Altair Movie Format
The Altair Movie Format (AMF) specification is a digital video format.
AMF has three requirements:
- Enabling arbitrary time coding. Any frame could exist at any point in time (sequentially) and frame rate would not necessarily remain constant throughout the file.
- Portable between PC and UNIX.
- Lossless images stored in an AMF would exactly match the images in the source image files.
Below is a description of how the AMF has evolved through the different versions.
Version 1.0 - Required Arbitrary Frame Timing and Lossless Compression
AVI format was chosen as a base because of its open format and capability for embedding extraneous data (our timing info). No existing coder/decoder (codec) module was found suitable because of the requirements for completely lossless compression and portability between PC and UNIX. Lempel-Ziv (LZ) compression was therefore chosen for its speed and portability.
Version 2.0 - Additional Embedded Data Added
- Title 1 and Title 2
- The first and second lines, respectively, of the header portion of a video page area in HyperGraph.
- Note
- Overlaid on the video image in the frame/time indicator box.
- Milliseconds display flag
- A switch indicating if the time values displayed in the frame/time indicator box should be tagged as milliseconds (ms).
- Support for JPEG Encoded AMF Files Added
- In the AMFBuilder, you can select the JPEG compression level, which allows you to create compact AMF video files. However, JPEG is a lossy format and requires much computing power to decompress. This results in reduced performance compared to LZ.
Version 3.0
- New Storage Format for LZ Compressed Files
- The newly improved storage format gives a performance increase and slightly smaller (10%) compressed output.
- New 8 bit LZ compressed format
- This new format performs quickly (as do all LZ formats) and has very high image quality, although it is not technically lossless due to color quantization. The high quality derives from new functionality in the AMFBuilder, which now has a highly effective color quantization routine to improve the fidelity of reduced color images to their high color source images. Quality can be further increased through the dithering process option, which makes motion video appear (subjectively) identical to high color video formats.