determining sharpness in photos

[jamn]

In a current thread ...

660 images were taken to allow comparison of the sharpest images in each condition.

I am curious to know how you reviewed 660 images for sharpness. Did you look at each image? Or use software? I am asking because I have recently been more interested in finding critical sharpness in a large number of focus stack shots. For the moment I am planning on using Fast Raw Viewer (which I do own) for this purpose, but have not dug into the program yet.

Thanks,

Jay

 

[HicHic]

Im probably the only person in the world to use my camera body, HDMI cable and large screen tv to cull 100s of ProCap shots.... 14X on a 56" screen from 10 feet away its pretty easy to see which eye etc is sharpest and marking them with either a lock or check. Plus , retaining muscle memory of the button or dials on the camera is always a good thing!

I try to cull my pro capture photos in the field during downtimes. It's very fast and convenient to do on the OM-1. It has a very good file delete system.

No it doesn’t. The OM-1 delete system is pathetically inadequate.

You can’t delete an entire burst. You can’t even IDENTIFY an entire burst.

For a camera that can generate 120 images in a second, to delete 120 images, you have to hold a button while endlessly spinning a wheel like a lunatic. To delete many bursts, hundreds of images, either risk a blister or do it on the computer. Your idea of “very fast and convenient” eludes me.

The OM-1 is even worse than its predecessor in this regard. The EM-1 mk2-3 wheels were open and accessible, and could be spun endlessly. The ridiculous recessed wheels on the OM-1 can only be rotated a short way before you must stop and raise and reposition your finger/thumb to start again.

The arrow buttons, which let you move through images 10 at a time, are unusable for selecting images for deletion, because OM evidently enjoys playing bad jokes on its users.

You can delete ALL the images easily, but that’s only safe if you religiously protect or rate the ones you want to keep first.

When it was released, the OM1 was the fastest camera in the world. But it has (still, because of a total lack of feature improvements via firmware) a delete system designed for one-at-a-time. Unless you’re doing an erase all, every single image must receive either an individual button press or a wheel click.

That is just not appropriate for a high speed camera.

You bring up a very good point. However, we all know OM isn't going to do sh*t to improve it in the next firmware update. If we even get one.

 

[Bassam Guy]

In a current thread ...

660 images were taken to allow comparison of the sharpest images in each condition.

I am curious to know how you reviewed 660 images for sharpness. Did you look at each image? Or use software? I am asking because I have recently been more interested in finding critical sharpness in a large number of focus stack shots. For the moment I am planning on using Fast Raw Viewer (which I do own) for this purpose, but have not dug into the program yet.

Thanks,

Jay

No matter what software you use, the process is also important.

I haven't used anything but Lightroom, so cannot say whether it is the best, worst, or middlest. However, I have had many occasions where I've had thousands of photos to sort through before "tomorrow morning".

(This was before Pro Capture, which now allows me to do most of my pruning with the shutter button.)

I doubt that you have 660 images of one single subject at one single focal length from one single angle (a "set"). So you don't have one big job but, rather, 66 small jobs (assuming ~10 similar shots per set).

Due to the nature of these jobs, my big pruning efforts are always of people, usually of a one person but sometimes a few more, so I'll immediately remove all out-of-focus, bad facial expression(s), obstructed photos within a set. Dramatic shots with wild facial expressions or interesting gestures, I'll keep. Of those that remain, I'll arbitrarily remove one of each pair that seem to have identically good AF, expression, and gestures. On average, I'd keep 25-30% of each set. I'll add a star to those that I deem worthy of consideration for delivery to the client.

 

[Trevor Carpenter]

I use Faststone to review my pictures

I load the pictures from my SD card to a new folder on my PC in their ORF form

The first thing I do mis look at all of them pictures and if I don't want to keep them I delete. A I understanding right in that what I am looking at and deleting is actually an embedded jpeg.

I have never used Fast Raw Viewer. Does that use the ORF which as I see it before processing will often look worse than the jpeg

Surely an out of focus picture will show up in either format

Why is Fast Raw Viewer so much better and quicker?

 

[philzucker]

I use Faststone to review my pictures

I load the pictures from my SD card to a new folder on my PC in their ORF form

The first thing I do mis look at all of them pictures and if I don't want to keep them I delete. A I understanding right in that what I am looking at and deleting is actually an embedded jpeg.

I have never used Fast Raw Viewer. Does that use the ORF which as I see it before processing will often look worse than the jpeg

Surely an out of focus picture will show up in either format

Why is Fast Raw Viewer so much better and quicker?

