Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
06-09-2019, 11:00 PM - 1 Like   #16
Pentaxian




Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 7,806
Original Poster
QuoteOriginally posted by dms Quote
dc = 5 ft, df = infinity,1/d = 1/2 (1/5 + 1/infinity) = 1/10 --> d = 10 ft (as expected for the hyperfocal distance)
That works.

QuoteOriginally posted by photoptimist Quote
Your results sound like a problem with choice of radius of the CoC (Circle-of-confusion).
Given the CoC is set according to user expectations. I can make set a smaller CoC value in the calculator :-)

---------- Post added 10-06-19 at 08:14 ----------

The "double the distance" rule that is recommended in videos , simply doesn't work. It only works is the double distance equals the hyper-focal that that condition is totally independent from the distance of the foreground subject. The "double the distance" recommendation is wrong, in these videos and also in the Photographylife.com site who wrote an article about it.

06-10-2019, 05:49 AM - 1 Like   #17
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 34,422
Looking at my old FA 50 and it's distance guide, if I set infinity at ƒ16 everything down to about 6 or 7 feet should be in focus. If my subject is say 10-15 feet away I set it fo ƒ11, further away I set it to ƒ8. That makes things pretty simple. A wider lens will produce more DoF so I can reduce the aperture, a longer lens produces less.

The depth of field defines where you should focus... 1/3 of your DoF is in front of where you focus, 2/3s are behind. I look for something to focus on 1/3 of the way between the closes point I want in focus and infinity.

On this one, the closest point in focus was about 30 feet away from me so I was able to open to ƒ8 the thing I focused on was the tree closest to the water about half way into the frame.



ƒ8 was probably way more than I needed, but my theory has always been don't waste time of calculation, just use massive estimated overkill, give yourself plenty of room for error with an AF lens. With an MF lens, use the hyperlocal DoF guide/

That's what works for me




When shooting for everything in focus, the sharpness advantage between ƒ8 and ƒ5.6 is negligible. There's just no reason to use anything wider than ƒ8 on FF. or ƒ5.6 on APS-c.

But ƒ5.6 can work on FF as well.

Last edited by normhead; 06-10-2019 at 09:22 AM.
06-10-2019, 08:39 AM - 1 Like   #18
Pentaxian




Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 7,806
Original Poster
QuoteOriginally posted by normhead Quote
The depth of field defines where you should focus... 1/3 of you DoF is in front of where you focus, 2/3s are blind. I look for something to focus on 1/3 of the way between the closes point I want in focus and infinity.
I agree with that. Focusing at the double distance of the nearest point in focus on camera side would produce mostly sharp foreground but the background would be out of focus (referring to "expert" recommendations that I believe are wrong).
06-10-2019, 09:55 AM - 1 Like   #19
Pentaxian
MJKoski's Avatar

Join Date: Aug 2016
Posts: 1,491
It really depends on the lens. Zeiss Distagon 28/2 for example has focusing field which is almost a sphere wide open. It allows to pull spherical subject fully into focus but blurring everything else outside the sphere. Stopping down flattens the field and the lens start to lose its "drawing qualities" - 3D-effect if you like to call it that. Some Zeiss lenses are similar except, well, Planar-macros. Focused area cannot always be just measured like |----| thick wall which turns into |------------------------------| stopped down.

Example of Zeiss 28/2 spherical focusing field wide open:



06-10-2019, 11:26 AM   #20
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 4,074
QuoteOriginally posted by biz-engineer Quote
I agree with that. Focusing at the double distance of the nearest point in focus on camera side would produce mostly sharp foreground but the background would be out of focus (referring to "expert" recommendations that I believe are wrong).
Yes, focusing at the double distance of the nearest point in focus on camera side only works if the nearest point in focus is 1/2 the hyperfocal distance or greater.

