jto555 Posted July 7, 2015 Share #21 Posted July 7, 2015 Advertisement (gone after registration) I am afraid that this is an Adobe bug and not Leica. I have had the same issue with Canon CR2 files. However, when I process the raw images and save a jpeg the thumbnail that is show is now correct, so no worries. Link to post Share on other sites More sharing options...
Advertisement Posted July 7, 2015 Posted July 7, 2015 Hi jto555, Take a look here Lost Images. I'm sure you'll find what you were looking for!
Guest JonathanP Posted July 7, 2015 Share #22 Posted July 7, 2015 I am afraid that this is an Adobe bug and not Leica. I have had the same issue with Canon CR2 files. However, when I process the raw images and save a jpeg the thumbnail that is show is now correct, so no worries. I have no doubt that you have seen similar symptoms with Canon file, but please don't dismiss a demonstrable Leica bug that myself and others would like fixed. Its easy to prove that the duplicate ImageUniqueId and RawDataUniqueId fields in the M240 DNG Exif data cause Lightroom (and Bridge) to incorrectly show the wrong preview from their cache. When this happens, if you use ExifTool or similar to randomise those fields you will find the problem disappears - I have done this on more than one occasion, rather than simply speculating that 'this is not Leica bug'. If you look at the DNG spec, you can see why Adobe's software gets confused by the duplicate ID fields (RawDataUniqueId, DNG v1.4): This tag contains a 16-byte unique identifier for the raw image data in the DNG file. DNG readers can use this tag to recognize a particular raw image, even if the file's name or the metadata contained in the file has been changed. If a DNG writer creates such an identifier, it should do so using an algorithm that will ensure that it is very unlikely two different images will end up having the same identifier. The M240 does not comply with that second paragraph, which is the bug that I would like to see Leica fix (they got it right in the M9!). Sorry if this sounds a bit of a rant, but its frustrating when you spend time analysing a problem, testing theories, producing a workaround (coded in my 'green shadows' plugin) yet then have it dismissed with no evidence. Guess thats the internet for you! Jonathan Link to post Share on other sites More sharing options...
jto555 Posted July 7, 2015 Share #23 Posted July 7, 2015 I have no doubt that you have seen similar symptoms with Canon file, but please don't dismiss a demonstrable Leica bug that myself and others would like fixed. Its easy to prove that the duplicate ImageUniqueId and RawDataUniqueId fields in the M240 DNG Exif data cause Lightroom (and Bridge) to incorrectly show the wrong preview from their cache. When this happens, if you use ExifTool or similar to randomise those fields you will find the problem disappears - I have done this on more than one occasion, rather than simply speculating that 'this is not Leica bug'. If you look at the DNG spec, you can see why Adobe's software gets confused by the duplicate ID fields (RawDataUniqueId, DNG v1.4): The M240 does not comply with that second paragraph, which is the bug that I would like to see Leica fix (they got it right in the M9!). Sorry if this sounds a bit of a rant, but its frustrating when you spend time analysing a problem, testing theories, producing a workaround (coded in my 'green shadows' plugin) yet then have it dismissed with no evidence. Guess thats the internet for you! Jonathan I'm sure you are correct. I was basing my opinion on the fact that the issue happening on two different camera systems for me. Either way it is annoying but I do not find it a major as the DNG file is still correct. By the way, I LOVE your Green Shadow Plugin and use it every day! John Link to post Share on other sites More sharing options...
Guest JonathanP Posted July 7, 2015 Share #24 Posted July 7, 2015 I'm sure you are correct. I was basing my opinion on the fact that the issue happening on two different camera systems for me. Either way it is annoying but I do not find it a major as the DNG file is still correct. By the way, I LOVE your Green Shadow Plugin and use it every day! John Sorry John about the tone of my post; it wasn't how I would normally comment and I apologise. Been out for a nice cream tea and walk around a lovely garden this afternoon so feeling much better now Thanks for your comments on the plugin - I have a new version almost ready which is much better at ISO400+800 and which I hope to release very soon. Jonathan Link to post Share on other sites More sharing options...
thighslapper Posted July 7, 2015 Share #25 Posted July 7, 2015 I'm guessing Adobe uses the unique ID as the index into the preview cache, which allows file moving and renaming without having to track those changes in the preview cache. correct. The issue appears to be the way Leica generates the 'unique file id' which is done by some proprietary algorithm which appears to have the sole purpose of stopping the user finding out how many actuations the camera has ....... but the generated id is 'random' rather than 'unique' judging from the problems with LR etc showing the wrong image initially. There was an interesting thread about all this when the M240 first appeared .... and to date no-one has figured out precisely what Leica is up to or how it works Link to post Share on other sites More sharing options...
CheshireCat Posted July 7, 2015 Share #26 Posted July 7, 2015 the generated id is 'random' rather than 'unique' I wish it was random. But unfortunately it is just stupid. A really random id is usually "unique enough". Link to post Share on other sites More sharing options...
