SuperResolution using Rawtherapee "Capture Sharpen" and Gimp Up-Sizing ~ Part One ~ Definitions

Living where I do, I _know_ very much how fortunate I am to live in peace.  There is mental space and physical safety to do the things I want.  Like write these little blog entries.  Not everyone has this option these days.  I wish peace for everyone.

-------------------------

Before delving once more into the subject of SuperResolution, I thought it might be helpful to provide a few links and to talk a little about the background that has led to my current understanding of these kinds of things.  This will be the foundation for the next couple blog entries.

When we consider up-sizing an image, there are (at least) two algorithms to consider.

• Bilinear interpolation - https://en.wikipedia.org/wiki/Bilinear_interpolation
• Bicubic interpolation - https://en.wikipedia.org/wiki/Bicubic_interpolation

I have been using the "cubic" up-size algorithm because of a post made by Mike Johnson on The Online Photographer many years ago.  The topic was the print method Stephen Sharf used to make tack-sharp 13x19inch prints from a 4 mpixel camera.  That's right.  4, count 'um and weep - 4 mpixels.

The image Mike Johnson shared is, indeed, fabulous from the point of view of exposure, technique, and execution.

The technical background Mr. Johnson shared included -

"... The picture was taken with a Canon 1D, 70–200mm ƒ/2.8L non-IS lens with Canon 1.4X tele-extender II, for a combined focal length of 280mm, at 1/400 sec. and ƒ/13, handheld, at ISO 200.

As Stephen describes it, "This rider was going about 75 mph when the photo was taken, and you can see every stitch, vent perforation, and the pebble texture of the leather with excellent detail and clarity."..."

The approach Stephen Sharf used was -

• Capture image
• Sharpen using Photoshop's "Capture Sharpen" detail recovery function
• Up-Size using Bi-Cubic smoother
• Sharpen the image (UnSharp Mask, or something else)
  
• Print
  

You can read the original post to see the details Stephen used at the time (19 December, 2008!!!).

I adapted Stephen's approach for use with the Open Source Software, the Gimp.  Here's what I came up with -

• Capture image
• Sharpen using one of the following
• Gimp UnSharp Mask - around 1 pixel
• GMic sharpen functions - there are several, Richardson felt less "severe" than several of the options
  
• Up-Size using Bi-Cubic smoother
• Sharpen using one of the following
• Gimp UnSharp Mask - around 1 pixel
• GMic sharpen functions - there are several, Richardson felt less "severe" than several of the options
• Print

Pixel peeping the up-sized images left me feeling like there was _apparent_ resolution being "left on the table", as it were.  The edges of 2x linear/4x area up-sized work appeared slightly "soft" to my eyes.

The "softness" of the Cubic (in Gimp terms, BiCubic in Photoshop parlance) very likely comes from what cubic/bicubic interpolation does.  Study the BiCubic link above and you'll perhaps see what it might mean to interpolate a 4x4 cube (16 nodes/pixels) to generate an interpolated node/pixel.

In the next post I will share the results of up-sizing a synthetic and particularly difficult image.

 

 

Rawtherapee "Capture Sharpen" ~ showing my homework

Here is the second of my "show your homework" assignments.  

This time I look at Rawtherapee's "Capture Sharpen" function.  

I recently wrote -

"... the Lens Turbo II induced field curvature "feels" very mild in this setup.  If I applied "Capture Sharpen" to the Lens Turbo II ...  it would appear to be sharper than the native Sony A7 full frame image out of the camera...."

The above is a quote from my  post that looked at field flatness of a Sony A6000+Lens Turbo II+Nikon Nikkor 85mm f/1.8 Ai K setup.

Relatedly, I'll repeat something:  

Image processing can be as important to how "sharp" a photograph appears as just about anything else we consider.  This includes "lens resolution", which is really, for most of us, system resolution where sensor size, lighting, subject all influence the result.  Rarely are we able to consider a standalone optics true resolution, unless you're Roger Cicala at Lens Rentals.

Using a set of images I'd taken with a rather nicely preserved Nikon Nikkor-P 105mm f/2.5 Xenotar pre-Ai lens I hope to illustrate by sharing my homework of my reasoning behind the first quote at the beginning of this blog entry.  

In addition to using Rawtherapee's Auto-Matched Tone Curve set the image levels, I applied "Capture Sharpen."   I will look at the effects of Rawtherapee's "Capture Sharpen" function on apparently "sharp" images (ie: Sony A7 and lens) and apparently slightly less "sharp" images (ie: Sony A6000, Lens Turbo II, and lens).

 

 

 

If I understand things correctly, this is very similar to Lightroom's "Enhance Details" function.  These functions try to overcome the effects of strong AA filters (see: Canon digital sensors) and slow shutter speed image blurring.  I find it interesting that this function can/should(?) be applied prior to any additional sharpening (ie: UnSharp Mark, etc).

Rawtherapee's website says -

"... The Capture Sharpening tool helps recover details lost due to in-camera blurring, which can be caused by diffraction, the anti-aliasing filter, or other sources of Gaussian-type blur.