The whole 1/3-2/3 rule also only works for specific distances and focal lengths. For greater distances and shorter focal lengths, the back focus is infinite. For short distances and longer focal lengths, the front and back zones are equal in size.
06-10-2019, 11:33 AM   #21
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 4,074
QuoteOriginally posted by dms Quote
1. This does not change the usefulness of the concept of the hyperfocal distance, it just means you can adjust the values for the degree of enlargement you use, or "quick and dirty" simply use the usual hyperfocal distance values but use a smaller aperture by a stop or two. Otherwise one does not have an easily used concept for zone focussing, and notwithstanding the above posts, I think simply using infinity cannot be right if there is nearer stuff of import--if for no other reason we tend to look at the larger (closer) things in a picture to assess sharpness--even if it is not the most important stuff.

2. Actually I make hundreds of prints--usually in the range 7"x11" to 8"x12"--for theatre producer/lighting designer/ costume ... That is pretty much what they expect to see and show. And for me (for FL of 28mm and below) I almost always use the hyperfocal distance--for theatre and outdoors.
If anything, knowing that DoF and hyperfocal distance depend on the photographer's goal/tolerance for CoC, makes the concept more useful. If you know the intended print size or required final required resolution, you know the right CoC to use.

The whole point of DoF calculators is to control what's in focus and what's out of focus in the final output by appropriate choices of aperture and focus distance. Sure, you can trial-and-error it with digital previews at various focus and aperture settings, but that will take a lot longer than just plugging a few numbers into a DoF calculator.

Even if you focus-stack, you still need some understanding of DoF (and CoC) to ensure the correct number and spacing of the focus slices.
06-10-2019, 12:43 PM - 1 Like   #22
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 34,422
QuoteOriginally posted by photoptimist Quote
If anything, knowing that DoF and hyperfocal distance depend on the photographer's goal/tolerance for CoC, makes the concept more useful. If you know the intended print size or required final required resolution, you know the right CoC to use.

The whole point of DoF calculators is to control what's in focus and what's out of focus in the final output by appropriate choices of aperture and focus distance. Sure, you can trial-and-error it with digital previews at various focus and aperture settings, but that will take a lot longer than just plugging a few numbers into a DoF calculator.

Even if you focus-stack, you still need some understanding of DoF (and CoC) to ensure the correct number and spacing of the focus slices.
IMHO, that's just not necessary for landscape as in the thread title. Maybe if you're focus stacking a macro where the back edge of your image is not near infinity. But most of the time I use hyperfocal settings it's with the horizon visible in the image. The far distance you want is infinity. All I need to know is what will be in focus at what distance at the front end of the image with the back edge set to infinity on the DoF guide. And that is pretty easily estimated with acceptable error built in.

For macro I use live view, manual focus and the focus confirm. No need to calculate anything. I can see it on the back screen.

But this also goes back to the need for resolution. You do need to know about the CoC if you plan to view a large print from too close to appreciate the composition. But as the print gets larger, your view from a further distance and the size of the CoC can be larger than they would have to be on a smaller print.

I always wonder about folks making things way more complicated than they need to be.

If you are concerned about minutiae then maybe these calculators and concerns will mean something to you. I've gone a lifetime using what I learned from my DoF guide on my 55 1.8 in film days without having to waste time with mathematical formula. You can, but in my experience you won't lose much if anything because you don't.

This is an area where you can develop a feel for what you are doing, and that's part of the skill of the trade. You can calculate stuff, but you'll never be as fast as those with a more experienced based approach. If you're a pro, time is money. If you're not, time wasted is shots missed.

The only way I'd recommend an actual DoF guide would be as training wheels while you're figuring it out.
But an FA 50 1.7 would be better. It's more intuitive and easier to understand and interpret. And you can take pictures with it.

I always find it so odd when people start recommending stuff I've never felt the need to do.
Sometimes it's not about how many technical calculations you bring with you and futz around with, it's about how confident you are in your ability to correctly gauge a situation. As longs I keep nailing the images the way I'm doing it, I'm not going to recommend something more complicated.

