Olympus-OM
[Top] [All Lists]

Re: [OM] What a shock re PTLens, was: I must admit

Subject: Re: [OM] What a shock re PTLens, was: I must admit
From: Jim Nichols <jhnichols@xxxxxxxxxxxxx>
Date: Wed, 10 Feb 2016 08:00:22 -0600
I had a surprise, also. I did the distortion computation for three lenses, two Fujis and the iPhone 6. The new Fujinon 35/2 has more barrel distortion than the Fujinon 27 Pancake or the iPhone 6. To me, this says that lens designers are relying less on optical perfection and more on firmware. The Pancake and the iPhone are actually pretty good.

Jim Nichols
Tullahoma, TN USA

On 2/10/2016 7:51 AM, Chuck Norcutt wrote:
Competing with the Google street view camera isn't very hard but I'm impressed with what they do considering the speed and volume of work.

But what was a shock here is that I actually went back and located these shots I did for PTLens. I exclaimed emphatically that I certainly didn't do this in February. But the truth is in the EXIF which says I did it on Feb. 19, 2008. There's snow on the ground. Brrr! I must have been very anxious to get these lenses tested for use with PTLens.

Another surprise was comparing the results of different lenses... something I had never done before. I was quite surprised to see that my Tamron 20-40/2.7-3.5 has far less distortion at 24mm than the Zuiko 24/2.8. The Zuiko has very pronounced barrel distortion, the Tamron has a very subtle moustache distortion. Apparently I didn't yet own the Tamron 24-135 (my favorite walk-around lens for the Canon 5D) since it's not in any of the results.

Chuck Norcutt

On 2/9/2016 7:17 PM, Jim Nichols wrote:
Looks like a very good test pattern.  I suspect that your photo was
better than theirs, which appears to be a poorly joined 3-panel pano.

Jim Nichols
Tullahoma, TN USA

On 2/9/2016 6:12 PM, Chuck Norcutt wrote:
And here's the actual school.  You may have to rotate around to find
it.  It has turquoise panels.
<https://www.google.com/maps/@42.1005948,-76.0495774,3a,75y,176.58h,85.68t/data=!3m6!1e1!3m4!1ssewFktFBNda38zwapVTxVQ!2e0!7i13312!8i6656!6m1!1e1>


On 2/9/2016 7:01 PM, Chuck Norcutt wrote:
I did that in Endicott, NY. You can be sure it wasn't done in February.
:-)

On 2/9/2016 4:24 PM, Jim Nichols wrote:
Hi Chuck,

It was too cold for me to go out and be as industrious as you.  I
settled for a framed painting on the living room wall. Shot square on
and centered, it worked pretty well.

Jim Nichols
Tullahoma, TN USA

On 2/9/2016 2:35 PM, Chuck Norcutt wrote:
I wasn't aware that Picture Window Pro was now doing that. I used to
use PTLens by ePaperPress.  It's a Photoshop plug-in that has
distortion profiles for lots of lenses which are provided by users of the software. I've provided profiles for most of my own lenses which
includes a lot of OM Zuikos as well as Canon mount Tokinas and
Tamrons.

The procedure involves photographing an architectural subject with
very square lines. The author was pleased with my test shots since he
said they were very easy to work with.  I chose a late 1950s 3-story
grammar school building whose exterior was largely straight aluminum
beams encompassing large windows and colored aluminum panels
separating the glass.  Vertical and horizontal lines all over the
place and very square.  He then does measurements from the images to
develop a distortion profile for the lens and at various focal lengths
for zooms.

PTLens is still available <http://epaperpress.com/ptlens/index.html>
You can get a free trial version that will correct 10 images.  After
that you need to buy a $25 license. Here's an interesting comparison
between PTLens and Photoshop lens correction.  PTLens is superior in
some cases. <http://epaperpress.com/ptlens/lensCorrect.html>

Chuck Norcutt


On 2/9/2016 11:15 AM, Jim Nichols wrote:
Thanks, Moose.  I use Picture Window Pro 7.0 as my image editor. I
dug
into the documentation and found that I can take a sample image made
with a given lens, pull a broken line to the edge that I want to
straighten, hit compute, and it will calculate polynomial
corrections to
remove the barrel distortion. This can then be saved for that lens,
and
recalled any time the correction is required.

With a few iterations, I got a perfect match to the in-camera JPG
image. While not an automatic solution, it only adds one step to my
work flow, and only for problem images.

It will even work with fisheye lenses.

Jim Nichols
Tullahoma, TN USA

On 2/9/2016 3:14 AM, Moose wrote:
On 2/8/2016 9:58 PM, Jim Nichols wrote:
If I may join this thread, I must admit that I learned something
new
(to me) today.  I have noticed that, starting with a RAW file, I
sometimes get some barrel distortion when capturing horizontal
lines
with the X-E1 and 35/2 lens.  When I look at the camera-produced
JPG,
most of the distortion is missing. I concluded from this that the
lens corrections are missing from the RAW files.

Is this correct?

Yep, it's a RAW file, and thus not processed to correct the linear
distortion, or anything else. The information to make the
corrections
is in the RAW file. The RAW converter makes the correction as
part of
the conversion from RAW to general purpose formats.

In the case of µ4/3 RAW files, Oly Viewer 3 and ACR/LR automatically
make the corrections, and in fact, that can't be turned off. So
their
outputs exactly match the JPEGs that were corrected in camera. DxO
Optics Pro doesn't use the manufacturers' settings, but uses their
own, from their own testing, and one may turn correction on and off.
In limited testing, their corrections were better than the
defaults in
ACR.

I know exactly nothing about Fuji RAW files and conversion, but
there
is undoubtedly a way to make distortion correction a part of the
conversion process.

Straight Lines Moose





--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz