focus/resolution tests

[rvaubel]

Advertisement (gone after registration)

another attempt, this time at f8, 50 lux, dng, processed in C1pro, noise red off, slight sharpening.

 

this was to determine if the better dof at f8 would crisp up the focus

 

looking at the properties, my 400 x 236 image is 144k

Carsten's 800x536 is 126k

????

 

Your getting there. With all the talk about this vs that, the ultimate test should be "does this look sharp to my eyes". I check all my lens to make sure they pass this "taste test". Its not too rigorous but if it looks good, its good enough for me. It is surprising how often a lens that seemed fine in the old days of film can't pass this teat in digital. But we did not have the capability of making a 16X20 print in two minutes either. The level of criticism at 100% raises the bar for all lenses.

 

Rex

[Advertisement]

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

[carstenw]

Hmm, I don't know C1LE well enough to know what is going on with the crops. In fact the previous image looks good in the full view, but the crops to me look over-sharpened. There is a slight high-contrast halo around the leaves and a disconnected kind of look to many details, like the holes in the pole of the sign. As long as that remains, the sharpening needs to be backed off, until the details get a nice, smooth yet sharp look to them, and the edges of the leaves and other details blend smoothly into their background, instead of having these slight halos. Do you see what I mean? It might be easiest to see somewhere along the midpoint of the left side of the pole, where there is a white-ish area just to the left of the pole, along the edge. That needs to disappear.

 

IMO, over-sharpening is one of the least attractive artifacts of digital photography, and I am normally very careful to avoid it, preferring a very slight softness if I can't get it just right.

 

Here is the same crop without and with sharpening in C1LE. The sharpened one used Amount: 10, Threshold: 2, which seems roughly right with my eyes. The third one is over-sharpened, with the standard settings of Amount: 25, Threshold: 3. To my eyes the detail is starting to break apart and there are some halos and a little stair stepping. The difference is subtle. Perhaps it is easiest to see it in the bottom right corner, on the blades of grass. The M8 images are already very sharp, and need less sharpening than most cameras, especially Canons.

 

Maybe I should even back the sharpening off a little from picture 2. It is borderline.

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!

[rvaubel]

Carsten

Your samples show that it is often(almost always) necessary to sharpen a M8 RAW photo, at least a little. I agree that the M8 needs less than most cameras as their is no AA filter. I also agree that oversharpening is more offensive than a slightly soft image. That being said it is hard for me to see much difference between the 2nd correctly sharpened image and the 3rd oversharpened image, at least on my monitor.

 

Rex

[carstenw]

I agree that most of the details look very similar, the second one just has a slightly harsher look, and a few edges here and there, some of the big leaves and some of the grass blades, and some of the pale stones in the back, have slight halos. I could be imagining it

 

These images also show how sharp Leica lenses are. It is perfectly possible to count the leaves, blades of grass and pebbles, from the fourth floor, in a photo taken with a 50mm lens. That is pretty incredible when you think about it.

[jlm]

carsten:

how do you post an 800 x 537 image? the attachment manager bumps me out if i go over 400 x 268. And your file sizes are smaller!

[carstenw]

There is a size (KB) limit as well as the dimension limit. I think it might be 220KB or 250KB or something similar. To export from C1LE, I use "JPEG Medium", Quality "High". I set Scale "To Width" and type in 800 and choose "px". sRGB profile. That should do it.

[Guest sirvine]

Advertisement (gone after registration)

I am a firm believer that any luminance noise reduction and sharpening detract from M8 images, esp. on the pixel-peeping scale. M8 images are super sharp right out of the camera.

[jlm]

a bit more: this was processed in C1 with no sharpening, no noise reduction and output as a medium quality 800 x 536 jpg.

 

it got bumped with this message:

Your file of 349.0 KB bytes exceeds the forum's limit of 244.1 KB for this filetype.

 

i opened it in CS2, did nothing but save it with a new name, jpg quality medium, and the corresponding re-compression sized it down to 104kb.

 

that I will load, but from now on, the processed images will be tiff.

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!

[jlm]

this all came about when I started doing focus tests at about six feet for another person's lenses; that led me to question my own lenses. i was able to improve on my 75cron at six feet by tweaking the roller, but that led to where this post started. i now have the roller back to where the 150yard test is in focus, but I'm afraid to look hard at the six foot test again. bugger.

[rvaubel]

this all came about when I started doing focus tests at about six feet for another person's lenses; that led me to question my own lenses. i was able to improve on my 75cron at six feet by tweaking the roller, but that led to where this post started. i now have the roller back to where the 150yard test is in focus, but I'm afraid to look hard at the six foot test again. bugger.

