Blogs - linfosoma

Lately I've been playing a few games I got on Steam sales that have tried doing things differently than everybody else, I got them mostly out of curiosity and after playing them I cant but wonder if the risk was worth it (spoiler alert: it wasn't!).

LA Noire:

I guess you could consider this game an adventure game with an open world and some questionable action thrown in for good measure from time to time.
LA Noire got its hype from the facial animation technology used in it, being a character modeler/animator myself I felt compelled to get it to figure out what the fuzz was all about.

As you can see in the picture above the actor's face is capture by several cameras surrounding him, then these images are merged (my guess) into a single video file and projected on the 3D model's head creating a very lifelike effect provided that you dont look the model too close (the video has a limited resolution, so you can forget about making close up shot of the character's face).
These "animations" also seem to interact with the game's lighting which makes me believe that they somehow transformed the video into a normal map as well (my guess is that it was converted from the base video, but there's always the possibility that the depth information was extracted using infrared cameras, that would be kinda rad actually).

The final result is convincing faces pasted onto bodies that sometimes fail to match this expressiveness. You also get to see some hilarious things such as the character's hair flickering due to video compression and whatnot (the hilarious part being the thought of a completely static body part using needless video/texture memory due to the engine's way of doing things).
On a final note, the character's face is mostly flat and motionless. The eyes for example dont have 3D eyelids, they are just a curved shape with the video of real eyes projected above. The only part that appears (surprisingly well) animated on 3D is the mouth (though you can sometimes see the character's tongue and teeth, which are also a 2D video projected on a flat surface, so it seems that only part of the lips and the jaw are animated).

As for the game itself, it's ok I guess. It's mostly entertaining, even though it features the most incredibly unlikeable main character ever featured in a video game and the game's plot and so called "pacing" are mostly terrible. But meh, some characters look pretty cool sometimes!

Veredict: If you think that the idea of filming thousands of videos, each for every one of your character's performances and for every single NPC in the world, and then jamming it into the game considering the resolution limitations and the amount of disk space used is a good idea then well, this horribly impractical technology is for you!

Rage:

Oh boy, are you ready to take the cake on the "horribly impractical technologies" competition? Rage uses the latest form of the Megatexture technology and it's quite possibly the most puzzling thing I've seen in a while.

The latest iteration of id's proprietary engine manages to be even more absurd than the previous one (id Tech 4, the one used in Doom 3, the game that made every single light dynamic and had it cast stencil shadows in a world where indirect lighting is nonexistent). The idea is "simple", make an incredibly detailed scene, and then bake everything (and I mean everything) into a single, giant texture. The result is that all the lights, specular highlights and even normal maps are baked into the texture (and therefore completely static).
The good: sometimes the game looks awesome provided that you look at it from a safe distance. Since the lighting is baked you often get fancy indirect or even bouncy lighting on the maps, it's pretty.
The bad: every single object has a unique texture, if you see the same asset used twice each one has a different texture file even if they look the same, resulting in a game that is 20 GB large yet has the overall texture quality of a game made in 2000 (Im mostly certain that the textures in Deus Ex looked way better than this, at least that game supported detail texture maps).

But it gets worst, as I said previously everything is baked, which is another way of saying that nothing is interactive. The light is completely static, this results in some things like rocks looking ok provided you dont get near and other things such as anything made of metal looking hilariously bad (baked specular highlights, in a game made in 2011, who thought this was a good idea?).

If there's one thing I have to praise, is that as usual id's art direction is fantastic. The characters are very interesting, levels have lots of style and scenes often look very cool from a design standpoint.
But after all the hype I cant help but feel cheated by a game that's technically no more advanced than the PC port of Resident Evil 4.

But hey, at least it's pretty fun to play. Simple mechanics and plenty to do. Ironic to think that an id title finally managed to have better gameplay than graphics but there you go, the world is full of surprises.

Veredict: Oh god, I hope they never use this engine again.

I can't really talk about any of the things I've been doing lately, but I really feel like things are moving along.
It's incredible to look back and think off where I started (and what a noob I was, I didn't even know how normal maps worked back then!), still got a long way to go, but I finally feel more confident and capable of taking on things when it comes to game development.

Depth has given me a lot, and Im really grateful of being a part of what I believe is something completely different (and yet so strangely solid and fun to play). I really hope to be able to see the project come to life just to watch the internet pick it apart (lol!).

My character modeling skills have really improved! Im still terrified to post anything on the Polycount forums, Im way behind that sort of level, but it's getting there.

