jrp Posted March 4, 2016 Share #61 Posted March 4, 2016 Advertisement (gone after registration) I've not noticed long end softness with mine, although I was expecting it from previous reports. Leaving to one side the accuracy of the placing of the focus point, focus shift seems to be a feature of almost every lens, particularly non-APO lenses. Jim Kasson has done some technical tests on a range of 85/90mm lenses on the Sony A7rII http://blog.kasson.com/ including testing how well the AF works. (Although the 90mm Summicron does poorly in the LoCA tests, his pictures (a week or two back in his blog) show that the Summmicron performs better in real life than the more modern lenses, other than the Otus, which is surprising for an older design. Link to post Share on other sites More sharing options...
Advertisement Posted March 4, 2016 Posted March 4, 2016 Hi jrp, Take a look here 24-90mm Focus Shift (Diglloyd). I'm sure you'll find what you were looking for!
tashley Posted March 4, 2016 Share #62 Posted March 4, 2016 OK, I have tested, quickly but I hope efficiently this morning, at 90mm using shutter speeds fast enough for it not to be an issue, with IS on, shooting several series of a target large enough for any minor discrepancies between cross hair and actual focus point to be irrelevant. I shot several series at f4 through f8 handheld, some with AF on and some with 2x magnified MF (refocus for each frame and aperture) and the results are very clear: the MF shots are always notably sharper than the AF shots, which are unacceptable, especially at this price but I would send back any mid range zoom that exhibited this poor level of performance. Here is a link to a zenfolio gallery for those interested. It contains representative shots taken from the series shot above. I tried to find the best focussed shots for each focus method and aperture but actually the results across different series were remarkably consistent. If you want to know which were AF which were MF, the MF shots are always the sharper ones. Additionally I now, when I have more time, need to test in more depth to see how much (if any) of this is focus shift. I suspect it isn't; partly because I also shot a couple series where I focussed manually and then re-shot as I stopped down and any focus shift is IMHO marginal and acceptable, and partly because the F4 shots (at which aperture the camera is shooting at the AF focussing aperture) are clearly weaker with AF than with MF. I also need to test to see where focus is actually falling in the AF shots, but I am pretty clear that it is falling closer to the focal plane than the subject. In other words: 1) this isn't a lens that is 'soft at the long end' (I shot some 50mm frames with the subject the same size in the finder as they were in the 90mm frames and the MF 50mm frames were only a little sharper than the MF 90mm frames. The lens is perfectly acceptable, good even, at the 90mm setting when correctly focussed) 2) this probably isn't focus shift caused by the fact that the camera focusses at an aperture different from that at which it exposes 3) this probably isn't much to do with a disparity between where the cross hair appears and where the actual focus point is 4) it is probably a bug in the focus routines, causing the camera to generally focus closer than it should at the longer end. In any event, I sort of don't care. If, when I have had more time to test in more depth it truly doesn't work - and I think that will prove to be the case - it's going back and I'm outta here! http://tashley1.zenfolio.com/p817622400/h6dc28143#h6dc28143 Link to post Share on other sites More sharing options...
trond Posted March 4, 2016 Share #63 Posted March 4, 2016 Dear tashley, I have the same experience, and attribute this to the contrast AF in the SL. Manual focus at 90mm is always tack sharp, and the SL 24-90 seem to be a really sharp lens at 90mm as the MTF suggests. However, AF is newer sharp at 90mm at far subjects. Switching from AF to MF, after first doing AF, you can see that the error is quite large when MF is used to pull in the focus. This also happens on flat subjects where the focus point bug does not exist. Best regards Trond Link to post Share on other sites More sharing options...
cpclee Posted March 4, 2016 Share #64 Posted March 4, 2016 How far away does the subject have to be for this problem to manifest? Dear tashley, I have the same experience, and attribute this to the contrast AF in the SL. Manual focus at 90mm is always tack sharp, and the SL 24-90 seem to be a really sharp lens at 90mm as the MTF suggests. However, AF is newer sharp at 90mm at far subjects. Switching from AF to MF, after first doing AF, you can see that the error is quite large when MF is used to pull in the focus. This also happens on flat subjects where the focus point bug does not exist. Best regards Trond Link to post Share on other sites More sharing options...
