21MP 1Ds Mark III Announced

"Annika1980" wrote in message
ups.com...
On Aug 21, 12:10 pm, SmartGuy wrote:

I read the specs at dpreview. I still wouldn't buy it. While the
(finally!) 2x
resolution increase would be nice over a 5mpx camera, I get the very same
features in my S3 IS with CHDK and can take the very same resolution as
the 1Ds
with a 4-panel stitched frame, plus so many more features that the 1Ds
can never
have. Minus all the hundred or so drawbacks that any DSLR has over any
P&S
camera.

They can keep it.

Only fools out to impress an uneducated client would buy it.

I think it's time to change your screen name.

To Totally Dumb F*cker?

acl wrote in news:1187693762.920561.175810
@57g2000hsv.googlegroups.com:

I didn't understand. How did you truncate, did you simply replace the
2 lowest bits by zeros or what?

Yes. Integer division, and then multiplication by the same number (4, 8,
and 16 for 12-, 11-, and 10-bit, respectively).

What someone who knows how to parse DNG files can do, is write a program
that does this within DNG files, for any camera with DNG. It would be a
more universal test, and doesn't depend on my manual conversion if you
don't trust it.

Uncompressed DNGs would be the easiest to do this with; the RAW data sits
in a very simple form after the headers. I wrote code that alters
uncompressed DNGs for the 20D only, but it's on a dead (or sleeping) HD.
For a specific camera, you can just write the code to read and re-write the
RAW data.

--


John P Sheehy

On Aug 22, 2:19 am, John Sheehy wrote:
acl wrote in news:1187693762.920561.175810
@57g2000hsv.googlegroups.com:

I didn't understand. How did you truncate, did you simply replace the
2 lowest bits by zeros or what?

Yes. Integer division, and then multiplication by the same number (4, 8,
and 16 for 12-, 11-, and 10-bit, respectively).


Thanks. I can't really see any difference between the 12- and 14-bit
results. It would be interesting to see how much difference to the 12-
bit result doing the interpolation in 12-bits would make. It'll also
depend on the algorithm used.

What someone who knows how to parse DNG files can do, is write a program
that does this within DNG files, for any camera with DNG.

For me, C is 2 lines a day, so I won't do it .

It would be a
more universal test, and doesn't depend on my manual conversion if you
don't trust it.

Uncompressed DNGs would be the easiest to do this with; the RAW data sits
in a very simple form after the headers. I wrote code that alters
uncompressed DNGs for the 20D only, but it's on a dead (or sleeping) HD.
For a specific camera, you can just write the code to read and re-write the
RAW data.

I've used dcraw -D to get raw data, and read the resulting file using
someone else's code (I'm not a programmer and this is too low-level
for my "abilities"). My routines are in Lisp. It was a miserable week.
Anyway, do you know for sure what dcraw does to the data before
outputting it if I use dcraw -D (actually dcraw -D -T)? I fourier
transformed some blackframes to look at the noise spectrum and got
strange results for some cameras. I'm not sure if it's dcraw, the
camera's processing, or the routines I used to read the resulting tiff
(but I can't see anything in them that alters the data).

On Aug 22, 3:03 am, acl wrote:

I've used dcraw -D to get raw data, and read the resulting file using
someone else's code (I'm not a programmer and this is too low-level
for my "abilities"). My routines are in Lisp. It was a miserable week.

The routines I used to read the files are in C, I write in Lisp, hence
the misery...

On Aug 20, 8:14 am, "David J. Littleboy" wrote:
"acl" wrote:

Hey David, did you notice the 1ds mkiii's pixel pitch?

Nope. Haven't computed it yet. Sheesh. I must be getting old. (Or fried by
the heat.)

Let's see. 21.1 MP implies that'd be 8.2MP in a 1.6x camera or 9.4MP in a
1.5x camera.

5616 x 3744 = 3744/24 = 156 pixels per mm, or 6.4 microns.

That's a Nyquist frequency of 78 lp/mm, and you'd like 50% MTF at around 75%
of that, so that's 58 lp/mm. 800/58 = 13.8, so the beast will probably be,
like the D2x, a tad soft at f/16.

(Has the heat fried my math???)

So for max sharpness, you're going to be shooting at f/8 and f/11 a lot with
lenses under 100mm.

Of course, my reading of figure 5 on this page indicates that there's no
point to bothering with a 14-bit A/D converter on either this or the 40D.
(Note that the 10MP D200 is a straight line down from its ISO 100
performance, which means that the ADC + electronics are adequate to pull out
all the sensor is giving at ISO 100.) It sure looks to me that 14 bits only
will be making sense on the 1DIII and 5DII (if such a creature appears at
16MP).

http://www.clarkvision.com/imagedeta...erformance.sum...