Im also doing some hard surface modeling and texturing, it's fun! I like how building things is like a puzzle, you have to really think about how you are going to do things to maximize the UV space you are going to work with.
I also learned how to rig and animate in Maya. It was surprisingly easy moving from Blender to that, I mean, the interface and tools are different yes, but once you know the basics it's easy to figure things out.

My next goal will be to learn those things in 3D Max, but I still have time for that. Blender's animation tool kick ass, and the rigging tool is still the best I've seen but man, the lack of any real (i.e. non-hacky) smoothing group support really pisses me off. Blender still got a long way to go in some sense.

Anyway, I think that's all for now. Godspeed.

This week we are releasing a bunch of new media. The game's finally doing great, most of the most important bugs are fixed and things are feeling nice and balanced.

One of the new things we are featuring is the Captain of the Salty Rose (the diver's boat):

The story behind this character is that we talked about having some sort of captain on the boat during one of our meetings. The idea was to have an NPC that wouldn't be playable but would be there and may or may not interact with the divers too.
Nobody knew at the time, but I really liked the idea and secretly started working on it on my free time.

This is the first model I've done completely on my own (on the divers, Alex helped me out with the texture work and Myles did the shaders), which is something Im particularly proud of!

Making this character was incredibly addictive an fun, Im guessing that in total making the whole thing probably took me about 8-10 hours, which when you consider that the divers took about a month each really shows that I've really learned a lot working on Depth (still got lots of more learning to do!). For that Im really thankful.

(These two images are taken directly from UDK's editor)

I learned a lot about texturing (something I used to hate and now love), and also about UV mapping and a few new rigging tweaks (he's got a fully functional face with teeth and everything). I also spent a lot of time working on a specular map for him, which is something completely new to me, and I think that at the end it paid off.

As usual, I made a simple reference mesh and sculpted as much as I could in Blender, later exported to ZBrush to add finer detail, retopologized the model and finally I imported it back to Blender to work on the UV mapping, rigging and animation.
Im doing the smoothing groups in 3D Max for the moment because UDK doesnt seems to like the ones I make in Blender.

I still got some work to do on him, you can see a nasty UV seam on the wrist and I could probably add some more detail here and there.

As a nice little tip I'll drop for ya, I tried something new with the rigging in the eyes: I rigged part of the eyelids to the eyes themselves, so when they move the flesh around it moves too creating a very organic effect.
This definitely enters into the real of "completely pointless details that nobody ever is going to notice", but that's my favorite kind of thing, so if you are into it, try it out!

That's it for now, our beloved Myles Lambert updated our website with tons of details, so feel free to check that out here:
Depthgame.com
Alex-

Work on Depth has been very intense for the last few months. We are now having playtests every day, which is incredibly useful since you can pick a lot of weird things and then work on them to fix them. It's been very useful for me since I was able to improve the animations a lot to a point where they look good but dont get in the way of gameplay.
There's also been a lot of tweaking and experimentation with different concepts.
The result is that the game's finally starting to feel really polished and balanced (something really hard to achieve in such an assimetrical multiplayer game).

It's great to be part of such a great project, and now that my apparent chronic depressive breakdown of the week is over Im also back to work! (lol).

Alex (our project leader) posted some nice info about what the game's about in our forums here:
Depthgame.com

It's a good read if you are interested in knowing more about the game.

Finally, for a while now I've been thinking about making a series of video tutorials about character modeling and animation from Blender to UDK. Is anyone insterested in such a thing? Let me know.

I've been working with Blender for a while now so I wanted to talk about the creation of models for UDK.

First of all, Blender is a free, open-source 3D modeling application that can be acquired at Blender.org.
There's several versions of Blender available, the stable version (the one Im using) is 2.49b and there's a new alpha version (currently 2.56) with new features such as particle simulations and a more advanced sclupting brushes.
That said, I dont quite like the new version just yet, it's nicer looking yes, but I've seen many people with issues exporting files and most of the new features are not really things that I personally need to use (also, most tutorials are made for previous versions).
If you are looking to learn how to use it, there's tons of great video tutorials at blenderunderground.com

If you are planning on making skeletal meshes (either characters or weapons), you'll need to install the psk/psa export plugin found here.

It works really well, any animations you make will be exported into the same .psa file (remember to change the name of the animtion from the default "Action" to something else).

The only real issue with this is smoothing groups, which are handled differently that in other tools like 3D Max or Maya. Here's a rundown of how it works, and the possible solutions link (I strongly suggest you read this).
The method Im using is basically splitting the mesh into the smoothing groups I need (using the Y key) and then rezise each one to 0.999.

It's a long and annoying process but it prevents UDK from merging models back together and in my experience there are no visible seams on that scale.
The result is exaclty the same that you would get with Max, so it's all good.

