Chasing Pixies ~ in the Real World

I've spent many years looking at various characteristics of photographic imaging systems.  

 

 

It started over 30 years ago with a question from a friend.  He wondered if we could use my USAF resolution test chart from Edmond Scientific to see how his large format lenses performed.  Kerry Thalmann and I looked at everything we could get our hands on.

After a physicist pointed out that "resolution" in common photographic use is really limited by the light sensitive materials (film, sensor) I turned my attention to understanding in to out of focus transition.  This, it appears, is one area where differences in optical design and implementation can be most easily seen.

 

 

Even more recently I've taken a look at how image processing software can influence the appearance of sharpness in an image.  Considering only the Rawtherapee Open Source Software application I've had a look at the detail enhancing "Capture Sharpen" function.

There is a lot to think about and learn.  I'm enjoying the journey.

At all steps I've tried to keep in mind how all these things relate to the real world of practical engaged enthusiast photographic craft in image generation.

In April 2021 while we were mid-pandemic Mike Johnson posted an article on the "Perfect Portrait Lens."  

 

 

Reading his article reminded me that I'd never fully explored the capabilities of 15cm, 21cm, and 210mm Voightlander Heliar large format lenses.  I'd owned several them over the years after hearing about how they were "magic" when shot at or very near wide open.  They had a veil of "softness" that overlaid a sharply rendered scene.  The "look" was distinct.

Heliars, along with Petzval, Verito, Darlot, Portland, Pinkham Smith, Gundlach,and Kodak Portrait lenses were considered by a small community of cognoscenti as the Holy Grails of the photographic optical universe.

 

 

Moving to Europe necessitated a serious downsizing of my collection of photographic materials.  Everything was sold at Fire Sale prices or given away to friends or the needy.  All found new homes except a pair of Canon DSLRs and Sony NEX5 (original) cameras and lenses that fit into a suitcase.

Now that I'm fully retired, settled in Europe, and have plenty of time on my hands (time on my hands? right! and if you believe that, boy! do I have a bridge to sell you) I've wondered if I could find lenses for my digital cameras that had the characteristics I enjoyed when I shot large format film.  I call it "chasing Pixies."

 

 

Taking everything I've learned and experienced thus far, and factoring Mike's comments on the "Perfect Portrait Lens" I finally feel I'm able to articulate in a clear manner what I'm looking for in lenses.  And more practically, I'm beginning to photograph in a style that hopefully has a stronger sense of coherency from image to image by selecting my cameras, lenses, and choosing specific processing steps.

Optical case in point?  Mike mentions the Nikon Series-E 75-150mm f/3.5 lens.  He writes -

"...In the 1980s, Nikon introduced a small series of budget lenses called Series E to go with the tiny EM beginner's camera. To Nikon's embarrassment, professional photographers began snapping up the simple 2X 75mm–150mm Series E zoom. The reason was simple: the lens had a smooth, pleasant look that made models look great. (Those were the days before Photoshop, of course.) It's not that it was unsharp, but it wasn't forensic..."

 

I just happen to have one that I picked up for shockingly little money and I've had a look at it over the years.  It's sharp from wide open, even at 150mm's.  Some people find this lens too soft for their liking.   Further, in the case of the lens I have, stopped down I can't tell the difference between it and my modern "wire sharp" lenses.

For me the "magic" is in the transition from in to out of focus behind the point of focus.  Nikon appears to have put some effort into making this lens perform well in this area.  It's only a 2x push/pull zoom.  Nikon made more than a few of them and they tend to be easily found.  This, I'm sure, is why mine came to me so easily. Perhaps people didn't really like them?

 

 

Of course I have to be different, right?  I'm looking to find and justify small format replacements to my old large format lenses.  So, to help prove to myself that this lens is "magical" in the ways I was looking for, I took it to the local farm show and deliberately shot as much as I could with it.  

The illustrating Sony A6000, Nikon E-series 75-150mm f/3.5 images here are from the Salon International de l'Agriculture 2022.  These were all shot at the supposedly worst, soft, ick! ack!! horrible, stay away from it!!! aperture - wide open.  I applied no, none, zip, zero "Capture Sharpen."  I adjusted the "curves" slightly and applied a G'Mic film emulation.  That's it.

I find it fun chasing Pixies.

2022 photo-ops ~ update

2022 isle de France Photo-opportunities - very heavily weighted toward things I enjoy

Salon International de l'Agriculture *DONE*  Photos here!

Retromobile - 16-20 March *DONE*  Photos soon

Foire Photo - Chelles - 20 March *Did not attend*

la traversee de Paris - 27 March *CANCELLED at the last moment!!!  Argghhhh*

Tour Auto - 25-30 April

Vintage Revival Montlhery - 7-8 May

Rallye des Princesses - 14-19 May

Paris - Rambouillet avec les Teuf-Teuf - 28, 29 May

Cafe Racer Montlhery - 18-19 June

le Mans Classic - 30 June - 3 July

 

Nikon Nikkor 50mm f/1.4 ~ S.C vs AiS ~ another look ~ chromatic aberration?

