P3D Too much info about JPEG

[Tom Deering <tmd@xxxxxxxxxxx>]

Regarding JPEG compression:

>>So what does the quality setting affect?

Oddly enough, I teach a graphics course at Stern College in New York.
Tonight's course covered JPEG and GIF encoding.  Lucky huh?  Here's more
than you want to know about the JPEG process:

1. Transform the image into a suitable color space. Usually this means that
the image is split into a brightness component and a color component.

2. (Optional) Downsample each component by averaging together groups of
pixels.  The luminance component is left at full resolution, while the chroma
components are often reduced 2:1 horizontally and either 2:1 or 1:1 (no
change) vertically. This step immediately reduces the data volume by
one-half or one-third. In numerical terms it is highly lossy, but for most
images it has almost no impact on perceived quality, because of the eye's
poorer resolution for chroma info.

3. Group the pixel values for each component into 8x8 blocks.  Transform each
8x8 block through a discrete cosine transform (DCT).  Thus you now have
numbers representing the average value in each block and
successively higher-frequency changes within the block.  The motivation for
doing this is that you can now throw away high-frequency information without
affecting low-frequency information.

4. In each block, divide each of the 64 frequency components by a separate
"quantization coefficient", and round the results to integers.  This is the
fundamental information-losing step.  The larger the quantization
coefficients, the more data is discarded.

5. Encode the reduced coefficients using either Huffman or arithmetic coding.
Notice that this step is lossless, so it doesn't affect image quality.

6. Tack on appropriate headers, etc, and output the result.  In a normal
"interchange" JPEG file, all of the compression parameters are included
in the headers so that the decompressor can reverse the process.  These
parameters include the quantization tables and the Huffman coding tables.

In summary, the image is divided into blocks.  The pixels are analyzed, and
blocks  that have more detail are crunched less; blocks with less detail
are crunched more. Add headers so the machine at the other end knows how do
decode.

I don't know anything about stereo photography.  This stuff I know.  Hope
this helps.

Tom



------------------------------

End of PHOTO-3D Digest 2688
***************************