>From: petertje@xxxxxxxxxxxxx
>
>I'm a fan of the following way of thinking. When you enlarge a picture that's
>been taken with film, there's a certain threshold where the grain starts
>showing up. If the film is any good, this appears pleasing to humans because
>it's irregular. When you do the same with digital, it appears unpleasing
>because of the regular raster. No in-camera or post processing can change
>this.
If you re-sample the image with the proper tool, you don't so much get the
"regular raster" pattern, but you begin to see sensor noise. To me, this looks
a bit like film grain.
There are also various ways of simulating film grain. I sell an image that was
taken with ASA 800 film, with a clear blue sky. In this case, the grain was
objectionable. So I sampled the sky at horizon and zenith, and filled the sky
with a gradient. But then it looked artificial, so I added some noise back in.
But I do agree that film grain has its own quality that can be quite pleasing,
and that would take considerable effort to simulate digitally.
>Another really bad thing is that some cameras store their pictures in JPEG
>format.
Yes. But it is generally a reasonable compression ratio. Once you get them off
the camera, the FIRST thing to do (assuming you want to mess with them, rather
than accept them as-is) is to convert them to a non-lossy format.
> > >Film still allows more flexibility in how the
>> >original image is recorded...
>>
>> Again, we must agree to disagree.
>>
>> Any given film has ONE way it reacts to light. A digicam has an infinite
>> number of ways it reacts to light. You haven't convinced me otherwise.
>
>I think the actual sensor has only one way that it reacts to light. What
>happens afterwards (in the camera or on PC), is like what happens in the
>darkroom.
Well, we're just playing with semantics here.
> > This is not rocket science. It is not my opinion. It is "Moore's Law,"
>> which has held true for 50 years of semiconductor technology advancement.
>
>I'm not entirely convinced Moore's Law holds for CCD. I've heard the story of
>how we'll all be using digicams within months from now, for far too many
>times. In the mean time, estimates are getting longer (2-3 years ?)
I've been on-track with my digicam predictions since my Apple QuickTake in
1996. At the time, I wrote (possibly on this very list) that digicams would
reach price-performance parity in about ten years. That's about 2-3 years from
now.
So let's see... 2003 minus 1996 is 7 years is 4 2/3rds times Moore's Constant
(18 months).
4 2/3rds doublings is a factor of about 25.
The Apple QuickTake of 1996 was 0.3 megapixels (640x480)
Times 25 is 7.8 megapixels.
Current high-end digicams are running 6-12 megapixels. I guess Moore wins again!
And in 2-3 years, we'll be seeing 12 megapixel cameras priced similarly to 35mm
film cameras (price-performance parity) and high-end jobbies at 24-48
megapixels that rival medium format film.
>Reminds me of the debate on how quickly computers will be able to think just
>like human beings.
You have to be careful who you listen to. The pundits on late-night talk shows
will say anything to sell their latest book.
As I've shown, my estimates are on-track for price-performance parity with
film, but you claim the "estimates are getting longer". Perhaps you are getting
your digicam-futures advice from late-night talk show "human thought computing
experts!" :-)
I believe Carl Sagan estimated the human brain's capacity at about 1 terabyte
in "Broca's Brain." Current computers max out at 2 gigabytes, a factor of 500,
or about nine doublings. Applying Moore's law implies that matching capacity
will take about 14 years.
This is not to say that I seriously believe computers will rival human thought
in 14 years -- there are so many other factors involved beyond simply
information capacity. But I don't think it will take place BEFORE 14 years from
now! :-)
>I'd hate to have to take a pile of memory cards, or a PC.
One hardly needs a "pile of memory cards, or a PC" to take a digicam on even an
extended trip. I can get over 2,000 snapshot quality images without changing
cards! And if I want something other than "snapshot quality," I can change it
on a per-image basis, rather than toting around an extra body or two, loaded
with different film.
>OM stuff is far smaller, too. (and the user interface is better, and it
>doesn't feel like goddamn plastic) :-)
Capice. Although the E-20 by no means feels like plastic (it is rather like the
single-digit OM in construction), it is a bit of a brute to haul around. I'm
anxiously awaiting 4/3rds to get quality digicams back to OM proportions.
Standard disclaimer: I still shoot a lot of film. I still love and use my large
OM collection. I am not saying anyone is a bad person for being a "digicam
resister." But I will continue to rebut inaccuracies and misstatements about
digital -- hopefully in a non-theatening manner... :-)
--
: Jan Steinman -- nature Transography(TM): <http://www.Bytesmiths.com>
: Bytesmiths -- artists' services: <http://www.Bytesmiths.com/Services>
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|