Living where I do, I absolutely know 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 amount to nothing important blog entries.  

Not everyone has this option these days.  We receive daily reminders of this fact and it's downright heartbreaking.  People are being killed for a man's out-sized sense of power, control, and entitlement.  I wish peace for everyone.

Here is what Peter Turnley has been doing in the Ukraine.  Here are some of his photos.

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

In a prior post I noted that the AiS version I have of a Nikon Nikkor 50mm f/1.4 exhibits a magenta color shift in broad/large high contrast transition areas.

Since the comparison images didn't quite match to my liking, I wondered how strong the color shift really was and I felt I needed to have another look at things.

I did this by taking the Rawtherapee "Auto Tone Matched" Nikon Nikkor-S.C 50mm f/1.4 image as the image to be matched and tried to process the AiS f/1.4 to match it as closely as I could.  

It seemed that if both lenses suffered from similar amounts of color shift in the extreme highlight to shadow transition areas that by rising the highlights to look like the S.C output that maybe, just maybe, the AiS color shift would diminish, if not outright disappear.

On the other hand, if the AiS' color shift was still present after doing what I could to match the luminosity of the two images, then I would be observing a real difference between these two double Gauss designs, or at least between the two individual samples I have on-hand.  

Further, if there was a difference, what would it look like in a human perception model color to black and white conversion?  I ask this because it is easy to see differences in colors of similar energy levels, but when converted to black and white certain out of focus effects change, to my eyes, for the better.  Said another way, we can see color changes across an out of focus disk, but if they are of the same energy level then the disk becomes even and flat when converted into black and white.

 

 

Setup - 

• Sony A7 - ISO100, +1EV, 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-S.C Auto 50mm f/1.4 pre-Ai
• Nikon Nikkor 50mm f/1.4 AiS
  
• Rawtherapee RAW to jpg conversion - Auto-Match function, AiS image processed to match the S.C photo by +3 Lightness, -10 Contrast
  

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 - 

Note: Please keep in mind I'm comparing single examples of lenses.  It's not like I have a deep repository of samples and examples of all these optics.  What I see in the lenses in my possession might not be seen in other people's copies.  And, as with my Lens Turbo II studies, not all effects are attributable to the primary optic.

Looking at the two lenses shot wide open, the AiS' magenta color shift is still visible after processing the image to more closely match the S.C.  The Nikkor-S.C image still looks good to my eyes.

Is this chromatic aberration in the AiS version?  Is this some effect of light bouncing off the sensor back into the lens back through the sensor?  Or is it something else entirely?  I'm not yet sure what it is.  Nor am I sure it would make much difference when photographing under less contrasty situations.  The contrast is rather extreme in this comparison.

The human perception color to black and white conversion is rather interesting.  I see no trace of the color shift converted to either lighter or darker gray tones, as both AiS and S.C images look equivalent.

So what do I have here?  I'm not entirely certain.  There must be other ways of looking at this, including recognizing the fact that the AiS design was used (if the internets are to be believed) in the AF-D 50mm f/1.4 Nikkor.  If the implementation is good enough for the AF lenses (yes, I realize there is a more recent G version), perhaps there's really nothing wrong with the lens I have? Maybe I'll have to go out and photograph with these two Nikkors and see how they behave in the outside world?

I'm chasing pixies.  Pixies, I tell you.

Nikon Nikkor 50mm f/1.4 lenses ~ Nikkor-S.C vs AiS ~ a comparison

Living where I do, I absolutely know 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 amount to nothing important blog entries.  

Not everyone has this option these days.  We receive daily reminders of this fact and it's downright heartbreaking.  People are being killed for a man's out-sized sense of power, control, and entitlement.  I wish peace for everyone.

Here is what Peter Turnley has been doing in the Ukraine.  Here are some of his photos.

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

After starting to clean up an old, well used Nikon Nikkor-S.C 50mm f/1.4 pre-Ai and seeing that it did pretty well, even though I haven't yet completed the CLA, I thought it would be interesting to see how it compared in its present state with a AiS version lens I recently picked up.

The Nikkor-S.C version comes from the early 1970's.  The Ais version comes from the early-1980's.  Optically it appears they are slightly different implementations of the base double Gauss design.  Compare the spacing between the 2nd and 3rd elements.  Scroll down the page to find a pdf cross-section of the Nikkor-S.  Scroll down this page to find a cross-section image of the AiS version.

In practice, what could the differences be, if any?  Let's have a look so see if we can find something to talk about.

 

 

Setup - 

• Sony A7 - ISO100, 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-S.C Auto 50mm f/1.4 pre-Ai
• Nikon Nikkor 50mm f/1.4 AiS
  
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening) to minimize processing effects
  

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 - 

In general the earlier Nikkor-S.C gives a slightly softer rendition than the AiS.  From wide open it's pretty clear that the newer AiS version gives more contrast.

