M8 unsharp at high speed ?

[t024484]

Advertisement (gone after registration)

I noticed in the past that pictures taken with shutter times above 1/2000 were not as sharp as I expected them to be, but I thought that I was in someway the cause of that and took no further notice of it.

 

This week I received my new Summilux 50mm, and because of bad experiences with a 28mm Cron (see also in this forum) and a bad (but succesfuly adjusted) 75 mm Cron, I took nothing for granted and started making test pictures.

 

Without having in mind to test the high speed shutter effects, I stumbled by coincidence across this effect again, but now in a very controlled situation.

I made pictures of a high building 400 meters away, during bright sunlight, starting at f 2,8 - 1/6000 going all the way to f 8- 1/750.

As from f 8 the pictures had their final sharpness, but with larger aperture they were gradually unsharper.

Pictures were taken with my M8 + Lux 50mm, from a tripod, with the self release timer and without any wind moving around the camera.

 

The crops below are 100% crops from the centre, taken in DNG, converted with LR, without any processing.

 

At first sight one would tend to think that the focus setting was wrong, but this is 100% not the case.

To check this, from the same position, with everything else exactly the same, I took pictures later that day with a dark clouded sky.

Now with a f 2.8 - 1/250 setting, the building, and especially the sun blinds were just as (razor)sharp as with f 8 - 1/30, and fully comparable in sharpness to the crop below with f 8 - 1/750. This proved both that the focus was correct, and that the lens behaved perfectly at different apertures.

 

Thinking about the possible cause I did some calculations.

Estimating that the shutter travels the vertical dimension of 18 mm of the sensor in 1/8 sec from bottom to top, it is easy to see that you need a 18 micrometer gap for an exposure of 1/8000 sec.

Since a pixel is 6,8 micometer, the gap at 1/8000 is only 2,6 times a pixel size !

Could it be that diffraction with such a small gap renders the picture unsharp ?

 

 

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!

 

M8 Summilux 50mm f 8 1/750

 

 

M8 Summilux 50mm f 2.8 1/6000

[Shootist]

What I think you are seeing is focus shift from different apertures.

Normally any lens will increase the DOF backward more then forwards when stopping it down. So in your shot @ f/8, and a lower shutter speed, the DOF covers what the RF is telling you is IN FOCUS.

When you open the lens up, increasing the shutter speed, the DOF get shallower and moves the true point of focus forward.

So what I think you are seeing is the lens moving the true point of focus in front of where the RF is telling you it is in focus.

The only way to check this is to do a test where you set the camera on a tripod and focus at something with the lens wide open. Take a shot and without moving the focus ring stop the lens down one stop at a time and take another shot at each f/stop.

 

I really don't think it has anythink to do with shutter speed. It does have something to do with the aperture used and the focus shift for that aperture.

[t024484]

What I think you are seeing is focus shift from different apertures.

Normally any lens will increase the DOF backward more then forwards when stopping it down. So in your shot @ f/8, and a lower shutter speed, the DOF covers what the RF is telling you is IN FOCUS.

When you open the lens up, increasing the shutter speed, the DOF get shallower and moves the true point of focus forward.

So what I think you are seeing is the lens moving the true point of focus in front of where the RF is telling you it is in focus.

The only way to check this is to do a test where you set the camera on a tripod and focus at something with the lens wide open. Take a shot and without moving the focus ring stop the lens down one stop at a time and take another shot at each f/stop.

 

I really don't think it has anythink to do with shutter speed. It does have something to do with the aperture used and the focus shift for that aperture.

Look again to what I said.

I made the same pictures later that day with with a dark sky and therefore longer shuttertimes.

In that case all pictures were just as sharp. The only thing that differed in both series of pictures was the shuttertime. So a very controlled way to compare.

 

As a matter of fact, since the building was 400m away, the focus was at infinity all the time, easy to control and to reproduce.

[M6J]

Why I have a bad feeling about this issue?

[Shootist]

Look again to what I said.

I made the same pictures later that day with with a dark sky and therefore longer shuttertimes.

In that case all pictures were just as sharp. The only thing that differed in both series of pictures was the shuttertime. So a very controlled way to compare.

