Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: BernardLanguillier on September 26, 2004, 07:00:01 pm
-
[font color=\'#000000\']Hi there,
A pity indeed, I was not aware that this function had been dropped.
It was in fact one of the items on my list of reasons for switching from Nikon to Canon... :-)
On the other hand, it is also possible to print out a small table of hyper focus distances for typical landscape focal lenghts (20, 24, 28, 35, 80, 120, 140, 170, 200, 300 and 400 for instance ) and apertures (f11, f16, f22, f32).
Checking it out and focussing manually in a slightly conservative way works fine and is probably faster than the methodology you recommended with the Canon implementation of the function.
By the way, I have been shooting a lot with my Hassy H1 and prime lenses recently, and it is such a pleasure to have clear indications of DOF written on the lens itself... It is obviously more difficult to do with zoom lenses, but not impossible. Most zoom lenses used to have at least some DOF information.
Now a more general comment. The problem with hyperfocal is that the size of the circle of confusion is part of the equation. For film, it was easy to use 0.03 mm and to forget about it, there might be a need to do some testing with a 1DsMKII to see if this value is still small enough, or if there is a need to go down to a smaller value (0.02) for instance. Interestingly, the result might look like you actually get less DOF with a 1DsMKII... which is not really the case. It is just that you need to close down more to make use of the added resolution of the sensor over 35 mm film.
Best regards,
Bernard[/font]
-
[font color=\'#000000\']Of course this may raise the issue of film versus digital but Harold Merklinger had a reasonable discussion on focus and depth of field in his book "The INs and OUTs of Focus". I believe he cited several origins to the 1/30 mm COC but mainly the Leica Manual. The goal was to view an 8x10 print from 10 inches distance using film of a quality available at the time. A link to the book is provided below. I myself use a smaller COC than 1/30 mm and printed out a simple table on a small card that I generated in Excel to bring with me on shoots even though my Canon camera has the DEP feature. Of course by now I've pretty much memorized most of the table. Thus far I have no complaints using this method but your mileage may vary.
Merklinger-Focus (http://www.trenholm.org/hmmerk/TIAOOFe.pdf)[/font]
-
[font color=\'#000000\']For what it's worth, I'm with Marty.
I learned to use the DEP mode on my 1Ds much as he described. I find it's fast to use, and it produces great results.
I will miss DEP on Canon's latest Pro bodies.
Enjoy!
-- Jim[/font]
-
[font color=\'#000000\']For measuring distance look out for an old time camera accessory: a shoe mounted rangefinder. I saw one for sale at a car boot sale recently and (stupidly) didn't buy it. If you need better distance graduations on your lens you can of course practice focussing on points at a known distance to get a feel of how the lens scale lines up[/font]
-
[font color=\'#000000\']Thanks Tim,
1. What then is the distance that is half way to infinity?
2. DOF does not have to assume a viewing distance of the print or the print size. I can easily select what I want the final print to look like and then calculate the rest.
DOF is a defined term. If I make a larger or smaller print, or view a print from nearer or farther, the DOF of the print changes. The closer I get, the less the DOF for a given print size. For a given viewing distance, the larger the print, the smaller the DOF.
I understand how the DOF preview button works. What I was trying to say is the dimmer the image in the view finder, the more difficult is to see what is and isn't in focus. Focus may appear acceptable in the view finder stopped down to f/22 partly because I can't see the image very well.
3. Agan, I understand how the DOF preview works. What I was wondering or asking was, does the previewed scene DOF really mimmic the results I expect to get on the final print?
4. I agree with the "relativeness" of the DOF. But if I want to really control what will be acceptably in focus and what will not appearto be focused, I cannot rely on the values of COC, image size, viewing distance, etc. assumed by my lense/camera maker. Nor do I have too. If I want an image that appears in focus between 50 feet and infinity (scene dimensions) on a 20x30" print viewed from 10', and I want subjects closer than 50' to appear out of focus, I can do that much more predictably than focusing half way between 50' and infinity (waht ever that distance is) and stopping down to f/11 or f/16.[/font]
-
[font color=\'#000000\']Jonathan and Tim appear to be discussing two different aspects of COC and hyperfocusing. I'd greatly appreciate clarification of the relationship between these two, and thank Tim and Jonathan in advance for their response:
Jonathan argues that the 1Ds demands a more restrictive COC than does film:
DOF with digital cameras should be calculated using the pixel pitch of the sensor as the Circle Of Confusion value. That will tell you what pixel-sharp DOF will be. Using larger values will result in files that are blurry and soft-looking, and smaller values would be rather pointless.
1Ds: .00888mm
1Ds-MkII: .00721mm
If you use a larger-than-pixel-pitch COC value, you're unnecessarily limiting the detail of your images by blurring to the point where there isn't much in the way of crisp detail in your image files. Digital sensors record much more detail per square millimeter than film can; that's the whole reason a 1Ds can compete head-to-head with 6x7 medium format and not be embarrased. The flip side of recording more detail per square millimeter is that the COC value has to be smaller. You can't have one without the other.
Tim explains that bigger enlargments demand a more restrictive COC:
That's the way it works. Any given DOF calculation makes assumptions about the viewing distance and the size of the print - this drives the COC number. DOF isn't binary - the larger you print and the closer you view the less-in-focus the image is going to appear regardless of the calculated DOF.
It is easy to see the dramatic difference that different COC values make with the Focus+ program for the Palm Pilot. In that program .03 (the old standard for film) appears to equate to 30 um and .008 is 8 um. There is a huge difference between these two in terms of the aperture that is required.
