show you

This weekend I was bored and tired so I have spent the whole two days at home basically doing nothing XDD. But I have compiled a DS application that will show you something very familiar from a N64 game :D
The application has two cameras you can change using "Select" in DS or "0" in Symbian. The first has been recorded in 3d max (by me, and that's why it looks so weird) and the other one is a new special one that I have developed for DS, you can move it with the touch screen and the control pad (like in the Metroid game)

Download it from here:
Symbian
Nintendo DS

Well, and now everything is ready to make a game. So I think it is about time :)

PD: don't think that I am working on a n64 emulator or somethimk like that. The scene has been exported from one of the n64 emulators using a plugging and then converted to my own format from 3d studio max.

Starting with 3d!!

2007-05-09T00:43:00.000+02:00

Yeah, that was something I wanted to do from the first day... how does 3d work in my DS? is it easy, hard... how far has libnds go with it... haha I have been investigating it during the last weeks. I am still on it, but now I got somethig ^_^

Well, when I started to read doccuments about all this stuff I realized that all this things were very familiar... they were like opengl... they were openGL!!. With a few differences. And then, after a box was painted on my screen I wanted more :D (if you want a box to be painted on your screen, just run one of the libnds examples)

Then I remembered that the last year I was doing something very similar to this... but with my other pocket console, the N-gage. Then I started to think... do I have to do everything I did again? All that stuff was done in Symbian and using openglES wich s very similar to openGl but not exactly the same. It was a pitty!! I had my exporter and a bunch of lines of code displaying animations, cameras...

And then I started to recode all that stuff to be a multiplatform 3d engine extracting everything that was only realated to Symbian and not C++ to a separated files that will have to be coded again for every new platform I wanted to port :D

Once it worked in Symbian I started DS port. Ds is wonderfull, but has some things really annoying. The first is that vertex has to be 4.12 fixed point... that mean that if a vertex goes beyond [-8.0f, 7.0f] there is going to be problems... XDDDD

I could fix this scaling the matrices... but then I had problems with the normals because DS doesn't have something like GL_NORMALIZE (basically what it does is auto-normalize your normals with is very useful when you have scalated matrix)

Well, maybe other day, when I know that someone is reading all this, I will write about all it. But now here is the first demo for DS and Symbian (series 60) of my multiplatform 3d engine!

By the way... do you know what is really good having a mulplatform whatever?... you cannot debug in NDS at all... but you can in Symbian!! ^_^ This is really something wonderfull!!!!

I will post again, when I am bored :P . Still need to fix some problems with big meshes. Here you have the files, if you want them. Source code? Of course... ask for it, like always ;)