One reason is that the embedded JPG of an OM-1 ORF is 3200 by 2400 (around 7.7 MP) pixels, compared to the 5172 by 3880 (20.4 MP) of the RAW file. So if you only use the embedded JPG you're basing your decisions e.g. about critical sharpness on a file only a third of the size of the original resolution.

Here a 100% crop of a picture I took recently - the embedded, probably internally denoised and sharpened JPG left (scaled up by FRV to match the RAW's original resolution of course), the unsharpened and not denoised RAW right:



The RAW IMO clearly shows more detail. Makes it easier for me to check for sharpness. And I often have to crop a lot, e.g. for "bees in flight" pics, so critical sharpness on pixel level is important to me.

AFAIK other manufacturers embed full sized basic (not fine) JPGs in their RAW files, e.g. Nikon. In that case sharpness can be judged fine on the basis of a well exposed embedded JPG. But since available highlight and shadow headroom cannot be judged and critical detail may lurk there, you IMO are still better of using FRV (or any other viewer showing the original RAW data).

YMMV of course.

Phil

--
GMT +1
Gallery: http://photosan.smugmug.com

 

[jamn]

I use Faststone to review my pictures

I load the pictures from my SD card to a new folder on my PC in their ORF form

The first thing I do mis look at all of them pictures and if I don't want to keep them I delete. A I understanding right in that what I am looking at and deleting is actually an embedded jpeg.

I have never used Fast Raw Viewer. Does that use the ORF which as I see it before processing will often look worse than the jpeg

Surely an out of focus picture will show up in either format

Why is Fast Raw Viewer so much better and quicker?

One reason is that the embedded JPG of an OM-1 ORF is 3200 by 2400 (around 7.7 MP) pixels, compared to the 5172 by 3880 (20.4 MP) of the RAW file. So if you only use the embedded JPG you're basing your decisions e.g. about critical sharpness on a file only a third of the size of the original resolution.

Here a 100% crop of a picture I took recently - the embedded, probably internally denoised and sharpened JPG left (scaled up by FRV to match the RAW's original resolution of course), the unsharpened and not denoised RAW right:



The RAW IMO clearly shows more detail. Makes it easier for me to check for sharpness. And I often have to crop a lot, e.g. for "bees in flight" pics, so critical sharpness on pixel level is important to me.

AFAIK other manufacturers embed full sized basic (not fine) JPGs in their RAW files, e.g. Nikon. In that case sharpness can be judged fine on the basis of a well exposed embedded JPG. But since available highlight and shadow headroom cannot be judged and critical detail may lurk there, you IMO are still better of using FRV (or any other viewer showing the original RAW data).

YMMV of course.

Phil

I might add to Phil's comments here. I am attracted to Fast Raw Viewer for two features I might have already mentioned. The first is FRV's ability to show which area of a photo is in critical focus with green lines similar to what the camera shows when using focus peaking. The other, and perhaps my favorite, is that the raw photos can be viewed from the memory card. So not only is this software very fast at producing the raw "previews", it saves time not needing to transfer unwanted photos to your hard drive only to delete them a short time later. I originally brought this up in the context of shooting macros. In that case say I shoot 200 photos because I tend to focus just in front of the subject then I can use FRV to find the first photo that has critical focus I wish to use. Then I can move to the end of the brackets to find the last photo I wish to use. So far I have been doing this either in LRc's import process or I just throw my hands up because that takes so long and import them all. I may end up using only 60 photos from set. Hopefully I will improve and not take so many unneeded photos.

Another consideration: I have generally not minded importing, culling, then deleting unwanted photos. However, while I am learning focus bracketing, I am taking a lot of photos that get deleted. In addition to the time issue my understanding is that read write cycles may partially determine the life of a SSD. So it appears to me that in my use case FRV is a win-win. My original question was asked to determine if I am missing something or what other approaches might be effective.

While I may be sounding like an experienced user, I am not. The above is based on initial observations of the software. I will do my best to get some macro samples up before to long.

Thank you everyone for the valuable information,

Jay

 

[Trevor Carpenter]

I don't think I understood that FRV works on the card rather than on the drive , am I correct here?

 

[meow]

I don't have an efficient way of doing it. I just stack them in Photoshop in many layers and go through them manually, hiding/showing each layer to compare. It's very time consuming, but it works.

