M8 Depth of field - reality v theory

[hollisterimages]

Advertisement (gone after registration)

I have just spent three days using the M8 exclusively on a landscape photoshoot. This has involved 4 lenses; Voightlander 15mm f4.5, Elmarit 28mm f2.8, Summicron 50 f2 and Elmar 90mm f4. Detailed examination of images each day has allowed me to explore the depth of field capabilities of the M8.

 

I am aware that hyperfocal distance focussing does not mean sharp focus across the hyperfocal range, but acceptable sharpness. On the Canon 5D I have found that using hyperfocal distance charts to give perfectly acceptable sharpness across the range.

 

However, starting off by using the 1-stop adjustment "proven" by the theory (resulting from the circle of confusion being .0225 ref: Barnack) simply did not give me acceptable results. Setting the infinity point at one f stop less than that the aperture being used resulted in distinct lack of focus on the horizon.

 

By experimentation I have found that acceptable front to back focus on landscape shots was only achieved when I made a 2.5 stop adjustment. i.e. for f16 aperture, set the infinity between f5.6 and f8. This results in images with the front to back sharpeness I have been used to when using hyperfocal distance charts on the Canon 5D.

 

This is clearly a bit restrictive in terms of depth of field (although it did not prevent me getting some images from the M8 that I am very pleased with!)

 

However, this implies a circle of confusion of 0.013, which is someway between the point and shoot cameras and the digital SLR's.

 

What is going on?? I have read somewhere about the Leica c of c being inaccurate and based on old data.

 

Has anyone any suggestions or thoughts about what is happening here?

[wattsy]

Has anyone any suggestions or thoughts about what is happening here?

 

Did you find the 2.5 stop 'rule' still applied when using an aperture of F8 or F11? I tend to avoid anything like F16 with Leica lenses because of the diffraction induced 'softness' but this might just be me adhering to the Leica mantra ("unlike the competition, Leica lenses are best at F2 to F4, blah blah").

[hollisterimages]

I have to say I have not tried this out at f8 or f11 but will do so when I am not risking the opportunities on a photoshoot.

 

however, I did not experience any problems with diffractions induced softness as you put it. Probably "mantra" as you suggest. The images from the m8 at f16 + are every bit as sharp as those from the Canon 5D using L series lenses, providing I use the 2.5 rule.

 

Your pictures are great. Did you have depth of field issues?

[tom0511]

Is it possible that slightly unsharp areas are more obvious because the M8 does not have an AA-filter? (vs the 5d which has an AA filter)

 

Are you saying with the 5d you could just take the numbers on the lens and they work for digital ? Strange-I allways though a bigger sensor would lead to less DOF.

[hollisterimages]

No, with the 5D I use the hyperfocal charts. With the Leica the hyperfocal charts based on the official circle of confusion result approximate to the one-stop rule.

 

Interesting point about the AA filters. Does that mean that although the detail is their in the M8 it appears less sharp as a result? Following on from that, is the direct comparison of unsharpened prints not fair to the Leica? Does the Leica image need a small amount fo sharpening before comparing sharpness with the unsharpened RAW 5D file?

[stunsworth]

My experience is that the M8 images are very sharp - I use less sharpenning than I ever did with my Canon 5D, but also that if you view at 100% the DOF markings don't give an accurate representation of what is captured.

[tom0511]

Advertisement (gone after registration)

No, with the 5D I use the hyperfocal charts. With the Leica the hyperfocal charts based on the official circle of confusion result approximate to the one-stop rule.

 

Interesting point about the AA filters. Does that mean that although the detail is their in the M8 it appears less sharp as a result? Following on from that, is the direct comparison of unsharpened prints not fair to the Leica? Does the Leica image need a small amount fo sharpening before comparing sharpness with the unsharpened RAW 5D file?

 

No, I meant the other way-since the M8 is very sharp its easier to detect areas wich are not totally sharp, or the other way around-an AA-filter might "smoothen" the sharp-unsharp transition. This is just an idea, I dont know if it is true.

[sean_reid]

Has anyone any suggestions or thoughts about what is happening here?

 

What is going on is that depth of field was in the beginning, is now and ever shall be, a relative concept. The various tables and formulas are of limited usefulness because there are too many variables that they cannot account for. The only way for one to really find out what the depth of field will be for a specific camera, lens, focus distance, aperture, etc. is to do just what you did - experiment first hand, make variations, observe, (preferably, print at the desired size) etc. That process, which you've started, will yield far more useful information that 100 pages of charts and formulas. Trust your first-hand experience.