What happens when both factors are combined? What should the COC value be for the 1Ds if you anticipate that a given shot on the 1Ds will ultimately be printed as a 16x20? When both are combined -- what is the resulting COC?
Does everyone agree that the 1Ds requires a more restrictive COC than film? For big enlargments like 16x20 that may be viewed at a distance of two to four feet?
If so, is it necessary to ratchet down to a COC of .008? Or is a compromise of 15 in the Focus+ program sufficient (.015?)
Thanks for the excellent discussion and for responses to the above.[/font]
-
[font color=\'#000000\']marty, I've had a few minutes to think about this and can give a rough estimate that should be close enough.
The circumference of a circle = 2*pi*r
where r is the radius of the circle will also be the viewing distance of the final print.
The human eye cannot tell the difference beween a point and a disc that is less than 1 degree, or 1/360th of the circle circumference. Thus,
COC (on the print) = 2*pi*r/360 = 0.0175*r
Now, divide this COC by the magnification factor of the negative, and you get the COC on the film. Use this COC in the standard optical equations to get the desired f/stop and hyperfocal distance.
Check the math. I have been known to make mistakes. Hope this is helpful.
I think I see the logic to the digital COC and agree. Using a COC smaller than the pixel pitch would not change the DOF on the print, but it would be less than you might be expecting?[/font]
-
[font color=\'#000000\']There is no need to be stuck with someone else's idea of a "standard viewng distance." If I want a big print for a small room, or a small print for a big space, why not? Why should I assume I will only make a 5x (or what ever sized enlargement) from a 35mm neg? I am the photogarpher (artist), I should be able to make what I want, not what Canon (or whoever) thinks I want to make. If I want to make a 30x40 print to be viewed from 2 feet, why not? But if I view that same print from 20 feet, the DOF will be different. I don't think you can refute that.[/font]
-
when I saw that DEP mode was removed, I was livid. Absolutely insensed. I was ready to take up arms and.... well, knowing I was going overboard, I went to the boards to see what the explanation was, and what others had to say.
Sadly, this didn't make things better. The "reason" for dropping it is reflected by people's indifference to it. Hardly anyone seemed to be using it. This makes sense, as it was never documented well, and that it was incompletely implemented from a user interface point of view. However, every time I showed it to people--whether pros or amateurs--the response was always, "WOW! I had no idea what that was!" Never have I showed it to someone that didn't immedicately become a convert.
What really bothers me, though, is that Canon just removed it completely, rather than just move it to a personal-function-enabled feature, even if it's just to see what the public response would be. I mean, programming the firmware to add features is dead-simple, it's extensible forever, and there's no cost to it. With the plethora of PF settings already, it's just a matter of clicking it on or off when you originally set it up, and then just leave it at that.
the other thing that bothers me is that the methods canon used for determining this feature's usefulness is so fraught with problems, that we may all find some other really important feature go by the wayside for the same reason. (Moral: fix the problem at the source.)
The last thing that bothers me is why this isn't mentioned in the review of the camera. I can understand if the reviewer himself doesn't use the feature, but by gosh--failing to note that it's gone completely borders on .... well, I don't want to go overboard again.
For me, and others who eventually figured out how to use it, this was one of the best features of the camera... that it's gone now will mean that I can't sell my older 1Ds.
...wouldn't it be great if we could just upgrade the sensor in the camera and leave everything else "as is?" :-)
-
Jonathan argues that the 1Ds demands a more restrictive COC than does film:
I don't think our positions are incompatible. What Jonathan is getting at (if I may take the liberty) is that you might as well maximize sharpness since the ultimate purposes for which most folks would use the 1Ds are not 8x10 (you don't need a 1Ds for 8x10s) but larger. If you're shooting for large framed, displayed prints the more conservative a COC you use the better.[/quote]
The 1-pixel CoC paradigm tracks very well with perceived print DOF at any print size where upsampling is required to achieve a print PPI of 300 or so. On the 1Ds this equates to any print size 8x12 or larger. I did some testing with the 135/2L wide open, and print DOF matched the spreadsheet predictions very closely.
But in the end there's no free lunch - the sweet spot is f8 - f11 +/- for most lenses, and as you move towards f32 other un-desirable stuff starts to happen. Ultimate DOF is going to be driven by lots of comprimises to the photographers intent. Is there motion? - problematical for long exposures. What's the composition, WA lends itself naturally to deep DOF, but not telephotos or macros.
The issue is important if you are using a calculator, since it will very likely use .03 or thereabouts for full frame and amend that for the smaller sensors (if it's a "good" calculator). If you need to factor in size of print etc. the "standard" coc is going to be misleading and you might as well use Jonathan's suggestion.
This is certainly a valid concern, but will vary depending on the lens. My DOF calculator will give you accurate predictions regarding what DOF will be given a sensor, focal length, distance, and aperture, but cannot tell you whether choosing a particular aperture will give you an overall sharp image within DOF due to diffraction or poor wide-open lens performance. Know your equipment, and know that some things are simply not possible with current technology.
-
Jonathan, it appears to me IMUO, that some confusion about DoF and CoC is created when your position is first made clear (and I believe accurately) when you state that your purpose is to make images with all parts within the DoF.