I have been trying to cull before sending to PS or Zerene. You might try Fast Raw Viewer. It highlights in Green lines what is in critical focus or you can choose to see what is in acceptable focus which is highlighted in red. One of the nice things about FRV is it reads the raw image directly from the card. No need to import photos that will not be used. The other is that it reads and presents the images incredibly quickly. Beats LRc by a mile. It has some other useful capabilities, but as I mentioned I am not familiar with it.

I thought one of the features of Zerene (or any stacking software) was to find the sharper images and stack only them. IOW, why not let Zerene sort through all the images while you have a coffee?

I'm pretty new to Zerene, but as far as I've noticed it doesn't tell you which images it uses and not. If it's in there somewhere, please tell me how to find it.

 

[john isaacs]

I am curious to know how you reviewed 660 images for sharpness. Did you look at each image? Or use software? I am asking because I have recently been more interested in finding critical sharpness in a large number of focus stack shots.

OM Workspace has a Focus Analyzer feature that can help you to pick the sharpest frames from a sequential burst. I've never used it myself and I don't know if it works with focus stack images, but it might be worth a try?

I tried it. It doesn't work.

I use the Lightroom compare tool to reject photos for lack of sharpness, but it is difficult to use.

 

[philzucker]

I don't think I understood that FRV works on the card rather than on the drive , am I correct here?

It works with any drive and directory you point it to. Working directly with the card just saves you the step of copying all files to the hard drive for culling.

Phil

 

[philzucker]

I don't have an efficient way of doing it. I just stack them in Photoshop in many layers and go through them manually, hiding/showing each layer to compare. It's very time consuming, but it works.

I have been trying to cull before sending to PS or Zerene. You might try Fast Raw Viewer. It highlights in Green lines what is in critical focus or you can choose to see what is in acceptable focus which is highlighted in red. One of the nice things about FRV is it reads the raw image directly from the card. No need to import photos that will not be used. The other is that it reads and presents the images incredibly quickly. Beats LRc by a mile. It has some other useful capabilities, but as I mentioned I am not familiar with it.

I thought one of the features of Zerene (or any stacking software) was to find the sharper images and stack only them. IOW, why not let Zerene sort through all the images while you have a coffee?

I'm pretty new to Zerene, but as far as I've noticed it doesn't tell you which images it uses and not. If it's in there somewhere, please tell me how to find it.

It doesn't tell you, but it pretty much ignores files without any sharp detail - as far as the end result is concerned. It just needs longer to process all unnecessary files, so I also try to feed it only those pictures that matter for stacking.

Phil

--
GMT +1
Gallery: http://photosan.smugmug.com

 

[meow]

I don't have an efficient way of doing it. I just stack them in Photoshop in many layers and go through them manually, hiding/showing each layer to compare. It's very time consuming, but it works.

I have been trying to cull before sending to PS or Zerene. You might try Fast Raw Viewer. It highlights in Green lines what is in critical focus or you can choose to see what is in acceptable focus which is highlighted in red. One of the nice things about FRV is it reads the raw image directly from the card. No need to import photos that will not be used. The other is that it reads and presents the images incredibly quickly. Beats LRc by a mile. It has some other useful capabilities, but as I mentioned I am not familiar with it.

I thought one of the features of Zerene (or any stacking software) was to find the sharper images and stack only them. IOW, why not let Zerene sort through all the images while you have a coffee?

I'm pretty new to Zerene, but as far as I've noticed it doesn't tell you which images it uses and not. If it's in there somewhere, please tell me how to find it.

It doesn't tell you, but it pretty much ignores files without any sharp detail - as far as the end result is concerned. It just needs longer to process all unnecessary files, so I also try to feed it only those pictures that matter for stacking.

Yeah, but that wasn't the question. I suspect the OP may have the same problem as I. I find it hard to see if there is a tiny sharp spot somewhere in an image in a focus stack towards the ends of the series. So I feed Zerene too many images to be sure and then I don't dare to delete the raws since I don't know which weren't used.

The ideal would of course be if the stacker told us what images it discards, but...

 

[ChuckB]

It doesn't tell you, but it pretty much ignores files without any sharp detail - as far as the end result is concerned. It just needs longer to process all unnecessary files, so I also try to feed it only those pictures that matter for stacking.

Yeah, but that wasn't the question. I suspect the OP may have the same problem as I. I find it hard to see if there is a tiny sharp spot somewhere in an image in a focus stack towards the ends of the series. So I feed Zerene too many images to be sure and then I don't dare to delete the raws since I don't know which weren't used.

The ideal would of course be if the stacker told us what images it discards, but...

