roger krueger

I have yet to see a RAW converter (in or out of camera) that behaves reasonably as the

number of channels blown out changes. In this image the blue channel has been

significantly darkened in the area where the red and green are blown vs. the area where

just the red is blown.

<br><br>

The orange ring is about having the red gamma too low--it falls of very fast once it stops

being saturated, giving this weird discontinuity.

<br><br>

When there's no detail anywhere near the problem area a big blur can be surprisingly

effective.

<br><br>

This is an 11 pixel blur, masked to preserve the detail at the bottom, shadow levels set for

each channel, a big red gamma boost (2.0) and a small blue gamma reduction:

<br><br>

<img src="http://www.punktures.com/photonet/FantauzziSunsetEdited.jpg">

<br><br>

Here is the layered Photoshop file:

<br><br>

<a href="http://www.punktures.com/photonet/FantauzziSunsetEdited.psd">Photoshop

file</a>

Not that I believe full-frame is the right answer for a lot of people. If you're mostly an f8

shooter, full frame doesn't hold anything special for you. Get a D2X.

Doesn't justify writing that veers from wild sloppiness to clear attempts to mislead like this

mess does though.

But if you do shoot near wide open a lot, or if you are in love with specific lenses that have

no APS equivalent (the 24/1.4 and Leica 19/2.8 for me), there is absolutely no substitute

for

full frame.

It's a blatent hatchet job, and he should be ashamed. Just a few of his sins:

He uses MTF graphs and refuses to disclose what lenses he used--what's he hiding? It's

also misleading because you need to be comparing equivalent (scaled for sensor size)

frequencies. It's all about resolution on the final print, not on the sensor. Show the crop

frame MTF graphs at 16 and 48 lp/mm and suddenly it's a lot closer.

He does a classic strawman fallacy, he rants about the insignificant difference in ISO 100

noise, and completely ignores the single biggest FF advantage, high ISO noise

performance. Some of the other stuff I could see as dumb mistakes, but I can't come up

with any explanation for this one besides being deliberately misleading.

His dopey excursion into angle-of-incidence is nearly as bad. He claims that this will

cause corners to be softer, especially with "wide angle, legacy optics". Vignette, sure, but

I've never seen ANY evidence of ANY wide being softer in the corners on FF digital than on

film. Yeah, pixel peeping sure can show up the inadequacies of many wides, but that's not

the sensor's fault. If angle-of-incidence caused softness, why did the mirror-up,

rangefinder based Voigtlander 15--with much, much greater angle of incidence than any

normal SLR lens--have such awesome corners on a 5D when 16-9.net tested it?

His 400D vs. 5D comparison complete ignores feature differences in his conclusion you're

paying $2600 for 2.6mp, but he then immediately uses features to justify paying another

$2k for a D2x that has 0.8 more megapixels than the 5D. Again, this is too convenient to

just be silly errors, this is being deliberately misleading.

He goes on and on about retrofocus like it's a good thing. Not that it matters much to the

FF vs. APS arguement, but retrofocus is NOT good, retrofocus is WHY most SLR wides

suck, and why they're expensive. Everyone goes on and on about how good the Contax

21mm Distagon is, but their 21mm Biogon for the G2 is a decidedly better lens for a lot

less money. And how about the $300 Voigtlander 15/4.5 that flat smokes $1000

ultrawides like the Nikon 15/5.6, and is considerably better than even Canon's $1700 14/

2.8?

Yes, absolutely.

But it also means considerably higher noise amplitude per pixel. The visual noise impact

ends up being similar for a given physical sensor size, although the large pixels should be

a little better at extreme ISOs.

There is a benefit to more pixels though--when you're using a noise reduction program

like Neat Image or Noise Ninja, the farther your noise frequency is from details-I-want-to-

keep frequencies, the harder you can filter without making a mess.

It probably defaults to f8 at 1/1000th because it can't deliver high shutter speeds at wide

open apertures. A leaf shutter is just another aperture, in many digital cameras they use

the same mechanism for both. It has a lot less distance to travel at f8 vs. f2.8, so it can

give shorter durations.

In fact, this "can't make max shutter speed wide open" effect is a dead giveaway of a

camera that completely controls the exposure with a leaf shutter, rather than doing

electronic shuttering on the sensor, and merely using the mechanical shutter for darkness

during readout.

A 90's design 28 Elmarit is next on my list. Wdies are the most interesting due to Canon's

weakness in this area, and because the Leicas' focusing scale is a massive improvement

over useless AF focusing scales.

I'm a huge fan of my 90's Leica 19/2.8, it beats my 24/1.4 senseless and leaves it crying

for its mommy. But I don't think earlier Leica glass is good enough to overcome 20 years

of progress in lens design. And even newer long Leica glass doesn't offer much/any

advantage, because long Canon "L" glass is so good.

