Infinity focus and DoF

[Colin Southern]

Hi Colin

It is a very beautiful shot. I don't know when would I be able to take shots like this. I believe you would have taken it at a fairly longer shutter speed.

Two questions

Where did you focus?

Where did you take the meter reading from?


Thanks

Thanks <person who's first name I don't yet know>

You'd probably find scenes like that are more common that you think -- it's just a case of learning to recognise them. In reality this is just a piddly little stream -- it's the fact that I can push in aggressively with a wide-angle lens that makes it look far more spectacular. He's another shot from the same shoot showing how the camera was positioned:



One of the best kept secrets of photography (shhhh - don't tell anyone) is to simply use a focal length that gives you a field of view that only records the attractive part of the scene; it doesn't matter how much crud is outside that field of view. Shoot that at the right time of day - process it carefully - and "hey presto" you'll have a keeper.

Focusing was about 1/3 of the way into the scene. Metering was probably just evaluative (it's a relatively low dynamic range scene, so nothing special required in terms of exposure) (I actually shot this at all apertures from 2.8 through to 22, but it was this F4 version that give a shutter speed that gave the water a nice movement).

[sdewal]

Thanks Colin.

I am reading a book called "Understanding Exposure - 3rd edition by Bryan Peterson". The writer has mentioned a metering technique called "sky brothers" in which we take the meter reading from the blue skies or dusky skies (whatever it is). He also stresses that after taking the meter reading when we recompose the scene the meter may go berserk and show us that the scene is under or over exposed, but we should ignore the meter and take the shot at the previously metered exposure (meter reading that we took from the skies).

For people like you, it's a child's play because you guys can judge most of the things instantly but for people like me, these guidelines become kinda gospels until we gain some experience.

Do you do any special metering ? In your posted picture, did you take the meter reading from sky or trees?


Thanks
Shail

[Colin Southern]

Thanks Colin.

I am reading a book called "Understanding Exposure - 3rd edition by Bryan Peterson". The writer has mentioned a metering technique called "sky brothers" in which we take the meter reading from the blue skies or dusky skies (whatever it is). He also stresses that after taking the meter reading when we recompose the scene the meter may go berserk and show us that the scene is under or over exposed, but we should ignore the meter and take the shot at the previously metered exposure (meter reading that we took from the skies).

For people like you, it's a child's play because you guys can judge most of the things instantly but for people like me, these guidelines become kinda gospels until we gain some experience.

Do you do any special metering ? In your posted picture, did you take the meter reading from sky or trees?


Thanks
Shail

At the risk of irritating Bryan Peterson fans, I do wonder if at times some of these things are strung out to be more complex than they really are just to sell another book.

Exposure really comes down to just a couple of things:

1. The technical requirements of what the sensor needs for an effective capture, and

2. The metering that gets one to that point.

The sensor itself is governed by the saturation point -- the point at which data is lost (due to over-exposure), and for the most part, it's what ultimately dictates the exposure of the shot (ok, I'm taking a few liberties here, but that's none-the-less the "rub" of it). So at it's simplest form, the main rule is "don't over-expose any highlights that contain important data" eg if you take a photo of a '58 Chevy and there's blinkies on the camera review screen due to the sun reflecting off the chrome bumper then "so be it" (not a worry if the rest of the shot looks OK), but if you take a shot of a bride and groom and get blinkies throughout the bride's dress, then you probably need to reduce the exposure somewhat. On the flip-side, if the histogram stops well short of the right-hand side and the image as a whole looks dark, then it's a reasonable indication of an under-exposure -- so it can be cranked up a little.

Modern cameras typically capture 11 to 12 stops of information, and yet we can typically only display around 6 on a monitor and 4 in a print ... so the camera is capturing a LOT more information than we can see, unless we manipulate it. So in that context, if you take a typical shot and the highlights aren't blown and the shot doesn't look overly dark, then chances are it's going to be just fine for subsequent processing (assuming shooting RAW here). Or put another way, "there's an awful lot of leeway".

In terms of metering, again, people often seem to make hard work of it. In general, you have a couple of controls to worry about - the first is the metering pattern (spot, partial, matrix/evaluative, and centre-weighted average). These simply tell the camera what parts of the scene to consider in it's calculation; for a landscape that will typically be everything (evaluative/matrix) -- in other situations it may be something smaller. So in terms of "what's the right mode", the answer is easy -- "the one that corresponds to what's important in the image". Part B to that is understanding how metering works - basically, it simply (VERY simply) assumes that what you're pointing at is middle gray. That's it. If you're evaluating the entire scene then it usually IS close to a middle gray (there will be an equal number of objects that reflect 0 to 2 stops above middle gray as there are objects that'll reflect 0 to 2 stops below middle gray) (that WILL be the case because the camera will MAKE it the case when it sets the exposure!). If you're shooting something that varies significantly from that (eg a white bear in the snow or a black cat on a black rug) then you need to tell the camera to make an allowance for that (because it doesn't know it's a black cat or a white bear). Obviously the more you narrow the scope of the metering (all the way down to spot-metering), the lesser the chance that what you're metering will be representative of the entire scene, so the more likely it is that you'll need to apply EC (Exposure Compensation) eg without EC if you spot-meter a black cat on a black rug you'll get a gray cat. Spot-meter a white bear in the snow and you'll get a gray bear in gray snow.

