RAW Conversion: Adobe Capture Raw vs. Canon Digital Photo Pro

Folks,

I have done alot of RAW conversions (20D) with ACR. I like ACR's
interface and the chromatic abberation correction. However, I seem to
keep coming across situations where ACR does not do a good a job as
Canon's RAW conversion (whether in camera to jpeg or via DPP). I have
seen the following issues:


- sudden light to dark tranasitions (e.g. a lampost against a bright
sky). ACR seems to generate color artifacts whereas Canon's conversion
significantly less so.


- very dark areas of the image. Canon's conversion seems to generate
smoother more 'photographic' or desireable results whereas ACR tends to

look noisy and/or posterized. I have noticed that sometimes the way ACR
converts,
it creates a spike at '0' in the Blue channel histogram. DPP with the
same image does not (even with a whole lot of tweaking, ACR still does
this). I wonder if this is why shadows appear noisier?


- color rendition. Canon seems to give a more neutral natural looking
result, whereas I have seen some casts when using ACR. I typically set
my WB to Sunlight when working outdoors, and the results seem to be
more what one would expect with Canon's conversion, even when I use the

'as shot' setting in ACR

- I have also occasionally noticed less detail with ACR but this is
hard to quantify (e.g. I had one
image where a building reflected in the glass of another building, and
there was noticeably less detail in the ACR image in this reflected
area. Sharpening was off, but even with the sharpening turned way up in
ACR, the detail was still less than the DPP conversion. Yet, other
parts of the image did not exhibit less detail...a strange one).


I really want to use ACR for the chromatic abberation and workflow. But
based on my experience, I now am using DPP. Also, it would seem not
even to make sense
for me to consider lightroom if it is built on the same RAW conversion
engine as ACR.


Has anyone else seen these or other issues with Adobe Camera RAW?

Thanks


W

 

Yes, clipping of shadows will create a noisy appearance. However, the
shadows may very well not really be clipped; what color space are you
converting to in ACR? A *LOT* of images will clip shadows going into
sRGB, but not with a wider space like Adobe RGB or ProPhoto RGB. You
may be able to squeeze the range into sRGB with adjustment, though the
quality will suffer for it.

Canon's conversion has the advantage of knowing what 'as shot' means.
ACR can only approximate. More adjustment will often be necessary as
a result of the simple fact that it's a third-party conversion versus
one designed to match the camera exactly.

You could also use both, choosing whichever one you feel will best suit
a given image.

 

Thanks for the response. I used Adobe1998 color space for both
converters. Even when I used 'exposure' and/or 'brightness' in ACR
making the image appear significantly lighter than the DPP conversion,
there was still a spike at '0' in the blue channel histogram on the ACR
converted image whereas the DPP converted image did not display such a
spike(and by the way, I am talking about the histogram as displayed in
photoshop on the converted image in both cases). That's why I suspect
something is 'not right' with the ACR conversion. I guess I will use
DPP as my default converter and ACR only when I feel a DPP result is
lacking in some regard.

 

What do you have the "shadows" slider set for in ACR, when you get blue
channel clipping? Do you get it even when you set it to 0?

 

one question from my side...i've read that these express photo labs which
(at least here) widely use Fuji 350....photo lab - by default use sRGB color
space, so if sending to develop pics, saved in other spaces, as Adobe or
ProPhoto may result in different colors in final developed picture...they
also can't read embedded color space (if it is embedded in jpg, that is)..
so, let's say i convert in ACR and use ProPhoto...then i get jpg in that
color space. What next? When final work on them is done, convert to sRGB?

 

yes even with the shadows slider set to 0

 

Typically it makes sense to give a photo-lab files in sRGB space unless
they state otherwise. So before saving your jpg, convert to sRGB for
files destined for such labs.

 

i wonder...never tried yet...is it possible to convert jpg images? I mean,
if i'd set to sRGB in ACR, what's the point of ProPhoto then... I guess i
should convert to jpg in ProPhoto mode, and then convert jpg's i want send
to lab into sRGB...?

