On 2/6/07 10:15 AM, "DMargulis@..." wrote:

We need to back off for a bit because I think Lee's explanation changed the
ground rules. As I understood what he said he is deliberately shooting in a
way
that would produce a bad JPEG, knowing that he has Camera Raw as a safety
net.

No, that¹s not necessarily correct and there have been no ground rules as
usual. There¹s always a Raw file and proper exposure is always a factor. The
JPEG is an after thought. It¹s only based, can only be based, on the initial
Raw capture. In film, which isn¹t linear encoded but has a H&D curve, we
expose for the shadows, develop for the highlights. In raw (or JPEG) we
expose for the highlight. It¹s as simple as that.

There¹s a huge misunderstanding by some on this list about how a digital
camera produces data. It¹s ALWAYS raw. You can decide to toss it and let the
camera bake the JPEG or you can do it yourself (or both). But what Lee
correctly points out is you need to know the true ISO (chip sensitivity) and
you have to expose properly for the scene you want to capture which IS an
artistic expression. We have cameras that have a good six stops of dynamic
range. If the scene exceeds this range, we have to decide where to place the
exposure to capture those six stops.

A primer or digital capture exposure:

http://www.luminous-landscape.com/tutorials/expose-right.shtml

This isn't over exposure, its correct exposure!

There's a lot of potential merit to that approach: digital cameras are a
marvelous step forward, but they have a couple of serious problems, most
pressingly, they don't handle underexposed (overly dark) images as well as
film
does--digital cameras, even the best ones, are very noisy under poor lighting
conditions. It can make getting a good result very difficult whereas a
similarly
underxposed piece of film could be corrected easily if drum-scanned.

I think you need to let us know what digital cameras you¹re referring to
because in the case of my Canon 5D, the images are spectacular with respect
to noise even at 3200 ISO and simply blow away film at that ISO in terms of
noise. What camera(s) are you basing this upon? I'm sure the list would like
to know this. If the noise does bug you, then check out Noiseware which will
reduce almost all the noise while maintain amazing fine detail.

When you accept my challenge, lets shoot some film (neg or chrome, your
call) at 1600 ISO, I'll shot the same scene at 3200 ISO digital and the
audience can decide if film did a better or inferior job with respect to
noise. We can dig up an old drum scanner, no problem.

Since I'm being ignored is anyone else interested in such a real world
challenge and if so, can you ask when proof will be presented to back up the
claims made here? I'm willing and ready. My suggestion is PhotoPlus in NY
(largest photo show in the country) in October. I'm pretty sure I can get
this on the agenda unless certain parties vacillates past April when the
show people start organizing the seminars.

There is, however, no similar difficulty that I'm aware of in correcting a
digicam capture that's *over*exposed (too light) unless detail has actually
been
blown out.

No, you can easily blow out highlights due to the raw to JPEG conversion,
which you have no control over, and then regain up to 1 stop back IF you use
the raw file in a good converter like ACR or LR.

For the more common type of exposure...

As a trained photographer, can you explain to the list what a 'common type'
of exposure is? There¹s correct exposure, based on the capture device (film
or digital) then there are varying degrees of incorrect exposure. And since
we have to capture a potentially huge scene dynamic range using a device
with a fixed capture dynamic range, the exposure is right when the person
creating the image decides what they want and where to place these six stops
in relations to the scene. I think you need to understand the concepts
designed by Ansel Adams (pre-visualization and the zone system) when
discussing Œcorrect¹ exposure. One of the roles of an image creator we
refer to as a photographers is not only to render the image as they seem fit
but prior to this, placing the limited dynamic range of any scene into that
which the camera (film or sensor) can realize. With that in mind, it would
be useful if you would explain what you mean by 'common exposure'.

however, if the person refuses to exit
Camera Raw at all, that means he has no access to channel blending, he won't
have channel-by-channel curves, he won't have the ability to select, or to
retouch, or to use LAB or CMYK, or to sharpen in a sensible way, or to correct
casts that aren't uniform.

No one is proposing this. What some of us are proposing is using the right
tool for the right job. You HAVE to convert scene referred data into output
referred data and build a pixel based color file before Photoshop can enter
the picture. As Lee and Mark have tried to point out, the toolset and
quality is vastly superior (and I am willing to prove it to you in front of
a live audience) compared to doing this later in Photoshop. The math is
undeniable but further, the people who have a better understanding of the
tools and are creating images are saying this here as well. And it has
nothing to do with polishing one or dozens of images in a converter.

Tell you what, even if I take 2x times longer to produce a preferred
rendering then you, I can apply that to similar images far faster than you
can drag and drop adjustment layers on the same number of files. Select,
copy/paste metadata settings. Done.

Example of this in LR:

http://pictureflow.fileburst.com/_Tutorials/Photoshop_LR/06/index.html

In short, he's racing on a tricycle against the jet
aircraft of Photoshop proper.

I think in this context, some are so unfamiliar with one mode of image
processing, the analogy falls very flat.

But again, I'm willing to prove to you that ACR/LR for the tasks it's
designed for is the jet and not Photoshop and that there ARE tools and
processes where the opposite is true. But making simple blanket statements
rather than proving them is par for the course I guess.

So anyone who (excepting Lee and people who shoot
in this way) says they will *never* exit Camera Raw is saying that time is
more important than quality.

No one is saying this. While the topic is photography, slanting this as
black or white is counterproductive. Its about using the tools properly and
if you¹re going to be real sloppy with exposure then raw rendering or JPEG
will suffer and begs the question, why did this sloppiness happen in the
first place?

On 2/6/07 12:13 PM, "Alan Klement" wrote:

When working on set as a tech on a big job, we capture in RAW, but the
aquisition software ( C1, LeafCapture...) will display a JPEG preview
of the capture. If one exposes for linear capture, this jpeg will
look overexposed, unless you apply a curve to the preview that makes
this preview look correct.

It shouldn¹t look like crap but there most certainly is a disconnect between
the JPEGs and the raws depending on the neutral raw rendering. And yes, the
Histogram and info we see on many DSLR¹s isn¹t based on the raw data but the
in camera raw rendering which presents a bit of a problem for the shooter. I
will say with the 5D once I nailed the ISO and exposure, the JPEGs on the
back of that LCD don¹t look bad but I am only using them to check focus.

It would be nice if the Leaf software would automatically apply a tone curve
to mitigate this crappy looking JPEG.

"I work with lots of very good, very high paid wedding
photographers. And those friends you mention above have compared these
JPEGs to the same images processed in raw?"

Not that I'm aware of, but why? They get brilliant results with their
JPEG workflow, and again, with output in mind, this is gonna be
printed by Joe Shmo on a Fugi-Lamda-Lightjet-whatever printer at 5x7
or MAYBE 2-3 8x10s and in an sRGB or colormatch color space.

I¹d be the first to agree that if the job is 500 widgets on a white bkgnd,
shooting raw under a tight deadline is probably overkill if the final is
going 2x2 in a parts catalog and that¹s that. A wedding? A different story.
And again, if the workflows are not compared, we can¹t place a metric on
Œbrilliant results¹ or what raw would bring to the party (and maybe take
away). Another reason I suggested a challenge which Dan will ignore since I
posted it but, I'm totally serious about doing this so that at the very
least, we can see the benefits and warts of both approaches.

Heck, comparing a 1.0 product to a version 10.0 product should be a slam
dunk.

Andrew Rodney
Author "Color Management for Photographers"
http://www.digitaldog.net/