It is applied to the raw file immediately after demosaicing and modifies the data in linear gamma to limit halo generation. This means that it will only work on raw files..." 

Perhaps this will turn out to be a "crazy good" function in the cases where you're looking for absolute over the top kind of image "sharpness."  It could be useful when combined with another function or two.  Perhaps I'll write about that?

In the meantime, take a look at the following and see what you think.

Showing my homework -

 

 

 

Contrast matching between APS-C/Lens Turbo II and Full Frame ~ showing my homework

 Earlier I said something that likely needs a little proof.

"...Though the A6000 images show less contrast and appear "softer" than the A7 center images, the Nikon Nikkor 24mm f/2.8 Ai looks reasonably sharp in the center wide open and at f/8.  When contrast is matched with the A7, the A6000 24mm images look fairly comparable..."

This is a quote from my  post that looked at field flatness (or the distinct lack thereof) of a Sony A6000+Lens Turbo II+Nikon Nikkor 24mm f/2.8 Ai setup.

When I wrote that comment I knew I'd have to come back to it in the future.  Well, the future is now.

Image processing can be as important to how "sharp" a photograph appears as just about anything else we consider.  This includes "lens resolution", which is really, for most of us, system resolution where sensor size, lighting, subject all influence the result.  Rarely are we able to consider a standalone optics true resolution, unless you're Roger Cicala at Lens Rentals.

Using a set of images I'd taken with a beautiful Nikon 75-150mm f/3.5 E-series lens I illustrate by showing my homework my reasoning behind the above quote.

 

 

In the following image I "matched" the look of an APS-C sensor'd Sony A6000 + Lens Turbo II + Nikon E image to a Sony A7 Full Frame image I took moments before using the same focal length and aperture.

I did nothing to the Sony A7 image other than let Rawtherapee's Auto-Matched Tone Curve set the image levels.

For the Sony A6000/LT2/Nikon image I went back into Rawtherapee, opened the A7 image in a photo viewer, and went between the two applications until the A6000/LT2/Nikon image looked the same as the A7. 

In Rawtherapee what I needed to do to get the images to match was to set the Auto-Matched Tone Curve'd APS-C A6000 image to +25 Contrast and +3 Lightness.  

The histograms are very nearly identical in the following comparison. 

Showing my homework -

 

Nikon Nikkor-P 105mm f/2.5 Xenotar-type ~ Lens Turbo II APS-C vs Full Frame

Today I look at an old Nikon Nikkor-P 105mm f/2.5 Xenotar-type pre-Ai lens.  This is the last in this series of field flatness comparisons.  

The original question came up when I realized the Lens Turbo II focal length reducer was introducing field flatness issues.  The matter has come down to determining which lenses are most effected so I can avoid using them focal reduced on APS-C Sony mirrorless.

 

 

I've owned several of 105mm Nikkors over the years.  Currently I own two.  One was sent to me by a friend back in the States.  The lens is the earlier Sonnar design and it is very sharp from wide open.  It suffers from just a hint of field curvature wide open.

A couple years ago there was a newer design pre-Ai Xenotar version that came up on That Auction Site.  No one bid on it.  The opening bid was low enough I could ask "why not?"  For not much money I scored another pretty little lens.

Wide open the Xenotar version is not as razor sharp as the early Sonnar version, though the field is in my case flatter than the older design.  Stopped down (from f/4, on) I can't tell anything between them.

A different friend says that owning lenses can be like owning shoes.  Who wants to own just one pair?  It's a form of madness, I'm convinced.  I can't help myself.

Setup - 

• Sony A7 with straight-thru adapter - ISO100, 2 second timer, in-camera levels used to square the whole plot up
• Sony A6000 with Lens Turbo II focal reducer - ISO100, 2 second timer
  
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor-P 105mm f/1.8 pre-Ai - shot at two f-stops - f/2.5 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function (no "Capture Sharpen")
  

 

Comparison - 

Here is the scene setup.  As you can see, I moved away from using the apartment scrims (curtains) to shooting a copy of le Canard Enchaine' tapped to a flat wall.  I'm looking more carefully at field flatness, right?  Even given the all-too-often shoddy work standards around Paris, this wall is sufficiently flat for my purposes.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

Comments - 

The Nikon Nikkor-P 105mm f/2.5 pre-Ai is sharp in the center wide open and at f/8 on both the A6000 APS-C with Lens Turbo II and the Sony A7 full frame cameras.

The extreme lower left corner of the scene at f/2.5 the Lens Turbo II APS-C image looks a tiny bit soft, but unlike other lenses, this combo looks very good at f/8.  

The Sony A7 with straight-thru adapter at the very corners are quite sharp wide open and stopped down.

Finishing up this sequence of comparisons with a bang! the 105mm Nikkor looks pretty darned good on both Lens Turbo II focal reduced APS-C and Full Frame.

