erlingmm Posted June 13, 2015 Share #21 Posted June 13, 2015 Advertisement (gone after registration) Let me be extremely politically incorrect here: How about an adapter lens to the Q to make it a 50mm ? Even the S system has a macro lens to attach to the S180. Link to post Share on other sites More sharing options...
Advertisement Posted June 13, 2015 Posted June 13, 2015 Hi erlingmm, Take a look here Q 28mm lens design. I'm sure you'll find what you were looking for!
jaapv Posted June 13, 2015 Share #22 Posted June 13, 2015 Isn't it one of the main attractions of Leica? I wouldn't call it Luddite. When image quality and lens design is concerned I expect Leica to lead the pack and apply any technology that is relevant. Link to post Share on other sites More sharing options...
Mike Rawcs Posted June 13, 2015 Share #23 Posted June 13, 2015 I must admit to knowing almost nothing about digital correction of a lens. Does it add data or take data away? Link to post Share on other sites More sharing options...
jaapv Posted June 13, 2015 Share #24 Posted June 13, 2015 Neither. It shifts data. Link to post Share on other sites More sharing options...
bernie.lcf Posted June 13, 2015 Share #25 Posted June 13, 2015 Depends on the correction, but distortion is handled by moving pixels (for the sake of argument, lets simplify). The RAW file is always RAW (except for Sony and others who apply lossy compression ). Correction is specified in the DNG and the application reading the DNG then processes the corrections as specified. For camera systems that output pure RAW files (DSLR come to mind), the correction needs to be done manually by the user with Photoshop (Gimp, etc) and therefore somewhat depends on the skills of the user. Also, the user needs to do this trial-end-error until the results are visually appealing. This may or may not be possible depending on degree and complexity of e.g. the distortion. So, where does that leave us? I think Leicas approach is not the worst. Lens design is always an act of balance: reduce some faults at the expense of increasing others. If you include software correction in your lens design, you can reduce the faults that cannot be sufficiently or adequately corrected in software at the expense of increasing the faults (distortion) that can be software corrected without leaving artefacts. You can then make sure that e.g. the distortion appears in a certain way for which you can calculate a precise correction formula that you can put in the DNG file for Lightroom to process. The corrected Q images are basically perfect. Compare that to a Canon L 24-105 at 24mm. Canon DSLR do not specify software correction and the lens was not designed to make distortion correction easy. Consequently, I very often pulled my hair trying to correct this beast. Link to post Share on other sites More sharing options...
pop Posted June 13, 2015 Share #26 Posted June 13, 2015 Neither. It shifts data. I'm afraid that merely shifting will not do. In order to compensate for the distortion the software has to interpolate the values of several adjacent pixels. This must result in a degradation of the image, even if it will not be readily apparent to anyone looking at the image. In that sense, the difference might be academic. Link to post Share on other sites More sharing options...
pico Posted June 13, 2015 Share #27 Posted June 13, 2015 Advertisement (gone after registration) Conversations like this make me wish to become a blacksmith. Link to post Share on other sites More sharing options...
gpwhite Posted June 13, 2015 Share #28 Posted June 13, 2015 160404136.tFLB6d0T.wleica_q_typ_116_18.jpghttp://www.fredmiranda.com/forum/topic/1370077/3 Leica Q: The first camera with fixed 28mm fisheye lens with an image circle that does not cover the sensor size. How to possibly achieve sharp edges after digital correction with such a lens design? Well, my Q does not achieve sharp corners, and it shows rather weak corners. PERIOD. But the more I shoot it, the more I am engaged and enthused about my Q. For me, the Q @ f/2 (seems it nicest opening) is at a large gap below the 28 Summilux @ f/2 on M. Yet, the Q is biting sharp and very good 3D... there seems to be field curvature or something behind the curtain with the Q. It is not pleasant. My hunch is that the image you posted, chrismuc, was somehow brought forward in the processing stream without in-camera distortion correction (software/ glitch). I shot a couple of tall buildings yesterday, and the lines are perfectly correctly. I not percieve any curvilinear distortion (on a monitor, as prints are a few days off). IMHO, IQ at distant focus on the Q, however, does not rival the laser-etched, straight lines of the 28 Summicron (at say 10m). I have not shot at a distance with the 28 SX yet. That's my download so far... Link to post Share on other sites More sharing options...
gpwhite Posted June 13, 2015 Share #29 Posted June 13, 2015 I must admit to knowing almost nothing about digital correction of a lens. Does it add data or take data away? It distorts it. So any gain in apparent correction of chromatic or spatial distortions comes at some cost in resolution, sharpness, etc. I leave my files alone in ACR. Perhaps the lines look wonky a little bit with wider lenses, but I prefer it. Just my view. Link to post Share on other sites More sharing options...
