DOF-question - D. K?

when shooting tabletops (archimodells), how is the relation from DOF and focale, when using different focale lenghts?

In practical terms:

We've a object, with, lets say 50 cm of width, that should fill the long side = 36 mm of the FF-sensor.

Now, I might use either a 50 mm macro, or a 28 macro.
Which of the two lenses will have the bigger DOF?

Chosing canons 50 mm will result in a bigger distance lens-object, therefore DOF will increase; as the 50 mm has a shallow DOF, it might diminish.

While using the macrofunction of the YCZ 28-70 @ 28, the lens will be closer to the object; less DOF, but then again, the 28 mm has a bigger DOF at a given f-stop, so that might balance it.

We might add the 100 mm macro in the game...

I'm aware, hat the MTF of the canon macro is better than the 28-70.

Any insights?

I'm sure Doug Kerr will know it, as I found a DOF-spreadsheet from him, dating 04, in my archives ;-)

A good question that I was pondering earlier this evening as I thought about shooting the lampshade in our hall!

I look forward to a knowledgeable exposition of dof and coc etc, with special consideration of the diffraction limit for sensors with different pixel spacing (seriously)


Mike

Do the math ;-)

Michael Fontana said:

when shooting tabletops (archimodells), how is the relation from DOF and focale, when using different focale lenghts?

In practical terms:

We've a object, with, lets say 50 cm of width, that should fill the long side = 36 mm of the FF-sensor.

Now, I might use either a 50 mm macro, or a 28 macro.
Which of the two lenses will have the bigger DOF?

Chosing canons 50 mm will result in a bigger distance lens-object, therefore DOF will increase; as the 50 mm has a shallow DOF, it might diminish.

While using the macrofunction of the YCZ 28-70 @ 28, the lens will be closer to the object; less DOF, but then again, the 28 mm has a bigger DOF at a given f-stop, so that might balance it.

We might add the 100 mm macro in the game...

I'm aware, hat the MTF of the canon macro is better than the 28-70.

Any insights?

I'm sure Doug Kerr will know it, as I found a DOF-spreadsheet from him, dating 04, in my archives ;-)

Click to expand...

Hi Michael,

You can do the math, it is easy. Short answer, if you keep the same subject and it will fill the frame fully, then the 50mm and the 28mm will both have the same DOF using the same aperture.

Let me demonstrate.
1) 50mm lens
The distance from the subject is 500mm (size of object) * 50mm (f) / 36 mm (frame size) = 69.44 cm
Go to the online DOF calculator and calculate the DOF using these values @ any aperture, eg f5.6

DOF = 6.1 cm

2) 28 mm lens.
The distance from the subject is 500mm (size of object) * 28mm (f) / 36 mm (frame size) = 38.88 cm
Go to the online DOF calculator and calculate the DOF using these values @ any aperture, eg f5.6

DOF = 6.1 cm

You can try various scenarios but as long as the image size on the sensor does not change, the focal length change will not affect DOF.

HTH,

Cheers,

Hi, Michael.

You rang?

For the magnification you mention (based on object and image size):

With a focal length of 50 mm, the distance to the object will be about 740 mm. If we assume an aperture of f/3.5, and an acceptable circle of confusion of 0.031 mm, then the total depth of field for that setup will be about 44.4 mm.

With a focal length of 36 mm, the distance to the object will be about 415 mm. If we again assume an aperture of f/3.5 and an acceptable circle of confusion of 0.031 mm, then the total depth of field for that setup will be about 44.6 mm.

(This all assumes that for both lenses, the entrance pupil is located at the first principal point, not a bad assumption.)

Indeed, until we get into the realm of really large magnifications, for a constant aperture and a consistent acceptable circle of confusion, the total depth of field varies about as the inverse square of the magnification, and if the magnification is constant, then the DoF is nearly constant. Focal length is not a big factor of itself (so long as those other factors, including magnification, remain the same).

Hi, Cem,

Cem_Usakligil said:

Let me demonstrate.
1) 50mm lens
The distance from the subject is 500mm (size of object) * 50mm (f) / 36 mm (frame size) = 69.44 cm

Click to expand...

Not quite, actually about 74.0 cm.

Your calculation for object distance is only valid for objects at a "great" distance from the lens, and in the example, that isn't really "enough" so.

But your conclusion is nevertheless correct about the DoF being nearly constant!

Best regards,

Doug

Doug_Kerr said:

Hi, Cem,



Not quite, actually about 74.0 cm.

Your calculation for object distance is only valid for objects at a "great" distance from the lens, and in the example, that isn't really "enough" so.