trond Posted March 4, 2016 Share #65 Posted March 4, 2016 How far away does the subject have to be for this problem to manifest? Dear cpclee, In my case this was at "infinity", meaning more than 100m, probably 2-300 meter landscape scene. The SL 24-90 works fine with AF on medium distances 20-50 meters or so. At short distance, up to a few meters, you need to take the focus point bug into account. Best regards Trond Link to post Share on other sites More sharing options...
tashley Posted March 4, 2016 Share #66 Posted March 4, 2016 My tests were at around 15-20 feet but in casual use yesterday I saw it at a range of distances. Link to post Share on other sites More sharing options...
cpclee Posted March 4, 2016 Share #67 Posted March 4, 2016 Advertisement (gone after registration) I did a quick test just now with a target about 45 feet away using the zoom at the 90mm position. The target was a bit larger than the cross hair when magnification was off. For the AF shots, I used the center cross hair and had the magnification set to off. (I don't believe the cross hair can be accurate when the magnification is on because the cross hair itself is not magnified.) The MF shots were taken using the highest magnification. The aperture was set to 4.0 the whole time. Generally, I feel my MF shots were a hair bit sharper. But the AF shots really were very closely behind. There was some shot to shot variation, whether amongst the AF shots or amongst the MF shots, but with the AF shots I had one that was noticeably less sharp and borderline "soft". The rest of the AF shots were very similar and couldn't be called soft. The MF shots had less shot to shot variation. Link to post Share on other sites More sharing options...
trond Posted March 4, 2016 Share #68 Posted March 4, 2016 Dear Tashley, In that range, I think that the focus point bug will be important depending on the kind of subject. Using manual focus, I find the SL24-90 to be very sharp at 90mm at any distance, and I see no trace of focus shift. Best regards Trond Link to post Share on other sites More sharing options...
tashley Posted March 4, 2016 Share #69 Posted March 4, 2016 Dear Trond, the target was easily big enough for the focus point bug as I understand it not to be a factor. Also, 'in the wild' of normal walkabout shooting I notice the 90mm shots with AF to be consistently noticeably softer than at other focal lengths and not as good as I would expect, with reasonable expectations based on wide experience of other systems. Link to post Share on other sites More sharing options...
thighslapper Posted March 4, 2016 Share #70 Posted March 4, 2016 4) it is probably a bug in the focus routines, causing the camera to generally focus closer than it should at the longer end. In any event, I sort of don't care. If, when I have had more time to test in more depth it truly doesn't work - and I think that will prove to be the case - it's going back and I'm outta here! http://tashley1.zenfolio.com/p817622400/h6dc28143#h6dc28143 I think it is a firmware issue. Surprisingly I have not found any images where it seems to have made a practical difference. It works fine for me ..... but I would expect Leica to correct it, whatever the ultimate real-world results..... Link to post Share on other sites More sharing options...
ramarren Posted March 4, 2016 Share #71 Posted March 4, 2016 I think it is a firmware issue. Surprisingly I have not found any images where it seems to have made a practical difference. It works fine for me ..... but I would expect Leica to correct it, whatever the ultimate real-world results..... Same here. When I've used the 24-90, it's been mostly at the long end and I have been very happy with the results. Link to post Share on other sites More sharing options...
thighslapper Posted March 4, 2016 Share #72 Posted March 4, 2016 Checked yet again ...... I can find no issues at f4 at infinity and 60m ....... MF and AF images look the same to me...... The mismatch between AF and MF is just apparent at about 5m....... and as my previous tests show becomes quite obvious at close distances if you look for it. Link to post Share on other sites More sharing options...
tashley Posted March 4, 2016 Share #73 Posted March 4, 2016 Dear Trond, the target was easily big enough for the focus point bug as I understand it not to be a factor. Also, 'in the wild' of normal walkabout shooting I notice the 90mm shots with AF to be consistently noticeably softer than at other focal lengths and not as good as I would expect, with reasonable expectations based on wide experience of other systems. Link to post Share on other sites More sharing options...