Having just said all that, generally, the camera will bias a shot to protect significant areas of "highlights" (not necessarily anything whitish, but more often incident light from the sky -v- reflected light from objects on earth. These types of scene can have a big dynamic range and one has to be careful about how they expose them (base ISO and pushing the highlights as far as possible) -- in these situations one may have to bias the camera's exposure suggestion, but once again, the blinkies and histogram tell you all you need to know.

So to answer the question (finally!), in shots like the one we're talking about, I simply had the camera meter the entire scene. The dynamic range of that scene is probably only 5 or 6 stops - and the camera is capturing 11 to 12 - so an appropriate exposure is pretty hard to miss. Nothing fancy required -- just let the camera automation do the job we paid it to do when we bought the camera

If you're shooting outside and you have the sky in a lot of shots a long with a lot of earth objects then usually metering is simply a case of leaving the camera on matrix / evaluative and (optionally) dialing in an EC of -1 if you want to have bluer and more saturated skies (at the expense of darker shadows that you'll need to raise a stop with the fill light control in processing.

So there ya go - beat Bryan by 173 pages (taking a few liberties along the way!)

[rambler4466]

Try this theory: Let the lens determine the focal length by doing the following. Divide the focal length by 3. Example:
(For a full frame sensor) 18mm divide by 3 = 6 Focus 6 feet into the frame. (For a cropped sensor DX) 12mm x 1.5 (magnification factor) = 18 divide by 3 = 6. Focus 6 feet into the frame. (credit to Jim Zuckerman)

Check out his gallery here:
http://jimzuckerman.com/?
page/56874/galleries

He got the focus right on this one! (The one with the town, steeple and mountains in the distance)

http://jimzuckerman.com/?page/115543/austria

[Andrew76]

Can you please give me the "Colin's Notes" to the next 4 questions I have?!?!?

[sdewal]

At the risk of irritating Bryan Peterson fans, I do wonder if at times some of these things are strung out to be more complex than they really are just to sell another book.

Exposure really comes down to just a couple of things:

1. The technical requirements of what the sensor needs for an effective capture, and

2. The metering that gets one to that point.

The sensor itself is governed by the saturation point -- the point at which data is lost (due to over-exposure), and for the most part, it's what ultimately dictates the exposure of the shot (ok, I'm taking a few liberties here, but that's none-the-less the "rub" of it). So at it's simplest form, the main rule is "don't over-expose any highlights that contain important data" eg if you take a photo of a '58 Chevy and there's blinkies on the camera review screen due to the sun reflecting off the chrome bumper then "so be it" (not a worry if the rest of the shot looks OK), but if you take a shot of a bride and groom and get blinkies throughout the bride's dress, then you probably need to reduce the exposure somewhat. On the flip-side, if the histogram stops well short of the right-hand side and the image as a whole looks dark, then it's a reasonable indication of an under-exposure -- so it can be cranked up a little.

Modern cameras typically capture 11 to 12 stops of information, and yet we can typically only display around 6 on a monitor and 4 in a print ... so the camera is capturing a LOT more information than we can see, unless we manipulate it. So in that context, if you take a typical shot and the highlights aren't blown and the shot doesn't look overly dark, then chances are it's going to be just fine for subsequent processing (assuming shooting RAW here). Or put another way, "there's an awful lot of leeway".

In terms of metering, again, people often seem to make hard work of it. In general, you have a couple of controls to worry about - the first is the metering pattern (spot, partial, matrix/evaluative, and centre-weighted average). These simply tell the camera what parts of the scene to consider in it's calculation; for a landscape that will typically be everything (evaluative/matrix) -- in other situations it may be something smaller. So in terms of "what's the right mode", the answer is easy -- "the one that corresponds to what's important in the image". Part B to that is understanding how metering works - basically, it simply (VERY simply) assumes that what you're pointing at is middle gray. That's it. If you're evaluating the entire scene then it usually IS close to a middle gray (there will be an equal number of objects that reflect 0 to 2 stops above middle gray as there are objects that'll reflect 0 to 2 stops below middle gray) (that WILL be the case because the camera will MAKE it the case when it sets the exposure!). If you're shooting something that varies significantly from that (eg a white bear in the snow or a black cat on a black rug) then you need to tell the camera to make an allowance for that (because it doesn't know it's a black cat or a white bear). Obviously the more you narrow the scope of the metering (all the way down to spot-metering), the lesser the chance that what you're metering will be representative of the entire scene, so the more likely it is that you'll need to apply EC (Exposure Compensation) eg without EC if you spot-meter a black cat on a black rug you'll get a gray cat. Spot-meter a white bear in the snow and you'll get a gray bear in gray snow.

