jonoslack Posted July 16, 2007 Share #1 Posted July 16, 2007 Advertisement (gone after registration) HI Everyone I have't been around much for a week or so, a local radio shoot, a big wedding and a family bereavement have rather filled up my time (elderly and splendid father after a log and happy life). Still - that's not the point - the wedding and commercial shoot meant using IR cut filters (no brainer of course), and afterwards I thought "why not just leave them on", as most others do. I did a batch of landscapes using a Canadian E60 pre Asph 21mm, which has been back to Leica for full service (expensive) and coding. I've been VERY pleased with it, sharp and generally lovely. However, using an IR cut filter today, I found that many shots had a cyan drift in the top right hand corner, so I went and took some more shots - some with the camera upside down . . . . . and the cyan drift moved to the opposite corner. So I tried shooting without the filter (but with IR set to 'on' in the firmware). This time I was getting red drift - but in the opposite corner. So I did a similar set of shots without filter and with the camera set to lens recognition 'on' (without IR filter). There isn't any obvious vignetting, and all colour casts are gone, with each corner behaving evenly. It should be said, all these photos were taken of evenly lit, drab, english grey skies, and the shots from the professional shoots have been all that you would expect. The clear implication is that there is a vignetting issue with the lens - but as it isn't visible when not using filters............ Every few weeks I think - 'Oh hell, I'll just leave the filters on', and every time I find some serious gotcha which makes me take them off again! (remember, we're talking nature/landscape here, not wedding/commercial/street). Link to post Share on other sites More sharing options...
Advertisement Posted July 16, 2007 Posted July 16, 2007 Hi jonoslack, Take a look here Filters and Wide angle and Cyan corners (help me Sean!). I'm sure you'll find what you were looking for!
jaapv Posted July 16, 2007 Share #2 Posted July 16, 2007 Even the filter brand can make a difference.... Link to post Share on other sites More sharing options...
jonoslack Posted July 16, 2007 Author Share #3 Posted July 16, 2007 Even the filter brand can make a difference.... Hi there Jaap Not in this case, as the result was duplicated (in reverse) with a red drift when NOT using a filter (with the lens recognition set to on+IR). Link to post Share on other sites More sharing options...
jaapv Posted July 16, 2007 Share #4 Posted July 16, 2007 Jono, I noticed that the cyan drift on my 15 mm Heliar was uneven too in some shots. I wonder what causes it. Maybe the direction the IR light is coming from? Link to post Share on other sites More sharing options...
sean_reid Posted July 16, 2007 Share #5 Posted July 16, 2007 Hi Jono, If you would, make me a list of what happens with various combinations: which filter (if any), menu setting for lens detection, which aperture, which focus distance. Scenario 1: XXXXX Results: XXXX etc. Maybe we can isolate what is going on. Cheers, Sean Link to post Share on other sites More sharing options...
jlm Posted July 16, 2007 Share #6 Posted July 16, 2007 the asymmetry is curious. if the cyan follows the corner when inverted, it would seem the sensor or some correction algorithm applied to the sensor is asymmetric as a test, did you try rotating the filter glass witn respect to the camera body to see if that moves the cyan? Link to post Share on other sites More sharing options...
Guest guy_mancuso Posted July 17, 2007 Share #7 Posted July 17, 2007 Advertisement (gone after registration) Well there still is IR on the sensor glass too that will affect the image, just not as strong with the filters . Also with the 21mm i would certainly be using the Leica brand here. The WATE is clean at 21mm with the leica filter but not the B+W. Like Sean said make some notes on what you are doing. Link to post Share on other sites More sharing options...