My goal is the final stacked image. After that I can delete all the other images if space is a concern. If I need to tweak the final image I'll hold off on deletions in case I need to run the stack again.

 

[meow]

I don't have an efficient way of doing it. I just stack them in Photoshop in many layers and go through them manually, hiding/showing each layer to compare. It's very time consuming, but it works.

I have been trying to cull before sending to PS or Zerene. You might try Fast Raw Viewer. It highlights in Green lines what is in critical focus or you can choose to see what is in acceptable focus which is highlighted in red. One of the nice things about FRV is it reads the raw image directly from the card. No need to import photos that will not be used. The other is that it reads and presents the images incredibly quickly. Beats LRc by a mile. It has some other useful capabilities, but as I mentioned I am not familiar with it.

The function that gives you green color actually detects contrast edges. So it can give you false positives in the same way as peaking does in the camera. And also false negatives in a low contrast area that happens to be sharp.

How "Fine details" (red color) works I don't know. But it doesn't detect all sharp bits.

For example this image isn't sharp anywhere.



But the contrast between the blue flower and the white background is big enough to make "Display contrast edges" see it as a contrast edge - but sharp it ain't.

The green edge FRW has added doesn't show so good on the screen cap as it did in the program, but it's there around the edges of the blue flower's petals that are against the white background.



I agree that these functions are helpful though. But one can't relay solely on them for finding sharp spots.

 

[meow]

It doesn't tell you, but it pretty much ignores files without any sharp detail - as far as the end result is concerned. It just needs longer to process all unnecessary files, so I also try to feed it only those pictures that matter for stacking.

Yeah, but that wasn't the question. I suspect the OP may have the same problem as I. I find it hard to see if there is a tiny sharp spot somewhere in an image in a focus stack towards the ends of the series. So I feed Zerene too many images to be sure and then I don't dare to delete the raws since I don't know which weren't used.

The ideal would of course be if the stacker told us what images it discards, but...

My goal is the final stacked image. After that I can delete all the other images if space is a concern. If I need to tweak the final image I'll hold off on deletions in case I need to run the stack again.

OK, but that still isn't what the OP asked about.

 

[ChuckB]

My goal is the final stacked image. After that I can delete all the other images if space is a concern. If I need to tweak the final image I'll hold off on deletions in case I need to run the stack again.

OK, but that still isn't what the OP asked about.

Since he was talking about finding such images in a series of stacked shots, my point would be why was he doing that since stacking software would do it for him. If there is another reason to do so he has not stated it.

 

[martinhb]

In a current thread ...

660 images were taken to allow comparison of the sharpest images in each condition.

I am curious to know how you reviewed 660 images for sharpness. Did you look at each image? Or use software? I am asking because I have recently been more interested in finding critical sharpness in a large number of focus stack shots. For the moment I am planning on using Fast Raw Viewer (which I do own) for this purpose, but have not dug into the program yet.

Thanks,

Jay

There is a program called 'Filterpixel' which runs on both Mac & Windows.

Also there is another program called 'OptiCull' which works only on macOS.

 

[jamn]

My goal is the final stacked image. After that I can delete all the other images if space is a concern. If I need to tweak the final image I'll hold off on deletions in case I need to run the stack again.

OK, but that still isn't what the OP asked about.

Since he was talking about finding such images in a series of stacked shots, my point would be why was he doing that since stacking software would do it for him. If there is another reason to do so he has not stated it.\\

I would like to cull in order to save time and resources. As previously stated Importing potentially hundreds of photos that are just going to be deleted takes time and shortens the life of an SSD. As I understand it the life of an SSD is affected by the number of write cycles, but I am not a computer geek and welcome being corrected on that if it is not correct. It is not a huge issue, modern drives are designed to work around that limitation to a large extent. However, as I am learning how to create the stacks I am after I find myself with a large number of rejects as well as some attempts that are a total failure. So the other consideration is the time involved to import and attempt to stack completely out of focus shots.

Jay

 

[jamn]

I don't have an efficient way of doing it. I just stack them in Photoshop in many layers and go through them manually, hiding/showing each layer to compare. It's very time consuming, but it works.

I have been trying to cull before sending to PS or Zerene. You might try Fast Raw Viewer. It highlights in Green lines what is in critical focus or you can choose to see what is in acceptable focus which is highlighted in red. One of the nice things about FRV is it reads the raw image directly from the card. No need to import photos that will not be used. The other is that it reads and presents the images incredibly quickly. Beats LRc by a mile. It has some other useful capabilities, but as I mentioned I am not familiar with it.