 

As a matter of fact, since the building was 400m away, the focus was at infinity all the time, easy to control and to reproduce.

 

Not really (Bold and Underlined). You moved the camera and tripod and you focused at two different times. That is NOT a controlled test.

In any event without a controlled test of the lens @ different aperture your results are not conclusive. I certainly don't see this with either M8 body I have.

I think your lens is front focusing at wide apertures.

Try some focus bracketing at the f/1.4 and see what you get. Also any lens will be more sharp as the aperture gets smaller, to a point.

 

 

Look I'm not tryinig to start a fight or tell you that what you are seeing isn't true. But I think it is for a different reason then what you think it is.

 

To Add 400m is NOT infinty. Infinity is the moon or a star.

[pascal_meheut]

I just did a quick test and shot the building in front of time at 1/1500 sec and 160 ISO, then 1/6000 and 640 ISO, using the 75/2.

 

Here is the comparison, 100% crop under Capture One Beta 4. No difference and as sharp as it gets handheld.

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!

[Shootist]

Advertisement (gone after registration)

Why I have a bad feeling about this issue?

 

I don't.

[t024484]

I don't.

 

The last thing I tried was to enter a fight, maybe my answer sounded a bit brisk without any intention.

In fact the tripod did not move at all, it was there all the time. So I think it was rather controlled.

I made some pictures this afternoon at 1 meter distance in full sunlight at 1/8000, and shielded the sun after that with an umbrella, and shot again with 1/500.

There was no difference in sharpnes this time, so I am still puzzled.

I will do some further investigations.

[M6J]

I just did a quick test and shot the building in front of time at 1/1500 sec and 160 ISO, then 1/6000 and 640 ISO, using the 75/2.

 

Here is the comparison, 100% crop under Capture One Beta 4. No difference and as sharp as it gets handheld.

 

BTW Pascal, have you noticed the artifacts in the sky next to the antenna in the ISO 160 picture?

[pascal_meheut]

BTW Pascal, have you noticed the artifacts in the sky next to the antenna in the ISO 160 picture?

 

Sure I did. Not the first time I had this problem with C1. Strangely enough, it occurs on Beta 4 and not 3.7.7 for some pictures, and the contrary with others.

 

P.S: I reported the problem to PhaseOne.

[stunsworth]

Strange that they both give the same effect but with different images. Have you tried ACR to see if it occurs with that?

[pascal_meheut]

Have you tried ACR to see if it occurs with that?

 

Never saw such problems with ACR. And especially not with this image. But I still prefer C1.

[sean_reid]

You're seeing something but I've shot a lot at 1/4000 with no problems.

 

Cheers,

 

Sean

[charlesphoto99]

I think what you are seeing is the focus shift, covered exhaustively on this forum , of certain asph (esp) lenses to shift focus between f2.8 and f5.6. That shift is then usually covered by depth of field by f8. I've never heard of a camera getting unsharper with higher shutter speeds.

[t024484]

Thank you all for the responses.

 

Personally I find a focus shift at 400 meter distance hard to believe because of two reasons,

First because I made those same pictures with less light and with the same apertures and infinity setting later that day with lower shutter times, and I had no difference in sharpness between f 2.8 amd f 8

Second because Leica specifies for the Summilux 50mm that the DOF at setting infinty and f 2.8 is from 28.82 meter to infinity.

So 400 meter is well within this range and seems far from critical.

 

On the other hand, my feeling that a very short shutter time made pictures unsharp, was proven to be wrong with the pictures I made this afternoon.

So if it is not the shuttertime that is making the pictures unsharp, it could be something like flare within the lense.

Maybe the strong reflection on the building itseld could be the cause of that ?

[colorflow]

Thank you all for the responses.

 

Personally I find a focus shift at 400 meter distance hard to believe because of two reasons,

First because I made those same pictures with less light and with the same apertures and infinity setting later that day with lower shutter times, and I had no difference in sharpness between f 2.8 amd f 8

Second because Leica specifies for the Summilux 50mm that the DOF at setting infinty and f 2.8 is from 28.82 meter to infinity.

So 400 meter is well within this range and seems far from critical.

