What's in Focus?

[xpatUSA]

I've been playing with Wavelet Processing lately. Amongst other stuff, it is possible to extract image data according to it's size - 2x2px, 4x4px, 8x8px, etc. Here's how to extract 1px data and thereby determine what was in focus and was not in focus.

Here's the victim:



And here's the victim de-composed into five wavelet detail levels in the GIMP:



This is the image you get for Scale or Level 1 only:



Looks just gray but a drastic application of the Levels tool gets you this:



Now we can see what was in focus down to the hair level.

We should remember that a scene detail that would be a pixel or two in the capture if in focus would be convolved into a blur circle if not in focus and so is unlikely to show at the 1px level. One caveat is that smooth or very low contrast parts of a scene that are in focus would not show in the above image - gotta have some texture.

Processing this shot told us what the DOF was by showing what was, in DOF terminology, "acceptably focused". Other wavelet levels could instead show what was almost in focus or indeed what was out-of-focus!

With a suitable scene e.g. grass, a procedure like the above could show a lens' field curvature, I reckon.

No doubt there are other ways to skin this particular cat but, for me, this is one of the best.

[mugge]

It you look very closely to a painting by van Gogh you'd see all the mistakes he has made. The brain sees a beautiful picture from 5m away and that is how the picture should be viewed.
Cheers Ole

[xpatUSA]

It you look very closely to a painting by van Gogh you'd see all the mistakes he has made. The brain sees a beautiful picture from 5m away and that is how the picture should be viewed.
Cheers Ole

Sounds like a rebuttal. Not sure what I said wrong, though.

[Tronhard]

Hi Ted:

I suspect, or surmise, that Ole is referring to the differences between a technically-inclined photographer and a general viewer.

Going back to the painting analogy that Ole referred to, the vast majority of people would stand back at a comfortable distance and appreciate the piece of art as a whole - and that is where the painter likely intended it to be viewed. On the other hand, those people who have got as close as possible to an artwork are usually students of painting (or painters themselves) who want to deconstruct the image for its technique. They are two different agendas, and have been around as long as we have had art to appreciate, I suspect.

I encountered the same thing at school when we studied and arguably destroyed Shakespeare - I would pose the question as to how many people today could look back and say they enjoyed reading and analysing the Bard's works. I was not terribly inspired myself, until my fifth-form teacher got us to put on a Shakespeare play and present it to the rest of the school. We suddenly had a context within which to examine the lines and gain the significance of them to be able to speak the parts. At the same time, those in the audience suddenly saw his works live and they gained a greater appreciation for the works: much more so that the traditional method of dismembering the work line-by-line, and examining the work with a microscopic view, but no context.

Personally, I want my photos to have quality, but my audience is not the pixel-peeper - it is the vast majority who will look at it as a whole. But that is just MY take on what I want to produce, and I don't pretend to impose my perspective on anyone else. In the other context, we will tend to link the quality of our work to the quality of our technology as the number of pixels, dynamic range and resolution etc. that are more functions of the technology that the photographer. It begs the question of why images of earlier photographers that are far less technically perfect in our modern context are still revered. I would like to think that my personal mantra of technique beating tech has something to do with that.

Some of my own favourite images are taken with older or more humble gear that would not stand detailed scrutiny at massive magnifications in post processing software. But again, I don't want to appreciate them that way...

Sounds like a rebuttal. Not sure what I said wrong, though.

[mugge]

Sounds like a rebuttal. Not sure what I said wrong, though.

You did not say anything wrong at all. In fact I think you are perfectly correct. To quote Trev:" Some of my own favourite images are taken with older or more humble gear that would not stand detailed scrutiny at massive magnifications in post processing software. But again, I don't want to appreciate them that way...", that is all I mean.
Cheers Ole

[arbits]

Sounds like a rebuttal. Not sure what I said wrong, though.

An image is what it is and I hope we don't need a tool to tell us if we like it or not.

A tool like this can be valuable when we have a lot of similar images and want to quickly and simply reject the soft ones. I am in that situation often with images more likely soft due to subject motion than mis-focus.

Fastrawviewer has focus peaking and edge detection displays and works quite well.

Photoshop has a focus area selection tool. I once played around with an action to run that on a batch of images to aid quick assessment.

