DxO says Adobe Lens profiling has "shortcomings"

I'm sure Adobe are thanking DxO for pointing all this out. ;-)

http://www.dpreview.com/news/1005/10...ensprofile.asp

Of course many of the DxO observations are correct. For example,
attempting these profiles for the full calibration of a wide zoom should
be done with more lighting (to achieve even illumination) and an awful
lot of care.

The "uneven sharpness" correction of DxO's claim seems a bit much.

But it's hard to believe that DxO don't have their own interests at heart!


--
gmail originated posts are filtered due to spam.

Alan Browne wrote:

I'm sure Adobe are thanking DxO for pointing all this out. ;-)

http://www.dpreview.com/news/1005/10...ensprofile.asp

Of course many of the DxO observations are correct. For example,
attempting these profiles for the full calibration of a wide zoom should
be done with more lighting (to achieve even illumination) and an awful
lot of care.

Lens testing is incredibly tedious work. I hope the lens manufacturers
do these tests and provide profiles. Alternately, DXO could sell
profiles to work with Adobe. Nobody really wants to load yet another
software to work on their images.


The "uneven sharpness" correction of DxO's claim seems a bit much.

I wonder what they do to address "uneven sharpness", apply more unsharp
mask to the corners? I'm not at all sure that's a good idea.

"volume anamorphosis"
http://www.dxo.com/us/photo/dxo_opti...s/anamorphosis
"Lens design is always going to remain a trade-off between correcting
for optical distortion and correcting for volume anamorphosis, which are
different but related image geometry phenomena.(1)

(1)This fact - that the laws of physics just don’t allow us to have our
cake and eat it too - is reflected in lens design. Fisheye-type lenses
have huge distortion but very limited volume anamorphosis. Rectilinear
lenses display much more limited (though very visible) distortion and
very visible volume anamorphosis."

Hmm, but they seem to be allowing people to have their cake and eat it
too???


But it's hard to believe that DxO don't have their own interests at heart!

Savageduck wrote:
On 2010-05-22 15:38:24 -0700, Paul Furman said:

Alan Browne wrote:

I'm sure Adobe are thanking DxO for pointing all this out. ;-)

http://www.dpreview.com/news/1005/10...ensprofile.asp

Of course many of the DxO observations are correct. For example,
attempting these profiles for the full calibration of a wide zoom
should be done with more lighting (to achieve even illumination) and
an awful lot of care.

Lens testing is incredibly tedious work. I hope the lens manufacturers
do these tests and provide profiles. Alternately, DXO could sell
profiles to work with Adobe. Nobody really wants to load yet another
software to work on their images.

Using Adobe Lens Profile Creator is relatively simple for anybody who
can follow instructions.

I was able to create a profile set for my Tokina 11-16mm f/2.8 for use
with my D300s with very little trouble.
I printed an Adobe target. (the largest I could with my Canon i9900,
13x19, 36 pnt with 23x35 black & white squares.)
I shot the 9 shots as directed for each set, created the profiles, and
loaded them into CS5.

I shot a test shot of a garage door. (a compliant, but not the most
attractive subject, with a fair number of somewhat straight lines.) Here
is what resulted;

Uncorrected;
http://homepage.mac.com/lco/filechute/Tok-11-16-A1.jpg
Corrected, Geometric Distortion, CA, & Vignette. Who knows about "Uneven
sharpness?"
http://homepage.mac.com/lco/filechute/Tok-11-16-B.jpg

Definitely an improvement and I don't mean to be overly critical but
that seems to be an example of 'mustache' distortion, which is not a
simple curve but a wavy result. I'm not sure DXO could do any better though.

Does anyone know if the Adobe thing works in Lightroom, or just PS?


I have been applying the corrections to various shots taken with this
lens, which is the widest in my bag. The results have been satisfying.

"Paul Furman" wrote:

Uncorrected;
http://homepage.mac.com/lco/filechute/Tok-11-16-A1.jpg
Corrected, Geometric Distortion, CA, & Vignette. Who knows about "Uneven
sharpness?"
http://homepage.mac.com/lco/filechute/Tok-11-16-B.jpg

Definitely an improvement and I don't mean to be overly critical but that
seems to be an example of 'mustache' distortion, which is not a simple
curve but a wavy result. I'm not sure DXO could do any better though.