With both lenses the scene is very sharp from f/2 on through f/5.6 (and beyond, though this comparison does not show that).  Further, in terms of resolution, there seems to be no practical difference stopped down between these old Nikkors and modern optics.

Exploring specific differences between the S.C and AiS version, some people have said that the S and S.C versions exhibit highlight "glow."  This is just about the point I feel I'm chasing pixies. I note that there are differences in how tonal values are recorded between the two lenses, but they are subtle.  Differences are easiest to see with both lenses shot wide open.

However, there is something curious about the way the AiS version renders color.  Have a look at the way the magenta color seems to wrap around the curtain cloth, particularly on the right side panel.  The S.C version doesn't show that.  Stopped down to f/2 the AiS color shift effect seems to disappear. 

The contrast range is pretty extreme between the out of focus background highlights and the backlit inside the apartment facing curtains.  I'm not sure it makes much difference in "normal" use and don't yet know under which conditions this AiS color shift occurs.  I've not yet had a chance to shoot both lenses long enough to know, one way or another.  Yet it is something interesting to think about.  

Is it the lens?  Is it the lens/sensor combination?  At the risk of not drawing any conclusions, I will take another look at the wide open color shift in my next blog entry.

Until then, please keep in mind I'm comparing single examples of lenses.  It's not like I have a deep repository of samples and examples of all these optics.  What I see in the lenses in my possession might not be seen in other people's copies.  As with my Lens Turbo II studies, not all effects are attributable to the primary optic.

Nikon Nikkor S.C 50mm f/1.4 ~ field flatness

Living where I do, I absolutely know 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 amount to nothing important blog entries.  

Not everyone has this option these days.  We receive daily reminders of this fact and it's downright heartbreaking.  People are being killed for a man's out-sized sense of power, control, and entitlement.  I wish peace for everyone.

And the best of luck to Peter Turnley.  He is in the Ukraine.  What he says about the refugees and the photographs he is making of them fleeing the war zone gives serious pause.

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

I'm starting to clean up an old multi-coated Nikon Nikkor-S.C Auto 50mm f/1.4 and thought it could be interesting to continue just a bit further down this Path to Insanity of considering oh so many many lenses for this, that, and another thing.

I realize that I'm not only being redundantly repetitive, but that I'm now looking for pixies.  The differences I've uncovered between various optics are difficult to discern.

Still, onward I march.

 

 

Setup - 

• Sony A7 - ISO100, 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
  
• Lens -
• Nikon Nikkor-S.C Auto 50mm f/1.4
  
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening) to minimize processing effects
• 
  

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 - 

This Nikon Nikkor-S.C Auto 50mm f/1.4 lens has a few problems.  

There was a bit of fungus which, fortunately, was easy to clean.  The coatings remain in immaculate condition.  As of this writing, I have 95 percent of the problem resolved.

A second problem is the grease is dry and the focusing ring is stiff.  In the process of trying to disassemble the lens I found that two screws are, with 50 years of service, welded in place.  I will have to drill out the old screws and re-tap these two locations.

To confirm the lens was worth the effort of completing a CLA I put it to my field flatness, "sharpness" test.  As you can see, yes, the field is flat across the entire field from wide open all the way down through f/5.6.

As for resolution, the lens is slightly soft wide open.  It sharpens up nicely at f/2, and is the equal of anything new or old from f/2.8 and beyond.

Once sorted, this could be an interesting lens.  There are hints of "greatness" in this one, as we will explore in the following two blog entries.

Nikon Nikkor 50mm, 55mm ~ Rawtherapee "Capture Sharpen" ~ a comparison

"Capture Sharpening" is a function that both the closed source Lightroom and open source Rawtherapee have.  I believe Lightroom sometimes refers to this as reveal details, or something similar.

To recap, 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.

Until recently I'd been using Rawtherapee's Capture Sharpen without really thinking about it.  My images were "sharper" looking.  Sometimes too sharp, in fact.  It is really something to look at a Capture Sharpened Sony file (incredible detail) and compare it against an old Canon 5D MkII file (soft, with noisy shadows).

Recently I looked Yet Again at Nikon Nikkor 50mm and 55mm lenses.  I thought it might be interesting to see what effect "Capture Sharpen" had on lenses shot wide open, and compare the apparent "sharpness" against these lenses closed down to f/5.6 (typically one of the sharpest apertures of full frame or APS-C lenses).

Setup - 

• Sony A7 - ISO100, 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 
• Micro-Nikkor 55mm f/3.5 pre-Ai
• Nikkor 50mm f/2 Ai
• Nikkor 50mm f/1.8 AiS
• Nikkor 50mm f/1.4 AiS
  
• Rawtherapee RAW to jpg conversion - Auto-Match function, Capture Sharpen applied to images shot wide open

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 - 

Across the board, "Capture Sharpen" takes somewhat soft wide open images and turns them into detail-rich, "sharp" looking photos.  It reveals so much detail that f/5.6 images straight out of the box look "soft" in comparison.

