DOF... where's it going?!

Davince

Davince

Messages
189
Name
Del
Edit My Images
Yes
I've been scratching my head over this last couple of weeks, and as embarrassing as it is in it's basic nature, I'm going to have to throw it out there in confusion.

A couple of nights a week I do some shots at a local bar, but I keep having issues with DOF - a prime example being this shot (cropped in to highlight focus issue):

SnL0604-4-4_zps3fbe9fb6.jpg


Taken at ISO 1250, 1/3rd sec, f/5.6, Focal lenght 50mm, Subject Distance 1.5m,

By my reckoning, and confirmed by various DOF calculators that should give me a DOF of about 30cm, not far off 50/50 split front and rear.

So how the hell is the bloke on the left so out of focus?! they were nigh on dead in line, certainly not with one even near 15cm fore of the other, I wondered if it was movement blur rather than focus, but I use this shot as an example as they were like statues in it.

Am I being thick? Am I missing something obvious looking for a complex answer?

Or is my expectation of what is 'acceptable' focus just to high? Personally don't think this is acceptable but I'm getting loads of them like this last couple of shoots!

One issue I have been having there is with the lighting system used which has been playing havoc with the camera but can't imagine it changing the focal perceptions of the camera?!!
 
I would reckon on about 200mm tops, probably 80 to 90 mm in front and about 100 behind..
That works out at about 3 1/2 inches ( in real money) in front of your focus point.
The guy on the right has his arm about the shoulders of the guy on the left which would put him (left hand guy) at about 100+mm in front - I'm talking about eyes here.

Therefore not in focus- just.

Try a different f stop.
1/3 sec is too slow in my opinion...
but that's just me!

Ken
 
Last edited:
It's not focus is it? There's ghosting (forgive the beginner terminology) around left-guy's jaw line and each side of right-guy's head.
 
As kendo has said the guy on the left looks a fair bit closer to me, it's possible the guy on the right's face is at the extreme end of the depth of field and the guy on the left being even slightly further forward has thrown him out of it.
It's also possible he moved more when the flash went off and it's motion blur or there's motion blur adding to the focus issue. They certainly weren't "like statues" (or you weren't) because there's a lot of ghosting.

Can you post a 100% crop of the left guy's face? Or part of it like an eye?

The DoF calculator in the Photo Tools Pro Android app gives me 25cm depth of field for 50mm f/5.6 1.5m on a full frame body.
 
Last edited:
To me the shutter speed is too low, 1/3rd of a second is very slow, hand held shot? I think a safer speed is 1/200th of a second, but it will require use of a flash I should think at night.
 
To be fair it's a pretty poor photo all round and think you would need a better example to show up any problem you might be having.
 
It looks like movement blur to me. Look at the 'sharp' fellow on the right, you can actually see another brow and ear to his right (our left). The one on the left held even less still is all, so he appears more blurred.

I usually try to stay above 1/125, but will reluctantly use 1/60 if I really don't have the light. 1/3 is really pushing it for shots of people I would think.
 
kendo1 said:
I would reckon on about 200mm tops, probably 80 to 90 mm in front and about 100 behind..
That works out at about 3 1/2 inches ( in real money) in front of your focus point.
The guy on the right has his arm about the shoulders of the guy on the left which would put him (left hand guy) at about 100+mm in front - I'm talking about eyes here.

Therefore not in focus.

Try a different f stop.
1/3 sec is too slow in my opinion...
but that's just me!

Ken
Click to expand...

Is that your experience on the DOF measurements? I assumed that it was just the distances hence and varied F stop to try and compensate, without success, trying a load of DOF calculators and they all came out with the sort of figures I put up, and there is no way he was more than 3" in front of the other bloke (I shift angle till people are as flat to the camera to give myself best chance) I wish I'd kept a better example with two girls, similar settings, their heads were touching and you could clearly see they were 2" at most fore/aft at the eyes, one was out of focus...

I vary between 1/20th and 1 sec in this bar to give a mix of atmosphere, levels of light trails etc.
 
Davince said:
Is that your experience on the DOF measurements? I assumed that it was just the distances hence and varied F stop to try and compensate, without success, trying a load of DOF calculators and they all came out with the sort of figures I put up, and there is no way he was more than 3" in front of the other bloke (I shift angle till people are as flat to the camera to give myself best chance) I wish I'd kept a better example with two girls, similar settings, their heads were touching and you could clearly see they were 2" at most fore/aft at the eyes, one was out of focus...

I vary between 1/20th and 1 sec in this bar to give a mix of atmosphere, levels of light trails etc.
Click to expand...