[xpatUSA]

Originally Posted by xpatUSA Sounds like a rebuttal. Not sure what I said wrong, though.

An image is what it is and I hope we don't need a tool to tell us if we like it or not.

Thanks and agreed. Elsewhere, I am in the habit of analyzing images posted by others - and any tool that helps that activity is of interest to me.

A tool like this can be valuable when we have a lot of similar images and want to quickly and simply reject the soft ones. I am in that situation often with images more likely soft due to subject motion than mis-focus.

Fastrawviewer has focus peaking and edge detection displays and works quite well.

Photoshop has a focus area selection tool. I once played around with an action to run that on a batch of images to aid quick assessment.

FastRawViewer doesn't open Sigma X3F raw data, just extracts the JPEG. So I haven't bought it yet. Tempting though; I have their RawDigger - a wonderful tool.

The Photoshop tool sounds good but my computer is Adobe-free ...

[Manfred M]

An image is what it is and I hope we don't need a tool to tell us if we like it or not.

Agreed! Focus is only one component of "sharpness" as a focused image that has camera movement is just as "bad" as an out of focus image.

On the other hand, selective focus is a powerful compositional too....

To quote Henri Cartier-Bresson; “Sharpness is a bourgeois concept”. He does have a point...

[xpatUSA]

Hi Ted:

I suspect, or surmise, that Ole is referring to the differences between a technically-inclined photographer and a general viewer.

Hello Trev!

Going back to the painting analogy that Ole referred to, the vast majority of people would stand back at a comfortable distance and appreciate the piece of art as a whole - and that is where the painter likely intended it to be viewed. On the other hand, those people who have got as close as possible to an artwork are usually students of painting (or painters themselves) who want to deconstruct the image for its technique. They are two different agendas, and have been around as long as we have had art to appreciate, I suspect.

Yes - enter pointillism such as some works by Seurat. But there is quite a number of folks who would carry a loupe into the gallery and study the dots.

...

Personally, I want my photos to have quality, but my audience is not the pixel-peeper - it is the vast majority who will look at it as a whole. But that is just MY take on what I want to produce, and I don't pretend to impose my perspective on anyone else. In the other context, we will tend to link the quality of our work to the quality of our technology as the number of pixels, dynamic range and resolution etc. that are more functions of the technology that the photographer. It begs the question of why images of earlier photographers that are far less technically perfect in our modern context are still revered. I would like to think that my personal mantra of technique beating tech has something to do with that.

"Quality" is one those words with many meanings in photography. It seems that you lean towards the aesthetic whereas I normally think of quality as comprising of measurable properties such as mentioned here:

https://www.imatest.com/solutions/cpiq/

Some of my own favourite images are taken with older or more humble gear that would not stand detailed scrutiny at massive magnifications in post processing software. But again, I don't want to appreciate them that way...

Yes, that first image shot out of someone's window wouldn't have passed any test in the 20th Century and onward to date.

I'm still shooting a 18-yr old 3.4MP camera and enjoying it tremendously.

[Manfred M]

"Quality" is one those words with many meanings in photography. It seems that you lean towards the aesthetic whereas I normally think of quality as comprising of measurable properties such as mentioned here

You continue to assume that quality is definable and measurable.

When it comes to something with tangible measures, that can prove to be easier than when we are looking at the intangible where opinion becomes part of the metric. If I have a specification for something and I measure perfection as to whether or not the item I am examining "meets spec", then we can probably do so.

When we add the intangible, this is a lot tougher (and less meaningful) to do.


Let's look at one of Robert Capa's well known images taken at Omaha Beach on D-Day.





From a technical (measurable) standpoint, the image quality is terrible; the shot is not particularly sharp (trying to stay on topic here). I would go a step farther and suggest it is not all that well composed either, as I would suggest that there are plenty of distracting elements, the positioning of items in the frame and of course the use of white space can all be questioned.

From those measures, we should trash this image because of the technical problems. Composition isn't great either, but it does have high impact. In face, one could argue that this impact is high because of the technical and compositional failings. So let's agree that this is not a good image, in fact it is a timeless great image, in spite of the measurable quality being on the poor side.

