imants Posted July 6, 2015 Share #61 Posted July 6, 2015 Advertisement (gone after registration) I think HCB (and countless notable others) Snow White got one from her stepmother and I know that to be true ................ ps HCB wasn't allowed to use one because other than being dead, God said no cameras are permitted due to all copyrights and patents are his! not HCBs 2 Quote Link to post Share on other sites More sharing options...
Advertisement Posted July 6, 2015 Posted July 6, 2015 Hi imants, Take a look here Thoughts on DigiLLoyd's Review of the Q. I'm sure you'll find what you were looking for!
mjh Posted July 6, 2015 Share #62 Posted July 6, 2015 This remains to be seen as there is some talk that the new backlight sensor may behave better with some of the wider RF lenses. Indeed the new sensor looks quite promising; it should be a much better match for M lenses than previous Sony sensors. How that works out in practice remains to be seen but there appears to be a lot of potential. Quote Link to post Share on other sites More sharing options...
firststream Posted July 7, 2015 Share #63 Posted July 7, 2015 (edited) I will be quite curious to see the distortion as I get my grubby paws on one next week it will be interesting how Raw Photo Processor 64 sees the files. Not buying just looking I'm finding excessive key-stoning whenever the camera is held off-axis in any of the x,y,z planes. The image is highly acceptable when only one is skew-if, but for two or more there is a distortion unlike that of any other 28mm lens. Most of the Q reviews so far show shots taken in a fashion that is either straight-on, or with only one axis off-plane. In those cases, the image quality is quite excellent, but: When you get the test camera, try shooting some architecture at a quarter-angle, with the camera tilted to one side, and pointed either up or down. Then, figure out if this impossible-to-correct distortion is something you could exploit creatively...which I have full faith you could do! Edited July 8, 2015 by firststream Quote Link to post Share on other sites More sharing options...
pico Posted July 8, 2015 Share #64 Posted July 8, 2015 (edited) Exactly ... it's too easy to end up accessing the bottom image. When I said they should do the correction in camera, what I mean that the correction should be baked into the DNG file so nobody ever has access to the uncorrected pixels. Doing that is entirely contrary to the purpose of DNG, and besides it is not possible . Edited July 8, 2015 by pico 2 Quote Link to post Share on other sites More sharing options...
alee Posted July 8, 2015 Share #65 Posted July 8, 2015 I'm finding excessive key-stoning whenever the camera is held off-axis in any of the x,y,z planes. The image is highly acceptable when only one is skew-if, but for two or more there is a distortion unlike that of any other 28mm lens. Every wide angle lens I've ever shot is prone to keystoning and perspective skews. There is nothing about the Q's 28mm that's any different that you wouldn't encounter on other cameras and lenses on a full frame. 2 Quote Link to post Share on other sites More sharing options...
kkonkkrete Posted July 8, 2015 Share #66 Posted July 8, 2015 (edited) Doing that is entirely contrary to the purpose of DNG, and besides it is not possible . Why is it not possible? Have you read the DNG specifications from Adobe? http://wwwimages.adobe.com/content/dam/Adobe/en/products/photoshop/pdfs/dng_spec_1.4.0.0.pdf The format allows lossy-compressed JPEGs as the image data! It allows cropping of pixels used for black point measurements. It allows scene or sensor referenced pixels. Why on earth would it not allow geometrical correction to be applied before writing the EFFECTIVE pixel values? Edited July 8, 2015 by kkonkkrete Quote Link to post Share on other sites More sharing options...
Guest JonathanP Posted July 8, 2015 Share #67 Posted July 8, 2015 Advertisement (gone after registration) It makes no sense to apply geometric correction to a Bayer encoded image as Michael explained earlier. Having spent some time working on algorithms to manipulate raw data I came to the conclusion that it is always best to apply corrections in the domain in which the faults were introduced. That is, if the error occurs before the bayer encoding then it needs to be corrected after the demosaicing, conversely if it occurs in the Bayer encoded domain (e.g the M240 non-linearity that leads to the green shadows) then correcting it on the raw data is best. This isn't a hard rule (e.g. coloured edge correction is possible on the raw data) but I've found no evidence to dismiss this as the most effective and efficient way to correct errors. Thats one of the reasons I believe the DNG spec includes the OpCode based geometeric corrections so that they can be applied after the demosaic. If you want geometrically corrected images out of the camera, then offering the option of 16bit TIFF or PNG demosaiced images rather than DNG would I think make more sense. Jonathan 2 Quote Link to post Share on other sites More sharing options...
firststream Posted July 8, 2015 Share #68 Posted July 8, 2015 Every wide angle lens I've ever shot is prone to keystoning and perspective skews. There is nothing about the Q's 28mm that's any different that you wouldn't encounter on other cameras and lenses on a full frame. It's not that there is keystoning in the images that I'm complaining about, it's the excessive amount evident in images I see from the Q compared to other 28mm field-of-view pictures. It makes sense: Uncorrected, there is a fish-eye like barrel distortion that, corrected, will show greater keystoning sensitivity in the off-axis plane. It would be nice if someone would compare this effect and match the Q against other 28s with the cameras held off-axis. Brick wall tests are no longer enough to test the image quality of lenses requiring in-camera corrections. Quote Link to post Share on other sites More sharing options...
kkonkkrete Posted July 8, 2015 Share #69 Posted July 8, 2015 OK, I get the reasoning. Perhaps an alternative 'raw' format would be better than DNG. In the end this is not strictly an engineering point so much as a marketing one. If I were Leica I wouldn't want there to be any way for end-users to access pre-corrected image data, because it could lead to confusion, disappointment or reputation issues. We've already seen several threads discussing this... Quote Link to post Share on other sites More sharing options...