 

Cheers,

 

Sean

[sean_reid]

Does the Leica image need a small amount fo sharpening before comparing sharpness with the unsharpened RAW 5D file?

 

No, the opposite is true.

 

Cheers,

 

Sean

[sean_reid]

No, I meant the other way-since the M8 is very sharp its easier to detect areas wich are not totally sharp, or the other way around-an AA-filter might "smoothen" the sharp-unsharp transition. This is just an idea, I dont know if it is true.

 

That's true in general, the M8 is relentless about accurate focus.

 

Cheers,

 

Sean

[Woody Campbell]

Adrian - thanks for taking such a careful look at this. There were a number of threads on this back in November - December. There was (and is) a lot of confusion on the subject of circles of confusion.

 

There are two factors at work here. First, the small sensor requires a smaller circle of confusion because you are magnifiying more to reach a given print size (for example 11x to print on ANSI letter, as opposed to 8x from 35mm). This results in at least one stop of the phenomenon that you describe. The small sensor has the same impact on the rule of thumb for minimum acceptable shutter speed. The second factor is that the M8 sensor is higher resolution than film - as you view images at 100% on screen or make A3 you need an even smaller circle of confusion. Add another stop.

 

I use a two stop correction from the depth of field numbers printed on the lens barrels.

 

Leica doesn't help on this - in the manual they say that you can use the depth of field scales as is, which is simply wrong.

[woodda]

Very interesting article. Are we saying that if I use f2 is will not get as depth of field as is indicated on the lense. IE no depth of field in the case of the 75f2.

 

Another example is if I want the depth of field I recieved with film at say f5.6 on a 35mm lense will I need to use F11

 

Thanks in advance

[Woody Campbell]

Very interesting article. . . .

Another example is if I want the depth of field I recieved with film at say f5.6 on a 35mm lense will I need to use F11

....

 

Yes. You could use f8 to get comparable letter-sized prints but since you are inevitably printing and viewing larger you would use two stops.

[woodda]

Thank you for the reply. I am probably slightly non technical but you refer to letter sized prints and 2 stops. I am not sure I understand the connection as I was thinking we were talking about depth of field . If would greatly appreciate if you elaborate on the letter and 2 f stops connection (was this F11 )

 

Thnks if this is a stupid question

[tom0511]

Interesting, but if this was the only reason small sensor cameras would show small DOF, but they show a lot of DOF. Why?

I allways thought: smaller sensor means using a shoter lens in order to get the same frame fromthe same distance, and this would lead to more DOF.

 

It seems there are really several effects which influence the DOF of digital cameras.

 

Adrian - thanks for taking such a careful look at this. There were a number of threads on this back in November - December. There was (and is) a lot of confusion on the subject of circles of confusion.

 

There are two factors at work here. First, the small sensor requires a smaller circle of confusion because you are magnifiying more to reach a given print size (for example 11x to print on ANSI letter, as opposed to 8x from 35mm). This results in at least one stop of the phenomenon that you describe. The small sensor has the same impact on the rule of thumb for minimum acceptable shutter speed. The second factor is that the M8 sensor is higher resolution than film - as you view images at 100% on screen or make A3 you need an even smaller circle of confusion. Add another stop.

 

I use a two stop correction from the depth of field numbers printed on the lens barrels.

 

Leica doesn't help on this - in the manual they say that you can use the depth of field scales as is, which is simply wrong.

[ho_co]

Adrian--

In addition to Woody's two points, remember that the Leica DoF engravings come from the earliest cameras.

 

As film and lenses improved, the rule became for most people: Use the engravings for the aperture next larger than the actual one. That is, if calculating for hyperfocal distance and shooting at f/8, use the engraved scale for f/5.6 for best results.

 

Now with a smaller sensor and increased resolution, as Woody said, we have reason to increase the leeway, just as you've shown.

 

Thanks for posting your results here! You've given us theoreticians a practical pointer, and--as Sean said in other words--since depth of field doesn't really exist except in our perception, and is completely dependent on many factors--you've done right in finding out for yourself!

 

--HC

[Jamie Roberts]

Very interesting.

 

In my own tests, admittedly with lenses wider than 50mm, I find the M8 has more DOF for a given focal length than my 5d or 1ds2. For a given FOV, it has much more DOF, since it's using consistently wider lenses.

 

This is pretty much what I expected, since the FF cameras have less DOF than the cropped bodies.