The function that gives you green color actually detects contrast edges. So it can give you false positives in the same way as peaking does in the camera. And also false negatives in a low contrast area that happens to be sharp.

Thank you for pointing out the false positives issue. I will keep that in mind when culling.

How "Fine details" (red color) works I don't know. But it doesn't detect all sharp bits.

For example this image isn't sharp anywhere.



But the contrast between the blue flower and the white background is big enough to make "Display contrast edges" see it as a contrast edge - but sharp it ain't.

The green edge FRW has added doesn't show so good on the screen cap as it did in the program, but it's there around the edges of the blue flower's petals that are against the white background.



I agree that these functions are helpful though. But one can't relay solely on them for finding sharp spots.

I appreciate this final comment - it brings a nice balance that is so often missing in a critique.

Jay

 

[Danielvr]

I asked an AI friend about Windows software that offers automated sharpness evaluation and he gave me a few suggestions.

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

Here are six Windows-compatible software options suited for selecting sharp images from large bursts/stacks, with pricing and a brief summary:

Aftershoot – US $120/year (Selects plan, culling only)
http://www.aftershoot.com/pricing
AI-powered culling and editing tool that detects blur, closed eyes, and facial expressions. It ranks images based on sharpness and other criteria to help photographers select the best shots efficiently.

FilterPixel – Starts at US $8.99/month (Pro plan)
http://www.filterpixel.com/pricing
AI-driven culling software that flags out-of-focus or blurry images and groups similar shots. It helps photographers identify the sharpest image in a burst or stack, streamlining the selection process.

Narrative Select – Free tier available; Paid plans start at US $10/month
http://www.narrative.so/select
AI-assisted culling and editing software that integrates with Lightroom. It evaluates sharpness, facial expressions, and other factors to assist photographers in selecting the best images.

Imagen AI – US $12/month (billed annually) or US $18/month (billed monthly)
http://www.imagen-ai.com/culling
Cloud-based AI culling software that analyzes sharpness and other technical aspects to help photographers select the best images. It integrates with Lightroom and offers personalized editing profiles.

Lightpanel by ON1 – US $99.99/year (Selects plan)
http://www.on1pro.com/lightpanel/buy
AI culling software designed for Lightroom Classic users. It detects sharpness, closed eyes, and facial expressions, allowing photographers to quickly select the best images from a shoot.

Photo Mechanic – US $149/year (Standard) or US $249/year (Plus)
http://www.camerabits.com/purchase
Fast image viewer and culling tool that allows photographers to quickly review and select images. While it doesn't offer AI-based sharpness detection, its speed and efficiency make it suitable for manual selection of sharp images.

ACDSee Photo Studio – US $59.99 (one-time purchase)
http://www.acdsee.com/en/products/photo-studio/
Comprehensive photo management and editing software that includes tools for organizing, viewing, and editing images. It offers features for evaluating image sharpness and making selections accordingly.

Excire Foto 2025 – One-time payment of approximately US $189
http://www.excire.com/en/excire-foto/
AI-powered photo management software that includes sharpness evaluation features. It helps photographers identify and select the sharpest images from a collection.

Optyx – Approximately US $7/month
http://www.optyx.ai
Cloud-based AI culling software that focuses on face and focus scoring. It assists photographers in selecting the best images based on sharpness and other criteria.

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-

AI is known to make mistakes, but if you're looking for this sort of functionality, I think this list would be a helpful starting point.

 

[ChuckB]

jamn wrote:I would like to cull in order to save time and resources. As previously stated Importing potentially hundreds of photos that are just going to be deleted takes time and shortens the life of an SSD. As I understand it the life of an SSD is affected by the number of write cycles, but I am not a computer geek and welcome being corrected on that if it is not correct. It is not a huge issue, modern drives are designed to work around that limitation to a large extent. However, as I am learning how to create the stacks I am after I find myself with a large number of rejects as well as some attempts that are a total failure. So the other consideration is the time involved to import and attempt to stack completely out of focus shots.

Jay

Well, I am not an expert on SSD life affected by writes. But, if you use stacking software on your images on the SD card out of the camera, there would be no write cycle importing them to your SSD. If your only interest is in the final product of the stacking software, then the only write to your SSD would be the final product. OTOH if you want the pre-stacked images for any reason, then you obviously would import all of them at some point. The stacking software won't tell you which ones are the ones it used.

Regarding write life, I gotta guess that my C drive SSD has a zillion writes more from everyday use than importing a few hundred images.