chrismuc Posted June 14, 2015 Author Share #30 Posted June 14, 2015 the file i uploaded is a jpg from an uncorrected raw in camera jpgs or using raw softwares that recognize the camera raw correction would not see a uncorrected images, so they look "straight" not "curved" I downloaded the jpg and raw f5.6 sample files from this dpreview site http://www.dpreview.com/previews/leica-q-typ116/7 and rawtherapee but i was not able to open the raw file (neither in PS CS 6 ACR 9 nor in Rawtherapee 4.2.171) this file would be a good sample to see the corner performance uncorrected (raw by rawtherapie) and corrected (jpg) maybe someone else can try and succeeds Link to post Share on other sites More sharing options...
edwardkaraa Posted June 14, 2015 Share #31 Posted June 14, 2015 Well, even though I personally don't like lenses that are designed with digital corrections, let's say that what matters most is the end result. So if we examine the end result here we see a fast lens with wonderful rendering, great bokeh, fast AF, but with somewhat weak corners and borders for landscape and architecture shooting. The package is priced very attractively by Leica standards. The Q will be a fantastic camera for general photography, street, candids, PJ, but not ideal for landscapes and architecture. Link to post Share on other sites More sharing options...
edwardkaraa Posted June 14, 2015 Share #32 Posted June 14, 2015 When image quality and lens design is concerned I expect Leica to lead the pack and apply any technology that is relevant. That would be my expectation from Leica as well. See my above post. Link to post Share on other sites More sharing options...
bernie.lcf Posted June 14, 2015 Share #33 Posted June 14, 2015 Well, I did some analysis on the dpreview picture (city scene, f11). It loads in PS and RawTherapee just fine. The remote corners are still dark (tiny bit of it)- it would be safe to assume that the lens just doesn't cover the entire circle. The corners are just dark, they are basically black. There is no information in there. By itself, this is a bit of a scandal for something made by Leica. Put in context, however, I cannot help but smile because this is actually not such a bad design choice. If you are going to apply software correction in the DNG file already, why bother with the black corners. Distortion correction will in any case eliminate them as it will transform/warp and crop the resulting image (cannot be done otherwise, you will always loose the outer corners of the image). By leaving those black corners, you will make the lens smaller, lighter and you can focus on e.g. center sharpness. This is perhaps a clever optical design choice for the overall camera, but for sure a terrible marketing choice as people will find out about it and then rant about it everywhere (as we can see happening already). We have seen this with the X Vario and it seems that Leica is following a strategy rather than the voice of the street - you may like it or not. Software correction in the case of the Leica Q does not cost a lot of detail as far as I can tell - I am finding it difficult to identify any meaningful loss, but this will be different on a calibrated test target vs. a real life image. There are reports that this lens is weak on the corners. I would disagree because that weakness is perceived relative to center sharpness. The Q is very similar to the X 113 because the lens is spectacular in the center (even wide open and when compared to the competition). The corners are not spectacular, just quite good and sadly never actually reach spectacular even beyond f11. Bokeh and 3D separation is quite amazing for a 28mm. I have seen and owned wide angle lenses (28mm and less) that showed more uniform performance at the cost of peak sharpness (Canon EF, Nikon F, Fuji X, MFT). All of them show distortion at similar levels, more or less - but most a bit more complex (as in: more difficult to correct). I have never seen a lens that shows black corners, but I suspect that most compact cameras (especially the ones that do not output RAW) will have the same problem. I have a Q coming in next week and, to be honest, the discussions have made me think about whether or not I should take delivery - after all, it is 4 grant and there is a lot of other stuff I could buy including taking a couple months off my house loan. In the end, however, I believe that the end result (what I have seen so far) speaks for itself. Some of the design choices may not be popular and this camera is not for people who are bothered by other people's opinion. Link to post Share on other sites More sharing options...
Daedalus2000 Posted June 14, 2015 Share #34 Posted June 14, 2015 But is it the case that the sensor is actually bigger than 24mp and the areas that are not covered are on part of the sensor that is not used for the 24mp ? I think I saw somewhere that theactual sensor was 26mp or something like this, I will see if I can find it. Edit: I used Picassa to open the Q files. I can see the same effect but Picassa says that the raw file is 6120 x 4016 and the jpeg is 6000 x 4000, so the sensor is a bit bigger than 24mp. I also found this quote from LeicaRumors: According to Leica the sensor is a CMOS sensor with 26.3MP total and 24.2MP effective pixels. Link to post Share on other sites More sharing options...