I have never said that, and that is simply impossible when shooting low-light concert photography at f/2.8 or wider apertures. What the CoC=pixel pitch premise does is gives me a very accurate idea of what range of distances will be sharply focused with detail resolved down to the single-pixel level. It tells me that when shooting with the 135/2L wide open at a distance of 8 feet, DOF is less than half an inch. This tracks very closely with practical portrait experience where the eyes are in focus, but the eyebrows, lips and beard are quite noticeably out of focus. Perceived DOF has been widened somewhat in this web-sized JPEG due to the downsampling, but you can still see the DOF is pretty narrow if you look at the beard--the edges (closer to the eyes) are in distinctly sharper focus than the center, which is closer to the camera:
(http://69.224.22.169:81/images/2004-09-24-0001.jpg)
In a decent sized print (like 8x12 or larger) perceived print DOF closely tracks predicted DOF based on CoC=pixel pitch. You can dispute the methodology if you like, but real-world results track predictions based on this method in the vast majority of print sizes and viewing conditions.
-
(Oh, sorry, yet another three-months-plus followup.)
Judging by the responses to the above posting, we are back to debating about how many angels can dance on the head of a pin. This thread has been filled with long dissertations on DOF charts, but the point is that DEP is far easier to use.
(...)
The real tragedy is the review of the MkII on Luminous Landscape. If anyone should be protesting the loss of DEP it is Michael Reichmann, since this web site is devoted to landscape photographers. In point of fact, Michael previously recommended DEP on this site and provided a primer on its use:
http://www.luminous-landscape.com/tutorials/dep.shtml (http://www.luminous-landscape.com/tutorials/dep.shtml)
Despite that, Michael does not even mention the loss of DEP in his review of the MkII.
Sic transit gloria EOS.
And yes, the loss of DEP should definitely merit mention one or two hundred places. It's not just landscape photography that benefits greatly from the feature, but also group portraits and several other kinds.
The DEPTH mode, as it was called on the 650 when the EOS series was introduced in 1987, was one of the most prominent marketing features I can recall in those cameras. (Of course there was the EOS electronic mount ...)
I used it extensively with my parents' 650, and got what I thought (and other laypersons claimed) were great results from it.
I've tried using the A-DEP feature of the EOS 20D, and to be frank, it's near useless to me. It just keeps on insisting to use wide apertures, at best one or two stops down from wide open, and that in the scenes where I'd really love for it to do it right.
With the smaller-size and dimmer viewfinder (as compared to the 650 at least), the stop-down button doesn't help me to decide how DOF actually works in any given scene. It's guesstimate and shoot, every time. Oh well, I'm getting better at guesstimation.
Hey, wait a minute, I know what to do!
I'll just compose the scene in the 650, lock the lens at manual focus, and shift it over to the 20D body!
-
[font color=\'#000000\']Canon has dropped an important feature in the 1Ds Mk II (and I believe in the 1D Mk II) that they previously included in all top-of-the-line models for over 20 years. This feature is of critical importance for landscape photographers and was a key reason why I bought the Canon system rather than Nikon 15 years ago. This feature helped differentiate Canon from Nikon and the dropping of the feature represents a major loss for the Canon system.
I am referring to the DEP mode. DEP allows Canon cameras to hyperfocus with zoom lenses. I recall that Chuck Westfall reported on the Galbraith forum that Canon dropped it in the 1D Mk II because professionals did not request it and didn't complain when it was dropped. That is probably because the implementation of DEP by Canon was dumb to start with and most photographers -- even professionals -- didn't figure out how to work around the poor Canon design and thereby properly utilize it.
Anyone who learned photography on manual cameras with non-zoom lenses appreciates the importance of hyperfocusing. It allows the photographer to maximize depth of field by altering the focusing point of the lens using the depth of field marks on the lens barrell. (See below link for an illustration.) DEP does the same thing -- when you use it you can watch the lens shift the focus.
Unfortunately Canon designed DEP to work with faster shutter speeds in order to hand hold the camera -- and it thus defaults to larger apertures that don't work for maximum depth of field. Most photographers would conclude that it doesn't work as advertised for depth of field -- and thus use AV at f16 rather than using DEP!
The solution is to set the camera to DEP and to allow the lens to hyperfocus by focusing on the near object, then the far object, then touching the shutter a third time and watch the lens hyperfocus. Now you need to override the large aperture setting while maintaining the focus. So you then set the lens to manual focus to lock it in. Set the camera on AV or manual and select a small aperture like F16 and then take the shot. You now have the advantage of both hyperfocusing and a small aperture for depth of field.
Many of us figured out this work-around on our own. We could see that the lens was shifting focus and thus hyperfocusing -- but had to utilize smaller apertures than what is allowed by DEP. Michael Reichmann follows the same system:
http://www.luminous-landscape.com/tutorials/dep.shtml (http://www.luminous-landscape.com/tutorials/dep.shtml)
If you've never used DEP -- I urge you to do just that. You will discover that DEP allows you to increase the depth of field and get landscape shots that are not possible when using F16 alone.
I am basing my report that DEP has been dropped based upon the Canon white paper and other reports on the 1Ds MkII. DEP is not listed as a feature on the camera.
This is an enormous loss for the Canon system. And an incredibly stupid move on the part of Canon. One of the key features that differentiated them from Nikon has been lost. I can only hope that if the small group of landscape photographers who used the feature complain it might be restored in the Mark III model two years from now.
The Mk II model can't be considered to the "absolute pinnacle of design" as Canon claims when they are dropping features and moving backwards in their design criteria.
I would have to think long and hard before selling my 1Ds and buying the mark II model without DEP.