David J. Littleboy
Tokyo, Japan

Expect Leica, Zeiss and Olympus OM WAs to get even more expensive on
Ebay.

On Wed, 22 Aug 2007 06:43:28 +1000, Pete D wrote:

Only fools out to impress an uneducated client would buy it.

I think it's time to change your screen name.

To Totally Dumb F*cker?

Too apt. I think that good ole anniker simply recognized that our
(not so) SmartGuy belonged on the Sock Puppet Troll List. Have some
pity for the sick puppets. Canon has taken steps to make sure that
their beloved CHDK software won't be compatible with any new
Powershots, so they're already living in the past. And now :

the all new . . .

**** CHDK / Photoline 32 / anti-DSLR Sock Puppet Troll List ****

Allan D., Baumbadier, BigBrother, Brad M, Bucky, CharleiD,
CoffeeTalk, CoolGuy, Craig Stevens, D. Farmington, Dartagnon,
DaveB, DOCJohnson, D-Rexter, Danny V., EdBancroft, DSLRs SUCK!
, FeastForThought, Fed-Up-With-Corel, FixItMan,
FrankLM, Gaile S., GilfordBrimly, Glen Bankwood, GnomeAlaska,
GoKiting, GreggAkin, GregoryH., Henry Hank, HatTrick, HokusPokus,
IdiotDetector, ImpressMe, JoeBS, Lurk, John Kaiber, M. Goode,
MoronDetector, NameHere, NameThere, New2_S3, ,
OptionsRus, OTPolice, RealityCheck, ReplyingToStupid, Rob Akins,
RockyZ, SayWhat, SelfImporantName, SelfImportantName, Siskel,
SmartGuy, Soujourner, spamless, SpamAlert!, TryinToHelp,
Wayne J.L., WillyWonka, X-Man and Yeti.

Have you tried a speech therapist, I believe they can fix that
It has to be 15-20 years since I programmed in LISP. Why LISP for this
application?

Cheers,

Wayne

Wayne J. Cosshall
Publisher, The Digital ImageMaker, http://www.dimagemaker.com/
Blog http://www.digitalimagemakerworld.com/
Photography and Art Forums http://www.dimagemaker.com/forums/index.php
Personal art site http://www.cosshall.com/



acl wrote:
On Aug 22, 3:03 am, acl wrote:

I've used dcraw -D to get raw data, and read the resulting file using
someone else's code (I'm not a programmer and this is too low-level
for my "abilities"). My routines are in Lisp. It was a miserable week.

The routines I used to read the files are in C, I write in Lisp, hence
the misery...

On Aug 22, 5:06 am, "Wayne J. Cosshall" wrote:
Have you tried a speech therapist, I believe they can fix that
It has to be 15-20 years since I programmed in LISP. Why LISP for this
application?


When all you have is a hammer...

I'm not a programmer. I taught myself various languages when I was a
child (talk about a misspent youth!), but of course in a completely
disorganised way, so I'm actually rubbish at coding. I also think in a
completely disorganised way, which doesn't help. Then I lost interest
(or transferred it to physics, to be precise). Some time ago, while
writing my phd thesis, I got bored and decided to teach myself lisp:
the perfect excuse to procrastinate. But my programs are clunky, to
put it mildly

I don't use it for number crunching, which I try to avoid, normally.

[A complimentary Cc of this posting was NOT [per weedlist] sent to
acl
], who wrote in article .com:
What someone who knows how to parse DNG files can do, is write a program
that does this within DNG files, for any camera with DNG.

For me, C is 2 lines a day, so I won't do it .

No C needed - if dcraw can MERGE a header from one DNG file with a
TIFF of "very raw" data.

a) Convert DNG to TIFF;
b1) use -fx of ImageMagick to do the transformation; or
b2) convert to .txt via ImageMagick, use any scripted tool to edit
.txt, convert back to TIFF;
c) Merge the modified TIFF into the DNG file.

(I have no idea whether 'c' is available).

Hope this helps,
Ilya

[A complimentary Cc of this posting was sent to
John Sheehy
], who wrote in article :
My padding of the zeros is probably key in getting almost exactly the same
results between 14-bit and 12-bit

I see no reason why any explanation is needed. According to
http://www.clarkvision.com/imagedeta...mance.summary/,
table 4, 1D mII has read noise about 17 electrons at iso100. Add to
this electron noise, and you get noise about 20 electrons.

With full well close to 64K, this gives noise of 20 units.
log2(20)4; There is practically no need to have more than 16-4=12
bits for storage. (Correct appropriately for mkIII; I do not know
relevant numbers.)

Hope this helps,
Ilya