Of course there are differences in the amount of "sharpness" "Capture Sharpen" provides depending on the lens.  

The Micro-Nikkor "Capture Sharpened" is just too sharp to my eyes.  It looks un-natural.  I'll have to be careful to modulate the amount of sharpening and not simply use the "Default" function settings.  

The three Xenotar/Plasmat design 50mm lenses go from ridiculously sharp with the f/2 Ai through to merely super sharp with the f/1.4 AiS wide open after "Capture Sharpen" has been applied.

As for how "Capture Sharpen" effects out of focus rendition, I haven't yet developed a well-formed idea. I like the "softness" of the un-treated images, actually.  I think I'll have to sit with this while and perhaps do a few more comparisons.  I believe there may be a change in rendition, but can't yet prove it.

Until then, here's a link to an article on the value of less than perfect lenses.  I found that there are some interesting things to consider.

Nikon Nikkor 50mm, 55mm ~ yet ANOTHER comparison ~ geez ~ will I EVER stop this MADNESS???

Living where I do, I absolutely know 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 amount to nothing important blog entries.  

Not everyone has this option these days.  We receive daily reminders of this fact and it's downright heartbreaking.  People are being killed for a man's out-sized sense of power, control, and entitlement.  I wish peace for everyone.

And the best of luck to Peter Turnley.  He is in the Ukraine right now.  What he says about the refugees and the photographs he is making of them fleeing the war zone gives serious pause.

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

Continuing with the series of posts on field flatness, I'd like to take Yet Another Look at four "standard" focal length lenses.

What brings this on is the arrival of very fine Nikon Nikkor 50mm f/1.4 AiS.  Previously I had a "K" version of the lens without realizing it's implementation was different than the early, original Nikkor-S Auto.  I only recently tripped over this gap in my understanding when I read Richard Haw's page on the 50mm AiS.

This time I'll look at field flatness using a Sony A7 full frame camera and will avoid comparing anything with a Lens Turbo II focal length reducer on APS-C.  I have other things on my mind than further exploring the limits of the focal reducer.

Setup - 

• Sony A7 - ISO100, 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 Micro-Nikkor 55mm f/3.5 pre-Ai
• Nikon Nikkor 50mm f/2 Ai
• Nikon Nikkor 50mm f/1.8 AiS
• Nikon Nikkor 50mm f/1.4 AiS
  
• Rawtherapee RAW to jpg conversion - Auto-Match function, but nothing further (ie: NO Capture Sharpening) to minimize processing effects
  

Since the most obvious differences between lenses, in my experience, show up at full aperture, it's the only aperture setting I'll consider here for each lens.

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 - 

The Nikon Micro-Nikkor 55mm f/3.5 pre-Ai is nice and sharp wide open straight across the field.  This macro lens can be used at distances normally optimized for by non-macro optics.  It's really a sweet lens.

The Nikon Nikkor 50mm f/2 Ai is fairly sharp wide open across the nicely flat field, though it is clearly not as sharp as the Micro-Nikkor.  The behind the point of focus out of focus rendition is "harsh" and looks like "soap bubbles".  This is the classic indication that a lens has over-corrected spherical aberration behind the point of focus.

The Nikon Nikkor 50mm f/1.8 AiS is slightly less sharp wide open than the f/2 Ai.  We begin to see a certain quality of "softness" at f/1.8.  This is attributable to the under-corrected spherical aberration behind the point of focus.  The field looks flat.

The Nikon Nikkor 50mm f/1.4 AiS is slightly less sharp wide open than the f/1.8 AiS.  Here that quality of "softness" that I mentioned the f/1.8 version having seems to be more pronounced in this f/1.4 lens.  And as with all the lenses seen here, the field looks flat.

Not that this comparison is able to show the following, but there is something that I've noticed about over-corrected spherical aberration lenses, and that is that with these older designs they tend to appear "sharper" at the point of focus than their under-corrected spherical aberration designed lenses.  The design layout might be identical, but the lens curvature calculations make one implementation behave differently than another.

The three 50mm Nikkor lenses here all derive from the famous double Gauss 6 element 4 group symmetrical design.  I mention this because from time to time you will read how such and such a lens might have a "classic" double Gauss design "rendition" or "drawing".  I'm not sure what any of that means, since lens designers can choose to control various optical aberrations, including control of coma, astigmatism, and out of focus rendition in any basic optical design (Sonnar, Gauss, Petzval, Tessar, Artar, Dagor, modern computer calculated, etc, etc, etc).  