You can't expect to keep moving objects in focus/blur free at such low shutter speeds.

A simple test would be to take some pictures of static objects at them shutter speeds with a tripod of to see if it is just blur affecting your shots.
 
Last edited:
Just to be clear, 1/3 is entirely intentional and is not the issue, I'm using flash and dragging the shutter intentionally to maximize ambient lighting from the club and give some background blur and atmosphere to the images. The 'ghosting' will be movement blur (and is intentional) but at the very low power setting of the flash the light burst is fast enough to freeze anything, it will freeze any motion out of them, my comment on motion was a bit tongue in cheek, apologies should have made that clear.

The specific problem is DOF I just can't work out why its so low.
 
Last edited:
There's no exif data, can you re-up?
 
I was thinking DX not FX. I've just noticed you use a D700....
And yes, from my experience rather than calculating.

Looking at the pic, it doesn't look like 1.5m with a 50mm lens on a full frame.
That's why I would like to see the exif data.
 
Davince said:
Ahhh, no as I said in the original post it's been cropped in, that's not the full image I got rid of extraneous area to highlight the problem.
Click to expand...

One day I'll learn to read...

DoF for D700, 50mm lens at 1m is about 130mm- 60 in front and 70 behind.

Of course, there could be something wrong with your sensor...:D
 
kendo1 said:
One day I'll learn to read...
Click to expand...

Lol ive never bothered :p

kendo1 said:
DoF for D700, 50mm lens at 1m is about 130mm- 60 in front and 70 behind.

Of course, there could be something wrong with your sensor...:D
Click to expand...

How about at 1.5m like in the photo? :p lol btw it's not a 50mm lens, it's a 24-70 set at 50mm, just in case that makes a difference!!!

Don't say that about the sensor - that would be just my luck at the moment, can't afford a month off while it is being dicked around under warranty!!!
 
At 1.5m you're talking about 300mm, like you said.

The problem/solution depends on a few things.
What is the exact distance?
Does it happen with other lenses?

If it is the lens, the elements may be out of alignment.

Sensor? The sensor may not be lying flat.

Have you dropped your camera/ lenses?
 
Davince said:
Just to be clear, 1/3 is entirely intentional and is not the issue, I'm using flash and dragging the shutter intentionally to maximize ambient lighting from the club and give some background blur and atmosphere to the images. The 'ghosting' will be movement blur (and is intentional) but at the very low power setting of the flash the light burst is fast enough to freeze anything, it will freeze any motion out of them, my comment on motion was a bit tongue in cheek, apologies should have made that clear.

The specific problem is DOF I just can't work out why its so low.
Click to expand...
Surly if you want to freeze their movement you will need first shutter speed? you need at least 1/50s to stop any hand shake movement ?
 
It's a combination of factors.

Your DoF calc for a 50mm lens at f/5.6 and 1.5m is correct, but calculators assume a non-cropped image. Cropping the frame changes everything, and you are effectively using a smaller format, which in this example will cut your DoF substantially. Check it out - if you have cropped half the frame area, you'll lose about a third of your estimated DoF.

There is a lot of movement, evident by the ambient ghosting. Sure the subject is frozen by flash, but either the camera or subject or both have moved after focusing. No more than you'd expect in a situation like that at 1/3sec hand-held. It's a long way back-focused, sharpest around the guy on the right's necklace.

Add up those things and it's easily explained.
 
hyeongsoong said:
Surly if you want to freeze their movement you will need first shutter speed? you need at least 1/50s to stop any hand shake movement ?
Click to expand...

artyman said:
Way to slow a shutter speed, looks more like a movement issue to me, the pattern on his shirt doesn't look too bad.
Click to expand...

Just Dave said:
(y) thats the problem
Click to expand...

Seriously guys, Photography 101... The flash fires a burst of light that, at the power I'm generally using it, lasts 1/10,000th of a second. Anything close enough to be lit by it, is frozen, Period. Dragging the shutter (Balancing ambient light to give atmosphere with slow shutter speed, getting the subject crisp with flash) is as old as the day is long. I've shot a 30 second exposure hand held with flash and got the image absolutely crisp. It is not movement.


HoppyUK said:
It's a combination of factors.

Your DoF calc for a 50mm lens at f/5.6 and 1.5m is correct, but calculators assume a non-cropped image. Cropping the frame changes everything, and you are effectively using a smaller format, which in this example will cut your DoF substantially. Check it out - if you have cropped half the frame area, you'll lose about a third of your estimated DoF.
Click to expand...

I'm not sure what you're saying Hoppy? I have a full frame camera, not cropped sensor.

Cropping the image after it's taken has no affect on DOF at all.