For static meshes, there's also an ASE exporter plugin but it sucks, just export your mesh as an OBJ and make the smoothing groups in Max or some other modeling tool (sorry, I guess you were expeting a better comment for this one!).

Finally, a few general random tips:
-When unwrapping a model remember to set the UV method "conformal" instead of "angle based" (default setting, tends to distort the UVs), you can also set if you want each UV island to be separate from each other, which might make texturing easier.
-You can grab, split, merge and resize faces or edges in the UV editor. This is helpfull to fix texture stretching and making sure you have a uniform UV distribution.
-Once a mesh is unwrapped, any face you duplicate will have the same UVs, which is useful to save texture space in symmetrical meshes.
-You can edit the mesh in any way you want after rigging. Any new geometry you add will need to be weight painted.
-The "mix" brush in the weight painting tool is godlike, very useful for cleaning messy areas and reducing texture stretching when a mesh is affected by bone movement.
-When rigging, if two bones have 100% control of the same area it will bend in a L shape, the smoother the weight painting is the smoother the resulting deformation will be.

So...yeah, feel free to ask if you want to know something.

We got in fifth place on the indieDB competition for upcoming indie game!
It's a little bizarre considering that we haven't released an update in a while and we really didn't promote the game for this contest, but hey! Thanks dudes!

Here it is:
Moddb.com
Also nice to see that Flayra (a dev from Unknown Worlds) is interested in the game as well, Im glad they won, the Natural Selection 2 beta is solid and visually impressive (even though I get my ass kicked all the time).

Damn, this is a good year-end. I just got a new rig built, which is like a little monster and we have been making some really big progress with Depth (which reminds me, I really need to get those LODs finished soon!).

2011 is looking up to be a great year!

PS: Steam sales are going to ruin me.

So, have you been playing your favorite game and stopped to think "hey, this is easy! I could do my own game!" and decided to start your own project?
You should probably recondiser.

Im no way an expert developer or anything like that, Im just getting started as a freelancer with the hopes to get good enough to join a decent indie team and stick with them, so bear with me for minute.

I often check what other people are working on and I also check recruitment threads, sometimes to see what kind of tools they are using and sometimes out of simple interest to see what's on the market.
This is where things go south.

First of all, video game devolopment is NOT easy, so get that idea out of your mind right now. Making a game is more than "having a good idea" (which is usually something regarding zombies in a "completely unique setting!"), in fact, I will go as far as saying that ideas are useless.
You heard me, ideas by themselves are worthless, what matters here is facts. You can have the best idea in the world (realistically, it's still probably just a zombie game) but unless you can turn it into something real it's just that, and idea.

Ok fine, so you decided to turn your idea into facts, you posted a recruitment form. That's a good start right? Wrong.

Most recruitment threads I've seen are completely out of touch with the reality of video game development. You are probably exited about your project yes, but suck it up and understand that you are going to be managing people's time, people who might have different interests and goals than you. So when you type in things like:

"Our project is a vision which will provide maximum immersion and set new standards for the co-op survival horror genre in delivering nail-biting tension and ulcer-inducing suspense"

I go bananas. And yes, that's a real quote, and it's terrible.

Let's put it this way, video games are composed by hundreds if not thousands of elements combined. You often spend several days working on an asset that nobody will notice, like achair or something like that, only to realize the model looks bad in-game and needs to be redone. You've just lost several days of work for something nobody is going to care about and you have to suck it up and get on it. Because if people notice the chair it's probably because it looks odd.

So, when you are trying to get people to join you, you should first let everyone know exactly what they are getting themselves into. That post I quoted didnt even have contact info, that's simply pathetic.
Knowing what the project is about, and what you as a team leader can do can tell me a lot about whether Im going to be wasting my time with you or not.

How big is the project?
How many people do you need?
What experience do they need? Which tools will they have to learn?
What are your goals?

Also, please try to make your project realistic. Yes, we all have an idea about making the best game evar but sometimes we need to understand that this is simply not possible. It's better to start small. Simple projects can be much more complex that they seem.
I've seen people that wanted to make MMOs in UDK, and let me tell you right now than that's completely retarded. Realtime Worlds couldn't succeed with an MMO that had a budget of over $100 millon and over 300 developers, chances are you wont be able to either. Especially on an engine that's not ready for that (there's a huge difference between UDK and a full Unreal 3 Lincense).

The same applies for people who try to acomplish too much. Trying to make a game with single player, multi player AND co-op components is an insult to everyone's intelligence. Do you even know how to code AI? No? Then stick to what know.

Last but not least, I've always felt that only coders should be team leaders. They are in average more organized, have more realistic goals and know what can be accomplished and what cannot.