Lens designers can take an exact same design, change the curvatures of the lens' elements and create a lens that has either over-corrected or under-corrected spherical aberration behind the point of focus.  Nikkor lenses, historically, have been designed to deliver under-corrected spherical aberration smooth out of focus rendition behind the point of focus.  Except when they did not, as in the case of the 50mm f/2 Ai.

I don't think a person can look at a lens design and tell how it will perform.  We either need the words of the lens designers to guide us, or have a look for ourselves.  Of the three 50mm Nikkors, one is over-corrected spherical aberration behind the point of focus (the 50mm f/2 Ai), and two have under-corrected spherical aberrations (both the 50mm f/1.8 AiS and 50mm f/1.4 AiS).

Monochrome images ~ processing vs sensor ~ a comparison

Living where I do, I absolutely know 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 amount to nothing important blog entries.  

Not everyone has this option these days.  We receive daily reminders of this fact and it's downright heartbreaking.  People are being killed for a man's out-sized sense of power, control, and entitlement.  I wish peace for everyone.

And the best of luck to Peter Turnley.  He is in the Ukraine right now.  What he says about the refugees and the photographs he is making of them fleeing the war zone gives me serious pause.

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

I have a well-used Sony NEX-5T that I'm not sure what to do with.  Should I sell it?  Well, yes, but there have been no buyers.

Could I convert it to monochrome?  Maybe.  What stops me is for the cost of conversion I could pick up two mint condition used Sony A7 full frame bodies.

So what to do?  I'm not yet sure.  Maybe I'll just keep the camera and keep smashing on the shutter release to see if I can't improve my "seeing."  To make a few more images.

While considering my list of options, including converting the NEX-5T to dedicated monochrome, I came back across Monochrome Imaging Services website.  They have a page that illustrates resolution gains seen by converting an RGB sensor to monochrome.

They make an offer.  "Download any of the RAW files with the links below."  So I did.

Since I've written recently and in more than a few posts about Rawtherapee's "Capture Sharpen" function, I thought I would take the opportunity to see how software "sharpening" compares against native sensor resolution in monochrome.

 

Comparison - 

Here is the scene the Monochrome Imaging Services starts with.  I've passed it through Rawtherapee, "Auto-Matched Tone Curve", and downsized to illustrate the starting point.

 

 

Now, the comparison image.

[As always, click on the image and open at 100percent to pixel-peep.  In fact, you may need to view this a 400percent to see some of the rather subtle detail differences.] 

 

Comments - 

The monochrome converted sensor image is indeed very sharp.  Adding "Capture Sharpen" to the monochrome image seems un-needed.

Working from the color version I took into consideration two things.  

The first was to confirm how a simple de-saturation worked.  Removing the Bayer color layer of a sensor could, to my way of thinking, make a camera work a lot like a pan-chromatic film.  All color sensitivities of equal energy would give equal exposure values.  

In concept a color de-saturated to B&W image should match a monochrome converted sensor photo.  This is what I see in the comparison.  Moving to a different method of color to B&W conversion is where I see subtle value differences.

To understand B&W tonal separation, one of the things I've been exploring recently is human perception color conversion to B&W.  In this case color values are set to match how humans experience various colors in monochrome.  For instance, we "see" blue darker than we do red or green of equivalent energy levels.

With color images we have various tools to achieve tonal separation when performing a B&W transition.  We can work on images after the shutter has been released.  

Working from a monochrome converted sensor image flips the image creation process back to traditional processes.  It requires us to think ahead on which filters to apply from our old traditional B&W filters set we used when shooting film before the shutter is released.

The second thing in this comparison was to "Capture Sharpen" the color converted into B&W images to see how close I might get to the resolution of the monochrome converted sensor image.

Both de-saturated and human perception converted images experience increases in "sharpness" using "Capture Sharpen."  How close they come to matching the monochrome converted sensor output is a matter of how I "see" things.

At 400 percent image size the "Capture Sharpen" images look "crunchy" compared to the monochrome converted sensor.  The difference is subtle but, to me, clearly visible.  The monochrome converted sensor image is very smooth.

I shot film for many years and in my youth worked as a B&W print technician in Hollywood and Irvine, California.  Using this background as a reference point I feel I have a sense of how current digital compares with film

For instance, based on the size of the noise and looking at light/dark transitions, and smooth fields of color and tone I feel images from a 1inch sensor "look" and "feel" like 120 format output.  The larger APS-C sized sensor can closely match what I used to do in 4x5inch format.  Full Frame digital seems to match my old 8x10inch format output.

Completing this line of thinking and looking at the monochrome converted sensor output I feel this could be like using an old Folmer and Schwing 12x20inch view camera.

SuperResolution using Rawtherapee "Capture Sharpen" and Gimp Up-Rez'ing ~ Part Five ~ image stacking

Living where I do, I absolutely know 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 amount to nothing important blog entries.  

Not everyone has this option these days.  We receive daily reminders of this fact and it's downright heartbreaking.  People are being killed for a man's out-sized sense of power, control, and entitlement.  I wish peace for everyone.