The DoF of even a 100mm lens is so narrow as to be pretty much useless for all in focus landscape if you're shooting objects form 10-15 feet way in the foreground. My DA*55 still has a DoF guide on top. If you apply 50mm DoF knowledge to a wider lens like a 20, 24, 28, 35, all you need to know is your lens is probably sharpest at ƒ5.6 and there's no reason to go below that. You portably have DoF over kill at ƒ5.6 or ƒ8.

Its the same with any skill, you train yourself to do what the amateurs need guides and other supports to accomplish. I also know how to file a chainsaw without a file guide, sharpen a chisel without guide to hold my angle and keep it precise without a guide, and file my teeth flat on a crosscut saw without a guide. That stuff is all part of learning your trade. And they make guides for all these things, I even own them , but they slow you down. My boss hated seeing guys using guides when he was paying them. You taking the extra time to use a guide cost him money, he'd be happier paying someone who was more efficient. Anything you can learn to do correctly without a guide increases your efficiency.

Maybe you need to work with a guide for a while, but your goal should be to leave it behind at some point. The look on my face right now is probably the look on one of the old cabinet makers seeing me use a sharpening guide to sharpen a chisel back when I was learning, Or the look a kid gets on his face when he sees a buddy using training wheels on his bike. Use them if you have to, you'll be happier when you learn to leave them behind, except in the most un-usual of circumstances.

Last edited by normhead; 06-10-2019 at 01:54 PM.
06-11-2019, 12:45 AM - 1 Like   #23
Pentaxian




Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 7,806
Original Poster
QuoteOriginally posted by photoptimist Quote
If anything, knowing that DoF and hyperfocal distance depend on the photographer's goal/tolerance for CoC, makes the concept more useful. If you know the intended print size or required final required resolution, you know the right CoC to use.
Well, by writing this post, I've come to realize that there is not such thing as a finite depth value for the so called "depth of field", DoF is only a range of distance where the sharpness is good enough*, along the axis perpendicular to the plane of focus.

(*) acceptable sharpness based on viewing distance and how good is viewer's vision.

In reality, every camera produce images that are only at their maximum resolution at the plane of focus, everywhere else the optical resolution falls off beside the plane of focus. Smaller sensors produce images that seem all sharp in depth because the sharpness diminished more slowly along the Z axis beside the plane of focus. Depth of field increased with reduction of aperture, but it is never infinite, at some point diffraction wins over increasing the DoF by stopping down the lens. So for cameras with larger sensors (e.g full frame and medium format), aperture / DoF / point of focus is a trade-off, and focus stacking is the method that breaks that aperture / DoF trade-off.

---------- Post added 11-06-19 at 09:49 ----------

QuoteOriginally posted by photoptimist Quote
Even if you focus-stack, you still need some understanding of DoF (and CoC) to ensure the correct number and spacing of the focus slices.
Yes, the number of exposures required will be related to how deep the is scene relative of the "DoF" thing. With our digital cameras, CoC shouldn't be based on human vision, CoC should be a function of sensor resolution I suppose it would make a lot more sense.

06-11-2019, 05:54 AM   #24
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 4,074
QuoteOriginally posted by biz-engineer Quote
Well, by writing this post, I've come to realize that there is not such thing as a infinite depth value for the so called "depth of field", DoF is only a range of distance where the sharpness is good enough*, along the axis perpendicular to the plane of focus.
If DoF is measured as the distance between the nearest and farthest points with acceptable sharpest, then it can be infinite in that everything from some near-ground object to the most distant stars is acceptably sharp.

QuoteOriginally posted by biz-engineer Quote
(*) acceptable sharpness based on viewing distance and how good is viewer's vision.
Acceptable sharpness at the print/screen level is determined by viewing distance and the view's visual acuity. Translating that to a physical CoC on the sensor also requires knowing the print size and the sensor size.