sandymc Posted July 17, 2007 Share #8 Posted July 17, 2007 What you're probably seeing is a lens that is decentered - i.e., its optical center is not "pointed" at the optical center of the sensor. I've seen this a few times, both with Leica and non-Leica lenses - to the point hat the optical center of the lens is pointed 10 to 20% off center of the sensor. Normally, this isn't an issue, either with film or with more conventional digital cameras. But stir in the cyan drift problem, and you get one corner more red or cyan than the others CornerFix can probably help somewhat in this situation, assuming that you are willing to build a set of lens profiles for your lens/camera combinations specifically. Reason why it helps is that (a) the compensation is tuned to your lens/camera combination, rather than being a one size fits all general correction, ( the illuminant compensation system helps a bit and © you can tune the chroma and luma corrections to escape the worst of the color difference. Sandy Link to post Share on other sites More sharing options...
luigi bertolotti Posted July 17, 2007 Share #9 Posted July 17, 2007 Intriguing issue... do the tests proposed by Sean and post... as first impression, the fact that asimmetry occurs with no filter BUT filter on set on M8, can effectively bring to think of lens center misalignement. (decentered, as Sandy points).. the software surely acts correcting corners SUPPOSED that center CCD = lens axis... but also Jaap's observation can be significant... uneven IR from the source... WAs capture of course ( ) a wide angle of scene... how were the light sources, expecially the angular position of sun ?... I'll wait with interest for Sean proposed tests, and I would reccomend very trivial sceneries... "critical" even colors at the corners with sun RIGHT (not angled) at the back of the camera. If I had to do it, I 'd use some dresses or fabric sheets, same color (napkins...?) attached to some "wall", at the corners of the image... choosing the right hour to have sunrays 90° to the "wall" Link to post Share on other sites More sharing options...
jonoslack Posted July 17, 2007 Author Share #10 Posted July 17, 2007 Hi Jono, If you would, make me a list of what happens with various combinations: which filter (if any), menu setting for lens detection, which aperture, which focus distance. Scenario 1: XXXXX Results: XXXX etc. Maybe we can isolate what is going on. Cheers, Sean Hi Sean Not so simple - I've been doing a little preliminary testing with a white board - it seemed easier to remove the filter altogether and look at the red vignetting (so that I could eliminate the filter from the calculations). I also tried the same lens with 2 different M8 bodies, and the results are fairly similar with each. It's most obvious at f2.8, but doesn't really change much from f4 onwards - red vignetting is more pronounced on the top right (when taking a landscape shot). The interesting thing is that it's really not by a huge amount, the trouble is that it really does show if you take a landscape shot with large expanses of sky (pretty tough to avoid around here). I suppose the lens could be de-centred, but it seems to behave extremely well otherwise (sooo much sharper than it was before Solms fixed it (and charged me £200)). Of course, we all know that they don't always get it right! For the wedding I was doing, I didn't notice any problems with it - it was only afterwards when trying the lens with a filter for some landscape work - because it's uneven, it's really difficult to sort out (not worth it). Any ideas? Link to post Share on other sites More sharing options...
Guest tummydoc Posted July 17, 2007 Share #11 Posted July 17, 2007 This appears as another instance where the average-value algorithm in the firmware fails to account for some phenomenon. It was another reason the professionals I lent my M8 were underwhelmed. They were willing to look the other way at the whole IR filter issue provided they performed invisibly, but in fact they present an additional postprocessing task in many cases. Link to post Share on other sites More sharing options...
jlm Posted July 17, 2007 Share #12 Posted July 17, 2007 is this what you did?: no filter, lens detection+IR on: red vignetting in corner (first post); no filter, red vignettoing in corner (last post, no indication of lens detection, or IR setting) and duplicated on a few M8 bodes. assuming lens detection+IR was on in both cases. it seems the correction factor is not matching what the lens is doing. Is that copy of the 21 listed in the Leica chart, or might it be an oddball 21? Bew nice to try another 21 and compare. Link to post Share on other sites More sharing options...