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

Some time back I wrote about trying to achieve a truer than simple up-sized sense of Super Resolution through image stacking.  The idea is to try and create a situation that emulates the little sensor wiggle super resolution function that certain Olympus, Fuji, Hasselblad, Phase One, and Sony cameras provide.  

For instance, Olympus takes a 16mpixel sensor, shifts it carefully in the X and Y directions, and builds in-camera a 40mpixel image. They are trying to add detail to an image with the goal of creating an image that equals a 40mpixel native resolution sensor.  

If we use cameras that don't have this ability, letting small camera movements during fast multiple exposures might give similar results.  If we are careful and do it right.  That's the promise, at least.

Having seen that Gimp CUBIC interpolation is actually "soft", I'm not surprised that my first super-resolution results were less than spectacular.  So with this post I would like to consider the use of NoHalo interpolation and see if it gives better results.

Reusing an image I shared in the last post, let's have a look at image stacking 2x up-rez'ing and see if we can't actually _add_ resolution to an image.  

It should be noted that no one has yet taken me up on my offer for a recommendation for some really good south Indian.  That's OK, of course.  Not everyone loves Indian food like my wife and I do.

 

Materials -

• Sony NEX-5T 16mpixel camera (handheld - multiple exposures)
  
• 19mm Sigma f/2.8 EX DN E lens
  
• Rawtherapee "Auto Matched Tone Curve" function
• Rawtherapee "Capture Sharpen" function
• Hugin to align images
  
• Gimp to -
• Stack Hugin aligned images as layers
• Set opacity levels so each layer gets equal representation
• Upsize using -
  
• CUBIC interpolation algorithm
• NoHalo interpolation algorithm

 

Comparison - 

Here is the scene again.  I didn't put much effort into making the image as my stomach was full and the chai had me jump'n and jive'n.  The Sony NEX-5T/Sigma 19mm output is rather fine, in spite of the caffeine.

 

 

[As always, click on the image and open at 100percent to pixel-peep.]

 

 

Comments -

Starting with the base "Capture Sharpen"ed image we see that it is nice and "sharp" across the entire field.  At low ISO and processed in this way this little Sony NEX-5T/Sigma 19mm f/2.8 EX DN combination is sharp.  If I can't pull a great print from this setup, I'm not trying hard enough.  They are simply glorious, these cameras and lenses.

Looking at the image stacked 2x (going from 4900 pixels on the long side to 9800 pixels) CUBIC up-rez'd images, they appear "soft" to my eyes, just as I've come to expect.  The 1 pixel UnSharp Mask doesn't look much better.  These two image sets (simple CUBIC up-rez and 1 pixel USM) look very similar to the kind of output I had in my earliest posts on the subject, too.

An image stacked 2x NoHalo interpolation up-rez, on the other hand, looks better than the 2x up-rez'd USM'd CUBIC image.  Adding a 1 pixel USM to the 2x image stacked NoHalo up-rez looks clean.  Very clean.

The question is, do these NoHalo image actually contain more information than a one image 2x NoHalo up-size?  To try and answer this I added another set of images at the bottom of the above comparison.  They are from a single exposure up-sized 2x using NoHalo interpolation.

What do I see?  There are two ways of approaching an answer.  In terms of apparent resolution the single image 2x NoHalo up-size appears to me to be "sharper" than the multiple-image stacked effort.

On the other hand, the image stacked 2x NoHalo rendition is "smooth."  The noise has been averaged out of the final composite.  This might be useful for something I'll talk about in the next post.

After all is said and done, however, I don't see it worth the effort to image stack to try and emulate the "sensor wiggle" functions some cameras come with these days.  I'm right back to thinking that if (and that's a really strong IF) I need super resolution that it's better to shoot over-lapping sections of a scene and to stitch them into a large photo.

SuperResolution using Rawtherapee "Capture Sharpen" and Gimp Up-Sizing ~ Part Four ~ a real image

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.

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

OK.  Enough background, theory, image simulations, and newspaper "tests".  It's time to see if this thing has any legs.

Let's have a look at a simple snap I made on the streets of Paris recently after lunch up in the area we call "Little India."  As an aside, if you need a recommendation for some really good south Indian dosa let me know.  I might be able to suggest a place or two.

 

Materials -

• Sony NEX-5T 16mpixel camera (handheld)
  
• 19mm Sigma f/2.8 EX DN E lens
• Rawtherapee "Auto Matched Tone Curve" function
• Rawtherapee "Capture Sharpen" function
  
• Gimp to take Rawtherapee output and upsize using -
  
• CUBIC interpolation algorithm
• NoHalo interpolation algorithm

 

Comparison - 

Here is the scene.  I didn't put much effort into making the image.  My stomach was full and the chai had me jump'n and jive'n, rather than concentrating on my surroundings.  Still, I think is sufficient for the task here.

 

 

[As always, click on the image and open at 100percent to pixel-peep.]