But your conclusion is nevertheless correct about the DoF being nearly constant!

Best regards,

Doug

Click to expand...

Doug I know you well enough not to go splitting hairs with you on such matters, especially when you are right . I forgot to add the focal distance of the lens to my equation, mea culpa.

Since the subject width of 50 cm is a theoretical given, my calculations are applicable to an object of a slightly narrower size than 50 cm but the end result is still valid. Because the question was not what exactly the DOF would be but whether it would be the same for different focal lenghts.


Cheers,

Hi, Cem,

In fact, the distance to the object, P, is given by:

P = (1+m)(f/m)

where m is the magnification and f is the focal length.

But in a case where P is large, m must be small, and then this becomes almost:

P = f/m

which is what you used.

Best regards,

Doug

Thanks!
Good to be here.....

yes, I had to smile Doug, when I noticed your name on the spreadsheet, I had forgotten for years....

My "photographic" feeling told me, that the 28 mm would have a bigger DOF.
Wrong.... but I admit that in school, during the optic lessons, I rather had a seat beside the window.

So thanks to both of you.

How about the macro beeing optimised to be stopped down further, having less diffraction?

In the mentioned case, diffraction is not a real problem, as the prints will not be big, therefore stopping down to f 16 will gain DOF to about 19 cm...

Interesting threads and answer!

Doug and Cem, do you have some aspirin for me? <grins:>)

Michael Fontana said:

How about the macro beeing optimised to be stopped down further, having less diffraction?

Click to expand...

Hi Michael,

Diffraction is mostly dictated by the aperture value (and the wavelength of light).

Macro lenses may be optimized for reducing residual aberrations (and flatness of field) at (very) close focusing distances, but diffraction is still a function of aperture(*).

In the mentioned case, diffraction is not a real problem, as the prints will not be big, therefore stopping down to f 16 will gain DOF to about 19 cm...

Click to expand...

It all depends a bit (a lot actually) on the final output magnification and viewing distance. With a small output size the unsharp areas that can be resolved by eye are also smaller, even to the point that they are a sharp as the in-focus and in-DOF areas.

One can do a reverse calculation.
Suppose that the output is magnified to 10x the sensor array size. That also means that to remain pin sharp at normal reading distance (say, 8 linepairs per mm) in output, the on-sensor resolution must be 10x higher, 80 line-pairs/mm. That would mean that the optimal COC or diffraction (whichever is largest) should not exceed 1/160mm line-width = 0.00625mm or 6.25 micron.

So that would theoretically require a sensel pitch of 6.25 micron and a diffraction or COC that doesn't exceed 1.5x that pitch, so 9.375 micron. The 1.5 factor is in my experience a close approximation of the threshold of visual degradation on a Bayer CFA sensor array. When you feed the 9.375 micron as COC into a DOF calculator, and set the distance to whatever is needed to fit the scene in the FOV with the focal length used, you'll get your magic f/number, and associated DOF. I use the handy offline (Windows) DOF calculator application by Bob Atkins.

If that is not enough for the subject depth at hand, then focus stacking will help to overcome the physical limitations.

Luckily we have some leeway due to sharpening and Raw converter capabilities, but the above is a close enough approximation to predict optimal sharpness. Of course there is not a sudden transition from in-focus to out-of-focus, so the calculation only shows optimal resolution (it can't get sharper than that). The MTF of the lens also plays a role, but let's not over-complicate things. One may also tolerate 'some' unsharpness at the foreground/background to give a sense of depth.

(*) BTW it's easy to calculate the diameter of the diffraction pattern (assuming a round aperture and green (555nm) light waves). Just multiply the Aperture (f/number) with 1.3542, so approx. f/7.0 or wider in the example above. Smaller output magnification or more distant viewing conditions allow smaller apertures.

Bart

Michael Fontana said:

when shooting tabletops (archimodells), how is the relation from DOF and focale, when using different focale lenghts?

Click to expand...

And there is another important factor to consider. Do you want the table top model to have a perceived perspective that's similar to the final construction, or is it okay for it to look like a model?

One has to scale the shooting postition (=perspective) of the model, to the viewing distance of the finalized construction to achieve a natural looking perspective.

Having established that DOF doesn't change significantly for a given magnification, one can determine the shooting distance of the model from it's scale, and from there derive the required focal length.

Bart

Hi, Bart,

Bart_van_der_Wolf said:

One has to scale the shooting postition (=perspective) of the model, to the viewing distance of the finalized construction to achieve a natural looking perspective.

