Test Data

Test Data

RGB 8bf p Data

SheerVideo Test Data
RGB 8bf p Data

RGB 8b is the original pixel format of the original published images. The RGB 8b format contains three 8-bit values per pixel, one each for:

  • (R) red
  • (G) green
  • (B) blue
Each of these channels uses an 8-bit full range of 256 possible unsigned values, [0..255], where 0/255 indicates black, and 255/255 = 1 indicates white. The color display on your computer, your game console, your digital television, and pretty much every other digital color display you might have nowadays all use the RGB pixel format. Analog color displays, such as your old analog television set, also end up lighting RGB phosphors on the CRT screen, even if they don't store or transmit the image in that format. High-end video and digital still cameras and video capture boards capture the image in RGB format too. The reason that the RGB format is so ubiquitous is that it's the closest digital representation we have to how the human eye captures the image.

Note that SheerVideo also supports the RGBA 8b format, which contains an 8-bit full-range alpha (A) component at each pixel in addition to the red, green, and blue components. Alpha represents the coverage or opacity of the pixel, which is used during compositing. An alpha value of 0/255 indicates that the pixel is completely transparent, while 255/255 = 1 indicates that it's completely opaque. When composited over a background image, a transparent foreground pixel lets only the background show through, while an opaque foreground pixel shows only the foreground.

To test the SheerVideo RGB 8b QuickTime codec, we repackaged the RGB images as a 24-frame QuickTime movie by importing the TIFF source image sequence into Apple's QuickTime Player Pro and exporting the sequence as a self-contained QuickTime movie using the None compressor set to Millions of Colors. We set the frame rate to 30 frames per second in order to easily relate the performance to standard definition video.

Inside QuickTime, the uncompressed RGB codec is known as kRawCodecType or 'raw '. On the Mac, the RGB pixel format is designated as k32ARGBPixelFormat or 32, with four components in the order ARGB, whether or not the alpha channel is actually used. In the less cooperative PC world, the components can be in just about any order, with the most popular pixel formats being RGB, BGR, ARGB, BGRA, ABGR, RGBA.

At three byte-size components per pixel, the intrinsic uncompressed size of each of these RGB frames is 3 x 768 x 512 = 1,179,648 bytes. Note that on the Macintosh, a raw RGB image is actually always represented with 4-byte pixels, as if it were RGBA, even though the 4th byte is unused. Thus to compare SheerVideo RGB 8b images to actual raw RGB 8b frames on the Mac, multiply the uncompressed sizes and the compression powers listed here by 4/3.

By the way, watching a slide show loop at 30 f/s is way cool. Unlike those subliminal tachistoscopic commercials that are banned in the U.S., or the single-frame Easter eggs in some old animations, where the frame is registered subconsciously but flits by too fast for people to consciously perceive it, in a high-speed loop of unrelated slides you can actually see each frame quite clearly if you concentrate on it.

previous pagenext page

Copyright © 2003..2011 BitJazz Inc. All rights reserved.
Site design by BitJazz Inc.