• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

A Mývatn Dream

Asher Kelman

OPF Owner/Editor-in-Chief
For some background and more images, please see my photo essay A Mývatn Dream.


So Rajan,

You are also hooked by the combination of the Canon 5DI and the 24mm TSE. For you picture her t works so well. Are you doing any post-processing.

I also looked at "Fog over Lake Mývatn" . This photograph s indeed deliciously wonderful with it's generous layering of pastels above the green landscape at the base of the picture. Why are you not also shooting overlapping images to cover the adjacent vew as this would be perfect for a panorama?

Asher
 

Rajan Parrikar

pro member
So Rajan,
You are also hooked by the combination of the Canon 5DI and the 24mm TSE. For you picture her t works so well. Are you doing any post-processing.
Asher,

Yes, the 5D Mark II and TS-E 24L II combination works beautifully. As you know, Canon's new version of TS-E 24 is a fantastic optic.

Apropos of post-processing: typically my first step is to do capture sharpening via Topaz InFocus which is a very good deconvolution programme once you get the hang of it. I hear that Topaz is currently involved in further improvements to their algorithm.

For this particular image, I worked in LAB space to crisply bring out the colour separation (with tips drawn from the wonderful book Photoshop LAB Color by Dan Margulis). Then seasoned it lightly with my favourite Nik plugin Color Efex Pro - needless to say, the defaults have to be modified, sometimes significantly.


I also looked at "Fog over Lake Mývatn" . This photograph s indeed deliciously wonderful with it's generous layering of pastels above the green landscape at the base of the picture. Why are you not also shooting overlapping images to cover the adjacent vew as this would be perfect for a panorama?
Asher
At the time, I was not mentally prepared to shoot panoramas. But with your prodding I have now come around to it (and I now also have the RRS basic pano kit). So don't be surprised if you see some panos in the coming months.
 

Asher Kelman

OPF Owner/Editor-in-Chief
Asher,

Yes, the 5D Mark II and TS-E 24L II combination works beautifully. As you know, Canon's new version of TS-E 24 is a fantastic optic.
Rajan,

This one optic has changed my shooting habits! Even now with Large Format 8"x10", I now go for my widest lens, the Schneider Super Symmar XL at 150mm dividing by 7.5-8 one gets 18.75mm to 20mm equivalent focal length, akin to stitching several overlapping 24mm shots on my 5DII!


Apropos of post-processing: typically my first step is to do capture sharpening via Topaz InFocus which is a very good deconvolution programme once you get the hang of it.
What's the logic of applying PK sharpener's "Capture Sharpening" (an implementation of the late Bruce Fraser's ideas) to Focus Magic. Do they advise sharpening so early in your workflow?

For this particular image, I worked in LAB space to crisply bring out the colour separation (with tips drawn from the wonderful book Photoshop LAB Color by Dan Margulis). Then seasoned it lightly with my favourite Nik plugin Color Efex Pro - needless to say, the defaults have to be modified, sometimes significantly.
I wonder whether the work in LAB space really pays off. I have to check if I have Margolis' book still.

At the time, I was not mentally prepared to shoot panoramas. But with your prodding I have now come around to it (and I now also have the RRS basic pano kit). So don't be surprised if you see some panos in the coming months.
For the benefit of others wanting to follow in your footsteps, given your distances from the landscape in the examples you share here, one can easily stitch 24 mm pictures, either handheld or simply rotating the head of your tripod! An expensive pano head is wonderful but not needed.*

I like taking overlapping pictures, especially where the travel distance and expense is high. There's a special thrill in discovering even better compositions once one has one's first print and can start the work of making the various elements work together. Having that extra tree or piece of a cloud can be all that makes the image superb for wonderful.

Asher

*A front back sliding Arca Swiss type stage to get the entrance pupil over the vertical axis, with your camera in portrait position, using an L bracket from http://ReallyrightStuff.com would work perfectly. This is all one has to have, unless making professional 360 degree virtual panos or else using a long lens and focusing to within 4-5 meters, there's often no call for a full pano head.
 
Even now with Large Format 8"x10", I now go for my widest lens, the Schneider Super Symmar XL at 150mm dividing by 7.5-8 one gets 18.75mm to 20mm equivalent focal length, akin to stitching several overlapping 24mm shots on my 5DII!
Off-topic, but bloody heck, Asher, I had no idea you were shooting the SSXL 150 on 8x10in - what a dream combination! Where can I see some of your work with this combination?
 

Asher Kelman

OPF Owner/Editor-in-Chief
Off-topic, but bloody heck, Asher, I had no idea you were shooting the SSXL 150 on 8x10in - what a dream combination!
It's the result of my lust for large prints packed with detail and great shading, but without the labor of stitching or the cost of a MF digital back!

Where can I see some of your work with this combination?
Hopefully, here, very shortly when my first negatives return to be scanned! This is all new to me as I've been locked into stitching and hours on the computer. Enough is enough!