[DanK]

What makes the Capa photo important is what it showed, not its poor technical quality. No one would have judged it less well if it were sharp. And many comparably low quality images of something less important than this would likely be considered discards. After all, most of us can't rely on photographing something like D-Day to make our photos valuable.

I think this discussion is trying to make something binary that isn't binary. Is sharpness all there is to image quality? Of course not. Does sharpness often contribute to image quality? Of course. Does the importance of sharpness vary depending on the type of image? Of course. For example, it's more important in macro photography than in street photography. Often one might find a not sharp street photo valuable because of its content. Unsharp macro photos--except for those that deliberately use narrow DOF--are discards.

It you look very closely to a painting by van Gogh you'd see all the mistakes he has made. The brain sees a beautiful picture from 5m away and that is how the picture should be viewed.

Perhaps, but many great painters went to great lengths to master specific types of brushwork, and while they contribute to the overall effect (just as sharpness does), you often have to get very close to see what was actually done. They weren't sloppy because people would view the image from a distance. For example, a considerable amount has been written about Renior's brushwork, and it was of course the core of Pointillism.

So can't we just say that sharpness is one quality that can contribute to a good image?

[xpatUSA]

You continue to assume that quality is definable and measurable.

I don't assume anything as regards device image quality. Credible people such as the IEEE do it for me:

"IEEE-SA Working group p1858

The Camera Phone Image Quality (CPIQ) standard seeks to standardize image quality test metrics and methodologies across the industry, correlate objective test results with human perception, and combine this data into a meaningful consumer rating system."

https://www.imatest.com/solutions/cpiq/

While posters here continue to talk about "quality" by itself without the word being preceded by a qualifier, there will plenty of controversy. Posting a blurry picture of an historic event counters my original post not a whit. I am rather surprised at the responses so far which keep bringing æsthetics into the equation.

When it comes to something with tangible measures, that can prove to be easier than when we are looking at the intangible where opinion becomes part of the metric. If I have a specification for something and I measure perfection as to whether or not the item I am examining "meets spec", then we can probably do so.

When we add the intangible, this is a lot tougher (and less meaningful) to do.


Let's look at one of Robert Capa's well known images taken at Omaha Beach on D-Day.





From a technical (measurable) standpoint, the image quality is terrible; the shot is not particularly sharp (trying to stay on topic here). I would go a step farther and suggest it is not all that well composed either, as I would suggest that there are plenty of distracting elements, the positioning of items in the frame and of course the use of white space can all be questioned.

From those measures, we should trash this image because of the technical problems. Composition isn't great either, but it does have high impact. In face, one could argue that this impact is high because of the technical and compositional failings. So let's agree that this is not a good image, in fact it is a timeless great image, in spite of the measurable quality being on the poor side.

[pschlute]

To be fair to Ted here, the title of his post and the content of it is all about "what is in focus and what is not". He does not attempt to make a quality judgement based on those findings, but to present them as they are.

I agree with Manfred that the Capa picture shows a powerful image, even though none of it is in what we would call "sharp focus".......but i don't think that is the point Ted is presenting here.

[xpatUSA]

To be fair to Ted here, the title of his post and the content of it is all about "what is in focus and what is not". He does not attempt to make a quality judgement based on those findings, but to present them as they are.

Thank you, Peter.

I agree with Manfred that the Capa picture shows a powerful image, even though none of it is in what we would call "sharp focus".......but i don't think that is the point Ted is presenting here.

Correct. It was certainly not my point.

[xpatUSA]

Here's a bonus - what's not in focus:



This image is not blurred by any means - wavelet detail levels were turned off ...

[Tronhard]

I think we come down to two areas of consideration here. One is what the camera can deliver, and the other is the impact of the image on the viewer. I perceive that Ted is leaning (in the context of this discussion) towards the former while the perspective that Manfred presents seems to me to lean towards the latter.

I often use the same Capa image myself to demonstrate the difference between technical performance and perceived value. With regards to that image - there were other, certainly much better technical images captured by Capa on D-Day, but this is the one that has gained traction and seen it as one of the 100 most significant images of the 30th Century by Time Magazine. To me, the technical flaws in the image create the sense of tension and drama that are lacking in many of Capa's other D-Day shots: there is a visceral aspect to them that portrays drama, tension and action. In fact, the result is almost like the charcoal drawings used to illustrate the opening credits for the Pacific TV war series.