Nikon Nikkor 85mm f/1.8 K pre-Ai ~ Lens Turbo II APS-C vs Full Frame

I'm continuing to work with various Nikon lenses to see if I can answer the question of which lenses when used in conjunction with a Lens Turbo II focal length reducer exhibit unacceptable field curvature.  Here is a look at an old Nikon Nikkor 85mm f/1.8 pre-Ai lens.

 

 

Another sad tale to tell, but this one has a relatively happy ending. 

I had a single coated 85mm f/1.8 H and thought I _needed_ a multi-coated version of the optic.  So I sold the H and found what I thought was a reasonably priced K.

Normally I put a lens through all the paces before I shell out the bux.  This time I checked everything _but_ the focus ring action.  Yes.  This time I got bit.  Or, should I say, I bit myself for not paying full attention to what was in play.

When I got home I realized there was sand in the focus mechanism.  It's the first time I have ever encountered such a thing. Hundreds and hundreds of lenses have passed through my hands over the years.  Yet, there I was.  Sand.  How it got there and did _not_ sand-blast the elements is beyond me.  Maybe someone dropped it in the sand when they were at the beach?  I'll never know and by now it doesn't matter.

So I did the expected and disassembled the lens, re-greased the threads, and now all is good.  Still, I should've checked more thoroughly.  I should have been able to pick up this lens for practically nothing, given the sand-packed problem.

My H version of this focal length was brilliant from Day One.  The focus was as smooth as any Nikkor I've ever own.  Further, the more I thought about it the more I liked the idea of single coated optics.  Something to give the highlights a little "sparkle."  Yes, I regret selling the H. I may have to find another one.  Someday.  But only at the "right" price.  And you can be sure I'll be checking for any consequences from trips to the beach.

Setup - 

• Sony A7 with straight-thru adapter - ISO100, 2 second timer, in-camera levels used to square the whole plot up
• Sony A6000 with Lens Turbo II focal reducer - ISO100, 2 second timer
  
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor 85mm f/1.8 pre-Ai - shot at two f-stops - f/1.8 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function (no "Capture Sharpen")
  

 

Comparison - 

Here is the scene setup.  As you can see, I moved away from using the apartment scrims (curtains) to shooting a copy of le Canard Enchaine' tapped to a flat wall.  I'm looking more carefully at field flatness, right?  Even given the all-too-often shoddy work standards around Paris, this wall is sufficiently flat for my purposes.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

Comments - 

The Nikon Nikkor 85mm f/1.8 pre-Ai is sharp in the center wide open and at f/8 on both the A6000 APS-C with Lens Turbo II and the Sony A7 full frame cameras.

The extreme lower left corner of the scene at f/1.8 are equally soft between the Lens Turbo II APS-C image an the A7 full frame with straight through adapter.  In the past I've tested to see if refocusing the corners helps, and it does, but as expected the center then goes slightly out of focus. 

As with other lenses I've used in this way, the Nikkor 85mm f/1.8 K on my A6000 with Lens Turbo II in the corners at f/8 don't fully "clean up".  On the other hand, Sony A7 with straight-thru adapter f/8 images at the very corners are quite sharp, just as we would hope.

As I've said before with a few other lenses, the Lens Turbo II induced field curvature "feels" very mild in this setup.  If I applied "Capture Sharpen" to the Lens Turbo II f/8 image it would appear to be sharper than the native Sony A7 full frame image out of the camera.

Nikon Micro-Nikkor 55mm f/3.5 pre-Ai ~ Lens Turbo II APS-C vs Full Frame

Working with various Nikon lenses to see if I can't answer the question of which lenses when used in conjunction with a Lens Turbo II focal length reducer exhibit unacceptable field curvature, here is a look at an old Nikon Micro-Nikkor 55mm f/3.5 pre-Ai lens.

 

 

Said tale to tell, when I was walking over to the port de plaisance in Saint Laurent du Var one early morning to take photographs of old cars, the Sony A6000 with this Micro-Nikkor slipped off my shoulder and hit the concrete sidewalk.  The camera took the brunt of the fall.  I couldn't believe it and didn't even feel the camera slipping down my arm.

Fortunately the camera only fell hard a foot or two.  One of the dials is a bit dodgy, now.  But other than a few dings on the camera, and one small chip on the focusing ring of the lens, there doesn't seem to have been much damage.  Both remain usable.

I bring this up in the event that the lens is now out of alignment.  I don't think it is as all my images remain clear and sharp across the field.  What's certain is that it's no longer absolutely pristine as it was before the accident.  *sniff*

Setup - 

• Sony A7 with straight-thru adapter - ISO100, 2 second timer, in-camera levels used to square the whole plot up
• Sony A6000 with Lens Turbo II focal reducer - ISO100, 2 second timer
  
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Micro-Nikkor 55mm f/3.5 pre-Ai - shot at two f-stops - f/3.5 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function (no "Capture Sharpen")
  

 

Comparison - 

Here is the scene setup.  As you can see, I moved away from using the apartment scrims (curtains) to shooting a copy of le Canard Enchaine' tapped to a flat wall.  I'm looking more carefully at field flatness, right?  Even given the all-too-often shoddy work standards around Paris, this wall is sufficiently flat for my purposes.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

Comments - 

The Nikon Micro-Nikkor 55mm f/3.5 pre-Ai is sharp in the center wide open and at f/8 on both the A6000 APS-C with Lens Turbo II and the Sony A7 full frame cameras.