Click to expand...

An excellent point.

Best regards,

Doug

Can't you shoot with one lens first and then, change lenses, shoot, place two images side by side and compare DOF etc... just how I would do it since I'm bad at math...

Bart_van_der_Wolf said:

And there is another important factor to consider. Do you want the table top model to have a perceived perspective that's similar to the final construction, or is it okay for it to look like a model?

One has to scale the shooting postition (=perspective) of the model, to the viewing distance of the finalized construction to achieve a natural looking perspective.

Bart

Click to expand...

Well Barth, that depends much on the way the model is build; and the type of modell;
I' ve been shooting between 500 and 1000 models, yet - stopped counting - and alwith struggling with that problem.

When having what we call a °urban model°, they' re often in gypsum and in 1/500-scale, you can't usually get a human perspective, due to the big downscale, and the surrounding buildings.

Usually, these shots are taken in kinda bird's view, here's large model example in a barbie-version; handheld stitch; 3 x 50 mm macro:





For getting the human perspective, I' ve got a Perspectar, a special lens - similar to a endoscope, but build for FF with a normal lens-FOV. It was build in a serie of about 100 copies, so you won't find any infos at all - maybe a university still runs one.

Here you see it on my 1 Ds-2:



You might notice the little mirror at its bottom, for a vertical use, between the buildings. Wide open it has a f-stop of 22 and closed down, it has 64! (sic)


Off course, it can be used horizontally - without mirror - as well, and with its nose diameter of 30 mm, it fits even in small working models, no chance with other 35 mm-lenses, and produces images like that, as this shot was taken in the center of the corridor!



If I recall correctly, the corridor was about 7- 10 cm in height. With this lens, taking photos is a surgery-job, using fiber-lites, I even build softboxes for them...

leonardobarreto.com said:

Can't you shoot with one lens first and then, change lenses, shoot, place two images side by side and compare DOF etc... just how I would do it since I'm bad at math...

Click to expand...

Yep, Leonardo, this would have been another solution.

Still, its nice to have such a great source here, to exchange and learn additional things from other OPFlers.

..."Its nice to have such a great source here,"... I know, isn't it? and they know so much about almost anything. ... There is even people with this ...

LOL, Leonardo

"about almost anything".
Almost is the keyword here, for tagging ;-)

Michael Fontana said:

For getting the human perspective, I' ve got a Perspectar, a special lens - similar to a endoscope, but build for FF with a normal lens-FOV. It was build in a serie of about 100 copies, so you won't find any infos at all - maybe a university still runs one.

Here you see it on my 1 Ds-2:

You might notice the little mirror at its bottom, for a vertical use, between the buildings. Wide open it has a f-stop of 22 and closed down, it has 64! (sic)


Of course, it can be used horizontally - without mirror - as well, and with its nose diameter of 30 mm, it fits even in small working models, no chance with other 35 mm-lenses, and produces images like that, as this shot was taken in the center of the corridor!

If I recall correctly, the corridor was about 7- 10 cm in height. With this lens, taking photos is a surgery-job, using fiber-lites, I even build softboxes for them...

Click to expand...

Michael this is really impressive!

I had a shoot yesterday -in studio- of a model of a yacht. Scale of model was 1/20.
Interestingly, this yacht (a power catamaran) as upper deck interiors built, the hell to shoot! almost impossible (DOF…)

I even tried the 70-200 at ƒ 32!

Below are 2 images of antother model (but scale 1/15), shot some months ago, for the same client:





Next time I'll ask you to rent me your combo!
Do you have any snap of the beast at work?

Nicolas Claris said:

Next time I'll ask you to rent me your combo!

Click to expand...

Hi Nicolas,

You might also want to search for the "EUROPEAN ARCHITECTURAL ENDOSCOPY ASSOCIATION".
There may be members in France as well.

As for impossible DOF, "Focus Stacking" will allow to solve that. I've read that Photoshop CS4 also allows to do that, although I don't know how well it's implemented in this first Photoshop version.

HeliconFocus (Pro) is a very user friendly application created for that specific purpose, TuFuse Pro (windows only) is producing good results, and there are several free programs around (but with their quirks).

Bart

Bart_van_der_Wolf said:

As for impossible DOF, "Focus Stacking" will allow to solve that. I've read that Photoshop CS4 also allows to do that, although I don't know how well it's implemented in this first Photoshop version.

HeliconFocus (Pro) is a very user friendly application created for that specific purpose, TuFuse Pro (windows only) is producing good results, and there are several free programs around (but with their quirks).