HoppyUK said:
There is a lot of movement, evident by the ambient ghosting. Sure the subject is frozen by flash, but either the camera or subject or both have moved after focusing. No more than you'd expect in a situation like that at 1/3sec hand-held. It's a long way back-focused, sharpest around the guy on the right's necklace.

Add up those things and it's easily explained.
Click to expand...

There isn't a lot of subject movement, the camera is intentionally shifted during the slow shot to mix and blur the ambient light and background image, and it's side to side which doesn't affect DOF.
 
Last edited:
It is not uncommon for sensors to come out of alignment. It doesn't happen often, but I've heard it mentioned on other sites.

Looks like you'll have to do a few tests with other lenses wide open and the same lenses on another camera body.

Good luck.
 
Davince said:
Seriously guys, Photography 101... The flash fires a burst of light that, at the power I'm generally using it, lasts 1/10,000th of a second. Anything close enough to be lit by it, is frozen, Period. Dragging the shutter (Balancing ambient light to give atmosphere with slow shutter speed, getting the subject crisp with flash) is as old as the day is long. I've shot a 30 second exposure hand held with flash and got the image absolutely crisp. It is not movement.




I'm not sure what you're saying Hoppy? I have a full frame camera, not cropped sensor.

Cropping the image after it's taken has no affect on DOF at all.
Click to expand...

It does, a lot actually.

There isn't a lot of subject movement, the camera is intentionally shifted during the slow shot to mix and blur the ambient light and background image, and it's side to side which doesn't affect DOF.
Click to expand...

There clearly is sideways movement, maybe fore/aft as well - perhaps not a major factor.

But the answer is clear enough - there is not as much DoF as you've estimated and the focus point is some way behind the faces. Simple as that really.
 
Why don't you setup a tripod, use the same settings on a subject like a pineapple or two to simulate the face's, then check the DOF, then if there is still a problem you can take it from there.

My money is on it just being a blurry image due to camera shake, but without testing it, well then it's your sensor:shrug:
 
Ivan S said:
Why don't you setup a tripod, use the same settings on a subject like a pineapple or two to simulate the face's, then check the DOF, then if there is still a problem you can take it from there.

My money is on it just being a blurry image due to camera shake, but without testing it, well then it's your sensor:shrug:
Click to expand...

This^

Forget the stupid (I understand what you're attempting but it's muddying the waters) low shutter speed and test the gear at the the specified distance under ideal conditions.

If you find it's faulty, you have your answer. If you find it isn't faulty, then you'll need to practice your technique some more.

Pointers (apologies if teaching my mum's mum to drink from an ovum)
Dragging the shutter to allow some ambient is often misunderstood. The ambient needs to be underexposed (anything between 1/2 a stop and a couple of stops) otherwise any subject freezing with the flash exposure will be superimposed with the shaky ambient image.

I'd only go as low as 1/3 if I wanted to show dj light patterns, and often move the camera on purpose to get that. For straight portraits I'd be looking at shutter speeds that fall within the usual handheld shutter speed guidelines.
 
By the way, Hoppy is correct about DoF. It sounds weird, but...

If the larger format is cropped to the captured area of the smaller format, the final images will have the same angle of view, have been given the same enlargement, and have the same DOF. (Stroebel, 1976).

Is the subject distance 1.5m on your exif data? The distance and the level of crop can be used to figure out the DoF.

And of course, we are assuming the two faces are parallel to the focal plane and central to the image... ?
 
Last edited:
HoppyUK said:
It does, a lot actually.
Click to expand...

Please explain how? DOF is a balance between focal length, distance from subject and aperture. How could cropping an image change that? Put it this way - if I print a picture out 8x10 and then use a knife to cut off an inch all the way round, how has that changed the DOF?

Cropping does one thing and one thing alone - it reduces the physical size of a file. Any change in perception of The image doesn't change the data acquired
 
Last edited:
Davince said:
Please explain how? DOF is a balance between focal length, distance from subject and aperture.
Click to expand...

That's how you work it out mathematically but it's really a function of magnification. If you crop an image, the magnification of it changes. As does printing it at different sizes (which is also a change of magnification) and the part you missed out was the circle of confusion size which varies depending on film/sensor size.

http://www.youtube.com/watch?v=Pdq65lEYFOM

If you have an image which is just a little bit out of focus, printing it small will often hide the fact.


Steve.
 
Last edited:
Phil V said:
This^

Forget the stupid (I understand what you're attempting but it's muddying the waters) low shutter speed and test the gear at the the specified distance under ideal conditions.

If you find it's faulty, you have your answer. If you find it isn't faulty, then you'll need to practice your technique some more.