When Canon does restore it they might consider modifying it so it defaults to F16 -- and warn users that it would then require the use of a tripod to maximize depth of field. But that is exactly how any landscape photographer would use it in any event.
Please Canon, restore DEP in future models.
I also would urge Michael to note that DEP has been dropped in his future review of the 1Ds Mk II -- both on this site and in a future review in the video journal. I doubt that Canon pays any attention to comments from users on this site or any other. I would presume that they do pay attention to Michael's reviews, however. In any event, comments from reviewers like Michael are probably our last hope that Canon will restore the feature in future models.[/font]
-
[font color=\'#000000\']I shoot mostly landscapes and changed to Canon about a year ago. I have experience with LF gear as well. I have never used DEP and wouldn't miss it on the 1DS MKII.[/font]
-
[font color=\'#000000\']If you use a larger-than-pixel-pitch COC value, you're unnecessarily limiting the detail of your images by blurring to the point where there isn't much in the way of crisp detail in your image files. Digital sensors record much more detail per square millimeter than film can; that's the whole reason a 1Ds can compete head-to-head with 6x7 medium format and not be embarrased. The flip side of recording more detail per square millimeter is that the COC value has to be smaller. You can't have one without the other.[/font]
-
[font color=\'#000000\']> As I said, if more photographers would try it
I tried DEP and found it awkward and slow. I prefer the method I've used for decades - just estimate the hyperfocal point and focus there sometimes using DOF preview to confirm or refine. What I miss much more that DEP is the lense markings that used to assist in this matter on all my older cameras (both Leica and Canon) that Canon started leaving out at some point. But I've learned to live without them.
Now if they made a mode that automatically focusing (with a single press) at the hyperfocal point relative to infinity I might find that useful particularly in fast paced situations. I never really understood the utility of DEP given the contortions you have to go thru to use it.
- DL[/font]
-
[font color=\'#000000\']Many thanks to EAD, mnogueira and Jonathan for their suggestions. I ordered one of the expodisc DOF pocket calculators just to see how it works. I'm guessing that it uses a COC of .03 mm.
The palm program allows the user to enter the COC directly.
Jonathan -- your excel spread sheet is great -- but having a Palm program in the field is a definite advantage. I don't plan on hauling around a notebook computer.
I assume that if your COC figures can be entered in the Palm program then the results should be the same as your excel program?
The palm program (Focus+) that mnogueira found on the Luminous site allows a user to input their own COC. It appears to be based on "30um." If that translates to .03 -- what would the corresponding figure be for your more conservative figures of:
1Ds: .00888mm
1Ds-MkII: .00721mm
0.88 or 0.72?
Jonathan, your COC figures do appear to be significantly more stringent and conservative than what is normally used. I typed .03 directly into your program for the COC and the difference is almost 4 stops -- your require the aperture to be 4 stops smaller. (Assuming that it is OK to input .03 directly into your program.)
A difference of 3 or 4 stops will make a huge difference.
I'm therefore wondering if all of the tech experts agree with your interpretation? Have the various DOF web sites and calculators adopted your more conservative figures for DOF for the 1Ds?[/font]
-
[font color=\'#000000\']1. What then is the distance that is half way to infinity?
****
Not completely sure I understand the question. The "half" refers to the distance between the camera and a point of focus which is at the hyperfocal distance. I'm not aware of any concept dealing with something half way to infinity.
*****
2. DOF does not have to assume a viewing distance of the print or the print size. I can easily select what I want the final print to look like and then calculate the rest.
***
Yes it does. viewing distance and print size are what drives setting the COC, which is a term of the DOF calculation -which as you point out next is clearly defined.
Maybe you could walk me though the process of "selecting what the final print should look like and calculating the rest"?
***
DOF is a defined term. If I make a larger or smaller print, or view a print from nearer or farther, the DOF of the print changes. The closer I get, the less the DOF for a given print size. For a given viewing distance, the larger the print, the smaller the DOF.
***
I agree.
***
I understand how the DOF preview button works. What I was trying to say is the dimmer the image in the view finder, the more difficult is to see what is and isn't in focus. Focus may appear acceptable in the view finder stopped down to f/22 partly because I can't see the image very well.
***
At f22 and I use the DOF preview, I can't tell what's in or out of focus either.
***
3. Agan, I understand how the DOF preview works. What I was wondering or asking was, does the previewed scene DOF really mimmic the results I expect to get on the final print?
**
Dimness aside, what's in focus in the viewfinder will be in focus on the sensor. What appears in focus in the view finder or the sensor may not appear in focus on a 40x60 inch print, but may very well appear in focus on a 8x10, as your point #2 indicates.
**
4. I agree with the "relativeness" of the DOF. But if I want to really control what will be acceptably in focus and what will not appearto be focused, I cannot rely on the values of COC, image size, viewing distance, etc. assumed by my lense/camera maker. Nor do I have too. If I want an image that appears in focus between 50 feet and infinity (scene dimensions) on a 20x30" print viewed from 10', and I want subjects closer than 50' to appear out of focus, I can do that much more predictably than focusing half way between 50' and infinity (waht ever that distance is) and stopping down to f/11 or f/16.
****
Technically you could rely on modifying the COC according to you printing/viewing intentions, but it would probably be more trouble than it's worth. I agree that there are better methods of determining what's in or out of focus, than what you described. Is the more predictable method you mention using the DEP mode?