Bart

Click to expand...

Wonderfull and incredible! that's OPF strength…!!!

When shooting yesterday, I had in mind some of your post here and told myself: "my guy shoot different focus point and we'll ask Bart about that stacking technique…"

And here you go! unbelievable ;-)!

I'll made a search for Mac sw… or if none, will load the Mac with Windoz!

Nicolas Claris said:

Wonderfull and incredible! that's OPF strength…!!!

When shooting yesterday, I had in mind some of your post here and told myself: "my guy shoot different focus point and we'll ask Bart about that stacking technique…"

And here you go! unbelievable ;-)!

I'll made a search for Mac sw… or if none, will load the Mac with Windoz!

Click to expand...

Here are the links:
HeliconFocus (also for MAC)
Tufuse Pro (Windows only)
And Photoshop CS4 ...

Bart

Thanks Bart

I'll give it a try soon, (I don't have the files with me for now…)

Nicolas

you' re right: unusual tasks require unusual methods.
Sometimes there's - apart from renderings - no option to get the desired perspective than having the lens in its proper place.

And yes, these "beeing-inside the space/model"-shots are much more difficult, than the aerial view of the external, aka "urban model" shots: moving the lens for 2 mm means in reality something like 2 meters .....

Sorry, no digitised version of the Perspectar in action. But I'll think about it, next time.
Here's another take with it, at the left is a stack of real, but thin glas, again the entire room was very small, maybe the glas stack was about 5 cm....





As for your examples of the Noah 76, they' re fine.
Actually, on the first image, I don't mind the fore- and background to be out of focus; it makes space more feelable, and emphasis the interior. To enhance the later, you could have cut of - when shooting or later as a crop - the upper deck.

The 2nd is just fine, having the foreground slightly unsharp, sliding into the sharp aerea and running at the end - infinity - into unsharp again.

As for focus stacking:

Having Photoacute, I used it rececently with success for focus stacking. Read the end of the linked thread, too.
There's a mac-version of Photoacute, as well.

Hi Michael
The images of Noah 76 shown here are part of the FF edited for final selection and for client's choice…
I agree with you for the suggested crop.
But these where just to show what kind of model I was talking… These shots where not done with stacking in mind (on the contrary of yesterday's shoot).
Though I'm terribly affraid that I haven't shot enough different focus points… we'll see!

Your last post is also impressive… what's the IQ with this lens, do you have enough resolution to get good "large" prints?

Nicolas

I see your intention with the Noah 76...

Perspectar: I wouldn' t use it for a billboard viewed in close distance....

meanwhile this lens has it merits, in my tests its far superior than the few samples of endoscope I tested, but due to diffraction - 64 is the f-stop used when exposing, it obviuosly hasn't as much resolution as a Zeiss-lens.

It has a nice bokeh, though... so you can - while accepting unsharp zones, quite go large. I should try to use it with focus-stacking.

You even can focus it, the nearest focus point - with the mirror on it - is 12 cm, so from 12 - 17 there's the sharp zone. Requires a lenshood, to avoid CA from lightsources glaring into it.

Still, there's no alternative, apart from rendering.

But I doubt you will be able to get a copy of it. A few only were build in the 80is or 90 is...

Michael Fontana said:

But I doubt you will be able to get a copy of it. A few only were build in the 80is or 90 is...

Click to expand...

I guess so! was curious;-)

Michael Fontana said:

There's a mac-version of Photoacute, as well.

Click to expand...

Yep! I just downloaded a trial version, but they don't have the right lens in their databass for 1DSMk3… :-{

Only:
EOS-1Ds Mark III
Canon EF 24-105mm f/4L (RAW / JPG)
Canon EF 70-200mm f/4L IS (RAW / JPG)

The 100 macro didn't had a profile either, so I had taken the profile for the 50 mm macro from the 1 Ds-2. Glad I screenshotted' em....

Just disable - in the prefs:
"Autodedect camera type from exif data"

and

"From Exif" in the Focal lenghts. Pull down that menu, and select Focal lenghts and f-stop manually.

Now load the images in the app, enable them in the checkbox, and highlight the midposition of the shots, as its the °pilot° - the anchor-image for the others.

Use these settings below:



As long as you don't correct for image distortion and CA, etc, it should work fine; only focus stacking will be done...

Read here, too

Wow! great help… thanks Michael!

huhuhu… crashing each time I click on the "Start" button (once images are uploaded and your settings validated…
I trashed pref and plist files, but still crashing! Will try tomorrow with office computer…