Im an artist, which means I always have these crazy ideas of super complicated systems that are probably going to end up being useless or annoying, which is why I choose to follow orders from someone who knows what he's doing.

And that's the point of my post really, if you are starting up on game development and you are serious about it, you should just try to stick to your guns. Find out what you are good at, join an existing project and improve your skills.

There's a lot to learn really, not only about development but also about dealing with other people, so stop posting about the incredible "cinematic experience " you want to give me and start posting about the real working experience you can provide.

Please!

There's a new update posted here:
Moddb.com

Lately I've been working on the shark's jaw that will be used when playing in first person view (both model and animations).

Here's a shot of the Mako's jaw:

It all makes sense when you are actually playing the game.

That's it for now!

So, lately I've been working on some environment meshes (I made a car!) but that is not really a whole lot of work, so I've decided to put my time into use and make some new characters for personal use (possibly an animation demo reel, which is something I don't have and definitely should).

I wanted to work on a concept I had for a game a while ago called "Runner". The idea is simple, one player is the runner, and other 3 play as hunters which, you guessed it, must hunt him down.
Hunters have short range weapons such as spears and swords, and the runner has no weapons. What he has though, is the ability to set magical traps which the hunters can't see.

So it's basically a cat an mouse game, the runner sets a bunch of traps (which have a spawn limit so you cant spam) and the hunters must try to guess where these are, avoid them and kill the runner.

Visually, I wanted the game to be sort of a mix of medieval fantasy and cyber-punk (I am aware that the combination doesn't makes sense, but it kinda does in my head).

As you can see, I originally made a simple sketch of the Runner being a white character and then decided to make him black, not sure why but it's fun stuff to make these models.

So, with the concept more or less done (even if it's rather simple) I decided to make a mesh.
So far I've been working on the general body proportions in Blender and when Im done the plan is to take the model into ZBrush and make a nice high poly model. I find Blender's sculpt tool MUCH better for low poly modeling and the camera controls don't make me wanna kill someone so I guess that kinda helps too :)
Anyway, here's what I've have so far after approximately 3 hours of work:

And another:

Im not exactly sure if Im ever going to do something with these models and ideas, Im definitely not interested in being the project leader of my own mod/indie game and Im no coder so this is all I can do, but hey! It's fun and it's good practice.
Im still surprised how much I've improved my workflow, I can't help but think that my first characters (Depth chars) took about one month each. Yikes!

So yay for me! :)
Anyway, I think that's all for now, if I ever continue working on this I'll post more screenshots.

I've been bored with the zombie genre for a while now, maybe due to overexposure or maybe because of how deranged the whole thing has become with them running zombies and endless bullet streams.
So, I decided to play the original Resident Evil.

I've got the Gamecube remake, Im playing it using an awesome emulator called Dolphin which works very well.
First of all, I've gotta admit it Capcom, you make some pretty games; this was released in 2002 and 8 years later it holds up REALLY well. The acting sucks, up to the point of being hilarious, but I guess there's no surprise there.

The game is puzzle oriented, more that I remembered -up to the point of being silly at times-, but puzzles still manage to be fun and keep you under the illusion that you are trying to both survive and solve a mistery.

What surprised me the most though, is the difficulty and pace, which are simply fantastic.
The game is very hard, and it's probably the only survival-horrol game to live up to it's name.
I've spent a large portion of the game with (Im not kidding) 2 bullets, that's it. There's practically no ammo, some enemies can take up to 11 shots to go down and when you find some it's the most happy event in your life. It gets to the point where I used a key item to get 6 shotgun shells, that's fucking hardcore.

Im still very much in the beggining of the game -mostly because modeling rocks is surprisingly fun- but Im amazed with the experience so far (sometimes it's great to forget about some games).
Dodging zombies, avoiding combat, these are things that became alien to me with time, and it's nice to have them back.

Even saves are limited, so you cant just spam the quick save key every time. At one point I ran out of Ink Ribbons (used for saving the game), each enemy encounter from then onto when I found some became terrifying.
You die, and you are fucked. you waste a single bullet, and you are dead. Happy times.

To make things worst, this remake introduced "Crimson heads". You see, killing a zombie is not enough, you have to burn it's body. Otherwise they randomly come back to life as a Crimson Head, a pissed-off zombie that runs, slashes AND throws up acid vomit on you (yes, both at the same time).
Did I mention that in order to burn the bodies you need an item that: 1-it's extremely rare and 2-takes 1 of the misserable 6 inventory slots you have.

Ahhh yes, the inventory; 6 slots, weapons and ammo use different slots, so do individual keys; isnt life pretty?

I love this game.