*****[/font]
-
[font color=\'#000000\']marty, it is my experience that I should select a COC size on the final print. That COC is determined by what I want to appear to be in focus and how far away from the print the viewer (usually me) will be. (I use a very different COC for a post card to be viewed at a foot and a half or so and a 20x24 intended to be viewed at 4 feet.) Then I work back to the film size to get a COC on the film. Using that COC on the film, I can select the focus distance and f/stop to give that result. Mathimaticaly, you can skip the film size issue so DOF does not depent on the film size.
I do not use digital cameras for this kind of work, so I cannot address the COC sizes for digital camera. I don't use the DOF preview button because the previewed scne on the view finder may have no relationship to my final print. I would viewing a dim 6x6 cm image with a magnifier. I simply cannot correlate that to what a 20x24" print in a well lit room at 5' will look like.
Depth of field and its use has been a topic with some controversy on this site for some time. What I have written above is how I do it for my own use. It works fine for me, but your methods may be different and equally effective for you.[/font]
-
[font color=\'#000000\']marty, I am in the process of moving and do not hae the materials to show the calcuation nor the calculator I use. So I cannot at this time tell you what I would recommend. However, this topic, with the formulae, has been presented on thi site several times. You might be able to find one of those threads. Basically, take the viewing distance and determine the size of disc at that point that will look like a point. The human eye resolves about 1 degree. Then, size that on the print COC down by the degree of enlargement for the negative to get a COC on the negative. From there, use the usual formulae for DOF. Sorry I can't be of more help at this time.
Tim, as you know, if you want to maximize DOF, back up (increase the focus distance), use a shorter focal length lense, and stop down as far as possible. (Note that the first two steps merely reduce the size of the focused image on the sensor, including the out of focus subjects, thus making more space appear to be focused.) Then, cross your fingers and take whatever you get. If you care what the DOF is on the final print, then you must take things into your own hands. I belong to perhaps the small school of photograhers who sometimes use out of focus elements as an intended element of the final print. Everyone has their own ideas and methods.[/font]
-
[font color=\'#000000\']
Jack, for the record, if I do as you suggest and make a print large enough (or crop severely enough) and stand close enough, the print will look out of focus where you might expect DOF to save you.
For the record, NO IT WON'T! --- At least NOT DIRECTLY from original DOF problems!
HOWEVER, since you are moving closer to the print, you alter the standard viewing distance which in turn alters the calculation for acceptable CoC. Thus you impart NEW constraints on your CoC limits, but these issues are NOT from original DOF limitations directly, only imparted after the fact by YOUR manipulation.
Since you also CROPPED, you actually CHANGED THE FORMAT of the sensor in question, again altering the appropriate CoC calculation. Again, this is NOT directly a result of any original DOF limitations, only imaparted after the fact by YOUR (careless) manipulation.
In summary, in both cases you altered the parameters that DIRECTLY affect the CoC calculation, and thus require CoC to be modified for a correct effective DOF calculation.
The 100x focal rule holds true for all formats and all lenses.[/font]
-
[font color=\'#000000\']That seems like alot to pay for a falty camera you should check out http://bargainfindsonebay.com/ (http://bargainfindsonebay.com/) and see about a sonycybershot maybe. [/font]
-
Jonathan, it appears to me IMUO, that some confusion about DoF and CoC is created when your position is first made clear (and I believe accurately) when you state that your purpose is to make images with all parts within the DoF. This is entirely consistent with using the smallest CoC. I also see that a CoC smaller than a pixel makes no sense because a point smaller than a pixel will appear as a single pixel anyway.
The confusion comes in when your rule of using a CoC equal to the pixel pitch is restated without all of the assumptions. Some of us, at least me, soemtimes wants to make an image with some elements out of focus for a reason. Then it makes sense to me at least to select a larger CoC, namely the one that will give the priny appearance I want.
A similar thing has happened to the DoF on some camera lenses. The lens maker assumed certain things about CoC, print size and print viewing distances. Then, over time, those assumptions are lost, forgotten, never learned, whatever, and the DoF scale on the lenses become fact. Well, it is fact, but for the assumptions only.
Before I get anybody all whipped up again, this is just my opinion.
-
Jonathan, thanks for the clarification. If I understand you correctly, you are saying that you select a CoC equal to the pixel pitch, and define any point so out of focus as to make a spot larger than a pixel as out of focus.
-
[font color=\'#000000\']Personally, I never used it when available on the previous 1 bodies anyway, so I don't miss it here.
Instead I prefer to set hyperfocal manually, using the DOF preview button to confirm the in-focus areas. Yes, it is sometimes difficult to determine actual focus when using DOF preview, but I am used to looking through the dimmed finder and always seem to make it work. This probably comes from years from looking through a dim LF groundglass to determine this very thing...
Furthermore, I never trusted the camera's pre-assigned programming to be more accurate than my actual viewing of even the preview-dimmed screen.
My .02,[/font]
-
[font color=\'#000000\']DOF with digital cameras should be calculated using the pixel pitch of the sensor as the Circle Of Confusion value. That will tell you what pixel-sharp DOF will be. Using larger values will result in files that are blurry and soft-looking, and smaller values would be rather pointless.