 

Comments -

Redundantly Noted: Keep in mind that we are _not_ adding detail to an image by performing an up-size.  In this way, I feel uncomfortable talking about Up-Rez'ing an image.  Up-Sizing, yes.  Up-Rez'ing, no.  If the information isn't there in the base file, then it certainly will not be there in the up-sized image, either.  

Another Note: I realize it may sound like I'm bashing Canon sensors in the following comments.  Photography, however, is seldom just about the sensor.  Witness the incredible volume of wonderful images that are made using Canon gear.  If it were that bad, who would use it?  But it's not bad and I'm only speaking in terms of comparison.  

The first thing is that when I saw the first images out of a Sony A6000 APS-C camera some years ago, I immediately switched systems and sold all my Canon full frame and APS-C gear.  

The second thing is a comment by Brooks Jensen noting that images readers sent to be considered for publication were sharpest when Nikon gear was used.  Out of curiosity he was looking at the EXIF data to see if there were any trends, and apparently there were.

Starting with the base "Capture Sharpen"ed image we see that it is nice and "sharp" across the entire field.  At low ISO and processed in this way this little Sony NEX-5T/Sigma 19mm f/2.8 EX DN combination is sharper than any of the old Canon DSLR L-glass'd gear I used to own.  If I can't pull a fabulous print from this setup, I'm not trying hard enough.  They are simply glorious, these old cameras and lenses.

Looking at the 2x (going from 4900 pixels on the long side to 9800 pixels) CUBIC up-size'd images, they appear "soft" to my eyes.  Even the 1 pixel UnSharp Mask looks like it could be better.  This confirms my earlier thoughts.  Perhaps the Gimp's CUBIC isn't the best up-size interpolation algorithm?  Well, if it was all we had, right?, but these days we have other options.

The NoHalo interpolation algorithm, on the other hand, looks better straight out of the up-size than the USM'd CUBIC image.  Adding a 1 pixel USM to the NoHalo up-size?  Well.  Yes.  This looks rather good, doesn't it? 

The 2x up-sized NoHalo image reminds me very much of how my old Canon 5D MkII images looked in the Canon RAW viewer I used years ago.  To be redundantly repetitive, this from an APS-C NEX-5T 16 mpixel image that is 2x up-sized to 9800pixels!

SuperResolution using Rawtherapee "Capture Sharpen" and Gimp Up-Sizing ~ Part Three ~ a "test" image

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.

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

The distance from my first post on "SuperResolution" to the present is rather long and the path is fairly twisted.  

In this latest series of posts I explore using the Gimp's new(?)/recent(?) NoHalo Image Scale up-sizing function.

Recall from Graphic Design that -

Gimp NoHalo interpolation - "...NoHalo level 1 consists of one diagonal straightening subdivision followed by bilinear interpolation..."

In my last post I looked at a black and white synthetic image and the various effects of up-sizing and UnSharp Mask sharpening.  I found that NoHalo is sharper than Cubic interpolation and had none of the artifacts that the Cubic function did.  And after applying a 1 pixel USM to the NoHalo interpolated image, things, frankly, looked quite good.

Test Image -

For this blog entry we will consider the NoHalo interpolation using my current favorite "test" target, a copy of the French Canard Enchaine newspaper.  It's a great read, if you're ever interested in French politics and all the in's and out's of what it's like to try to govern a First World nation.

To make things more "interesting" I will present two comparisons, side by side (as seen below).

The left hand image was taken using a pretty Nikon Nikkor-P 105mm f/2.5 shot wide open at f/2.5.  I felt this might be a good "test" to see how the slight softness seen wide open behaved when "Capture Sharpened" and then up-sized.

The right hand image was created using the very same lens, but this time shot at f/8.  I felt it could be interesting to see what might happened when I up-sized a demonstrably pixel-sharp base image, and even to compare this against f/2.5 images.

[As always, click on one or both of the images and open at 100percent to pixel-peep]

 

 

Comments -

Note: Again, keep in mind that we are _not_ adding detail to an image by performing an up-size of this kind.  You can see this by my use of the word "up-size" in place of "Up-Rez."  If the information isn't there in the base file, then it certainly will not be there in the interpolated larger image, either.  So saying any of this is found under the heading of "SuperResolution" can be mis-leading.  In general, I've found that if I _really_ need extremely high resolution that it's better to shoot overlapping scene sections and then to stitch them together into a larger image.

We can easily see that RawTherapee's "Capture Sharpen" function works the treat on a base image.  In both the f/2.5 and f/8 cases, the 100 percent pixel-peep'd base images look and feel sharper after "Capture Sharpen."  Additionally, there's not much difference between the "Capture Sharpen" base images shot at the two apertures, either.

Starting with the 1.5x linear Cubic Up-Rez'd images, the edges are slightly "soft" after the interpolation to 9000 pixels.  Then, with a 1 pixel wide UnSharp Mask applied, it appears to me as if the image "cleans up" fairly nicely and is quite usable.

