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.
|
|
Re: Anaglyphs
- From: P3D John Ohrt <johrt@xxxxxxx>
- Subject: Re: Anaglyphs
- Date: Wed, 20 Aug 1997 23:57:56 -0600
P3D Webmaster wrote:
> I don't believe anyone has made the point that images presented on the
> web
> are formatted as jpg or gif. If you compare a tiff file of an anaglyph
> to a jpg version derived from the same file, there's a very noticable
> difference in the stereoscopic effect between the two. There seems to
> be a sacrifice in sharpness and depth information when compressing and
> decompresing anaglyph images. Unfortunately, the nature of the web
> makes small, compressed graphic files a necessity.
>
For a discriminating viewer of a photograph, you really need 36 bit
photometric data, i.e. 3 channels at 12 bits each. This has been
repeatedly demonstrated by test. The so called "true color" of 3 channels
8 bit deep has always been a compromise, primarily due to cost, just as
most displays are limited to 3 color channels primarily due to cost.
What the heck, most of us still view anaglyphs primarily because of cost.
While some people will argue that the en/decoding of .gif is lossless,
this is only true of the actual en/decoding technique itself. The
compromises that have to be made to fit a image of photographic quality
into the 256 color table of a gif are very lossy for all but 8 bit mono
data. In general, the .jpg image is compressed in a lossy fashion.
Please note that lossless compression is possible under the .jpg spec and
was used by NASA to distribute Clementine data, regrettably none of which
was 3D. So .jps can be used to distribute lossless hi-res data.
When ever lossy compression occurs, nor only is data lost, but "noise" is
injected into the image. Since and stereo imaging system is more
sensitive to noise than a 2D image, the degradation appears much sooner.
While small compressed graphic files are often an advantage for many web
applications, no one has ever suggested that they are mandatory. Quite
the opposite, .pbm and .png are specifically targeted to pass high
resolution files. While .tif suffers from portability problems, you won't
have a problem with the most common formats, especially if you avoid
compression.
My experience is that a zipped .tif file using optimum compression option
is more effective (marginally) than the built in compression options.
Also, most can run real-time hardware compression on the modem link, via
v.42bis/lap-m/arq so that even uncompressed files can be transferred with
little loss of time. Furthermore uncompressed images load faster and big
hard drives are real cheap these days.
Higher quality photo sites often use lossless .tif or uncompressed .tif
for those who really want a high res copy and smaller .jpg to serve as a
"super" thumbnail.
Remembering the days when I downloaded 2 Mbytes images via 2400 bps modem,
downloading at X2 (typically 44 kbps or higher) is scarcely a burden if
the image is worth downloading.
I am thankful for the 3D sites that are provided, even if some don't
provide high res data. If more would provided such data, it would be
nice. But many of these sites are provided at the personal expense of the
maintainer, and many such sites are charged by the MBytes of storage
required. If the site provider is financially limited, then I will take
the best available.
After all, this is a hobby and most of use have some limits on what we can
spend. :-)
Regards,
--
John Ohrt * Toronto * ON * Canada
------------------------------
|