1Ds: .00888mm
1Ds-MkII: .00721mm
I have no idea what COC value Canon used in DEP mode; I've never used it. Actually I have it disabled via a Personal Function so that my 1Ds and 1D-MkII behave exactly the same way when switching modes.[/font]
-
[font color=\'#000000\']Distance measurement isn't as nigglingly important as it might seem. The distance scales on the lens are sufficient in most cases. For example, the 17-40/4L has the infinity mark just after 1 meter on the scale. The reason is that it starts getting into hyperfocal territory shortly after 1 meter and DOF is correspondingly wide. In general, the more fussy DOF is, the greater the precision of the distance markings on the lens will be. You can use a laser rangefinder if you like (I happen to have one, but I use it for other purposes) but in practice it's overkill. One thing you might consider is taking a laptop on shoots with you. You can use it to dump memory cards to, and you can have your spreadsheet handy for any shot you might want to take. If you're shooting tethered, the laptop is mandatory anyway. Another helpful thing to remember is the DOF relation between lenses of differing focal lengths. If you double the focal length, the DOF at any given aperture/distance combination will be divided by four until you approach hyperfocal distance and the numbers hit infinity. So if you print up a chart for a 50mm lens, you can use the figures for a 100mm lens if you remember to divide the DOF amount by 4. Play around with the spreadsheet and you'll see what I mean.[/font]
-
[font color=\'#000000\']Read this:
http://www.luminous-landscape.com/reviews/palm_pilot.shtml (http://www.luminous-landscape.com/reviews/palm_pilot.shtml)
so close it could bite you ! :laugh:
I quite like it, of course you need to have a pda, but then again, who doesn't these days ?
I think it solves the problem of carrying your pc with you.[/font]
-
[font color=\'#000000\']At small aperatures like f11 or 16, I focus to a point about 1/3 the distance into a landscape, and get sharp focus front to back on my 20"x30" landscape prints (using wide lenses on a 1Ds). Simple, but it works for me.
Dave[/font]
[font color=\'#000000\']Amen.
More to the point, focus 1/3 of the way BETWEEN the forground and background elements you want in focus in the final image, confirm with the DOF preview button. If you can only confirm only one point in focus due to the dim finder, the other is probably good too.[/font]
-
[font color=\'#000000\']I know DOF is a well defined and understood topic, but I have some questions.
1. What is the distance that is 1/3 of the way between 20 feet and infinity? (The distance scale on the lense barrel is very non-linear, so simply 1/3 of the distance between 20 feet and infinity isn't really 1/3 of the distance between 20 feet and infinity.)
DOF should be determined on the final print. (Do I really care what the image DOF is in the camera?) That is a function of, among other things, the viewing distance of the final print and the size of the final print. For me, it is even dependent n wether I amwearing my glasses while I view the print.
2. How does this relate to the image you see in the view finder? The viewer is very close to the very small image, and the image in the view finder may be magnified. If I view the small image on the back of my Sony F-707 from three feet away, the entire image usually looks sharp. If I get really close, the image starts to fall apart.
The DOF preview shows a dim image with increased DOF on the view finder. I believe the dimmer image will also increase the apparent DOF, the same as viewing a print in a dim room will makesit look sharper.
3. Does the DOF preview really "preview" the final print?
4. Does the described method allow me to predictably make a portion of the foreground or background out of focus?[/font]
-
[font color=\'#000000\']Howard --
So, for film, what COC do you recommend for a 16x20 print that would be viewed fairly close up at a distance of a few feet?
Hopefully someone else can address the digital issue. The COC recommended for the 1Ds is far more restrictive than any number proposed for film as far as I can tell.[/font]
-
[font color=\'#000000\']Jack, for the record, if I do as you suggest and make a print large enough (or crop severely enough) and stand close enough, the print will look out of focus where you might expect DOF to save you.[/font]
-
[font color=\'#000000\']Jonathan argues that the 1Ds demands a more restrictive COC than does film:
I don't think our positions are incompatible. What Jonathan is getting at (if I may take the liberty) is that you might as well maximize sharpness since the ultimate purposes for which most folks would use the 1Ds are not 8x10 (you don't need a 1Ds for 8x10s) but larger. If you're shooting for large framed, displayed prints the more conservative a COC you use the better.[/font]
[font color=\'#000000\']I'd say that is a fairly accurate characterization of my position. Using the sensor-pixel-pitch-as-COC means that the resolution limit of the sensor is the ultimate limiting factor to image detail rather than the focus of the lens. There's no benefit using a COC less than the sensor pixel pitch, but using COC = pixel pitch will maximixe the enlargability of the image.[/font]
-
when I saw that DEP mode was removed, I was livid. Absolutely insensed. I was ready to take up arms and.... well, knowing I was going overboard, I went to the boards to see what the explanation was, and what others had to say.
Sadly, this didn't make things better. The "reason" for dropping it is reflected by people's indifference to it. Hardly anyone seemed to be using it. This makes sense, as it was never documented well, and that it was incompletely implemented from a user interface point of view.
Judging by the responses to the above posting, we are back to debating about how many angels can dance on the head of a pin. This thread has been filled with long dissertations on DOF charts, but the point is that DEP is far easier to use.
Many of you have said that you'd rather carry and consult DOF charts. Talk about an imprecise and time consuming alternative. I don't have the skill to accurately judge distance at 15, 20 or 40 feet and then determine where to precisely focus the lens to obtain the hyperfocus point. And that is EXACTLY what is required when using charts. You must accurately determine the near distance (assuming the far focus point is infinity). Then you stop shooting and consult the correct chart for the focal length of your lens. Obtain the hyperfocus distance. And then focus at exactly that distance to set the focus on the lens. There is no other way to do it since detailed distance scales are no longer included on lenses. The distance scale on the 24-70 is useless in setting such distances. You can't set the focus using the scales -- you must focus at that distance to set the hyperfocus. DEP is a snap compared to the above, and I am very surprised that more photographers haven't seen the light. I agree with Dan!
