Focus shift on new 35/1.4 FLE

[MSUSpartan]

Advertisement (gone after registration)

I've had my suspicions because since I've started to shoot with smaller apertures in my street photography I've seen more and more of my shots out of focus with my M9. Yesterday was a sunny day here in NYC and I did a bunch of testing outside.

 

f/1.4 is spot on, but from f/2.8 on I can easily see the depth of field expanding farther back than forward. Is this something that can be fixed by Leica? I thought the whole point of this new lens was to eliminate this problem, but so far it's pretty disappointing for a $5,000 lens. I love the one lens solution of the summilux to shoot from dawn into the night, but this makes me consider going back to a summicron because the DOF markers on the lens can't be relied on.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Voigt]

Please clarify. In my experience, all lenses increase their depth-of-field more behind the focus plane than in front of it. Are you saying that the opposite is occurring?

[MSUSpartan]

Please clarify. In my experience, all lenses increase their depth-of-field more behind the focus plane than in front of it. Are you saying that the opposite is occurring?

 

Thanks, exactly what you mean. I guess it's normal then, but it's really annoying when at f5.6 the markers say something should be in focus from 2 to ~5.5 meters for example, and a subject at 2.2 meters might not be in focus.

[horosu]

I've had two samples of the this lens. Both exhibited significant focus-shift by backfocusing as the aperture is stopped down. I even went so far as to send the first of them to Leica for calibration !!.Received it back, same problem. I've since sold them both.

 

I found a nice example of the first-version double ASPHERICAL (at a pretty reasonable price), which doesn't (at least obviously) suffer from this problem.

 

When receiving the new FLE lens I took some (crappy) side-by-side shots with both lenses. I must say that the results were quite clear: the older lens has a center-sharpness which is way better than the new one (at the intended point of focus) and maintains it throughout the aperture range. On the other hand, the older lens has a slightly warm color rendition which is un-natural, to my eye.

 

Just my 2 cents.

 

Horea

[colorflow]

Thanks, exactly what you mean. I guess it's normal then, but it's really annoying when at f5.6 the markers say something should be in focus from 2 to ~5.5 meters for example, and a subject at 2.2 meters might not be in focus.

 

All lenses follow the 1/3-2/3 rule. The depth of field increases by 1/3 in front of and 2/3 in the back of the intended focal plane as you stop down the aperture. The dof markers on the lens indicate ranges that may be acceptable in general. What is acceptably sharp to some may not be acceptable to you. The focal point is still the distance that is in sharp focus. It is advisable to read 1 or 2 stops wider than the actual setting of the aperture if you want to use the dof markers. i.e. if you set the lens at f5.6, use the f4 or 2.8 reading of the dof scale.

[MSUSpartan]

Thanks for all the quick responses! You guys are great.

 

I didn't know it was normal behavior and didn't know about the 1/3:2/3 rule. Everyone talks about Leica precision I'll just learn how to live with it, because I do love the lens very much otherwise.

[pop]

Advertisement (gone after registration)

That's not how depth of field works.

 

Depth of field is a convention, not a technical property of a lens.

 

Any lens which is properly designed, built and adjusted will make a sharp picture of anything which is exactly at the distance the lens is focused at. Anything closer and farther will be less sharp than that. There's no way any lens designer can avoid this. That's how lenses work.

 

Several decades ago the industry decided how large the pictures taken with a 35mm camera usually will be and how people will look at them and what the acuity of their vision was. The industry then decided that for those people anything would look like a sharp point if it was depicted as a circle with a certain diameter or smaller.

 

From this, it should become clear that the manufacturer of the lens has no way to influence the depth of field for any lens. Again, that's how lenses work.

 

Using a digital camera and looking at the resulting pictures at a screen at large magnifications makes all that invalid. What would look sharp on a 8x10 print viewed at arm's length will look blurred if viewed at most magnifications on a computer display.

 

From what you wrote above, I infer that you focused your camera at 3m. You then expected anything between 2m and about 5.5m to be utterly sharp and are dismayed to find that something at 2.2 is not utterly sharp.

 

For one, it would be very hard to tell the sharpness of a point taken at a distance of 2m from one taken at 2.2m. Then, it is not possible to read the depth of field scale on the barrel of the lens with that accuracy. Lastly, no 35mm lens in the whole world will depict something a 2m sharp and at 1.9 unsharp when focused at 3m. Both will be "somewhat sharp" or "somewhat unsharp", depending on the viewer and the viewing distance.

 

If you're interested, I'm sure there are lots of sources in the internet which do a better job at explaining what the depth of field is about. I'd give the Wikipedia a try.

[01af]

All lenses follow the 1/3 - 2/3 rule.

The so-called "1/3 - 2/3 rule" is just a myth that doesn't really exist.

 

Still, when stopping down, the depth-of field will always grow towards the background more than towards the foreground.

 

 

Depth of field is a convention, not a technical property of a lens.

As a matter of fact, depth-of-field is both. You decide what exactly is supposed to be considered the borderline between sharp and not sharp, and from there it's all technical properties.

[mjh]

I didn't know it was normal behavior and didn't know about the 1/3:2/3 rule.

Then you should better forget you’ve ever heard about it since there is no such rule. Depth of field varies from 1:1 at the near limit to 1:infinity from the hyperfocal distance onwards.