QuoteOriginally posted by biz-engineer Quote
In reality, every camera produce images that are only at their maximum resolution at the plane of focus, everywhere else the optical resolution falls off beside the plane of focus. Smaller sensors produce images that seem all sharp in depth because the sharpness diminished more slowly along the Z axis beside the plane of focus. Depth of field increased with reduction of aperture, but it is never infinite, at some point diffraction wins over increasing the DoF by stopping down the lens. So for cameras with larger sensors (e.g full frame and medium format), aperture / DoF / point of focus is a trade-off, and focus stacking is the method that breaks that aperture / DoF trade-off.
Correct. If you 3-D plot sharpness as a function of BOTH distance and aperture, it is shaped like a triangular bowl. At the widest aperture, the walls of the bowl are steep across the distance dimension and the deepest part is not very deep because lenses are almost always softer wide open. As the aperture narrows, the deepest parts gets deeper up to the "sweet spot" and then gets shallower due to diffraction. As the aperture narrows, the walls of the bowl in the distance dimension grow shallower and shallower.

If "infinite DoF" means that everything from 0 to ∞ is acceptably sharp, then very large format pinhole cameras do approach that ideal.

QuoteOriginally posted by biz-engineer Quote
Yes, the number of exposures required will be related to how deep the is scene relative of the "DoF" thing. With our digital cameras, CoC shouldn't be based on human vision, CoC should be a function of sensor resolution I suppose it would make a lot more sense.
CoC is a function of the sensor if the image is going to be pixel-peeped or printed at maximum size and minimum viewing distance. CoC is a function of human vision if the final print or screen image is smaller or seen at greater distances.

Don't forget that sometimes the photographer's goal is to ensure some things are out-of-focus in order to isolate the subject from foreground, background, or both. If you use the sensor-based CoC to do this but then make a small print or web-resolution version of the photo, the DoF in the small image will be much too large and the image will fail to provide the desired subject isolation.
06-11-2019, 06:32 AM - 1 Like   #25
Pentaxian




Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 7,806
Original Poster
QuoteOriginally posted by photoptimist Quote
Don't forget that sometimes the photographer's goal is to ensure some things are out-of-focus in order to isolate the subject from foreground, background, or both. If you use the sensor-based CoC to do this but then make a small print or web-resolution version of the photo, the DoF in the small image will be much too large and the image will fail to provide the desired subject isolation.
Yes, of course. Here we've discussing for full frame and larger sensor sizes, that can produces subject isolation easily (blur) but can be difficult for all in focus scenes. What's interesting is that most test charts used for reviews are flat charts, they don't include the depth of field aspect of a camera system. In practice, yes there are some times where the scene is flat, and many more situation where the scene is 3D.
06-11-2019, 06:54 AM - 1 Like   #26
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 34,422
QuoteOriginally posted by photoptimist Quote
Don't forget that sometimes the photographer's goal is to ensure some things are out-of-focus in order to isolate the subject from foreground, background, or both. If you use the sensor-based CoC to do this but then make a small print or web-resolution version of the photo, the DoF in the small image will be much too large and the image will fail to provide the desired subject isolation.
The one thing you're missing here is that once the CoC become overlapped which is most out of focus areas, they are no longer distinct. So you can use a smaller print to make border areas look sharp, but the higher percentage of the OOF areas will never be sharp no matter how much reduced So you're actually addressing only the transition zones with smaller COC. Once you've done the subject isolation thing and turned your out of focus areas into CoC soup, they aren't coming back. Image size is irrelevant for much of the image. But you may be able to shrink those areas that are on the edge of focus into appearing to be sharp. When you look at the visual field of a moving lens and see how little is in acceptable focus, and then work out how many can be brought back into sharpness by reducing the image it's not very much, and that to work the CoC have to be relatively small. Once you have COC overlapping the CoC of their neighbours two or three pixels deep, as in almost all ƒ1.4 images, there's simply no coming back. IN my now work with small birds I've notice an eye that isn't sharp 1:1, never looks as good as an image that is, no matter how much you reduce it, because of the lack of detail. This is a very limited concept and often in practice, you try it and it simply doesn't work. But every now and then you get away with it, Its not something you can build into a standad practice.