Looking at the extreme lower left corner of the scene at f/3.5 we see the edges are slightly out of focus.  It looks like they are equally soft between the Lens Turbo II APS-C image an the A7 full frame with straight through adapter.  

As with other lenses I've used in this way, the A6000 with Lens Turbo II corners at f/8 don't fully "clean up".  On the other hand, Sony A7 with straight-thru adapter f/8 images at the very corners are quite sharp, just as we would expect for the flat field macro design lens.

The Lens Turbo II induced field curvature "feels" very mild in this setup.  If I applied "Capture Sharpen" to the Lens Turbo II f/8 image it would appear to be sharper than the native Sony A7 full frame image out of the camera.

Nikon Nikkor-O 35mm f/2 pre-Ai ~ Lens Turbo II APS-C vs Full Frame

Continuing to work through the question of which lenses when used in conjunction with a Lens Turbo II focal length reducer exhibit unacceptable field curvature, here is an old and still rather pretty Nikon Nikkor-O 35mm f/2 pre-Ai lens.

 

 

I recently confirmed that this 50 year old Nikkor-O can, on the surface of things at least, keep up with newer current Wowy-Zowy optics.  If you think about it long enough, there is really no reason why old lenses can't produce images indistinguishable from the modern stuff.  As a bonus, old lenses can be purchased many times inexpensively.  If memory serves, I paid less than 50Euro for this 35mm f/2.

Setup - 

• Sony A7 with straight-thru adapter - ISO100, 2 second timer, in-camera levels used to square the whole plot up
• Sony A6000 with Lens Turbo II focal reducer - ISO100, 2 second timer
  
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor-O 35mm f/2 pre-Ai - shot at two f-stops - f/2 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function (no "Capture Sharpen")
  

 

Comparison - 

Here is the scene setup.  As you can see, I moved away from using the apartment scrims (curtains) to shooting a copy of le Canard Enchaine' tapped to a flat wall.  I'm looking more carefully at field flatness, right?  Even given the all-too-often shoddy work standards around Paris, this wall is sufficiently flat for my purposes.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

Comments - 

The Nikon Nikkor-O 35mm f/2 pre-Ai looks sharp in the center wide open and at f/8 on both the A6000 APS-C with Lens Turbo II and the Sony A7 full frame cameras.

Checking the extreme lower left corner of the scene at f/2 we see they are out of focus.  I believe this is minor field curvature in both the straight-thru and Lens Turbo II adapters.  If I focus the edges, they are sharper than they appear here.  

The A6000 with Lens Turbo II corners at f/8 don't fully "clean up".  There is still a small amount of field curvature.   On the other hand, Sony A7 with straight-thru adapter f/8 images at the very corners are quite sharp.

The Lens Turbo II induced field curvature "feels" relatively mild in this setup.  It certainly is nothing nearly as bad as the field curvature seen when the Nikon Nikkor 24mm f/2.8 Ai is used with the Lens Turbo II focal length reducer.

Nikon Nikkor 24mm f/2.8 Ai ~ Lens Turbo II APS-C vs Full Frame

I've begun this latest round of "let's have a look at things" by considering a well-used Nikon Nikkor-UD 20mm f/3.5 pre-Ai objective.

It's now the turn of a Nikon Nikkor 24mm f/2.8 Ai.

 

This is the very optic I complained suffered from field curvature.  At least this was the case when I used it with a Lens Turbo II focal reducer on my APS-C format NEX-5T, A5000, A6000, and NEX-7 cameras.

This blog entry is to confirm or deny that the field curvature is the fault of the lens.

Setup - 

• Sony A7 with straight-thru adapter - ISO100, 2 second timer, in-camera levels used to square the whole plot up
• Sony A6000 with Lens Turbo II focal reducer - ISO100, 2 second timer
  
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor 24mm f/2.8 Ai - shot at two f-stops - f/2.8 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function (no "Capture Sharpen")
  

 

Comparison - 

Here is the scene setup.  As you can see, I moved away from using the apartment scrims (curtains) to shooting a copy of le Canard Enchaine' tapped to a flat wall.  I'm looking more carefully at field flatness, right?  Even given the all-too-often shoddy work standards around Paris, this wall is sufficiently flat for my purposes.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

Comments - 

Though the A6000 images show less contrast and appear "softer" than the A7 center images, the Nikon Nikkor 24mm f/2.8 Ai looks reasonably sharp in the center wide open and at f/8.  When contrast is matched with the A7, the A6000 24mm images look fairly comparable.

Checking the extreme lower left corner of the scene we can see what I observed years ago.  For me it's the Smoking Gun (as it were, I'm really a peaceful kind of guy).  Right here is why I questioned the focal reduced 24mm Nikkor.  