Asher
 

Rajan Parrikar

pro member
Rajan,

What's the logic of applying PK sharpener's "Capture Sharpening" (an implementation of the late Bruce Fraser's ideas) to Focus Magic. Do they advise sharpening so early in your workflow?
Asher,

I don't understand your question. I don't use PK Sharpener. That software, I understand, does not use deconvolution. Whereas Topaz InFocus does. Deconvolution helps recover some of the sharpness/detail lost during capture due to the anti-aliasing filter on the DSLR sensors.

I wonder whether the work in LAB space really pays off. I have to check if I have Margolis' book still.
It does not always pay off (Margulis spells out instances where it does). In fact, I use LAB sparingly, only in those situations where I think it might offer more purchase than RGB space.
 

Asher Kelman

OPF Owner/Editor-in-Chief
Asher,

I don't understand your question. I don't use PK Sharpener. That software, I understand, does not use deconvolution. Whereas Topaz InFocus does. Deconvolution helps recover some of the sharpness/detail lost during capture due to the anti-aliasing filter on the DSLR sensors.
Rajan,

The custom of so-called "capture sharpening" came from Bruce Fraser and the PK software which packages his ideas. So I wondered why you would use the deconvolution at the early stage immediately after translating the RAW file. That's what capture sharpening is.

Asher
 

Rajan Parrikar

pro member
Rajan,

The custom of so-called "capture sharpening" came from Bruce Fraser and the PK software which packages his ideas. So I wondered why you would use the deconvolution at the early stage immediately after translating the RAW file. That's what capture sharpening is.

Asher
Oh, I see.

Perhaps Bart can add more to this discussion.
 
Rajan,

The custom of so-called "capture sharpening" came from Bruce Fraser and the PK software which packages his ideas. So I wondered why you would use the deconvolution at the early stage immediately after translating the RAW file. That's what capture sharpening is.
Capture sharpening makes sense to do as early in the process as possible. The reason is that the image data is still relatively pure, neither contrast altered nor noise altered. Ideally, one would like to perform the math involved on the linear gamma Raw data, but doing it on the gamma adjusted data is also possible. Keeping the data in 16-bit/channel (or better, e.g. floating point) helps avoiding losses which can hamper the operation.

Depending on what is being deconvolved, it's best to address it right after it is created. So optical aberrations (lens aberrations, defocus, diffraction, AA-filter, camera shake, subject motion) would be candidates to address early in the Raw processing pipeline. The demosaicing itself is a compromise between detail and artifact generation. The demosaicing process adds its own interpolation blur and should be addressed early as well.

The challenge for a successful deconvolution is that each of the blur contributing sources has their particular Point Spread Function (PSF) and they're weighted differently with each change of settings. The more accurate one can estimate the combined PSF, the more successful the restoration of the original image data will be.

Cheers,
Bart
 

Rajan Parrikar

pro member
Bart,

Would it help the software developers if Canon, Nikon etc made available the transfer function data of the AA filters for every body? Wouldn't that provide a good initial point?
 
Bart,

Would it help the software developers if Canon, Nikon etc made available the transfer function data of the AA filters for every body? Wouldn't that provide a good initial point?
Hi Rajan,

The problem is that we are faced with a number of blur sources, and the AA-filter is just one of them. To give an idea, the sensor array's sampling density of Green is higher than for Blue and Red. It therefore depends on how the Raw converter handles that. Furthermore the different blur sources interact with each other.

Luckily, the added result of the different types of blur PSFs usually resembles something that looks a bit like a 2D Gaussian curve. So with an adjustable shape Gaussian PSF as input, with visual preview, we can come a long way with restoring the most significant detail. Another task remains, and theat is to differenciate between noise and detail. We do not want to increase noise.

Of course, when there is a dominant blur source, e.g. defocus or motion blur, then it's best to remove as much as possible of that even if the required PSF shape is not optimal for the other blur contributions. Infocus' estimate function work resonably well to guess the complex PSF, as long as the subjects are in the same focus plane. Out side that plane the correction will generate artifacts.

So, for images taken with good technique, and a good lens, and not too much diffraction, the InFocus generic settings with a small radius (07-0.9) already remove a significant amount of (lens+AA-filter) different blurs.

Cheers,
Bart
 

Asher Kelman

OPF Owner/Editor-in-Chief
So Bart,

As of today, what would be your recommendation for capture sharpening v. output sharpenng and how would you vary that according to whether you used DXO, Capture One, Adobe or other favorite RAW convertors?

Asher
 

Jerome Marot

Well-known member
This is a very nice image. The colors, but especially the road encircling the bottom of the landscape. Impressive!

However, I wonder if you should not have left a very little bit more space at the bottom, so that the yellow grass surface would be continuous from left to right. If this is cropped in post, you may want to try it.
 
Top