Think of a simple lens like a magnifying glass with the sun. Where you raise it up and it goes from a sharp y focused point, to a circle 3 or four inches wide. Only the areas right around the central point in focus can benefit from reducing image size. Only a few pixels right after sharp focus can be made sharp by reducing image size, because the CoC don't have much overlap. Once the CoC size means each pixel in the OOF areas are illuminated by 9 different circles of confusion, no sharpness is possible no matter what you do. Reduce it as much as you want, it will still be CoC soup.

By the time you stop down to reduce the size of the circles of confusion to make print reduction more practical to make up for lack of acceptable sharpness, diffraction starts to set in. You're talking about a very low percentage of your images (and a very small percentage of the ares within those images) that this could even be applied to. For the most part, OoF areas remain out of focus, even when reduced in size. I'd guess the order of that the area of the average image that can be rescued by reducing the size of the print is less than 5%. it only works if you didn't need more than that.

Last edited by normhead; 06-11-2019 at 07:16 AM.
06-11-2019, 07:29 AM - 1 Like   #27
Pentaxian




Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 7,806
Original Poster
The commonly set CoC for 35mm is 0.030 mm, or 30 um at the sensor level, based on viewing distance of a 10" print, is too large for 36Mpixels resolution, it's like 10 pixels wide. Changing the CoC from 0.03 to 0.01 pushes the hyperfocal distance way further (3x).

Last edited by biz-engineer; 06-11-2019 at 07:39 AM.
06-11-2019, 07:30 AM   #28
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 4,074
QuoteOriginally posted by normhead Quote
The one thing you're missing here is that once the CoC become overlapped which is most out of focus areas, they are no longer distinct. So you can use a smaller print to make border areas look sharp, but the higher percentage of the OOF areas will never be sharp no matter how much reduced So you're actually addressing only the transition zones with smaller COC. Once you've done the subject isolation thing and turned your out of focus areas into CoC soup, they aren't coming back. Image size is irrelevant for much of the image. But you may be able to shrink those areas that are on the edge of focus into appearing to be sharp. When you look at the visual field of a moving lens and see how little is in acceptable focus, and then work out how many can be brought back into sharpness by reducing the image it's not very much, and that to work the CoC have to be relatively small. Once you have COC overlapping the CoC of their neighbours two or three pixels deep, as in almost all ƒ1.4 images, there's simply no coming back. IN my now work with small birds I've notice an eye that isn't sharp 1:1, never looks as good as an image that is, no matter how much you reduce it, because of the lack of detail. This is a very limited concept and often in practice, you try it and it simply doesn't work. But every now and then you get away with it, Its not something you can build into a standad practice.

Think of a simple lens like a magnifying glass with the sun. Where you raise it up and it goes from a sharp ly focused point, to a point 3 or four inches wide. Only the areas right around the central point n focus. Only a few pixels right after sharp focus can be made sharp by reducing image size, because the CoC don't have much overlap. Once the CoC size means each pixel in the OOF areas are illuminated by 9 different circles of confusion, no sharpness is possible no matter what you do. Reduce it as much as you want, it will still be CoC soup.

By the time you stop down to reduce the size of the circles of confusion to make print reduction more practical to make up for lack of acceptable sharpness, diffraction starts to set in. You're talking about a very low percentage of your images (and a very small percentage of the ares within those images) that this could even be applied to. For the most part, OoF areas remain out of focus, even when reduced in size. I'd guess the order of that the area of the average image that can be rescued by reducing the size of the print is less than 5%. it only works if you didn't need more than that.
What you say is true -- very blurry areas never become sharp -- but the ratios between sensor pixels and print pixels are much larger than you think with today's high-resolution sensors. That makes the sharpening effect of downsampling much greater than you think.