jaapv Posted July 8, 2015 Share #70 Posted July 8, 2015 DNG is the most future-proof image format, which is the reason Leica uses it. The drawback is, obviously, that it is more accessible, being open source. I don't think there is any reputation damage, only those who not understand that modern compact camera-lens design is an integrated digital-optical process, not just by Leica, but by the whole industry, will get upset. 1 Quote Link to post Share on other sites More sharing options...
alee Posted July 8, 2015 Share #71 Posted July 8, 2015 It's not that there is keystoning in the images that I'm complaining about, it's the excessive amount evident in images I see from the Q compared to other 28mm field-of-view pictures. It makes sense: Uncorrected, there is a fish-eye like barrel distortion that, corrected, will show greater keystoning sensitivity in the off-axis plane. It would be nice if someone would compare this effect and match the Q against other 28s with the cameras held off-axis. Brick wall tests are no longer enough to test the image quality of lenses requiring in-camera corrections. We're not comparing cropped sensors and "equivalent views" vs full frame. This is a proper full frame 28mm with a full frame sensor. Regardless of the quality of the optics, the image at 28mm, with its keystone and perspective skew quirks off-axis, is the same you get with every 28mm on a full frame sensor. The in-camera correction is inconsequential to the output when it comes to field of view. You can debate corner sharpness and straightness of lines with in camera correction, but the optical physics remain unchanged and you see the world the same way. 1 Quote Link to post Share on other sites More sharing options...
firststream Posted July 8, 2015 Share #72 Posted July 8, 2015 We're not comparing cropped sensors and "equivalent views" vs full frame. This is a proper full frame 28mm with a full frame sensor. Regardless of the quality of the optics, the image at 28mm, with its keystone and perspective skew quirks off-axis, is the same you get with every 28mm on a full frame sensor. The in-camera correction is inconsequential to the output when it comes to field of view. You can debate corner sharpness and straightness of lines with in camera correction, but the optical physics remain unchanged and you see the world the same way. We'll have to wait for some comparative images showing the degree to which the Q keystones planes off-axis in order to settle our disagreement. Intuitively, for every degree of skew in the Q, it appears as a magnitude of what I'd expect in any other 28mm lens, although I wouldn't want to peg a number to it. I'd certainly like to be proven wrong, as I think otherwise the Q is superb. Perhaps Barjohn or Imants can look into this with a fixed-on-tripod comparison? Quote Link to post Share on other sites More sharing options...
batmobile Posted July 9, 2015 Share #73 Posted July 9, 2015 I agree with thighslapper: if you like the results, then they are good. If you don't, then they are not. Lloyd can deconstruct things as much as he likes, but it does not alter the fundamental importance of the 'do you like the output or not' question, or the answer. 1 Quote Link to post Share on other sites More sharing options...
imants Posted July 9, 2015 Share #74 Posted July 9, 2015 Imants can look into this with a fixed-on-tripod comparison?...... love to but I never have owned a tripod 1 Quote Link to post Share on other sites More sharing options...
digitalfx Posted July 12, 2015 Share #75 Posted July 12, 2015 Fortunately we have choices. If we want a 28mm Lens that is designed so as to not need corrections in software, we have that in the Summilux-M 28mm lens...but if we want to save a considerable amount of money, and get a spectacular 28mm lens that performs extremely well, but does this via some software corrections to save cost and size we can get that in the Q's 28mm Summilux and get a really bad ass AF camera included in the package...whats not to love? Personally I find the Q a spectacular little package...one of the best digital cameras Leica has ever produced. I really don't care if the image is corrected as I only care about the end result...and the results speak for themselves. If you have any doubts, look in the image thread. Congrats Leica on an extremely well designed camera and lens...you have hit a home run with the mighty Q. Hopefully this is just the beginning of the new Leica that innovates rather than being two steps behind the technology curve. 1 Quote Link to post Share on other sites More sharing options...
Surelythisnameisfree Posted July 12, 2015 Share #76 Posted July 12, 2015 He's posted again about a bad experience with a faulty Q and calls it a POS (and I don't think he means point and shoot). Seems to be a bit of a vendetta Quote Link to post Share on other sites More sharing options...
barjohn Posted July 12, 2015 Author Share #77 Posted July 12, 2015 I am certainly hoping that the complete death of his AF system and the rear screen clicking (poor adhesive?) are not harbingers of more to come. I'm sure Leica must be using to build these cameras and QA can slip under such conditions. It appears that the Q and the monochrome are the two cameras in high demand and lacking adequate production. Quote Link to post Share on other sites More sharing options...
jaapv Posted July 12, 2015 Share #78 Posted July 12, 2015 Well, given that any camera with such defects will likely be reported on the forum, I seriously doubt that AF failure is the norm. The LCD issue has been reported here and I understand leica is looking into it. Hardly a reason to call the camera a POS, on the contrary, it will do his site good;).... The response of a professional reviewer would have been to record Leica's reaction and to investigate the incidence rate of the problems he found - resorting to name-calling gives a rather odd impression. 2 Quote Link to post Share on other sites More sharing options...
IkarusJohn Posted July 12, 2015 Share #79 Posted July 12, 2015 I am certainly hoping that the complete death of his AF system and the rear screen clicking (poor adhesive?) are not harbingers of more to come. I'm sure Leica must be using to build these cameras and QA can slip under such conditions. It appears that the Q and the monochrome are the two cameras in high demand and lacking adequate production. Dont buy one, John. You know what will happen! 1 Quote Link to post Share on other sites More sharing options...
AlanJW Posted July 12, 2015 Share #80 Posted July 12, 2015 John, if you buy one, you will have done so in spite of the warnings from DigiLloyd. Then when you come here to complain, you will only have yourself to blame. The rest of us will point that out to you and then somehow find a way to go out and take pictures. 2 Quote Link to post Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.