 

You have hit the nail on the proverbial head with your concern at the six foot test. Adjusting that roller thingy makes a global adpustment to the whole focal range. There are three major focus adjustments available on the m8 and the roller is just the most accessible. I am a avid tinkerer but, until I understand the ins and outs of the adjustment proceeder, I would leave well enough alone. Don Goldburg spent some time with me explaining the whole procedure back in my RD1 days and it is not easy or intuitive. For now I just check to see if everything is hunky dory and send it to him if it needs adjustment.

By the way, just establishing a foolproof method of testing the lenses is a major chore in itself. That's why I stopped trying to be rigoreous about it and just settled for causal test at infinity and 3 or 4 ft. Someday I would like to do something better but that is goimg to take a lot of time and effort.

 

Rex

[jlm]

amazing. set a target at 8 feet: some printed text at 12point. had to tweak the roller and make about ten shots to get it right. both the 50lux at 1.4 and the 75 apocron at f2 were able to focus perfect. even the quotation marks were distinct.

did this by setting the focus on a vertical line on the page, taking successive shots by not changing the focus, but moving the tripod back, chimping with max zoom, until I found the shot in focus. then tweaked the roller and repeated.

 

Of course, now when I go over to the window and shoot the parking sign...it looks like the very first image I posted.

 

so the question is, can this camera/rangefinder be adjusted so near and far will both be in focus, of do I have to follow Guy's method of setting the roller for excellent near focus at max aperture and to hell with far, assuming you never shoot landscapes wide open?

[London Lad]

I think what we need to understand is that the roller adjustment will have the same influence on the arm / lens cam relationship at every point throughout the focusing range.

 

So if your focus is correct at say 3 meters and out at infinity, then adjusting the cam will always have an effect on both (and every where in between) and never fix the problem.

 

Now the arm pivot adjustment effectively lengthens the arm and (I think Mark mentioned) changes the 'gain' of the system. I doubt that this change would be linear and suspect that this is where the ability to have an effect at one distance and less at another lies.

 

For effective adjustment both of these would always need to be adjusted together.

 

It has been mentioned that the roller adjustment is for infinity focus. This is not true, it may be that you start by adjusting this for infinity but it will always effect the whole range.

 

I strongly suspect that getting these spot on at all ranges without the use of the correct alignment equipment would be nothing more than very good luck!

[jlm]

i agree with that, although several have posted that tweaking the roller for infiinty brought back focus over the full range.

 

I figure i can always get the roller back so infinity focus will return and that should bring the camera close to its original state. I'm not too keen on going benind the red dot to fiddle sith bnear focus. What pisses me is the cost/perfomance ratio for what is an absolutely essential function.

 

the question remains, can this camera be adjusted to attain both near and far focus?

[cme4brain]

I have been going cross-eyed trying to do some focus test with the M8 and a few lenses, trying for sharp, crisp, what ever you call it, to judge backfocus. this has led to tweaking the roller a bit to get the image to match what the rangefinder is set to.

And I am not happy with what i am seeing at long distance about 150 yards in these shots, shot as jpg fine

 

lens: new 75 apo cron, shot at f4 on tripod

attached are:

the full frame and a crop around the no parking sign

I also did focus bracketting, tsking a series of shots with the rangefinder set to focus a bit closer, and a bit farther, both were worse, indicating my rangefinder is on target

 

I am no expert on sizing for the forum, so put them in photoshop:

for the full frame, reduced file size to 400 x whatever to get it small enough

for the crop, set the view to actual pixels, then cropped.

 

please enlighten me for a better way to deiplay these images.

 

I then took the same shot with my 5d and 24/105, f4 framing it about the same for a reality check. the 5d is quite a bit better.

 

pleae advise; I am about to shoot the same with my 50 lux

 

i see it sort of worked: (the 5d shots show the pavement cooler grey),

 

 

It is tough to compare the 5D and the M8. The chip in the 5D is better at ISO's greater than 800, but the M8 has access to better lenses. See Sean Reid's articles for just this type of comparison. Each camera has its strengths and weaknesses.

[Tony C.]

While I have far less experience and technical expertise than many other contributors, I must say that, after initially adjusting the focus to infinity, I have had no problems yet with respect to M8 focussing or sharpness. I have been using several lenses, and, for example, have found the 35mm Summilux and 75mm Summicron to be superb in both respects.

 

I've attached an original size jpg and (screen-grab) crop taken with the 75 'cron. It would, in my view, be hard to ask for a crop to much more sharp than this.

 

Regards,

 

Tony C.

 

 

 

[carstenw]

John, behind the red dot is the vertical offset. To set the near-far focus balance, you need to loosen the screw at the pivot-end of the arm, and then gently move the little washer with the notch which it holds. Then you need to adjust the wheel for infinity again, and then check near focus, repeating until done. Leica has a jig for this. I did do mine personally, but it is rather fiddly, so I am not sure I recommend it.

[jlm]

sorry to be rude, for the third time I could give a s about comparing the M8 and the 5d. I used another digital camera i had at hand for a point of reference. my digilux 2 was at home and the only other choice was my cell phone.

