LocalHero1953 Posted July 30, 2016 Share #21 Posted July 30, 2016 Advertisement (gone after registration) For anyone worried by the SL's spot AFs capability with the 24-90. Posted on the SL image forum http://www.l-camera-forum.com/topic/253192-leica-sl-image-thread-post-your-examples-here/?p=3087861 Link to post Share on other sites More sharing options...
Advertisement Posted July 30, 2016 Posted July 30, 2016 Hi LocalHero1953, Take a look here Front focusing errors with the 24-90mm?. I'm sure you'll find what you were looking for!
steppenw0lf Posted August 1, 2016 Share #22 Posted August 1, 2016 Now that the 2.1 firmware is available - which improves AF especially regarding tracking - could you please reevaluate your findings ? Most of the remarks in this thread were from a time before the latest release. So can you please check and either confirm the problem you found (that it still exists), or explicitely declare the problem as solved. For me it is a complete mess. And completely unclear if Digloyd found a real problem and if it still lingers. Or if it never existed. Or if there are now new problems. I have no 24-90. But I can add that I found no problems with the 90-280 (some people mentioned incorrect focus with AFs, but I am not able to reproduce this.) (No comment from me about tracking or face recognition - I am still testing). Link to post Share on other sites More sharing options...
kikouyou Posted August 1, 2016 Share #23 Posted August 1, 2016 I tested my new lens yesterday with 2.1 I found no obvious signs of front focusing The manual scale on the LCD works like a charm, you can very precisely create front and back focusing manually, brilliant. I tried all stop test shots and got no sign of focus shift for close focus and infinity. I did not check tracking yet. It was poor with 2.0 Link to post Share on other sites More sharing options...
jonoslack Posted August 1, 2016 Share #24 Posted August 1, 2016 Now that the 2.1 firmware is available - which improves AF especially regarding tracking - could you please reevaluate your findings ? Most of the remarks in this thread were from a time before the latest release. So can you please check and either confirm the problem you found (that it still exists), or explicitely declare the problem as solved. For me it is a complete mess. And completely unclear if Digloyd found a real problem and if it still lingers. Or if it never existed. Or if there are now new problems. I have no 24-90. But I can add that I found no problems with the 90-280 (some people mentioned incorrect focus with AFs, but I am not able to reproduce this.) (No comment from me about tracking or face recognition - I am still testing). The way that Lloyd did the testing simply showed up that the camera was focusing with the bottom part of the focusing square rather than the middle of the cross hair when using zoom focus - As far as I'm aware this was to do with different focus point 'grids' when you zoom in I understood that there was some discussion about this, and that Leica feel that it's fixed (certainly by 2.0). I managed to duplicate the behaviour at the time, but I never ever thought that it was to do with focus shift - but to do with the size of the focus box and the inevitability of 'granularity' in the grid. Whatever, the point was that even then I didn't think it ever caused a problem in real life shooting. It's not a mess - but don't expect a conclusive definition of the situation - Lloyd has criticised almost all recent Leica products in his blog - I guess that you can always find fault if you try hard enough and that bad news generates subscriptions (but perhaps I'm too cynical). The trouble is that it's really easy to say "This Lens Has Focus Shift", but extremely difficult to prove definitively that it doesn't! Surely - the fact that there are no complaints around here about the focusing suggests that . . . there isn't a problem with the focusing. I've put the two zooms through some seriously testing situations (1200 images of moving tractors in a day - low light band shoots - weddings etc. etc.) and as far as I'm concerned the focusing is cracking - much more accurate than any camera I've used before, of course, that doesn't mean that it's perfect, but I firmly believe that it is functionally excellent. Does that help? I guess not! all the best Link to post Share on other sites More sharing options...
jrp Posted August 1, 2016 Author Share #25 Posted August 1, 2016 I can well imagine that the focus point would be in the wrong place in the image if it did not properly take account of the distortion correction, or movement caused by OIS, or if you are taking a picture where something in the box around the focus point is closer or has more contrast than the centre of the box itself. I assume that Leica, or whoever programs the focus module would, however, have been smart enough to to correct for any of these situations, unless they found the cure to be worse than the disease in the field. In my own case, I have found successive pictures of the same static scene to show different levels of sharpness, but put it down to taking pictures from shaky bridges or other causes. Link to post Share on other sites More sharing options...
thighslapper Posted August 2, 2016 Share #26 Posted August 2, 2016 I can well imagine that the focus point would be in the wrong place in the image if it did not properly take account of the distortion correction, or movement caused by OIS, or if you are taking a picture where something in the box around the focus point is closer or has more contrast than the centre of the box itself. I assume that Leica, or whoever programs the focus module would, however, have been smart enough to to correct for any of these situations, unless they found the cure to be worse than the disease in the field. In my own case, I have found successive pictures of the same static scene to show different levels of sharpness, but put it down to taking pictures from shaky bridges or other causes. The big 'unknown' is what size the 'spot focus' grid actually is in terms of pixels ..... no info at all anywhere, and clearly whether it is very discrete or averaged over a larger area makes a big difference in speculation on 'errors' versus the 'functional resolution' of the system as it is designed. If you can take photos in 'normal' controlled conditions in a repeatable manner and the results are in focus then it doesn't matter. The problem with Digilloyd and a lot of other review sites is that they don't use 'normal' conditions at all ..... they tend to use extreme situations where it is unsurprising that the odd few quirks show up ...... but they rarely have any bearing on real world use. They do sell copy, however, which is why these people just keep digging...... Link to post Share on other sites More sharing options...