...

I'd only go as low as 1/3 if I wanted to show dj light patterns, and often move the camera on purpose to get that. For straight portraits I'd be looking at shutter speeds that fall within the usual handheld shutter speed guidelines.
Click to expand...

Indeed now I've tried a different lens and got the same problem I'm going to have to.

With ref. to the shutter speed Phil, at 1/3rd, ISP 1250 f5.6 ambient is between 2/3rds and 2 stops under depending on where you are in the bar and what phase the lighting mixer is at. During a night there I take a range of shutter speeds because of three factors:

-the bar requested a mix of styles throughout the night, but there are limited possibilities due to the nature of the bar

-it is a very, very awkward shaped and 75% of the shots have a very very cluttered and close background that you can't lose, drop out of focus or out of reach of the flash. Straight portraits look awful. I needed a way of concentrating the focal point better

- they asked for shots that portrayed energy and enthusiasm, nothing says that less to me than a bright flash with zero ambient lol, I would however be VERY interested to hear some alternatives Phil as they'll get old very quickly! In fact I spend at least half my time there liking for alternatives.
 
Steve Smith said:
That's how you work it out mathematically but it's really a function of magnification. If you crop an image, the magnification of it changes. As does printing it at different sizes (which is also a change of magnification) and the part you missed out was the circle of confusion size which varies depending on film/sensor size.

http://www.youtube.com/watch?v=Pdq65lEYFOM

If you have an image which is just a little bit out of focus, printing it small will often hide the fact.

Steve.
Click to expand...

Yes but I'm not changing output magnification by digitally cropping - I'm just selecting a smaller are of visibility. Had I shown the original it would have just occupied more of the screen space and looked identical in focal range.
 
Davince said:
Please explain how? DOF is a balance between focal length, distance from subject and aperture.
Click to expand...
....and image magnification. You forgot image magnification. Cropping in lowers DOF merely because DOF is actually a constant set by the lens/focal length, but the more the image from the lens is magnified by a narrower field of view, the greater those effects become. Hence DOF being inversely proportional to format size.

If you, as you said, print a 10x8, and then physically trim an inch from all the way around, then no, it will change nothing, as you've not magnified the image... you've just turned a 10x8 print into a 8x6 print. The image magnification will remain the same. However, trim an inch from all the way around a digital file in photoshop, then print that on a 10c8 page, what you have done is ENLARGE a 8x6 print up to a 10x8 print, and then yes, you will be magnifying the effects of DOF as well.


...however. That's not your problem here. He's out of focus. You have misjudged the fact that they are not parallel to the focal plane. Simple.

I'd up the ISO, avoid the close in head shots.. back off to head and shoulder shots... and shoot around f8ish with ISO3200. Backing off a bit will be your biggest ally, as the differential between focal distance and comparative subject distances of the people will be so much greater... and hence give a much larger margin of error.
 
Pookeyhead said:
....and image magnification. You forgot image magnification. Cropping in lowers DOF merely because DOF is actually a constant set by the lens/focal length, but the more the image from the lens is magnified by a narrower field of view, the greater those effects become. Hence DOF being inversely proportional to format size.
Click to expand...

Absolutely, but as you go on to say cropping has nothing to do with output whatsoever, it is merely chopping bits of an image off.

Pookeyhead said:
...however. That's not your problem here. He's out of focus. You have misjudged the fact that they are not parallel to the focal plane. Simple.

I'd up the ISO, avoid the close in head shots.. back off to head and shoulder shots... and shoot around f8ish with ISO3200. Backing off a bit will be your biggest ally, as the differential between focal distance and comparative subject distances of the people will be so much greater... and hence give a much larger margin of error.
Click to expand...

I don't think it is that simple, because its not just one shot that's the issue. I think Ken has the answer and will be running some tests later with something less animated than a couple of drunks lol.

This shot pre-crop is head and shoulders btw, very much agree that head shots are a tad tricky in this situation!
 
Davince said:
Absolutely, but as you go on to say cropping has nothing to do with output whatsoever, it is merely chopping bits of an image off.
Click to expand...

If you were physically cropping the print, yes, but you're not.



Davince said:
I don't think it is that simple, because its not just one shot that's the issue. I think Ken has the answer and will be running some tests later with something less animated than a couple of drunks lol.

This shot pre-crop is head and shoulders btw, very much agree that head shots are a tad tricky in this situation!
Click to expand...

Ken and I are essentially saying the same thing. He's asking you to check your focus distance to see if it is what you think it is, because the closer you go, the more critical it becomes.. and he also suggests the two people are not parallel to the focal plane.


There simply is no other reason for this happening.
 