 

Anyway, there is just one distance that is truly ‘in focus’. The depth of field is defined as the area in front and beyond that point where sharpness is still acceptable, according to a somewhat arbitrary standard of acceptability that isn’t shared universally. A subject near the limits of the depth of field would be acceptably sharp, if only marginally so, and if you should set a higher standard of sharpness (as many do) it would already be regarded as slightly blurred.

[colorflow]

The so-called "1/3 - 2/3 rule" is just a myth that doesn't really exist.

 

Still, when stopping down, the depth-of field will always grow towards the background more than towards the foreground.

 

Of course its not meant to be precise. The "myth" becomes reality when stated as " when stopping down, the depth-of field will always grow towards the background more than towards the foreground by about 2/3 to 1/3"

[zlatkob]

I've had two samples of the this lens. Both exhibited significant focus-shift by backfocusing as the aperture is stopped down. I even went so far as to send the first of them to Leica for calibration !!.Received it back, same problem. I've since sold them both.

 

I found a nice example of the first-version double ASPHERICAL (at a pretty reasonable price), which doesn't (at least obviously) suffer from this problem.

 

I am quite surprised to read this. I only have the older lens, and everything I've read so far seemed to confirm that the new FLE version is improved in this respect.

 

Of course its not meant to be precise. The "myth" becomes reality when stated as " when stopping down, the depth-of field will always grow towards the background more than towards the foreground by about 2/3 to 1/3"

 

That feels true from my experience. I don't understand 01af's comment that this is "just a myth". Is it a myth only in that it is not an exact representation of the phenomenon? If so, it is still a good guideline or rule of thumb.

[pop]

There's a calculator for the depth of field at Understanding Depth of Field in Photography.

 

It calculates for an M9 with a 35mm lens focused at 3m with an aperture of 1:5.6 a depth of field ranging from 2.029m to 5.289m. That's 0.971m in front of and 2.289m behind 3m. In this particular situation the distribution of the depth of field is not all that far from 1/3 to 2/3 if it's to be taken as a rule of thumb.

[horosu]

100% center crops

 

Summilux 35 FLE at 2.8

 

Summilux Double ASPH at 2.8

 

Summilux 35 FLE at 4

 

Summilux Double ASPH at 4

 

In my opinion the results speak for themselves. However, these were all hand-held shots so this is totally unscientific.

[horosu]

The scene:

 

 

 

If anyone wishes, I can also post the other samples at f1.4-f4 made with both lenses, so that the focus-shift is more apparent

[zlatkob]

The photos are not appearing.

[mjh]

I don't understand 01af's comment that this is "just a myth". Is it a myth only in that it is not an exact representation of the phenomenon? If so, it is still a good guideline or rule of thumb.

To qualify as a rule it should be universally true or at least for most cases. In reality this ‘rule’ is generally wrong; the 2/3 to 1/3 distribution is only found at a single distance (that also varies with the aperture and the focal length, so generally one wouldn’t even know where it is true).

 

The following graph shows the ratio of the depth of field beyond and before the chosen distance (for 50 mm and f2.8, so this is just an example). At the near limit the ratio is 1:1 which up to about 3 or 4 m continues to be a useful rule-of-thumb. At 10 m it reaches 2:1 (the infamous 2/3 to 1/3 ‘rule’), 3:1 at 15 m, and 5:1 at 20 m.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[horosu]

They do here (Google Chrome). Strange

[adan]

Just as a reminder, the focus/footage/meters scale on lenses is not linear. That is, the markings become more spaced apart at closer settings.

 

As example, on my 35 f/1.4, to change focus by two feet from 8 feet to 6 feet moves the focus scale 4.5mm; to change focus another two feet from 6 feet to 4 feet requires 9mm of movement.

 

This difference in spacing is built in to how lenses focus in and out, and has nothing to do with DoF directly - but it does roughly coincide with the 1/3-2/3 guideline. Which allows lens makers to engrave nice symmetrical DoF scales. Focused at 6 feet, one needs f/11 to get 4 feet "sharp" but only f/5.6 get 8 feet "sharp." So say the scales - so says the final print (for a given print size).

 

As Pico and other have alluded to, there is no such thing as "depth of field" UNLESS and UNTIL one also specifies how big the image will be printed or viewed (including viewing distance). What "looks" sharp in a 4 x 6 print or viewed at 16% on the computer screen from 2 feet away may be visibly NOT sharp in a 14" x 20" print or viewed at 100% on the computer from 2 feet away.

 

Unless one backs away from the big print or 100% view until the whole image appears as small as a 4 x 6.

______________

 

To check for actual focus shift, forget DoF. Focus on a 3D array of items at different but close distances (or a sloped ruler or one of those testing devices) and shoot shots at f/1.4, f/2, f/2.8, f/4, and f/5.6.

 

Then view those images really big. If the SHARPEST point changes as the aperture changes, then there is "Focus shift" present. If the point of maximum sharpness stays in the same place, there is no focus shift.

[Advertisement]

Hallo,

Look here: Leica M

[01af]

... when stopping down, the depth-of field will always grow towards the background more than towards the foreground ...

That's right.

 

 

... by about 2/3 to 1/3

That's wrong. And it won't get any better by repeating it.

 

 

If so, it is still a good guideline or rule of thumb.

It's like saying, "shooting 100 frames a day is a good rule of thumb" ... no, it's not. It's just nonsense. See mjh's diagram above.

[pop]

They do here (Google Chrome). Strange

 

They may on your computer. They do not on ours.