Certainly, there is a constituency of photographers who love the distinctive results of using optics like the Petzval lens design of the last century, while others will seek utter sharpness and lack of any distortions - it seems to me that there is room in the world for these and all the others, as long as we appreciate each within its own context.

[xpatUSA]

I think we come down to two areas of consideration here. One is what the camera can deliver, and the other is the impact of the image on the viewer. I perceive that Ted is leaning (in the context of this discussion) towards the former while the perspective that Manfred presents seems to me to lean towards the latter.

Sorry, Trev, I must disagree. This thread, placed quite deliberately in the 'processing' forum, was not intended to initiate any discussion of the æsthetics of focus. The original post simply presented a method for showing the degree of focus in any image, albeit historic, arty, crafty, test, etc. Any images posted by myself in this thread are illustrations of the method and are not related in any way to the works of known artists or photographers.

I disagree because your post continues to conflate or group my method with stuff (the second area) that is not relevant to this thread.

Of course, I have no objection to a sub-discussion about the æsthetics of focus or lack of it - so long as it is not related to or made relevant to the original post.

[Manfred M]

Ted - focus is achieved at capture, so discussing it in the post-processing and printing thread is a bit of a stretch, unless one is working in the traditional darkroom, where once could have focus issues on the enlarger.

That being said, we can agree that in many instances sharp focus is an important attributed of a strong image in many cases. I suspect that we can also agree that in most cases, a soft image has less impact than a sharp one and being sharp in the wrong places does not work well either.

In the post processing example where you showing methods of compensating for lack of sharpness due to a focusing error is something that has been done (with varying degrees of success) for many years. I've know a couple of old B&W darkroom photographers who rebuilt images that were not sharp enough using the darkroom unsharp mask technique to save historically important images with out of focus issues. These images were taken on large format sheet film cameras, so making film masks and registering them took a lot of skill and hours of work.

The modern sharpening algorithms. like the ones that you mention can mitigate some focus issues. The best tools I have worked with is Topaz Labs Sharpen AI, which they claim uses machine learning technology to do its work, rather than the more traditional Wavelet or USM techniques from the past. Development is ongoing as I receive updates to the software about once a month and there are definitely improvements as the software matures. What's more it supports focus, camera movement and general softness corrections.

[xpatUSA]

Ted - focus is achieved at capture, so discussing it in the post-processing and printing thread is a bit of a stretch ....

Thank you, Manfred.

As a Moderator, please move the thread to a more appropriate forum, or better yet, just delete it.

[xpatUSA]

Ted - focus is achieved at capture, so discussing it in the post-processing and printing thread is a bit of a stretch, unless one is working in the traditional darkroom, where once could have focus issues on the enlarger.

That being said, we can agree that in many instances sharp focus is an important attributed of a strong image in many cases. I suspect that we can also agree that in most cases, a soft image has less impact than a sharp one and being sharp in the wrong places does not work well either.

In the post-processing example where you [are] showing methods of compensating for lack of sharpness due to a focusing error is something that has been done (with varying degrees of success) for many years. I've know a couple of old B&W darkroom photographers who rebuilt images that were not sharp enough using the darkroom unsharp mask technique to save historically important images with out of focus issues. These images were taken on large format sheet film cameras, so making film masks and registering them took a lot of skill and hours of work.

I showed no methods of compensating for lack of sharpness due to a focusing error !!!

The modern sharpening algorithms. like the ones that you mention can mitigate some focus issues. The best tools I have worked with is Topaz Labs Sharpen AI, which they claim uses machine learning technology to do its work, rather than the more traditional Wavelet or USM techniques from the past. Development is ongoing as I receive updates to the software about once a month and there are definitely improvements as the software matures. What's more it supports focus, camera movement and general softness corrections.

I find this trivial mention of sharpening methods almost insulting.

I mentioned no sharpening methods, either !!!

You seem determined to destroy this thread by talking about anything but the original subject matter!

I am forced to request that this thread be deleted due to your continuing misunderstanding of both the original post's content and it's intent and to your failure to move it to a more appropriate forum.