chrismuc Posted June 14, 2015 Author Share #35 Posted June 14, 2015 Hi Bernie, can you upload two corner crops in 100% with and without the in-camera-raw correction to compare the corner sharpness? Thx. My concern simply is that if such strong digital correction is applied, the lens might have great (uncorrected) corner sharpness, but this sharpness is just wasted because it is recorded by the 24MP 6um sensor and then stretched and interpolated by in-camera post, so the corners always will be less sharp then the center. It would be a different case if the same lens would be attached to a Sony or Canon 36/ 42/ 50 MP sensor and that image first would be digitally corrected and then down sampled to 24 MP. Then the whole image would have the same resolution of 24 MP (assumed that the lens has over the whole 36x24mm area a resolution of let's say more than 30 MP). Link to post Share on other sites More sharing options...
jaapv Posted June 14, 2015 Share #36 Posted June 14, 2015 Isn't it a bit silly to rant about dark corners that are not even included in the image that the camera produces? Every lens has light fall-off that reaches zero towards the edge of the image circle, obviously. So Leica designed a lens that covers the image area of the final output by various means of correcting, both optical and digital. Yes, and? Take away one of the corrections and it is not well corrected any more. If this were not so the correction would be senseless. Surprise,surprise... :rolleyes: Link to post Share on other sites More sharing options...
jaapv Posted June 14, 2015 Share #37 Posted June 14, 2015 Hi Bernie, can you upload two corner crops in 100% with and without the in-camera-raw correction to compare the corner sharpness? Thx. My concern simply is that if such strong digital correction is applied, the lens might have great (uncorrected) corner sharpness, but this sharpness is just wasted because it is recorded by the 24MP 6um sensor and then stretched and interpolated by in-camera post, so the corners always will be less sharp then the center. It would be a different case if the same lens would be attached to a Sony or Canon 36/ 42/ 50 MP sensor and that image first would be digitally corrected and then down sampled to 24 MP. Then the whole image would have the same resolution of 24 MP (assumed that the lens has over the whole 36x24mm area a resolution of let's say more than 30 MP). Aren't you confusing sharpness and resolution here? The corners may well have less resolution, but if the interpolation is done properly the per-pixel acuity can be unaltered. I think digital correction of distortion may well produce sharper results than optical correction would have. Corrected it must be in any case. It is all about tradeoffs and compromises. Considering most comments on this lens/sensor combo the choices appear to have been not too shabby. Link to post Share on other sites More sharing options...
viramati Posted June 14, 2015 Share #38 Posted June 14, 2015 Okay so I don't have my Q yet but I do have the Sony FE28/f2 and the Sony A7II and this is a lens that has some severe distortion unless digitally corrected and wide open the corners can be pretty soft but by F5.6 that are very good and from what I see from looking at older shot taken with my M and 28 summicron are so close at not to matter (if not better). I should imagine and hope the the Q can better the Sony FE28 a lens that only cost me £350. Personally I don't tend to do landscape work where I want the background or whole scene in sharp focus as I am often more interested in isolating some foreground subject to some degree or other and from what I am hearing this is the perfect lens for me for documentary work where I want to have a certain amount of subject isolation in say environmental portraits. I suppose though that it would be nice to access the unprocessed DNG file if you wanted it and then you could crop it to 28mm (I say this as it would appear from other posts that is in fact a wider lens) and this might give you sharper corners and could work in landscape photographer where the distortion is not so much of an issue but obviously wouldn't work with Architectural work Link to post Share on other sites More sharing options...
Daedalus2000 Posted June 14, 2015 Share #39 Posted June 14, 2015 Isn't it a bit silly to rant about dark corners that are not even included in the image that the camera produces? Every lens has light fall-off that reaches zero towards the edge of the image circle, obviously. So Leica designed a lens that covers the image area of the final output by various means of correcting, both optical and digital. Yes, and? Take away one of the corrections and it is not well corrected any more. If this were not so the correction would be senseless. Surprise,surprise... :rolleyes: Personally I am more interested in understanding how such a correction works. I have looked at the Q images and I am very happy with their sharpness at the corners and overall. My main problem is learning how to use the 28mm fov (which is in part why I bought the Q), and not the sharpness. I am very happy with the lens performance. Link to post Share on other sites More sharing options...
bernie.lcf Posted June 14, 2015 Share #40 Posted June 14, 2015 Hi Bernie, can you upload two corner crops in 100% with and without the in-camera-raw correction to compare the corner sharpness? For all I can tell, there is basically no loss... first picture is uncorrected via RawTherapee, second picture is via LR6 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! Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! ' data-webShareUrl='https://www.l-camera-forum.com/topic/246312-q-28mm-lens-design/?do=findComment&comment=2834328'>More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.