The real tragedy is the review of the MkII on Luminous Landscape. If anyone should be protesting the loss of DEP it is Michael Reichmann, since this web site is devoted to landscape photographers. In point of fact, Michael previously recommended DEP on this site and provided a primer on its use:
http://www.luminous-landscape.com/tutorials/dep.shtml (http://www.luminous-landscape.com/tutorials/dep.shtml)
Despite that, Michael does not even mention the loss of DEP in his review of the MkII.
If a reviewer with the clout of Reichmann does not urge Canon to restore DEP in future models, then I fear that all hope is lost. And one of the most practical features that Canon has included in its cameras for the last 20 years will disappear.
Reichmann has promised his readers one more installment on the MkII, this time with his assessment of actual prints from the 4000. (I hope in a direct comparison with the camera that it replaces.)
I can only hope that Michael will strongly protest the loss of DEP in his final installment on the MkII. If he doesn't do so -- as one of the leading advocates for landscape photography -- the a wonderful feature for landscape photographers may be gone from all future models.
That would be a tragic blow for landscape photographers, given how easy it would be to continue to include DEP as an option through the functions or firmware.
-
[font color=\'#000000\']> What then is the distance that is half way to infinity?
Well mathematically infinity/2 = infinity
But I would point out that the initial impulse to focus landscape shots at the "halfway" point or even the "thirdway" point is often not optimal, particularly with WA shots where the lens has a tremendous DOF. For instance with a typical DSLR, 17mm, f11, your hyperfocal point is 3.7 feet! I sometimes focus "infinity" landscape shots by focus locking on my feet .
- DL[/font]
-
Yes. That is a useful predictor of what appears "in focus" and "out of focus" in most combinations of display/print/viewing conditions. With very small prints or greatly downsampled image files, perceived DOF increases, but not always by as much as you'd think.
-
[font color=\'#000000\']I'd recommend trying both methods and comparing the results. The Canon DEP hyperfocusing system works very well. (I assume Michael shares that assessment or he wouldn't have recommended it.)
Looking through a dim viewfinder at F16 using depth of field preview and trying to determine whether everything is in focus is difficult at best.
As I said, if more photographers would try it (using the method I described and also by Michael) there would be quite a few requests that it be restored in future models.[/font]
-
[font color=\'#000000\']DOF with digital cameras should be calculated using the pixel pitch of the sensor as the Circle Of Confusion value.
1Ds: .00888mm
1Ds-MkII: .00721mm[/font]
[font color=\'#000000\']My recollection was that using the pixel size as a surrogate for the COC was somewhat controversial in that a coc that is smaller (or more conservative) than necessary is implied.
As the sensor size goes down, the coc becomes smaller proportionately, but I've never seen it suggested that a ff 1DS coc should be significantly different than the standard film coc of .03mm (I've seen ranges of .027 to .035 suggested). Any references to the contrary would be appreciated.[/font]
-
[font color=\'#000000\']Question for Jonathan and jpoll:
I'm assuming that DEP is dead. Canon has left it off the last two top of the line models it released. And very few photographers appear to miss it, judging by the responses to my posting.
So I'm planning for life-without-DEP. Several postings refer to taking "excel" spreadsheets to figure out depth of field (DOF). That led me to Jonathan's excellent excel calculator/spread sheet posted on his web site.
However, I'm totally puzzled as to how you use it in the field. Let's assume that you can access the excel spreadsheet on a calculator -- or that you print out those that you need based upon focal length and DOF to infinity. And do it for a variety of focal lengths and only for the apertures that apply for 35 mm. (That results in a one inch stack of print outs, even for only 24, 35, 20, 50, 70, 100 and 135 mm -- because I printed them out.)
You still have to be able to precisely measure distance. For example, let's assume that I've framed a shot with the near object 20 feet away and the far object in infinity. To use your excel chart -- or any DOF chart -- I'd first have to measure the distance to the closest object. And do so with some precision. That might be a rock in the middle of a stream -- which rules out a measuring tape. Then, using your excel chart, I'd obtain the actual point to focus for hyperfocus and DOF -- which might be 40 feet. You'd have to determine what object in front of you is at 40 feet and focus on that -- and then lock the focus and take the shot.
All of the above requires the ability to measure distances with a fairly high degree of accuracy in the field. And you can't use the distance markings on a lens for that. For example -- the Canon 24-70 lens only includes distances for 3 feet -- 5 feet -- 10 feet -- infinity. It is impossible to accurately focus at 30 feet with such a lens.
The only method to accomplish the above (that I've figured out) is use a laser finder of the type used by real estate agents to measure distance -- and they cost between $300 and $500. (The cheaper distance finders use sound and would not work for this purpose.)
Am I missing something?
If you ask me DEP is a lot faster -- even using the method I described in my original posting. But Canon has already deleted the feature, and no one appears to miss it.
So -- back to life without DEP -- and on to hyperfocus charts.
Exactly how are they used in the field?[/font]
-
[font color=\'#000000\']This may be of some help, though I haven´nt used it myself...
In this pageExpodisc (http://www.expodisc.com) you have a "expo depth of field guide" which seems to be a pocket-size disc with several options regarding aperture and focal distances, capable of accurately calculating the most apropiate combinations....Anybody with experiene?
