Mailinglist Archives:
Infrared
Panorama
Photo-3D
Tech-3D
Sell-3D
MF3D

Notice
This mailinglist archive is frozen since May 2001, i.e. it will stay online but will not be updated.
<-- Date Index --> <-- Thread Index --> [Author Index]

P3D Re: RE: JPG and destruction


  • From: roberts@xxxxxxxxxxxxxxxxx (John W Roberts)
  • Subject: P3D Re: RE: JPG and destruction
  • Date: Mon, 20 Apr 1998 14:35:05 -0400


>Date: Mon, 20 Apr 1998 04:44:59 -0600
>From: Dylan The Hippy Wabbit <spacey@xxxxxxxxxxxxx>
>Subject: P3D RE: JPG and destruction

>John Roberts answered Ole Hansen saying:-
>>>> JPEG retains the spatial resolution of the original image....
>>>It certainly does not !
>>I agree. Whatever the process is called, some spatial information is lost
>>in the process of lossy JPG compression.

>What it does is transform the Red-Breen-Blue description of the colours
>into a Grey-Red-Blue one.  The green is implied by the difference between
>the greyscale and the combined red-blue values.  Then, based on good
>psychological principles, the greyscale retains its full resolution but
>the colour components are reduced by half.

Is that based on the idea that spatial resolution of the human visual system
is much better in green than in blue or red? (I recall being told that blue
resolution is about half that of green, but I didn't recall a specific
number for red.)

>Yes, this will be more obvious in stereo than in mono.

Of course stereo has two challenges in this respect - depth is encoded in
fine horizontal details, and images are often magnified (when used in viewers,
projection, etc.).

>No, using the minimum compression/maximum quality setting will make no
>difference.  That affects a later stage of the compression.

So what does the quality setting affect?

>BTW, I've cross-posted to sd-3d since I believe we'll be asked to move
>across if we go into any more detail.

It never got there - are you sure you're properly subscribed and using the
correct address? (Perhaps calcite inhibits cross-posting?)

>Dave Spacey

John R


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