sean_reid Posted July 17, 2007 Share #13 Posted July 17, 2007 Hi SeanNot so simple - I've been doing a little preliminary testing with a white board - it seemed easier to remove the filter altogether and look at the red vignetting (so that I could eliminate the filter from the calculations). I also tried the same lens with 2 different M8 bodies, and the results are fairly similar with each. It's most obvious at f2.8, but doesn't really change much from f4 onwards - red vignetting is more pronounced on the top right (when taking a landscape shot). The interesting thing is that it's really not by a huge amount, the trouble is that it really does show if you take a landscape shot with large expanses of sky (pretty tough to avoid around here). I suppose the lens could be de-centred, but it seems to behave extremely well otherwise (sooo much sharper than it was before Solms fixed it (and charged me £200)). Of course, we all know that they don't always get it right! For the wedding I was doing, I didn't notice any problems with it - it was only afterwards when trying the lens with a filter for some landscape work - because it's uneven, it's really difficult to sort out (not worth it). Any ideas? Hi Again, Lay out the full combinations for me please just so I can understand the variables. My first blush guess is to concur with Sandy. The lens may show some de-centering and this might not otherwise effect your pictures in most respects. But it would help to know the specifics so that I'm sure I understand them. I'd start trying to isolate the variables (as you did by removing the filter). Also be sure that the lighting is absolutely consistent across the test chart. As I think you tried before, turn the camera upside down and see if the effect moves to the lower two corners. I'm assuming here no filters and no lens detection. If the effect moves to the lower corners, the lens is probably de-centered. If it doesn't move then your test board is probably not receiving even amounts of visible light or IR. With a bit of scientific method, you'll be able to isolate this pretty quickly. Again, I'd test with no filter and lens detection turned off. If its feasible, indirect daylight (cloudy day best but shadow can work as well) can be great lighting for your test board. Cheers, Sean Link to post Share on other sites More sharing options...
jonoslack Posted July 17, 2007 Author Share #14 Posted July 17, 2007 is this what you did?: no filter, lens detection+IR on: red vignetting in corner (first post); no filter, red vignettoing in corner (last post, no indication of lens detection, or IR setting) and duplicated on a few M8 bodes. assuming lens detection+IR was on in both cases. it seems the correction factor is not matching what the lens is doing. Is that copy of the 21 listed in the Leica chart, or might it be an oddball 21? Bew nice to try another 21 and compare. Hi John Pretty much, but I've done quite a lot more since, which has confused me a great deal more. As for the lens - it's a pre-asph Canadian lens 3507xxxx, it's in perfect condition, coded by Leica, and also serviced and recalibrated by Leica. This is what I did: evenly lit white card, shot from both sides i) f2.8, ii) f5.6 and iii) f11. I took shots a) lens recognition/IR on, WITH filter (B&W 60 - I don't have a leica) lens recognition/IR on, Without filter c) Lens recognition only - without filter So, I get 18 shots from each of two M8 bodies. What do I find? I find that in although there is a little variation from shot to shot, everything is pretty much even, and notably more neutral in situation a) above - very little variation between left and right shooting. On the basis of this test the filter is doing pretty much as it should do . . . . . . . which doesn't alter the fact that I had cyan top right corners in my landscape shots yesterday Incidentally John - I'd like to get one of your WATE adaptors - please could you drop me an email with the current situation - I'm in the UK, and I don't have a spare 49 filter at the moment (I do have one on order - but but but . . . . . ) Link to post Share on other sites More sharing options...
Guest guy_mancuso Posted July 17, 2007 Share #15 Posted July 17, 2007 I took shots a) lens recognition/IR on, WITH filter (B&W 60 - I don't have a leica) lens recognition/IR on, Without filter c) Lens recognition only - without filte Bingo this maybe the issue right here is the B+W filter on a 21mm it still has some cyan cast with a B+W but the leica filter it is free from the cast completely. i would not go by this testing with that filter , it throws a curve ball right into it. We have to remember the M8 firmware is designed specifically for the Leica filters , i was told this directly from Leica and with lenses from 24mm down they highly recommended the leica filters as Sean and myself have tested this and the B=W still leave the cyan cast but the leica's it is completely clean. Not to say you may not have a decentering problem but i would use the Leica filter here Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.