parallax

“3D clip” made by myself just for fun, strictly speaking, this is not 3D video but rather a pseudo – 3D effect, parallax, (tiếng Việt: thị sai) created totally from still images (see on the left). My video is partially inspired by this excellent CSS – only parallax. It’s quite easy to figure out how to acquire this, but let read about the definition from wiki.

This is also a popular technique in game programming, some early (and current) games choose to reduce a full 3D scene to several layers, some use parallax – mapping to acquire a pseudo – spatial effect, all helps eliminating excessive 3D computations. Clip made from random still images captured at Khe Gà, Phan Thiết.

3d graphical user interface

The code is written in C++ base on Clutter 1.2.4 library. It shows basic widgets: tabbed container, scrolling container, list box, combo box, check box, edit box, button, label, radio button…

1.   If you just want to add some eye-candy effects with the cost of a much more complicated GUI with “unorthodox” ways of representing information, that would be a very bad idea! The best way to understand graphics is not digging into math books, but rather computer graphics history, the ideas behing the NeXT computer, PostScript, PDF, NeWS & X windows system, Quartz & Cocoa…

2.1  Hardware with only a framebuffer and some blitting, blending devices.

2.2   Hardware with some 2D vector calculating abilities.

2.3   Hardware with different 3D calculating abilities supported in their GPU. Take a look into the Mesa3D code to understand how the graphics pipeline hardware & software stack works.

3.   I suggest that a graphics developer should try coding a complete module by himself from scratch, for example: orthographic or perspective projection module, bump – mapping implementation, light ray – tracing model… to gain the working experiences himself as reading theory from books won’t be enough. Eventually game & graphics requires lots of mathematics understanding & coding skills, and are not for average coders who only got used to “formal” (shallow) software methods, out – sourcing or such things!

his is my personal project in which I tried to evaluate some new ideas and concepts in using 3D techniques for GUI (Graphical User Interface). It’s quite tiny indeed (about 6000 LOC – line of code), all written by myself in about a month (mostly in my free times at weekends). So please don’t blame me on some un-completed or un-polished features, they are just for testing the ideas only. I would try to examine the trend of applying 3D graphics to GUI, but first, let have a look at the video below.

1.   The first thing to consider is not 3D engines or hardwares, it is about usability. Many traditional (2D) GUI out there are already complex and obstructive. GUI is about information presentation and presentation should be really simple and clear so that even my grand mother can understand and use. I’ve been seeking, trying to explore many new ways invented to represent information in 3D (or 2.5D) space, and I could say only a very very small percent of them could make a usable value. Users have long been familiar with 2D, and yet 3D hasn’t been very much persuasive.

2.   3D hardwares can be categorized into 3 groups (as listed on the left), and the 3D engine can be configured to off-load calculations to hardware. With the 1st type of device, only bitmap operations could be off-loaded. With the 2nd type of device, many drawing operations (the path_xxx functions) can be off-loaded. The 3rd type of hardware is most valuable since we can have its finish for ourselves a lot of work.

It’s important to understand the target device our software stack would be running on. If it is an out-date machine with just some blitters, we should use paint-like (aka: bitmaps) operations, while on most modern PC, draw-like (aka: 2D vector) operations are more encouraging. While bitmaps may offer nicer and more customizable GUI, the cons is it lacks the scalable ability that vector has.

3.   The last is about software implementation. Many of the graphics concepts are first introduced in software, which then embedded to hardware, which then standardized by software (like OpenGL or DirectX) again. Thus, many elements of a graphics pipeline is the production of a long history of interactive evolution. A graphics pipleline is not a general framework, it depends on very detailed, in-depth techniques to be operational in real-world application.

Game & graphics are the domains where most formal software methods would easily failed. 3D graphics developers should have good knowledge on graphics in general (bitmap and vector drawing), geometry, linear algebra & discrete maths, deep understanding on data-structures and algorithms… are strong pluses, advanced tips & tricks in coding and optimization is a must (game & graphics programming has always been a hell of tricks from the age of dawn) and finally good understanding on 3D techniques (model, scenegraph, projecting, shading, clipping, lighting, effects…)