ssr is a reader for the bi image format.
ssw is a writer for the bi image format, which takes desktop screenshots and saves them as .bi image (which can later by read by ssr).
ssmeta shows the header data for a bi image.
bi stands for blosc image.
ss stands for screenshot.
A bi file/image is just 0) a 64-byte header followed by 1) pixel data.
-
The header is uncompressed.
-
The pixel data is compressed by blosc.
Currently supported pixel formats:
- RGB565 (16-bit pixels, 3 channels)
- RGB888 (24-bit pixels, 3 channels)
- RGBA8888 (32-bit pixels, 4 channels)
- BGR565 (16-bit pixels, 3 channels)
- BGR888 (24-bit pixels, 3 channels)
- BGRA8888 (32-bit pixels, 4 channels)
Note that the RGB888/BGR888 formats are packed. That means that pixels are NOT 32-bit aligned, by 24-bit aligned.
If you want a 24-bit image that's 32-bit aligned, that counts as a 32-bit image.
bytes [0x00 .. 0x08): magic
bytes [0x08 .. 0x10): version
bytes [0x10 .. 0x18): image width
bytes [0x18 .. 0x24): image height
bytes [0x24 .. 0x2c): image format code
bytes [0x2c .. 0x40): unused
0: UNKNOWN
1: RGB565 (16-bit pixels)
2: RGB888 (24-bit pixels)
3: RGBA8888 (32-bit pixels)
4: BGR565 (16-bit pixels)
5: BGR888 (24-bit pixels)
6: BGRA8888 (32-bit pixels)
ssr can currently read/render 16-bit images on a 16-bit X server, 32-bit images on a 32-bit server, and 32-bit images on a 16-bit X sever.
ssw only considers 2 color depths: 16-bit and 32-bit.
Notice that what the X server calls 24-bit color is actually 32-bit, because each pixel occupies 32-bits anyway.
The header magic is 0x7a5ac811bd4476be, if you write it in BIG ENDIAN (ie. if you read the number from LEFT to RIGHT, then the most-significant nybble comes first).
You can use the header ordering as a proxy for the endianness of the header (and maybe the entire image).
Eg. if the first byte in the file is 0xbe, then the header is probably little endian (least-significant byte first).
Eg. if the first byte in the file is 0x7a, then the header is probably big endian (most-significant byte first).
ssr: zoomssr: panssr: 16-bit to 32-bit image conversion