 

You can convert Adobe RGB to sRGB in Photoshop, but you won't get any
more colors going the other way. So it makes sense to use Adobe RGB (or
ProPhoto) and convert later to sRGB. Some printers do use Adobe RGB
like a Chromira or Lightjet, so alway ask your printer. I did have some
vacation 4x6s made at CVS recently, they said they had an sRGB printer,
so I converted my files, everything looked flat, sent some back in
Adobe RGB ancd these came out looking like my files, pays to
experiment.

Tom

 

My recommendation for "what next" is to avoid those labs like the plague
on photography they are.

sRGB butchers images. Almost all print processes are capable of gamuts
exceeding sRGB, especially photographic printing processes. Use a lab
that (preferably) supports color managed workflow, or (at least) doesn't
require sRGB input.

In addition to the smaller gamut, there is an issue with doing the
conversion in working space -- if you convert from a wider-gamut working
space into sRGB, out-of-gamut colors are simply clipped. You can't use
"perceptual" rendering here, because there is no such thing as a perceptual
conversion to sRGB (Photoshop will let you choose it, but it will do a
relative colorimetric conversion anyway). If, however, the lab itself
does a conversion from your working space into its output space, the
conversion can be perceptual, and out-of-gamut colors can be scaled back
rather than clipped, resulting in better output quality.

So, avoid sRGB and the labs that require it.

 

It sounds like you may just have images that are better suited to the Canon
software, since it doesn't produce that problem on output. Sorry I couldn't
be of more help...

 

yeah, but I question why an image should be 'better suited' to a
particular RAW converter. If I had to guess, I would guess that there
are subtleties particular to the 20D sensor that maybe only Canon knows
about. ACR must work for a whole boatload of different cameras. Perhaps
whatever public data that Adobe gets from the various vendors does not
completely describe how the sensor captures the image. Perhaps the
transmissive characteristics of the red green and blue filters in the
Bayer array vary among manufacturers and are not completely described
in the public data. Or maybe they are but it is asking too much of ACR
to accomodate such sublteties for all sensors. I really don't know, but
I do know that I have seen deficiencies in some ACR
conversions...enough to (at least for now) switch me over to DPP as my
default converter. I did find a post in some other photo site
newsgroup with similar 'complaints' about ACR vs. DPP. So at least I
know I am not completely nuts! It is an interesting topic, but it takes
away from photography time .

 

I just did a quick check of Lightroom Beta 4....it seems to do the same
thing as ACR

 

I just did a quick check of Lightroom Beta 4....it seems to do the same
thing as ACR

 

from i saw it does quite a bit more than ACR...

 

Yes, featurewise it does alot more. I was referring to the fundamental
RAW conversion issues I mentioned in an earlier post on this thread.

 

that's similar, yes. In fact, maybe is lightroom for normal user even too
rich with features which will rarely be used...

 

I have similar experience. I experiemented with RAW+JPG using PS CS2,
ACDSee Pro 8, and Canon's Zoombrowser 5.6c. Photos converted from Raw
to Jpg by Zoombrowser is more vivid and smooth while PS CS2 and ACDSee
produce flat and darker pictures. Zoombrowser produced Jpg is very
close to the camera's JPG output.

 

Yes it is unfortunate that ACR has this problem. It kind of eliminates
the usability of Lightroom since it seems to exhibit the same 'noisy
shadow' problem. All the features in the world don't help when there is
such a fundamental shortcoming in the RAW converter

 

The sad thing for me is I have been using ACR for the last couple years
for converting my 300D raw photos. At first, I thought it is my camera
that produce darker photos as I have read many threads indicating 300D
are deliberately set to use lower exposure. Adjusting higher exposure
in ACR will make the photo look flat. When I upgraded to 20D, the
exposure improved but still photos showing in ACR is not as vivid as my
friend's Nikon D200. He suggested to me to try using Canon's own
solftware to display the RAW because he has similar experience that
Nikon's own software shows NEF photos much better than ACR. Now I have
to go back and re-convert my last 2 years of RAW. Fortunately, I
didn't delete those RAWs. I am really pissed with ACR!!!