wlaidlaw Posted August 2, 2016 Share #27 Posted August 2, 2016 Advertisement (gone after registration) Focus has definitely been improved by the 2.0 or 2.1 firmware. I am sure it was focusing at or even slightly below the bottom edge of the green box with FW1.2 and now is as close as I can estimate, at the centre of the green box. I have tested at long range and close range. Turning OIS on or off on a tripod does not seem to make a lot of difference to my surprise - the images seem equally sharp either way. In preparation I had programmed the viewfinder select button to OIS on/off on a long press but I am not convinced it is required. I will run a couple more tests on this because logic says it should make a difference. Wilson Link to post Share on other sites More sharing options...
jonoslack Posted August 2, 2016 Share #28 Posted August 2, 2016 The big 'unknown' is what size the 'spot focus' grid actually is in terms of pixels ..... no info at all anywhere, and clearly whether it is very discrete or averaged over a larger area makes a big difference in speculation on 'errors' versus the 'functional resolution' of the system as it is designed. .. HI There I hope you're flourishing! The problem which Lloyd discovered (as I understand it, and of course I'm only too capable of being wrong) is that the density of the grid changed when you zoomed in and zoomed out (rationally enough otherwise there wouldn't be much point in zooming in). It was this which caused the impression of focus shift with Lloyd's technique of focusing zoomed at widest aperture in and then zooming out and changing aperture. He interpreted the difference as focus shift, but actually it was because of the change of density of the grid when zoomed in and zoomed out . . . .. . . . . and by the way, I completely agree with you that there are many people who get traffic on their website by being controversial (Ken Rockwell started it all) Sean Reid is one of several glowing exceptions. I guess I'm as dishonest as the next man, but as I don't earn money from the stuff I write it doesn't make much difference to me how many people read it and I don't have nearly a good enough memory to take a chance on lying Link to post Share on other sites More sharing options...
stephan_w Posted August 2, 2016 Share #29 Posted August 2, 2016 I just jump in as I haven't been here often over the last year. I have made some testing over hundret frames last week end as I was a bit worried about the result of a shooting I made two weeks ago, and found more or less significant focus errors on more than the half of the tests I made, mostly front focus of about 2-6 cm in average. This tests were made on a lens testig target. Of course, the issue is more visible on my S-lenses than it is on the zoom, as on the SL zooms the depht of field is a bit bigger and you might not see it without close inspections of the files. I repeated the tests with another target that has more microcontrast, and the results were a bit improved. I have no explanation for this, as with contrast focussing (and all shots were made wide open) we cannot speak of lens shift, it must be another phenomenon. I suspect that what we see on screen (the focus cross) is not precise enough to define the correct focus areas, or that flat 2D targets are not sufficient to test the focus correctly. What ever it may be, I'm actually very concerned about this issue as for me focus is essential for having this camera. For comparison, on my S2 with the same lenses I have zero (0!) focus variation. Link to post Share on other sites More sharing options...
jrp Posted August 2, 2016 Author Share #30 Posted August 2, 2016 It's hard to comment, without further and better particulars of your focus mode settings, distance to target, target angle, firmware version, etc. Link to post Share on other sites More sharing options...
stephan_w Posted August 2, 2016 Share #31 Posted August 2, 2016 It's hard to comment, without further and better particulars of your focus mode settings, distance to target, target angle, firmware version, etc. yes, true, but of course it was AFs with a single crosspoint, distance to target 4 and 7 meters and of course a rectangular angle. Firmware version 2.0 and 2.1 (it doesn't make any difference), I just tested as i would have photographed portrait and fashion, and I would say you can exclude user issues while testing :-) i did the same tests with the S2 and will reapeat the tests with my A7r2 when I have more time. Link to post Share on other sites More sharing options...
thighslapper Posted August 2, 2016 Share #32 Posted August 2, 2016 When I tested this originally ...... I assumed it was something to do with the inaccurate placing of the 'spot' cross so that the actual focus point was slightly below the cross, which when using typical sloping lens targets yielded some front-focussing. However using the camera upside down had virtually the same result and suggested that this was only a minor part of the problem. The algorithm that sets the focus point of the lens seems to place it at the very front of the DOF area ....which is fine but leaves very little room for error. With the modest apertures of the zoom lenses this effect is mostly hidden. I haven't re-checked whether this has changed ..... mainly because I've taken thousands of photos and not really noticed any issues ...... or at least none that were not explainable by the inherent issues with contrast AF, such as not locking accurately onto 'infinity' when focussing in diffuse low contrast landscapes where using the sky/horizon will cause trouble. I've done several days of portraiture purely with the 24-90 and 90-280 (thousands of shots) ..... and that is one area where focussing errors show up very quickly .... and I can't say there were any that weren't caused by the mug behind the viewfinder. If I get time this week I will revisit this and see what turns up...... Link to post Share on other sites More sharing options...
wlaidlaw Posted August 2, 2016 Share #33 Posted August 2, 2016 It is a pity we couldn't pre-program an "infinity" button for low contrast landscapes, instead of having to fiddle about changing to MF and then twiddling the focus ring. Wilson Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.