Even stopped down, the corners at f/8 remain out of focus.  When used with a Lens TurboII the 24mm's field is very strongly curved.  It is so strongly curved, in fact, when the center is focused at infinity the corners of the frame remain out of focus.

On a straight-thru adapter on a Sony A7 the lens is beautifully sharp across a nicely flat field.

The Nikon Nikkor 24mm f/2.8 Ai does not work well when mated to a Lens Turbo II Nikon F to Sony E focal length reducer. 

I observed something very similar with a 35-105mm f/3.5-4.5 AiS that I briefly owned.  Not realizing what was going on, I blamed the extreme field curvature on the cheap little zoom and quickly sold it.  Could it be that I jumped too soon?  Maybe that zoom's field, too, isn't curved under normal circumstances?

Which raises yet another question - does the Lens Turbo II work better with some lenses and not very well with others?

In the meantime, I freely admit that I was rather slow to realize that field curvature I'd observed in an otherwise beautiful Nikon Nikkor 24mm f/2.8 Ai was the fault of the Lens Turbo II focal length reducer and not the Nikkor lens.

Nikon Nikkor-UD 20mm f/3.5 ~ Lens Turbo II APS-C vs Full Frame

I recently groused about how slow I was to realize that field curvature that I'd observed in an otherwise beautiful Nikon Nikkor 24mm f/2.8 Ai was the fault of the Lens Turbo II focal length reducer and not the Nikkor lens.

Once I worked through more than a few "DUH!!!! How stupid can I be?" moments, I thought I'd confirm my updated understanding and have a careful look at several lenses of various focal lengths to see how the Lens Turbo worked.

I'll begin by looking at a well-used Nikon Nikkor-UD 20mm f/3.5 pre-Ai objective and in future blog posts work my way up to 105mm's.

Setup - 

• Sony A7 with straight-thru adapter - ISO100, 2 second timer, in-camera levels used to square the whole plot up
• Sony A6000 with Lens Turbo II focal reducer - ISO100, 2 second timer
  
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor-UD 20mm f/3.5 pre-Ai - shot at two f-stops - f/3.5 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function (no "Capture Sharpen")
  

 

Comparison - 

Here is the scene setup.  As you can see, I moved away from using the apartment scrims (curtains) to shooting a copy of le Canard Enchaine' tapped to a flat wall.  I'm looking more carefully at field flatness, right?  Even given the all-too-often shoddy work standards around Paris, this wall is sufficiently flat for my purposes.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

Comments - 

The Nikon Nikkor-UD 20mm f/3.5 pre-Ai looks sharp in the center wide open and at f/8 on both the A6000 APS-C with Lens Turbo II and the Sony A7 full frame cameras.

Checking the extreme lower left corner of the scene, however, reveals that the Sony A6000 with Lens Turbo II focal reducer does introduce a focus shift.  The corners at f/8 don't "clean up" as well as they do on the Sony A7.

The induced field curvature "feels" minor in this setup.  I'm not sure how obvious the induced field curvature would be in "real life", but based on years of using Lens Turbo II adapters I doubt there would be much of a problem.

Nikon 75-150mm f/3.5 series-E Ai ~ Yet Another Look

Having looked at Nikon Nikkor 24mm f/2.8 Ai, an old Nikon Nikkor-O 35mm f/2 and a Nikon Nikkor 50mm f/1.8 AiS for field flatness on full frame Sony A7, here is a look at a Nikon 75-150mm f/3.5 series E AiS.

 

 

Setup - 

• Sony A7 - ISO50, 2 second timer, in-camera levels used to square the whole plot up
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon 75-150mm f/3.5 AiS - shot from f/3.5 through f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening)

Comparison - 

Here is the scene setup.  It's just a pair of closed gaze scrims in our apartment.  The details are interestingly small, so therefore useful for this kind of "wee look-see."  The center section, upper left corner, and lower right corners were used in the comparison.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

 

Comments - 

I know I'm beginning to sound like a broken record, but, wide open the 75-150mm Nikon is a bit soft.  

At f/5.6 and f/8 this un-loved under-appreciated ancient 2x good-for-nothing $25 lens is plenty "sharp" to my eyes, even though I think I see just a hint of field curvature at the extreme edges.  

I think there might be a bit of magic in this cheap optic. 

After having worked with manual focus Nikon lenses for many years I think I finally understand that this softness is deliberate.  I think it's how the Japanese express their sense of "delicateness" in photographic images.

Nikon's "Thousand and One Nights" says -

"...Although the E 75-150mm f/3.5 was manufactured as a popular lens for use with the Nikon EM, the finished lens was, in fact, sufficiently well executed as to find favor with the experts...

... the background blur is delightful, and makes the lens suitable for portraits... -–in line with the original concept for the lens. As on the AI Nikkor 135mm f/2 lens, the secret behind this is the deft use of close-range aberration fluctuations-–the correction for spherical aberration being slightly attenuated at close range in order to smooth the background blur..." 