With the 2x linear Cubic up-sized images we begin to see the phenomenon that started me down the path of considering other image scaling interpolation algorithms.  The edges of the transition zones look slightly "soft."  This remains true after applying a 1 pixel USM as well.  

Looking at NoHalo straight out of the interpolation step, the 2x linear up-sized image looks sharper than the Cubic Up-Rez.  This, by going from a 6,000 x 4000 pixel image (24mpixel sensor) to a 12,000 x 8,000 pixel image size.  Crazy, isn't it?

Adding a 1 pixel USM to the NoHalo up-sized image seems to help it go from looking good to looking great.  And there are no obvious artifacts.  

This is what I was looking for when I started down the path of increasing an image's file size.

SuperResolution using Rawtherapee "Capture Sharpen" and Gimp Up-Sizing ~ Part Two ~ starting with a synthetic image

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.

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

Years ago I wrote my first blog entry on the topic of "SuperResolution."  A lot has changed since then with my knowledge on the topic having been greatly improved.

For example, I learned that when using the Gimp's Cubic (BiCubic in Photoshop parlance) interpolation using 300dpi or 360dpi sample (resolution) slices produced blocky, pixelated  images and were no better than the simplest linear up-size function.

Setting image size resolution to 1200 dpi sample slices for 1.5x linear/3x area (ie: 6000x4000 to 9000x6000 pixel) up-sizing can give a fairly nice, crisp looking image.  

2x linear/4x area (ie: 6000x4000 to 12,000x8000pixel) up-sized work, however, remained a challenge and were "soft" to my eyes when using the Cubic/BiCubic interpolation algorithm mated with the UnSharp Mask sharpening function.

Recently I noticed that the latest Gimp version includes two new to me image size change interpolation algorithms.  They are "NoHalo" and "LoHalo".  

How long have they been there?  I have no idea, but I decided to give them a try just in case I was missing something.

From Graphic Design  I learned that -

"...NoHalo and LoHalo are the new GEGL samplers, developed by Nicolas Robidoux to reduce the blur in the resampled images..."

Gimp NoHalo interpolation - "...NoHalo level 1 consists of one diagonal straightening subdivision followed by bilinear interpolation...

...LoHalo is described as a sigmoidized EWA (Elliptical Weighted Averaging) with the Robidoux bicubic for upsampling, blended with non-sigmoidized EWA Robidoux for downsampling..." 

So I decided to see how NoHalo compared with Cubic interpolation in up-sizing by 2x (linear) a synthetic 500 x 1000 pixel image. 

Since LoHalo seemed to be dedicated to down-sampling I took only a cursory look and returned quickly to working with NoHalo.

I created the synthetic image in the Gimp knowing that pixelation and related artifacts would raise their ugly heads and felt it would be interesting to see how the two up-size algorithms, Cubic and NoHalo, compared with each other.

 

Synthetic Image -

The image consists of three blocks of solid color.  They are oriented at 0, 23, and 45 degrees of inclination.  I added several 1 pixel wide straight lines oriented along the X, Y axis and two lines randomly oriented.  As you can see, there is some heavy pixelation going on with the randomly oriented lines. There is a dot in the scene as well.

I took the resultant 500 x 500 pixel image, inverted its color, and put the two images side by side.  It is this 500 x 1000 pixel image that I up-sized.

As you can easily see, I performed a Cubic up-size, then added a 1 pixel USM.  These are followed by a NoHalo up-size and then a 1 pixel USM.

[As always, click on the image and open at 100percent to pixel-peep.  In fact, open this at 400percent to see some rather interesting things that I will describe below.]

 

 

Comments -

Note: Keep in mind that we are _not_ adding detail to an image by performing an up-size.  In this way, I feel uncomfortable talking about Up-Rez'ing an image.  Up-Sizing, yes.  Up-Rez'ing, no.  If the information isn't there in the base file, then it certainly will not be there in the up-sized image, either.  

Looking carefully at the Cubic up-sized images, the edges are slightly "soft" in all orientations.  This continues with when a 1 pixel wide UnSharp Mask applied.  The randomly diagonal lines are not crisp and clear.  Black lines appear smaller than up-sized white lines.

I think we can account for these effects when we consider how a Cubic interpolation operates.  It uses 16 pixels to generate one interpolated pixel.  This I imagine is good when dealing with gradient areas, but going from pure white to pure black, and vice versa, we see the influence of averaging across a comparatively wide area (ie: the 16 pixels).

Looking at the NoHalo interpolation, I feel the simple up-sized image is already "sharper" looking than the 1 pixel USM'd Cubic up-size.  Adding a 1 pixel USM to the NoHalo image, to my eyes (and, yes, they are getting old, so YMWV) is correct and sharp.

In Part Three of this series, we'll have a look at how NoHalo can perform in the "nearly real world."