I'm always amused by people who buy thw 24/1.4 for stopped down use. For similar

money you could get the 24-70 that's actually better at 24. I love my 24 as a low light

lens, there's just nothing even close. But at f11 it's collossal waste of money.

My 19/2.8 is embarrassingly better than my Canon 24/1.4. Retrofocus wides are hard to

make really good. Especially since Canon and Nikon only have two classes of wide lens--

economy, and bleeding fast. Only Leica, Contax and Olympus have made the slow-but-

first-rate wide lenses that are really the most useful.

Even with Leica it's only the recent (1990 and later) designs that are markedly superior to

Canon. Something like the designed-in-the-70s-by-Minolta 24/2.8 is very unlikely to be

worth the hassle.

For things like landscape and architecture, where corner sharpness matters, lots of people

will be able to see the difference. A sharp lens is no substitute for talent, but no use

handicappinng what talent you have with a lens that's got soft corners.

Well at least someone's thinking about a big-sensor single-focal length P&S. Pity it's

Sigma.

But f4 is ridiculously slow for a fixed lens, and if a viewfinder costs too much at least give

us a cold shoe. I'd rather have my nice external Voigtlander 28mm finder over a lets-save-

every-penny finder anyway.

I also have zero patience with Foveon's blatently misleading MP calculations--even if the

technology was decent I still wouldn't buy something from someone trying to pull the wool

over my eyes like that.

But this does show it's possible, someone else has to try it sooner or later. Pity Konica's

gone--A digital Hexar AF would be so sweet.

I'm mostly not worried about using my real name, but I can certainly see why some folks

would not want to. Particularly, more and more employers are looking at the online

activities of applicants. If I were an executive type I wouldn't dream of having a bunch of

hobby-related posts come up when my next job Googled me. Especially since I do street

photography, something some people find very weird and upsetting.

As long as photo.net knows your real name I fail to see any relevance in posting under it.

If you're a jerk you'll be banned with no easy route to circumvention, whether you were

called Bob Smith or Beelzebub.

P.S. the italic behavior noted above only occurs when HTML is selected, it does not happen

to "Plain Text" posts

The confirm page is somewhat broken, looks like when you italicized the warning at the

end you didn't close the italic tag, leaving the user's entire post italicized.

It posts O.K., it's just the confirm preview that's hosed, but this is quite disconcerting,

especially when you're trying to use italics yourself.

It's also vaguely alarming given the problems there have been in the past where an

unclosed italic in one post dooms the rest of the thread to all-italics. I though that was

what I was fighting until after I changed my <i> passages to <q>s and finalized the post,

and everything looked O.K.

The RD-1 is a D-70 sensor shoehorned into a $500 rangefinder body, going for $2000. It

is by far the worst digital camera value anywhere. And a D-70 sensor isn't going to beat a

D200 at anything.

<br><br>

<q> framing is inaccurate compared to an SLR. </q>

<br><br> only marginally so, and being able to see beyond the edge of the frame makes

anticipating action easier in som einstances.

<br><br>

<q>In lesser cameras, the viewfinder is barely good enough to point in the direction of

the subject.</q>

<br><br>

The viewfinders on my Zorki's are about as bad as it gets, and they're still better than

entry-level DSLRs

<q>The touted lens quality may be true, but the issue is moot. Any decent lens, including

many zoom lenses, so far exceeds the capability of film or digital sensors,</q>

<br><br>Hogwash. Name a single lens that delivers 80% MTF in the corner of the frame

at 50 lp/mm wide open, or even one stop down. Name a lens wider than normal (ie 50 on

FF) that can deliver 80% MTF at 50 lp/mm in the coirners at ANY f/stop.

<br><br>

<q>that there is no practical difference between the best rangefinder and the best SLR

lenses. </q><br><br>

On the wide end there's a huge difference. The cheap Voigtlander 12 and 15 beat the snot

out of SLR ultrawides costing 3-5 times as much. As you get longer the gap narrows, but

even the great SLR wides, the Contax 21 and Leica 19, although much better than what

Canon and Nikon have in this area aren't as good as their rangefinder equivalents.

<br><br>

<q>I saw that 20 years ago, comparing my son's Nikon FG to my Leica (the Nikon won,

hands down). Incidently, I use a D2h and a D2x now, and I'm not looking back.</q>

<br><br>

O.K., maybe if the Leica was way out adjustment, or was using glass a lot older the the

Nikon. The Nikon should have been able to keep it close from 35 up, but no way it could

"win hands down" against same-era Leica glass without some sort of trouble with the

Leica,

<br><br>

All that said, I love rangefinders, yet neither the RD-1 nor M8 tempt me in the least. Where

a rangefinder is really a better tool than an SLR is at the wide end, something very much at

cross purposes with a crop sensor. Give me a FF rangefinder for no more than a 50%

premium vs. the 5d and I'll be all over it. 'Til then, no thanks.

http://www.procamerarepair.com/

These people have done CLA's on my 150/5.6 and 50/6.3. For my taste the helicoid

grease they put in the 150 was way too stiff, but the work was otherwise competent.