Having just said all that, generally, the camera will bias a shot to protect significant areas of "highlights" (not necessarily anything whitish, but more often incident light from the sky -v- reflected light from objects on earth. These types of scene can have a big dynamic range and one has to be careful about how they expose them (base ISO and pushing the highlights as far as possible) -- in these situations one may have to bias the camera's exposure suggestion, but once again, the blinkies and histogram tell you all you need to know.

So to answer the question (finally!), in shots like the one we're talking about, I simply had the camera meter the entire scene. The dynamic range of that scene is probably only 5 or 6 stops - and the camera is capturing 11 to 12 - so an appropriate exposure is pretty hard to miss. Nothing fancy required -- just let the camera automation do the job we paid it to do when we bought the camera

If you're shooting outside and you have the sky in a lot of shots a long with a lot of earth objects then usually metering is simply a case of leaving the camera on matrix / evaluative and (optionally) dialing in an EC of -1 if you want to have bluer and more saturated skies (at the expense of darker shadows that you'll need to raise a stop with the fill light control in processing.

So there ya go - beat Bryan by 173 pages (taking a few liberties along the way!)

Thanks Colin

[Polar01]

Frank your theory is not correct, 18mm divided by 3 as you state is 6, so you focus 6ft into the frame at what f/stop settings will allow you to get you infinity is it 1.8?, 2?, 2.8?, 3.5?, 4?, well none it is F/6.3. Using your method would mean that the image Colin shot at f/4 would not work. If Colin had used your method then he would have everything from 3.7ft to 20.2ft in focus not exactly infinity. As I use a wide angle a lot, I know that if I focus on something 10ft away with the lens set at 18mm I can focus out to infinity at f/4 to f/32. Once I get greater than 28mm I have to me more careful so now I am up more in the f/11 to f/18 range this is where the an app become useful, if it tells me that everything from 5ft away to 3500ft is in focus thats' close enough to infinity for me. Have you every seen anything with the human eye in focus at the distance, at lot of us have a hard time at 3ft away without glasses.

Allan

[ajohnw]

Thanks Colin.

I am reading a book called "Understanding Exposure - 3rd edition by Bryan Peterson". The writer has mentioned a metering technique called "sky brothers" in which we take the meter reading from the blue skies or dusky skies (whatever it is). He also stresses that after taking the meter reading when we recompose the scene the meter may go berserk and show us that the scene is under or over exposed, but we should ignore the meter and take the shot at the previously metered exposure (meter reading that we took from the skies).

For people like you, it's a child's play because you guys can judge most of the things instantly but for people like me, these guidelines become kinda gospels until we gain some experience.

Do you do any special metering ? In your posted picture, did you take the meter reading from sky or trees?


Thanks
Shail

A problem with books is that the are written to earn income and if some one wants to write another for even more income they have to come up with something "new" otherwise people wouldn't buy them. The techniques mentioned often have a name as well and that needs to be new too.

To me this sound like a poor variation of an ancient venerable technique of metering for good skies by including more of it when the exposure is set and then reframing before taking the shot. Poor because I think that many cameras would under expose the sky making other areas darker than they need be and wasting dynamic range. So when you read something like this you have a camera and hopefully the sky to so go try it rather than just accepting it. Any old scene will do for this. Most people could just open a window and take a shot. Skies usually vary from day to day so over a period of time all sorts could be tried.

A better method of obtaining good skies has gained the name shooting to the right. Matrix metering and an exposure using the cameras built in histogram. There is a tutorial on histograms in the tutorial link at the top of the home page. Also one on shooting to the right

https://www.cambridgeincolour.com/tu...techniques.htm

Some people fiddle with the camera jpg settings as the real dynamic range of the camera from raw is higher than the jpg range the histogram shows. The gains at the bright end can vary from next to nothing so personally I would leave things as they are. One other way of handling that is to deliberately clip a certain EV value off the highlight end once you know what the camera does or bracket exposures, take several shots etc. The dark end is often more realistic than people tend to think.

DOF? Please bear in mind it's all based on getting 1500 points / dots across the diagonal of a 10x8 print from a full frame film camera. Zeiss decided it should be 1750. This shot has a diagonal of around 40in on my screen so 6000 points are needed according to most or more according to Zeiss.

http://backup.cambridgeincolour.com/...98480/original

A link as it's a bit big. Click when opened and it will expand to full size , 100% resolution. The PP could be better. It's one of many many taken trying a camera out. Nothing special at all.