 

Whether or not the DOF lookup charts are in order is another question, but I've found the 1-stop rule to be just about right on the M8.

 

So here's the predictable question: could a slight mis-alingment of lenses / RF / infinity focus account for this being off a bit? I think it could...

 

FWIW, I just spent a weekend with the M8 and the CV 15, 21 and 35Lux photographing people in the Rockies (so a lot of DOF necessary), and they were all right on the money (except when I inadvertently moved the CV 15 and forgot to hyperfocus it back!).

[steve_l]

"Small sensors have larger perceived DOF" comes from thinking in terms of the effective focal length, not the actual length - e.g., on the Digilux 2 we compared the 28mm FOV to a 28mm (actual) lens, due to equal FOV, even though it actually is a 7mm lens, with the DOF that any 7mm lens would get......

 

I've heard two other points of interest:

1. The transitions from in- to out of focus occur more suddenly in digital sensor-based images (vs. film, which is thicker), allowing you to see more clearly where things start to get some blur. thus requiring extra aperture to compensate.

 

2. This one I've been trying to verify, but haven't been able to - I think it came from an Erwin article - that Leica lenses in particular, different from other makers, won't actually show sharp focus at infinity unless the focus is set at infinity.

[charlesphoto99]

I have to say I have not tried this out at f8 or f11 but will do so when I am not risking the opportunities on a photoshoot.

 

however, I did not experience any problems with diffractions induced softness as you put it. Probably "mantra" as you suggest. The images from the m8 at f16 + are every bit as sharp as those from the Canon 5D using L series lenses, providing I use the 2.5 rule.

 

Your pictures are great. Did you have depth of field issues?

 

I would not risk f16 on a photoshoot! It's been said many times here and elswhere that the CV 15 is at its sweetest around f6.3 - f8. True for most lenses (though some Leica glass we've been finding the sweet spot is wide open!). Why do you feel you need to use f16 to shoot something at infinity? And if it's at infinity why not just set the lens at infinity?

 

The M8 is great compared to the earlier M's in the ability to shoot with higher shutter speeds thereby being able to use the f-stop that's best.

 

Best of luck next time.

[Advertisement]

Hallo,

Look here: Leica M

[lars_bergquist]

DoF—theory and reality.

 

Way back in the 1920's, when Leica lenses got their first d.o.f. scales, the 35 mm format was a newfangled one, regarded with suspicion by respectable photographers. What was the competition? Normally one of those folding cameras for 6x9 cm roll film. All these negatives were usually printed contact. Yes, I remember when people were as satisfied with a 6x9 cm contact print, as most people are with a 10x15 cm enlargement today! – Enlargers were extremely primitive. Leitz made some of the first really useable models for the home darkroom.

 

Now Barnack and Berek figured that in order to compete, a 35 mm neg should be capable of being enlarged about 3 x linear, i.e. to around 7x11 cm, and still look as sharp as a contact print. At normal reading distance, the human eye was supposed to resolve about .1 mm, i.e. distinguish two lines or edges 1/10th of a millimeter apart. With 3 x enlargement, this meant that the negative should resolve 1/30 or .33 mm. So this was the number that the d.o.f. tables and scales were founded on. At the plane of best focus, an Elmar lens resolved much better than that, but anything in front or behind that plane that still was resolved at the 1/30 mm standard, would still look acceptably sharp. And this became not only the Leica but the industry standard.

 

Already ten years later bold souls made much larger enlargements, so that the standard really was obsolete, but by then the sales department did not dare rock the boat. But fastidious photographers began using a one stop correction, so that for instance if they would use f:12.5 they read d.o.f. at f:9 instead. This was a stricter criterion.

 

The sensor of the M8 measures 18x27 mm, so that it has to be 'enlarged' even more in order to produce a specified print size. Many workers had already begun to use a new criterion, that of a two stop correction: when using f:11, read at 5.6! This means that acceptable resolution must be 1/60th of a mm, or c. .017 mm. So this is the standard I and many others are using nowadays. It is in practice sufficient for an 16x24 cm print. It is not sufficient for a 100% pixel peep – nothing is! Keep this in mind: use practically relevant criteria of judgment.

 

All this means of course that a calculated use of d.o.f is possible only with wide angle lenses – and with the M8, this means 28 mm or shorter! For the rest, use point focusing. I hope this will clear up some of the apparent confusion.

 

The old man from the Age of the Roll Film Contact Print