P3D Re: JPG and destruction

[Jonathan Gross <jonathan.gross@xxxxxxxxxxxx>]

I wrote (and subsequently edited to make more understandable after
reading a number of comments):

> Sometimes, after quantization, the spatial cue gets
> converted into a color difference.  That's why JPEG may work great for
> monoscopic images, but is a poor choice for electronic stereoscopic
> images.  JPEG retains [the size in pixels of] the original image, but
> homogenizes the colors of adjacent pixels to achieve high data
> compression ratios.

Dave Spacey wrote:

> The bitmap is veiwed as a 2D waveform, and transformed into a set
> of cosines which sum into an approximation of the original pattern. Then
> the least significant frequency components are rounded down to zero. All
> those zeros compress well at the next stage.  The quality/compression
> setting affects how significant a component has to be to escape being
> dumped in that fashion.

Thank you, David, for pointing out the fundamentals of the method. What
may not be obvious from your restatement of the process, however, is
that the subtle horizontal boundaries which are cues to our visual
system and provide the physiological stimuli that are MAJOR COMPONENTS
IN THE STEREOSCOPIC EFFECT are encoded in JPEG (and JPS) as MINOR HIGH
FREQUENCY COMPONENTS, the first ones to be "rounded down to zero", i.e.
fuzzed out!  FINE for monoscopic images, BAD for stereoscopic images.

As for the argument, "what else is there": I don't accept that there are
no alternatives because several have been mentioned here by other
members of P3D.  I've got my own favorite, but after the pen lashing
that I took last time I tried to explain this, I'll just wait until
somebody else ventures forth; then I will try to support them. I hope
that I have been supportive of you.

Jon Gross


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