Does anyone know if the Adobe thing works in Lightroom, or just PS?

The Adobe thing will work in Lightroom 3.0 when Lightroom 3.0 is released. I
understand that there are third-party plug-ins for the current version of
Lightroom that do at least the mustache distortion correction stuff.

--
David J. Littleboy
Tokyo, Japan

On 10-05-22 18:38 , Paul Furman wrote:
Alan Browne wrote:

I'm sure Adobe are thanking DxO for pointing all this out. ;-)

http://www.dpreview.com/news/1005/10...ensprofile.asp

Of course many of the DxO observations are correct. For example,
attempting these profiles for the full calibration of a wide zoom
should be done with more lighting (to achieve even illumination) and
an awful lot of care.

Lens testing is incredibly tedious work. I hope the lens manufacturers
do these tests and provide profiles. Alternately, DXO could sell
profiles to work with Adobe. Nobody really wants to load yet another
software to work on their images.


The "uneven sharpness" correction of DxO's claim seems a bit much.

I wonder what they do to address "uneven sharpness", apply more unsharp
mask to the corners? I'm not at all sure that's a good idea.

It's not - it's just DxO using "negatives" marketing and grasping at
diminishing points to sell their product.


"volume anamorphosis"
http://www.dxo.com/us/photo/dxo_opti...s/anamorphosis

"Lens design is always going to remain a trade-off between correcting
for optical distortion and correcting for volume anamorphosis, which are
different but related image geometry phenomena.(1)

(1)This fact - that the laws of physics just don’t allow us to have our
cake and eat it too - is reflected in lens design. Fisheye-type lenses
have huge distortion but very limited volume anamorphosis. Rectilinear
lenses display much more limited (though very visible) distortion and
very visible volume anamorphosis."

Hmm, but they seem to be allowing people to have their cake and eat it
too???

Compromise, or "perfection is the enemy of good enough."


--
gmail originated posts are filtered due to spam.

On 10-05-22 19:13 , Savageduck wrote:
On 2010-05-22 15:08:12 -0700, Alan Browne
said:


I'm sure Adobe are thanking DxO for pointing all this out. ;-)

http://www.dpreview.com/news/1005/10...ensprofile.asp

Of course many of the DxO observations are correct. For example,
attempting these profiles for the full calibration of a wide zoom
should be done with more lighting (to achieve even illumination) and
an awful lot of care.

The "uneven sharpness" correction of DxO's claim seems a bit much.

But it's hard to believe that DxO don't have their own interests at
heart!

No kidding!

You and I have had this discussion privately,

We did? Are you sure?

and we are both creating
profiles for our lenses, without the added expense of DxO.

Now that I have upgraded to CS5, and have made sets of lens profiles for
my Tokina 11-16mm f/2.8, rather than making fixes manually with the Lens
Correction Filter. I have to ask myself, "Why would I spent all that
extra cash on DxO Optics Pro?"

I would never have bought their s/w anyway. Well, maybe if I did more
wide angle work.


I think for many DxO is going to have to come up with very convincing
reasons to buy Optics Pro, and they know it. Certainly it takes a little
time and patience to create the lens profiles using Adobe Lens Profile
Creator, but it isn't rocket science. All you are paying for with DxO is
the the time it takes to create a whole bunch of lens profiles for
lenses you will probably never own.

Right now they have the Standard Edition on sale until June 15, for $99,
down from the regular $169.
The Elite edition regularly at $299 is on sale for $199. I just can't
see spending that, when all I did was upgrade to CS5 for $199, and get
satisfactory results for my needs.

This reflects the usual "picking off" of specialized s/w and packaging
it into more general s/w.

I just wish CS5 had GPS tagging integrated into it (or into Bridge -
better). The various free/shareware programs are hoaky.

--
gmail originated posts are filtered due to spam.