cpclee Posted March 4, 2016 Share #74 Posted March 4, 2016 I tried another test at 90mm -- this time with a target that was about 7.5ft away, so quite close. The target was a tissue box. I focused on the fine text printed on it so focus quality was easy to see. All shots (AF or MF) were taken at f4.0 at 1/125s. Again when I focused manually I brought on the biggest magnification, and when focusing with AF I turned any magnification off. Still no consistent difference between AF and MF. If anything this time around I wouldn't call the AF shots a hair behind. They basically were neck and neck with the MF shots. There was some shot to shot variation amongst the AF shots just like there was with the MF shots, and the worse ones looked just a bit out. But no systematic difference between MF and AF and the only conclusion is that yes, there is tiny shot to shot variation no matter which way you focus and that variation is more noticeable than any systematic difference between AF and MF. My camera is set to 49 AF steps rather than 37 and I don't know if that makes any difference. Overall I really am very happy with the speed and accuracy / consistency of AF on the SL. Checked yet again ...... I can find no issues at f4 at infinity and 60m ....... MF and AF images look the same to me...... The mismatch between AF and MF is just apparent at about 5m....... and as my previous tests show becomes quite obvious at close distances if you look for it. Link to post Share on other sites More sharing options...
thighslapper Posted March 4, 2016 Share #75 Posted March 4, 2016 The issues here and those pointed out at Digilloyd are (most ... if not only) apparent with single point focussing ...... the point of best focus is just in front of the point that the green cross is fixated on...... and is masked to a greater or lesser extent by the differences in DOF with distance and smaller apertures. It is not due to the green cross on the EVF/monitor not mapping correctly to the equivalent point on the sensor ....... as using the camera upside down with a sloping target yields basically the same results. For some reason the AF system is fixing the focussing point optically in the lens slightly in front of the optimal point ..... quite how this can happen with a contrast detection system is beyond my simple understanding of how this works ..... Link to post Share on other sites More sharing options...
cpclee Posted March 4, 2016 Share #76 Posted March 4, 2016 I used 1-point AF (central crosshair selected), a close target, and the widest aperture. Still I wasn't able toe replicate the issue. What version of firmware are you using? The issues here and those pointed out at Digilloyd are (most ... if not only) apparent with single point focussing ...... the point of best focus is just in front of the point that the green cross is fixated on...... and is masked to a greater or lesser extent by the differences in DOF with distance and smaller apertures. It is not due to the green cross on the EVF/monitor not mapping correctly to the equivalent point on the sensor ....... as using the camera upside down with a sloping target yields basically the same results. For some reason the AF system is fixing the focussing point optically in the lens slightly in front of the optimal point ..... quite how this can happen with a contrast detection system is beyond my simple understanding of how this works ..... Link to post Share on other sites More sharing options...
cpclee Posted March 4, 2016 Share #77 Posted March 4, 2016 BTW, I don't think when using AF you want to aim your crosshair while the magnification is on. From an engineering perspective, the precision of AF has to be fixed and independent of whether the magnification is on or off. (The CDF algorithm always sees pixel level regardless of what magnification you set the EVF to. The magnification is what YOU see, not what the CDF algorithm sees.) Therefore the AF crosshair you see when magnification is on is misleading because it is not enlarged in proportion withe everything else you're seeing. So if you have been using your AF with magnification on that could explain the kind of results you have been seeing. The issues here and those pointed out at Digilloyd are (most ... if not only) apparent with single point focussing ...... the point of best focus is just in front of the point that the green cross is fixated on...... and is masked to a greater or lesser extent by the differences in DOF with distance and smaller apertures. It is not due to the green cross on the EVF/monitor not mapping correctly to the equivalent point on the sensor ....... as using the camera upside down with a sloping target yields basically the same results. For some reason the AF system is fixing the focussing point optically in the lens slightly in front of the optimal point ..... quite how this can happen with a contrast detection system is beyond my simple understanding of how this works ..... Link to post Share on other sites More sharing options...
tashley Posted March 4, 2016 Share #78 Posted March 4, 2016 I have not been using cross hair with magnification. Link to post Share on other sites More sharing options...
thighslapper Posted March 4, 2016 Share #79 Posted March 4, 2016 ditto. latest firmware. CPC .... try using a sloping target like my test shots a few pages previously .... that will show you where the point of best focus lies relative to your target ...... Link to post Share on other sites More sharing options...
tashley Posted March 5, 2016 Share #80 Posted March 5, 2016 Curious about the comments from people who aren't seeing this problem at longer distances in the real world, I banged off a quick comparison this morning. Distance to subject is very close to 235 metres. Guess which one is manual and which auto-focus? It was this sort of thing that I experienced on walkabout shooting that encouraged me to start reading up and then testing. http://tashley1.zenfolio.com/img/s4/v9/p1782968203.jpg http://tashley1.zenfolio.com/img/s3/v42/p1646823769.jpg Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.