Each pixel of 900x600 pixel social media image covers 64 pixels of the K-1 sensor. A blur circle 8 pixels across on the sensor or the pixel-peeper's screen becomes a sharp 1-pixel point in the tiny print or screen image. The DoF in a 900x600 pixel social media image is 8X deeper than the DoF of pixel-peeper crop or max-size 36 MPix K-1 print. And good luck having any subject-isolating blur in an 80x80 pixel avatar -- the DoF for 80x80 pixels is 60X that of the original K-1 image.
06-11-2019, 08:47 AM   #29
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 34,422
QuoteOriginally posted by photoptimist Quote
That makes the sharpening effect of downsampling much greater than you think.
That's extremely unlikely, I try and produce good images from slightly blurred small bird images all the time. I can think of one or two instance out of thousands where it actually worked. We're talking experience versus untested notions here as far as i can tell.

QuoteOriginally posted by photoptimist Quote
Each pixel of 900x600 pixel social media image covers 64 pixels of the K-1 sensor. A blur circle 8 pixels across on the sensor or the pixel-peeper's screen becomes a sharp 1-pixel point in the tiny print or screen image. The DoF in a 900x600 pixel social media image is 8X deeper than the DoF of pixel-peeper crop or max-size 36 MPix K-1 print. And good luck having any subject-isolating blur in an 80x80 pixel avatar -- the DoF for 80x80 pixels is 60X that of the original K-1 image.
You are completely ignoring the effect of the overlap of the circles of confusion.

QuoteQuote:
the DoF for 80x80 pixels is 60X that of the original K-1 image
.

So what?
2017-03-18-Pentax-tog by Norm Head,

Did you really think that would hold up?

Once the circles of confusion are blurred by enough overlap, they will never be even distinct, no matter how much you reduce their size. A clear example of a theory sounding good, that doesn't include enough parameters to be in any way useful. You are assuming that the circles of confusion exist in a form from which clear data can be extracted, when in fact, done properly, its all one big stirred up soup.

This slightly soft image reduced from a K-3 did gain some sharpness being reduced, but not enough to affect subject isolation. And at the original size it's not that bad. At the reduced size it's not that bad either. You are misrepresenting the value of this effect.

Last edited by normhead; 06-11-2019 at 09:00 AM.
06-11-2019, 08:48 AM - 1 Like   #30
Site Supporter
Site Supporter




Join Date: Feb 2016
Posts: 541
Should you really need or wish to use a DoF calculator there are two that I can recommend:

Lumariver DoF (Android and iOS) - Has many customisation settings including airy disk and or pixel pitch and of course CoC. Anders Torger also has an excellent profile designer
Lumariver Depth of Field Calculator
And
True DoF. iOS only
Intro to TrueDoF-Pro

Then there is Harold Merklingers method not using hyperfocal
DoF Merklinger
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
aperture, camera, cameras, coc, depth, distance, distances, dof, focus, focus distance, frame, ft, landscape, length, pentax, photography, plane, post, resolution, rule, scene, sharpness, technique, videos, youtube
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
New to photography - looking for advices on the Pentax system Tanguy Troubleshooting and Beginner Help 59 08-28-2018 12:55 AM
K1 setting for landscape and hyperfocal biz-engineer Pentax K-1 32 08-07-2018 10:17 AM
Nature Wrong Lens, Wrong time, Wrong…………. lukulele Post Your Photos! 4 05-08-2014 03:11 AM
Confirm my understand of metering w/ manual lenses focus hyperfocal focus DeadJohn Pentax SLR Lens Discussion 4 03-25-2011 12:31 PM
Understanding Hyperfocal Focus nandystam Troubleshooting and Beginner Help 9 02-09-2011 09:43 AM



All times are GMT -7. The time now is 10:13 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top