That could be an answer...[/font]
-
[font color=\'#000000\']At small aperatures like f11 or 16, I focus to a point about 1/3 the distance into a landscape, and get sharp focus front to back on my 20"x30" landscape prints (using wide lenses on a 1Ds). Simple, but it works for me.
Dave[/font]
-
[font color=\'#000000\']1. The 1/3 is a recognition that when you focus on any given point the depth of field you get will go roughly from 1/3 in front to 2/3 behind the point of focus (ie 1/3 of the depth of field) - for finite DOFs (ie not dof to infinity). The DOF is not equally in front or behind the POF. This also might be a reference to the hyperfocal distance which is defined as the distance at which 1/2 of that distance to infinity is in focus.
2. That's the way it works. Any given DOF calculation makes assumptions about the viewing distance and the size of the print - this drives the COC number. DOF isn't binary - the larger you print and the closer you view the less-in-focus the image is going to appear regardless of the calculated DOF.
The dimness/sharpness is totally unlrelated to the dark room scenario you mention. The DOF preview simply stops the lens down to the selected aperture. At f22 the hole is pretty dang small so appears very dim, but the higher the f stop the greater the dof.
3. No, as noted above the DOF preview simply stops down the f stop.
4. You can certainly rely on the "relativeness" of the DOF, ie the longer the FL, closer the subject and wider the aperture, the more the foreground and background will be blurred. As I said earlier, what you see in a print adds the size of the print and the viewing distance to the mix.
Maybe part of the confusion comes in that what is in focus (technically) is the 2 diminsional plane at the point of focus. The farther you diverge from that plane the more you will tend to see blur. It's not true that everyting within a DOF is in fact "sharp" or "in focus", only that based on the assumptions reflected in the chosen COC it will "appear" sharp.[/font]
-
[font color=\'#000000\']IMHO, the underlying formula uses "limits" as we approach infinity. Here's an empirical proposal that will get you there. For grand landscapes, assume your distant point is 5000 ft. As far as the lens is concerned, lim(f->infinity) ~ 5000 ft is certainly close enough. Now you have a useful far point you can crank into your DOF formula.[/font]
-
[font color=\'#000000\']
Jonathan and Tim appear to be discussing two different aspects of COC and hyperfocusing. I'd greatly appreciate clarification of the relationship between these two, and thank Tim and Jonathan in advance for their response:
Jonathan argues that the 1Ds demands a more restrictive COC than does film:
I don't think our positions are incompatible. What Jonathan is getting at (if I may take the liberty) is that you might as well maximize sharpness since the ultimate purposes for which most folks would use the 1Ds are not 8x10 (you don't need a 1Ds for 8x10s) but larger. If you're shooting for large framed, displayed prints the more conservative a COC you use the better.
But in the end there's no free lunch - the sweet spot is f8 - f11 +/- for most lenses, and as you move towards f32 other un-desirable stuff starts to happen. Ultimate DOF is going to be driven by lots of comprimises to the photographers intent. Is there motion? - problematical for long exposures. What's the composition, WA lends itself naturally to deep DOF, but not telephotos or macros.
The issue is important if you are using a calculator, since it will very likely use .03 or thereabouts for full frame and amend that for the smaller sensors (if it's a "good" calculator). If you need to factor in size of print etc. the "standard" coc is going to be misleading and you might as well use Jonathan's suggestion.
My only caveat is that even being conservative in using a smaller COC, I'm still not convinced that a 1 pixel COC is appropriate. I don't have the math, I'm simply recalling other discussions.... The question is under what size/viewing conditions is a 4 square block (next logical size up from 1 pixel) perceived, not as a point, but as a disk - that's a testable question.[/font]
-
[font color=\'#000000\']FOR THE RECORD:
In optics, the practical "infinity" is defined as 100x the focal length of the lens. I cannot find the reference, but I believe this standard was established for f2 and smaller apertures. Thus, set the lens on infinity and everything from 100x focal to infinity will fall within CoC limits.[/font]
-
[font color=\'#000000\']Howard:
Your position and crops do not change the effective DOF -- it simply changes the CoC YOU are now willing to accept...
And infinity is still 100x the lens' focal length.[/font]
-
Much as I agree with Marty's statement about the obvious ease of use that DEP mode gave us (and it applies to far more photogs than just landscape shooters, BTW), discussing the various merits of the fuction at this juncture is like warning the captain of the Titanic that there are ice bergs ahead, all while the ship is sinking.
In other words, it's too late for that. Let's move on.
To me, the answer is simple: custom functions. There's already one for turning ON and OFF all the other shooting modes, like "P", "TV", "Bulb", etc... Why not just have DEP mode "off" by default so as to make the next version of the camera look exactly like this version? Those who don't use it don't have to be bothered with it... whereas, those of us who need and depend on it, will be made whole again. Doesn't this seem like a perfectly reasonable win-win resolution?
If Canon claims that focus groups said that they don't use DEP mode, what did those same people say about the mode where the shutter doesn't release until something moves into the focus plane at an AF point? Sure, I can see the one time in a million that I could use it, but can you really say that MORE people want that over DEP mode? No one's complaining that it's in the CF list--in fact, no one's complaining that the CF list is even too big.
There simply must be a rational, reasonable discussion on the matter moving forward.
One thing that I must also agree with Marty about: this isn't going to happen unless photogs with clout--and who have the ear of decision-makers at Canon--say something. If it's product reviewers, fine by me.
-
Many thanks for this excellent thread. I certainly miss DEP on the 1Ds2 and with others regret that it could not simply be disabled as a personal function, should one so desire.