What I think I see is that Nikon has developed a lens design approach that favors under-corrected spherical aberration behind the point of focus when a lens is shot wide open.  They seem to do this as a way of lending an image the Japanese "delicateness" I mentioned at the top of this section.

Nikon Nikkor 50mm f/1.8 AiS ~ Yet Another Look

Having looked at Nikon Nikkor 24mm f/2.8 Ai and an old Nikon Nikkor-O 35mm f/2 for field flatness on full frame Sony A7, here is a look at a Nikon Nikkor 50mm f/1.8 AiS "pancake".

 

 

Questions of field flatness arose when I used a Nikon Nikkor 24mm f/2.8 Ai in conjunction with a Lens Turbo II focal reducer on Sony APS-C mirrorless cameras.  I couldn't find a way to get that combination sharp in the corners, even stopped down and I couldn't believe that Nikon would ever design a lens with that much curvature.  

After moldering around for four or five years, the idea finally hit me that maybe the problem was the Lens Turbo II focal reducer and not the Nikkor.  Duh.

For completeness, I mounted a Nikon Nikkor 50mm f/1.8 AiS "pancake" on my low-mileage Sony A7 using a straight-thru adapter and pointed the lens at a fine-detailed subject.  The fine-detailed subject just happened to be our daylight back-lit apartment gaze curtains.

Setup - 

• Sony A7 - ISO50, 2 second timer, in-camera levels used to square the whole plot up
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor 50mm f/1.8 pre-Ai single coated - shot from f/1.8 through f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening)

Comparison - 

Here is the scene setup.  It's just a pair of closed gaze scrims in our apartment.  The details are interestingly small, so therefore useful for this kind of "wee look-see."  The center section, upper left corner, and lower right corners were used in the comparison.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

 

Comments - 

Wide open the 50mm Nikkor is a bit soft.  What was important to me was to confirm there was a similar level of softness across the field from the center all the way to the edges.  This appears, in fact, to be the case.

From f/2.8 on down through f/8 this little "pancake" lens is plenty "sharp" to my eyes.

I believe the Nikon Nikkor 50mm f/1.8 AiS "pancake" is the same design as the Nikon 50mm f/1.8 series-E and the more recent Nikon Nikkor AF-D 50mm f/1.8.

Nikon's "Thousand and One Nights" says -

"...the Series E 50mm f/1.8, sold in some regions including the U.S., and the AI Nikkor 50mm f/1.8S were actually sister lenses designed with the same basic optics, but different outward appearance and coatings...

What's more the same basic optical design was reborn as the AI AF Nikkor 50mm f/1.8S (coatings and such were modified), and is still available today (2016) in the AI AF NIKKOR 50mm f/1.8D. When this lens was reborn as an AF lens, it lost its slim, pancake design, but its rendering characteristics remain the same..." 

I've owned several series-E 50mm f/1.8 lenses over the years, and they performed for me the same way the AiS Nikkor version does.  All were/are slightly soft wide open and with a pleasing out of focus rendition.

This compares with my experience of using a Nikkor f/2 pre-Ai and Ai 50mm.  These are wickedly sharp wide open, but the out of focus rendition can be very harsh "soap bubble."  For those who believe that lens formulas hold the secret to great out of focus performance, both the harsh f/2 and beautiful f/1.8 50mm lenses are the exact same 6 element, 4 group plasmat layout.  The difference is in the calculation of the curves of the various lens elements.

Coming back to the original question, I don't recall seeing field curvature issues with the 50mm Nikkor f/1.8 AiS lens when combined with a Lens Turbo II focal reducer.  So, for whatever reason, there might not be the set of circumstances that lead to the effects I saw clearly with a Nikkor 24mm with a Lens TurboII.

In any event, this confirms the 50mm Nikkor is respectably flat field across the field and very sharp when shot on full frame.

Old vs New ~ Nikon Nikkor 50mm f/1.8 AiS vs Sony 50mm f/1.8 SEL OSS (APS-C)

Continuing in the style of recent posts and asking a question about how old lenses compare with current day computer designed optics, I take a quick look this time at a Nikon Nikkor 50mm f/1.8 AiS "pancake" lens and compare it against an APS-C format Sony 50mm f/1.8 SEL OSS.

Setup - 

• Sony A7 - ISO50, 2 second timer, in-camera levels used to square the whole plot up
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Lenses -
• Nikon Nikkor 50mm f/1.8 AiS - shot from f/1.8 through f/8
• Sony 50mm f/1.8 SEL OSS (APS-C) - shot from f/1.8 through f/8 shot Full Frame which shows the edge of the frame clipped due to the lack of FF coverage
  
• Rawtherapee RAW to jpg conversion - Auto-Match function
• "Capture Sharpen" used on the f/1.8 images to see how they compare with other apertures without "CS"
  

Comparison - 

Here is the scene setup.  It's just a pair of closed gaze scrims in our apartment.  The details are interestingly small, so therefore useful for this kind of "wee look-see." 