Bit strange that there is only 4,500 pixels across the diagonal of the shot. Worse still those pixels are in blocks of 4 in an RGB arrangement, Even worse it's a micro 4/3 shot so the circle of confusion should be 0.015mm not the one I used above so supposedly 7,600 are needed.

Just wait till you get to max size of print.

Basically the software involved with pixel type sensors has an effect as well but by all means carry on worrying about it rather than taking some shots to find out.

John
-

[William W]

A comment on old lenses. In use it wasn't a good idea to make use of the markings on the lens for focal distance as they were not accurate enough. Focusing was done by eye.

Another comment on the use of old lenses and also on new lenses:

The distance scales on the lenses were indeed often used and used very accurately.

But maybe only used by a specific group of photographers, being those engaged in street photography or similar styles.

Especially Newspaper Photographers used the distance and the DoF markings on the lenses for Zone Focussing Technique, particularly when they would shoot “Hail Mary” (e.g. at Doorstops)

Street Photographers would also use Zone Focussing when: Shooting From The Hip.

Even with modern cameras, equipped with Auto Focussing Lenses, MANUAL Zone Focussing Technique is quite often more accurate for any shot made, when the eye is not to the viewfinder (and there is no LCD view).

I use the Zone Focussing Technique with the: Canon EF 35 F/1.4 L USM; EF 24 F/1.4L USM and sometimes even the EF 50 F/1.4. on my 5D Series cameras.

These lenses have enough data on the barrel to allow Zone Focussing Technique to be used accurately enough.

Here you can see my 35/1.4:


The overlay is of tape with the black markings is where I need to hide the white numeral “1” that appears on the METERS distance scale to manually set the lens at the three, most used, Focussing Distances.

***

There is enough travel in the focus turret of the EF50/1.4 to make reasonable results using Zone Focussing, with a bit of practice and even though the EF 50 F/1.4 has DoF aperture markings only at F/22 and it is possible to set up an overlay of tape suitable for Zone Focussing at another aperture: typically F/8.

***

An adaption of the (manual) Zone Focus Technique is using AF to nail a particular distance and then to maintain that particular Shooting Distance.

This technique assumes that one has rote knowledge of DoF for Typical Framings at a few key apertures. For example, I know F/2.8; F/5.6 and F/11.

This is an example of the result of that particular adaption of the Zone Focussing Technique.


As I moved to camera left I was holding the camera extended to camera right and I kept the same shooting Distance to the Subjects.

WW

[ajohnw]

My experience having used a number of makes is that if the distance scales were used then a slower F stop than the scale indicated was needed to take care of the inaccuracies. As there has been an increase in the use of manual lenses on AF cameras I have also seen comments asking why when they set infinity for instance it wasn't in focus.

35mm Cameras I have used and others have used with the same problem are Zenith, Nikon, Topcon, Minolta, Mamiya. Olympus and Pentax. 4 of those myself. The other 2 belonged to close friends. I also owned a Minox 35GT so am well aware of the problems of using F2.8 and a distance scale. Even the use of a tape measure was dubious.

John
-

[HJMCD47]

Just focus about 1/3 into the scene. With a WA or UWA lens the DoF will be huge at medium to minimum apertures.

I recently purchased a second hand 11-22 Olympus wide angle lens and I cant get the whole photo in perfect focus. I thought perhaps I was focusing at the wrong point but I tried 1/3 into the scene and it is still not clear. I am beginning to think it may be something wrong with the lens but it could be that I need to get to know the lens and where the sweet spot is for focus.

[jcuknz]

The metering off the sky sounds very much like the original Weston Master meter which I started with with its A-?-C on the dial and you took a reading off the back of your caucasion hand and put the 'C' to it .... C being one stop closed down from the average of the scene ..... B&W film is very forgiving Todat with digital and looking for a simple solution after reading the Luminous Landscape article 'Expose to the right' I found my camera has 'blinkies' ... areas flash when the camera notices they are over-exposed ... so all I have to do is decide if I amd bothered by that and make an adjustment or not.

Since I am not a landscape photographer and probably never will be DoF is of little interest to me becuase most of my subjects are what I want to be sharp ... I focus on them and press the trigger ... then in editing I can put the background OOF ....

You make the choice ... life can be infurnally complicated or you can make it simple ... I prefer the second option.

[Colin Southern]

I recently purchased a second hand 11-22 Olympus wide angle lens and I cant get the whole photo in perfect focus. I thought perhaps I was focusing at the wrong point but I tried 1/3 into the scene and it is still not clear. I am beginning to think it may be something wrong with the lens but it could be that I need to get to know the lens and where the sweet spot is for focus.

What aperture are you shooting at? What shutterspeed? On a tripod? Any samples you can post for us?