They also repaired a damaged rangefinder coupling adjust screw in my 100/2.8, but that's

more of a mixed experience--the repair worked well for about 3 years, but eventually

failed catastrophicly, causing irreperable damage to the brass attachment on the helicoid

that the adjuster screw rides on.

Really, these lenses aren't rocket science, I was able to break down the 150 and relube the

too-tight helicoid myself. It's not like you have a lot to lose, used 100's are hardly more

than the price of a CLA. Note that there is a specific lube for helicoids, motor oil and

Vaseline are not substitutes :-)

The cambridgeincolor link is generally very good, but he completely misunderstands

Bayer--his conclusion that the airy disk can be the size of 2 pixels in a Bayer camera

before diffraction degrades the image is NOT correct.

Elevator has been doing it for years, they convinced a Lambda to eat regular fiber paper.

I'm guessing they're livid about this new paper, it'll kill their ability to charge a huge

premium as the only digital-to-fiber service available.

I don't think we'll see wide adoption, just because fiber requires much more care in

washing and drying than RC.

Not that I know anyone's plans, but Ken Lieberman would be a logical place for it to show

up--they already run a Lightjet, and already do traditional fiber, so they wouldn't need any

new equipment or skills, unlike most Lightjet owners.

Mr. Benton is entirely mistaken, the 1dsII sync circuit is rated for 250v.

Mark pretty much has it nailed, although I'll mention that the Pocket Wizards add a small

amount of delay too. But even 1/200th will likely work

I use a Leica 19/2.8 on my 1dwII, and it completely smokes my 24/1.4.

I had to file down the back of the lens some to make it clear the mirror. Using a rented 5d

when my 1dsII was in the shop I found that the 5d's longer mirror still didn't quite clear.

There are several Leica 19/2.8s, the earlier ones with filter threads are the least

desirable--still better than Canon, but nothing spectacular. Lenses made after the 1990

redesign have 3 integral filters on a turret, and no threads. The few pre-ROM copies of

this lens go for about half as much as the ROM version, but they function identically in

adapter use.

What it neglects to explain is that the larger format can be stopped down more before

hitting diffraction.

If you stop down to a given diffraction-limited resolution on equivalent size and

composition final prints, DOF is the same, from the tiniest cell-cam to a 16 x 20 monster.

It just happens at different f-stops.

Pick a winner? Nah, far more interesting to gripe about what's been left out...

Biggest omission is the Canon 24-70/2.8, a lens that in Castleman's tests beat the 24/1.4

by a little and the 24/2.8 by a lot.

I'd also like to see the Contax 25/2.8 and Leica 24/2.8, even though both are elderly

designs that in no way represent the best of what these companies can do.

I will pick a lens to vote against, expecially since someone else voted for it--Canon's 24/

1.4 is unique, there's nothing else that fast, wide, and at least decent wide open. It's far

and away my most used lens (I shoot a lot of live music, specifically punk). But as a

stopped down lens it's nothing special, Even by 5.6 the corners are a little soft..

This is kind of the ugly-stepchild focal length--both Contax and Leica have nice, recent

28s, as well as stellar wider lenses (the 21 and 19), But the 24 and 25 are 70's holdovers.

Canion also has an antique 2.8 holdover, as well as a lens that sold its soul to be

screaming fast.

I guess my money would be on the 24-70/2.8 if you included it, the Nikon 17-35

otherwise.

Shoooting live music I'm always at least a stop under (because there's so much that wants

to blow out) and often two or even three under to get a useful shutter speed/aperure.

Noise can be fixed by software, noise can be buried by letting the shadows block, but

motion blur is forever.

<i>Even if Krueger's claims are true...</i>

<br><br>

What, I need to post my repair receipts/work orders/etc.?

<br><br>

Make no mistake, I love my 1dsII. It can do things in low light utterly impossible

with film. My long Canon glass (135/2, 200/2.8, 400/5.6) has been absolutely

impeccable, from both an image quality and reliability standpoint.

<br><br>

My point was merely that Canon stuff breaks too, that Leica is far from unique in being

fragile. If I want rock solid reliability I'll drag out my old Mamiya Universal--drop that

sucker on the sidewalk, and you worry about cracking the concrete, not the camera.

<br><br>

Are Leica lenses overrated? Yeah, a lot are. I can't think of any Leica SLR lenses designed

before 1990 that offer much image quality improvement over current Canon glass.

Continuing to make the designed-in-the-70s-by-Minolta 24/2.8 is shameful enough,

selling it for thousands is just insane. But the recent designs are incredible, and worth the

money--the latest 19/2.8, 28/2.8, 100/2.8 APO macro and 28-90 zoom--are

astonishingly good lenses, albeit at equally astonishing prices.