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

 

Comments - 

At f/1.8 the Nikkor 50mm is slightly softer than the Sony.  I figured this might be the case based on a webpage that has long disappeared that compared the Sony against Leica's latest 50mm and won.  Though I see nothing wrong with the Nikkor, this Sony optic is very special.

From f/2.8 on down through f/8 I have a difficult time telling the difference between the Nikkor 50mm lens and the Sony 50mm.

This comparison includes "Capture Sharpen"'d f/1.8 images.  It's interesting to see how much "crisper" these images become with a simple push of a button.

As I've suggested in the past, I suspect "Capture Sharpen" removes the character of a lens when shot wide open.  It's at the widest apertures where I've seen the designed-in effects of spherical and chromatic aberrations of old Nikkor manual focus lenses.

After seeing a large show of Steve McCurry's glorious images at the Musee Maillol where I studied his images with an eye toward such things, I may have more to say about "Capture Sharpen" and optical effects in the future.

Old vs New ~ Nikon Nikkor-O 35mm f/2 vs Sigma 30mm f/2.8 EX DN E

From a recent post, the Nikon Nikkor-O 35mm f/2 pre-Ai looks pretty good, regardless of age.  For me an obvious question is how it compares with current day computer design optics?

I have a Sigma 30mm f2.8 EX DN E that is wickedly sharp from f4 on down.  So I mounted the Sigma on my Sony A7 and shot it in full frame mode (no APS-C cropping).  I did this to keep things more or less equal between the newer and old lenses.  All I needed to do was move the tripod to account for the 5mm focal length difference.

Setup - 

• Sony A7 - ISO50, 2 second timer, in-camera levels used to square the whole plot up
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Lenses -
• Nikon Nikkor-O 35mm f/2 pre-Ai single coated - shot from f/2 through f/8
• Sigma 30mm f/2.8 EX DN E - shot from f/2.8 through f/8 shot Full Frame which shows the edge of the frame clipped due to the lack of FF coverage
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening)

Comparison - 

Here is the scene setup.  It's just a pair of closed gaze scrims in our apartment.  The details are interestingly small, so therefore useful for this kind of "wee look-see." 

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

 

Comments - 

At f/2.8 the Nikkor-O is just a touch softer than the Sigma.  Not by much, mind you, but just enough that I can see it pixel-peeping.

From f/4 on down through f/8 I have a difficult time telling the difference between the 50+ year old Nikkor-O 35mm lens and the 10 year old Sigma 30mm.  In terms of how I "see sharpness", the images are equivalent.

Though I don't show it here, by applying Rawtherapee's "Capture Sharpen" to any of the images, Nikon or Sigma, the "apparent sharpness" will visibly increase.

Coming to subtle differences between how the Nikon and Sigma lenses "render" an image, I found it interesting the way the tree in the background appears to be "softer" and "smoother" with the Nikkor-O than with the Sigma.  I could've predicted this based on earlier comparisons of point light sources out of focus.  Nikkor lenses, historically, have been designed for this kind of smooth out of focus rendition behind the point of focus.

So what, if anything, has been gained in commercial optics between the time of the Nikkor-O seen here and the Sigma Art DN?  Not much optically, maybe.  Additionally, perhaps not all lens manufacturers have Nikon's sensitivity to how the out of focus regions are rendered, either, but, a lot of other closely related things _have_ changed in 40 years.

We now have autofocus, as well as very small, sharp lenses like this little 30mm f/2.8.  It weighs all of 140grams compared with the Nikon 35mm f/2 at one aperture stop wider and 285grams, which comes without autofocus (of course).  

I'm certain that vast improvements have been made in "manufacturability" over the years.  It takes fewer steps to make a lens these days and with higher assembly accuracy (though I have a Sony 16mm f/2.8 SEL that might suggest otherwise).

Nikon Nikkor-O 35mm f/2 pre-Ai single coated ~ yet another look

After recently reconsidering what turns out to be a brilliant little Nikon Nikkor 24mm f/2.8 Ai, I thought it might be time to have yet another look at my old Nikon Nikkor-O 35mm f/2 single coated optics.

 

 

For this look at things, I wanted to confirm the field of this focal length lens is as flat as the newer 24mm Nikkor.

I mounted it up on the Sony A7 using a straight-thru adapter and pointed the lens at a fine-detailed subject.  The fine-detailed subject just happened to be our daylight back-lit apartment gaze curtains.

Setup - 

• Sony A7 - ISO50, 2 second timer, in-camera levels used to square the whole plot up
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor-O 35mm f/2 pre-Ai single coated - shot from f/2 through f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening)

Comparison - 

Here is the scene setup.  It's just a pair of closed gaze scrims in our apartment.  The details are interestingly small, so therefore useful for this kind of "wee look-see."  The center section, upper left corner, and lower right corners were used in the comparison.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

 

Comments - 

Borrowing from my Comments on the 24mm Nikkor, nearly everything I saw with that lens can be applied to this Nikkor-O 35mm f/2.