 

On the other hand, my feeling that a very short shutter time made pictures unsharp, was proven to be wrong with the pictures I made this afternoon.

So if it is not the shuttertime that is making the pictures unsharp, it could be something like flare within the lense.

Maybe the strong reflection on the building itseld could be the cause of that ?

 

Very interesting. Have you tried f2.8 at different speeds by varying the ISO without waiting or changing the setup or readjusting focus?

 

Alan

[ndjambrose]

Thank you all for the responses.

 

Personally I find a focus shift at 400 meter distance hard to believe because of two reasons,

First because I made those same pictures with less light and with the same apertures and infinity setting later that day with lower shutter times, and I had no difference in sharpness between f 2.8 amd f 8

Second because Leica specifies for the Summilux 50mm that the DOF at setting infinty and f 2.8 is from 28.82 meter to infinity.

So 400 meter is well within this range and seems far from critical.

 

On the other hand, my feeling that a very short shutter time made pictures unsharp, was proven to be wrong with the pictures I made this afternoon.

So if it is not the shuttertime that is making the pictures unsharp, it could be something like flare within the lense.

Maybe the strong reflection on the building itseld could be the cause of that ?

 

I'm unsure why you think it's anything other than focus shift? Plenty of people here saying that they don't have the problem. ( I agree with them, BTW - I tend to shoot wide open a lot of the time and have never noticed any form of unsharpness in the higher shutter speed ranges). Plus, there's already a lot of documentation on the net about fous shift, and even Leica has agreed that the Summilux lenses are particularly susceptible to the problem. (See article in LFI about 3 months ago where there is an article that documents exactly what you're describing and explains how and why it occurs).

 

As the saying goes, if you see hoof prints don't look for zebras....

[orjanf]

I agree that shutter speed would make no difference but I doubt that focus shift is the problem.

Focus shift when stopping down is usually away from the camera and in my experience not noticeable at the infinity setting.

 

Perhaps your lens has a problem and does not focus at infinity when set to this setting?

 

In that case the increased DOF at f8 would account for the sharper image.

 

Do you notice a problem at shorter distances?

 

Incidentally, when pixel peeping at 100% even objects well within the DOF look soft, the limits are a bit too liberal for this day and age.

[mjh]

Estimating that the shutter travels the vertical dimension of 18 mm of the sensor in 1/8 sec from bottom to top, it is easy to see that you need a 18 micrometer gap for an exposure of 1/8000 sec.

How did you arrive at the 1/8 sec figure?

[Advertisement]

Hallo,

Look here: Leica M

[t024484]

How did you arrive at the 1/8 sec figure?

 

That's a very valid question, that also kept me busy afterwards

 

With no lens on the camera, I looked at the shutter when firing at different speeds.

Although quite difficult to see, I tried to find the shuttertime where, when the movement of one of the two shutter halves was at its end, the other halve started moving.

At this shuttertime, the shutter can be seen as 18 mm gap moving along the sensor with the shutterspeed just found.

At a shuttertime below 1/8, the whole sensor showed for some time with no shutter movement at all, so this was too slow.

At 1/8 sec it looked as if this was the shuttertime, where when one shutter halve came to a halt, the other halve stated moving.

I used this value of 1/8 to do my calculations.

 

Because I was anxious to see if I could find a more accurate way to determine the speed of the shutterblades, I photographed the Leica M8 shutter with my Canon 1D, set to 10 fps 1/160 sec.

Here I found that the blades were moving quite a bit faster than my initial guess, because with the M8 set at 1/100 sec, I could still see the complete sensor, meaning that the blades are moving at least at 1/100 sec or even faster.

Here I stopped, because it was extremely difficult to get results at faster shutter times.

 

Making my previous calculation again with 1/100 sec instead of 1/8 sec, gives already a gap width of 32 pixels at 1/8000 sec, but I realize that the blade speed can still be significantly higher, and because of that, the gap width will be wider accordingly.

 

So all in all, my feeling that the small gapwidth could be the cause of diffraction unsharpness, is hereby proven to be not the case, just because the gap is quite a bit larger than I thought.

 

If you happen to have more accurate information, I would be very interested to hear.