You won't win the DoF debate with Hoppy... it's a war of attrition and he's a veteran. I've tried and failed like many others... ;) :LOL:

There's always a possibility of equipment failure (sensor misalignment and/or optic mis-alignment) so you might as well do tests in controlled lighting using brick walls and focus charts and all that jazz.

But like others say, if that isn't the issue then you need to look at your technique and how you approach the situation in terms of exposure values, to give you the best shot at nailing it every time.
 
Last edited:
Pookeyhead said:
If you were physically cropping the print, yes, but you're not.

Ken and I are essentially saying the same thing. He's asking you to check your focus distance to see if it is what you think it is, because the closer you go, the more critical it becomes.. and he also suggests the two people are not parallel to the focal plane.


There simply is no other reason for this happening.
Click to expand...

Erm... yes I am? The screen displays at 72ppi, (okay some 100 but you know waht I mean) so output is set at that for digital display. If I crop an image in photoshop, without resampling, then I am doing EXACTLY that, I am digitally 'cutting bits of the print off'. Nothing else. I am just removing pixels, not altering the remaining in any shape or form.

Ken also suggests elements not in alignment and sensor being loose, two things I'm pursuing.

The whole basis for this thread was that I started checking DOF calculators and found that given the settings I was using, depth of field being seen in results do not tally up. The setting in this picture give a focal depth or 300mm, that's a foot, split 140 in front, 160 behind. Focal point is spot on the RHS persons eye. The person on the left was in no way 6 inches in front of the other, that is a HELL of a lot to misjudge at this distance from the subject.
 
Last edited:
Davince said:
Erm... yes I am? The screen displays at 72ppi, (okay some 100 but you know waht I mean) so output is set at that for digital display. If I crop an image in photoshop, without resampling, then I am doing EXACTLY that, I am digitally 'cutting bits of the print off'. Nothing else. I am just removing pixels, not altering the remaining in any shape or form.
Click to expand...

Agreed.. but the minute you then zoom in a bit to get the image the same size on the screen, you have magnified it. Let's say you sell the client 8x6 prints... and one of the images, you cropped as described above... would you then sell that print as say... 6x4 to keep the magnification the same, or would you print them all at 8x6? If the latter, you have enlarged one of the images.

Academic anyway, as the level of crop you're doing will have minimal effect. I was just clarifying that Hoppy is indeed correct.

Davince said:
Ken also suggests elements not in alignment and sensor being loose, two things I'm pursuing.
Click to expand...

If that were the case then you'd have serious issues with all your images.

Davince said:
The whole basis for this thread was that I started checking DOF calculators and found that given the settings I was using, depth of field being seen in results do not tally up.
Click to expand...

Then you're probably misjudging the distances from subject to camera. You think they're both square on to the camera, but they're not. In the shot you posted up, it' apparent to me that the guy on the left is clearly nearer the camera than the guy on the right.

Davince said:
The setting in this picture give a focal depth or 300mm, that's a foot, split 140 in front, 160 behind. Focal point is spot on the RHS persons eye. The person on the left was in no way 6 inches in front of the other, that is a HELL of a lot to misjudge at this distance from the subject.
Click to expand...

Well.. misjudged it you must have... as there's no other explanation other than something in the camera being out of alignment. That's highly unlikely. Easy to solve though.

Place your camera on a tripod. Face it at a brick wall square on (or any wall with pattern or texture and detail)... take a shot at the same settings you used for this one. If the left hand side of the image is blurred, then you have a camera problem. If it's sharp across the frame, then you have a technique problem.

Until you have done this, you, and everyone else in this thread is guessing. So go and do that... it takes 5 minutes tops, and you'll have your answer.
 
Last edited:
Pookeyhead said:
Agreed.. but the minute you then zoom in a bit to get the image the same size on the screen, you have magnified it. Let's say you sell the client 8x6 prints... and one of the images, you cropped as described above... would you then sell that print as say... 6x4 to keep the magnification the same, or would you print them all at 8x6? If the latter, you have enlarged one of the images.
....

.
Click to expand...

I never get this - why introduces arguments that are not part if the equation to a debate?

Zooming in will affect it, of course it will, but I'd have said 'cropping and zooming' or 'cropping and magnifying'

Hoppy said that cropping changes DOF. IT does not. I haven't changed magnification, I've not resembled it, I've not done anything else that would suddenly make his or your argument valid, I've just cropped it, the pixels on screen are exactly as they were, there are just a few missing around it.

While in doing my test you get two footballs and place one 6" in front of the other. These two lads are NOT that far out of alignment.
 
You must log in or register to reply here.
Back
Top