 

I suggest anyone interested try this test so we can gather some sort of larger response. Print out some times roman text at 12pt. a paragraph or two and some quotation marks, (times roman has serifs, good for testing) tape it to the wall, camera on a tripod 8 feet away and take a few shots setting the rangefinder focus on the text chimp in the viewer with the magnifier to check actual focus. move the tripod back and inch and re-shoot, withoout changing focus. re-chimp and repeat, until you zero in on the right distance. if the image got worse, you either had it right or you are front focusing.

use a longer lens, 50 or 75 and shoot wide open for the toughest test.

 

then shoot something the size of a parking sign or license plate 150 yards away. check that focus

 

I want/expect a camera to get both right

[rvaubel]

i agree with that, although several have posted that tweaking the roller for infiinty brought back focus over the full range.

 

I figure i can always get the roller back so infinity focus will return and that should bring the camera close to its original state. I'm not too keen on going benind the red dot to fiddle sith bnear focus. What pisses me is the cost/perfomance ratio for what is an absolutely essential function.

 

the question remains, can this camera be adjusted to attain both near and far focus?

 

Yes. But you are probably not going to do it yourself. Without the proper tools and training, it will be nothing but shear luck to get it perfect. About the only thing you cvan do yourself, is what you discovered. Pick one distance or another and adjust the roller for that exact distance. Assuming all the other adjustments are close, this is the best compromise you can do yourself. And that is just for that one lens...I hope you picked one that wasn't slightly wonky to begin with.

 

If you want to do it right you have to start with confirming that the flange to sensor distance is within manufacters tolerance. If not then the sensor needs to be shimed to meet, or preferrable, exceed that tolerance. That is just the beginning.

 

I aim to be able to do the whole schmel someday. But there are really only a handful of independent technicians that can do it now. I actually would trust one of them rather than Leica itself because they are not subject to Monday morning blahs and other afflictions that plague the world of paid employees, German or not.

 

Anyway, carry on. The only way to find all the pitfalls is to do it yourself. Plus, You have to know how to test your own lenses in order to cull out the bad ones and get them (or your camera!) calibrated. Once I was pretty well satisfied that my camera was OK and my previous believed best lenses to be OK also, I test new lenses informally by taking a lot of pictures and feviewing them at 100%. It doesn't take that long to be satisfied that the lens is OK or that it needs to go off to DAG or Sherry, or I have a local guy I test. I would never send a lens to Sohm for fear I wouldn't see it for months. And anyway most of my lenses aren't even Leica

 

Rex

[Advertisement]

Hallo,

Look here: Leica M

[London Lad]

i agree with that, although several have posted that tweaking the roller for infiinty brought back focus over the full range.

 

I figure i can always get the roller back so infinity focus will return and that should bring the camera close to its original state. I'm not too keen on going benind the red dot to fiddle sith bnear focus. What pisses me is the cost/perfomance ratio for what is an absolutely essential function.

 

the question remains, can this camera be adjusted to attain both near and far focus?

 

Jim, it can.

 

My first M8 back focused with all my lenses at close and medium range but was pretty good at infinity. I used 3 tests, focus chart at 45 degrees, yard stick horizontal on wall and a sea horizon. I tested all my lens (only 5) with all 3 tests at a range of apertures and bracketed. The camera was undoubtedly out of adjustment or faulty.

 

My dealer exchanged the camera and the replacement focuses perfectly at all ranges with all lens (both new and old) it's not 'near' or 'acceptable' it's spot on, so it can be done!

 

Unfortunately the new camera has a fault with the L.E.D. display and is going to be exchanged on Wednesday. Needless to say I will be doing a few tests at the dealers !!!

[carstenw]

I suggest anyone interested try this test so we can gather some sort of larger response. Print out some times roman text at 12pt. a paragraph or two and some quotation marks, (times roman has serifs, good for testing) tape it to the wall, camera on a tripod 8 feet away and take a few shots setting the rangefinder focus on the text chimp in the viewer with the magnifier to check actual focus. move the tripod back and inch and re-shoot, withoout changing focus. re-chimp and repeat, until you zero in on the right distance. if the image got worse, you either had it right or you are front focusing.

use a longer lens, 50 or 75 and shoot wide open for the toughest test.

 

John, using text is not such a great test, at least if you shoot at the edge of the resolution. A lens test chart is better, because those thinning, converging lines allow you to see where exactly the resolution runs out. With text, the camera either has enough resolution to show it well, or it doesn't. Another photographer on another board used text to compare the 5D and M8, for example, but the 5D's extra resolution was just enough to make the text look nice, whereas the M8 had too little. The visual difference was huge, in spite of the small effective resolution difference.

 

P.S. I did see that you don't care about comparing the 5D and M8 specifically, I am just making a related point