The Center images at f/2 looks a bit soft.  Is it the lack of contrast?  Or perhaps a bit of spherical aberration?  Both, probably, as it's certainly softer and less contrasty than at f/2.8 or anything further south than that. 

From f/2 down to f/8, the extreme corners clean up incrementally.  The corners are certainly not bad, even at wide open, but the improvement as the lens is stopped down is apparent.

Thinking about this, isn't it remarkable how a lens manufactured in the early 1970's, is single coated, and designed sometime in the 1960's can still look half way decent on modern digital cameras?  Standing alone without a reference, the lens looks nice.

It makes me wonder what useful gains have been made in optical design between then and now?  I'll consider this quesion more fully in a future post.

Nikon Nikkor 24mm f/2.8 Ai ~ AA filter "sharpening" effects

Writing a recent entry on rediscovering the Nikon Nikkor 24mm f/2.8 Ai and the fact it actually has a flat field afterall, I stumbled on something that I'd like to consider here.

Capture Sharpening.

This is something that Lightroom and Rawtherapee provide.  It is a function that attempts to offset the effects of Anti-Aliasing filters.  AA filters are used to combat moire in digital sensors.  Canon is famous for using strong AA filters.  Sony, from what I understand, has much lighter AA filters in many of their mirrorless camera offerings, including the NEX/A6xxx and A7 products.

I've been using Rawtherapee's Capture Sharpen without really thinking about it.  My images are "sharper", so what's not to like, right?

A couple years ago I acquired a very low mileage (ie: <700 click) Sony A7 full frame light grabbing mirrorless monster.  While this is only the Mark I version of a long line of great Sony 24 mpixel cameras it really performs.  The sensor is very quiet (ie: no noise) and is far sharper than any Canon DSLR I ever used (and I used a lot of them). Do I really "need" Capture Sharpen to crisp images up???

Using images from a prior post I applied Capture Sharpen to images at two apertures to see what effect the function had on the basic off the sensor image.

Setup - 

• Sony A7 - ISO50, 2 second timer, in-camera levels used to square the whole plot up
• Manfrotto tripod - it's capable of securing an 8x10inch view camera, so it's sturdy enough for this
  
• Nikon Nikkor 24mm f/2.8 Ai - shot at f/2.8 and f/8
• Rawtherapee RAW to jpg conversion - Auto-Match function plus Capture Sharpen

Comparison - 

Here is the scene setup.  It's just a pair of closed gaze scrims in our apartment.

 

 

[As always, click on the image and look at it to 100percent file size to see whatever there is to be seen.]

 

 

Comments - 

Looking at the images at f/2.8 we can see where Capture Sharpen function lives up to its name.  Capture Sharpen'd images at f/2.8 are "sharper" looking that images out of the camera at f/8.  And at f/8, Capture Sharpen takes images yet another step further into "sharp."

I am "good" with all this, until I start looking at the contrast in the out of focus regions (that little peek between the scrims).  It's there that something interesting is taking place.  The "soft smoothness" of the out of focus rendition is starting, to my eyes, become a little harsh.  The light/dark transitions have become rather "steep" and "contrasty."

Now I'm confronted with a reality of software processing "enhancements."  I ask myself several things - 

• Is this really how I want my images to appear?
• Does software intervention go some ways toward eliminating optical effects?
• What if native lens rendering, defects and all, are what attracted me to a set of lenses in the first place, only to be removed by software?

That last question is something rather interesting.  

For instance, let's say a lens manufacturer like, oh, let's take Nikon as an example, _deliberately_ designed their old manual focus lenses to be soft wide open and sharpen up as the aperture is stopped down.  

Let's say Nikon knew their Japanese customers _loved_ soft rendering under certain circumstances to create a "delicate" look.  

Let's say Western customers know little to nothing about any of this and just want "sharp", "sharp, "sharp" at all costs and at _all_ apertures.  Or, let's say that modern AF systems require a certain level of resolution wide open to focus accurately.

Which leads me to a related conversation made some years ago around the photo-forums.  That was the subject of "focus shift" between wide open and stopped down.  In particular, I remember comments about Nikon Nikkor 50mm f/1.4 lenses being prone to this dastardly defect.  Oh the horror.  Oh the hopelessness.  Oh the hand wringing.

Only... if a lens is designed to be deliberately soft wide open it will give the Japanese markets desired "delicate" effect.  It will exhibit under-corrected spherical aberration behind the point of focus.  One of the things about under-corrected spherical aberration is that there is a region, not just one point!, where an image will appear to be in focus.  

The implications of this are several fold and it might take several tumblers of the "Good Stuff" to think through all this and to come to an enlightened understanding of what's really going on.  I'll comment a bit more on these things in a future post.

Coming back to the effects of Capture Sharpen on softer images shot wide-open, I see where it's possible to make such an image appear as "sharp" as anything shot at other apertures.

The questions are: Is this really what I want?  Or do I want to retain the "character" of a lens as it was originally designed?