Peter Branch Posted July 9, 2015 Author Share #27 Posted July 9, 2015 Advertisement (gone after registration) It looks like the latest FW update has not fixed this particular problem. Link to post Share on other sites More sharing options...
tobey bilek Posted July 9, 2015 Share #28 Posted July 9, 2015 When importing , I use Bridge, get files camera. There is an option "advanced renaming". Generate the same identifier format year-date- image #. If the camera will not make unique id #, do it yourself. Change the folder # as the old one fills. If you keep using the "factory one", there will be duplicates at some point. I would never import to LR directly either. Import to a pictures file, then copy to LR library, not move. To me there is no advantage to Lr as it does nothing PS/bridge can not do and there is no reason fool around with a library. Link to post Share on other sites More sharing options...
coupdefoudre Posted April 13, 2016 Share #29 Posted April 13, 2016 Shirley not, Jaap - it's the Adobe bug where old previews are used if the file name has been seen before. Zooming in and out of an incorrect preview should show the correct picture.... Thanks for this info, posted almost a year ago. I have had an image from an earlier shoot show up, again, later, as a DNG file 3 times now and was getting quite curious and alarmed. But a simple clicking to enlarge it in Lightroom eventually brought up the correct image (and kept it when I went back to the regular size.) I automatically number my photographs via the in-camera set-up and Format cards regularly so I could not figure out what was happening. Glad I took the time, at last, to search the forum to discover an answer! Link to post Share on other sites More sharing options...
rafael_macia Posted April 14, 2016 Share #30 Posted April 14, 2016 I had a similar issue. But my problem was that I was deleting images in camera. You say you do not delete, so I agree with Jaapv, it is probably the card. San-Disk is the most counterfeited SD card brand. The only problems I ever had was with San-Disk. I use for 8 years only Transcend cards ...... never a problem Make sure you buy from a reputable dealer. Rafael Link to post Share on other sites More sharing options...
bencoyote Posted April 14, 2016 Share #31 Posted April 14, 2016 Shirley not, Jaap - it's the Adobe bug where old previews are used if the file name has been seen before. Zooming in and out of an incorrect preview should show the correct picture. It's a cosmetic problem only as the dng is not damaged. I think you could avoid seeing it by: Copy the images to your computer. Put the card back in the camera. Switch on the camera and format the card ready for use. If you want another copy of the pictures for security, don't keep them on the card, re-copy them to your computer after step 1. This method avoids duplicate file names; your procedure guarantees them. Are you sure it is just a filename thing? I have been watching it and trying to catch it long enough to make the associations for sure (I do this kind of bug tracking for a living) and I think it may be related to the "Raw Data Unique ID" not actually being globally unique. I ran a test a while over my 25000 Leica M shots and had a few duplicates. I think that Lightroom uses the Unique ID as a hash into a database to find thumbnail previews quickly. And so when there is a duplicate, it gets the wrong preview. Being a Leica apologist for a second - making a globally truly unique ID is a difficult problem. Being a Leica critic for a second - there are several papers explaining how to do it correctly. Link to post Share on other sites More sharing options...
pico Posted April 14, 2016 Share #32 Posted April 14, 2016 Being a Leica apologist for a second - making a globally truly unique ID is a difficult problem. Yes it is, and I developed several accepted time stamps with additional ID data over the past thirty years. For others who might not appreciate the difficulty, how would you create a unique ID? Remember that date/time is problematic unless it conforms to internet time-stamp criteria but the Leica is not so connected. How else? From an internal counter hashed with the serial number? That would be an approach but it seems Leica has ignored such. We will not know until we decode their byzantine convention. Has anyone figured it out yet? I'm not smart enough. . Link to post Share on other sites More sharing options...
CheshireCat Posted April 14, 2016 Share #33 Posted April 14, 2016 I had a similar issue. But my problem was that I was deleting images in camera. You say you do not delete, so I agree with Jaapv, it is probably the card. It is not the card. Please read this thread from the beginning. Link to post Share on other sites More sharing options...
CheshireCat Posted April 14, 2016 Share #34 Posted April 14, 2016 Yes it is, and I developed several accepted time stamps with additional ID data over the past thirty years. For others who might not appreciate the difficulty, how would you create a unique ID? Remember that date/time is problematic unless it conforms to internet time-stamp criteria but the Leica is not so connected. How else? From an internal counter hashed with the serial number? That would be an approach but it seems Leica has ignored such. Even a pseudo-random number is better than what the M does now. Really, we are not talking about difficulty here. It is a godod thing Leica does not have a division in the nuclear plants field. Link to post Share on other sites More sharing options...
bencoyote Posted April 14, 2016 Share #35 Posted April 14, 2016 Even a pseudo-random number is better than what the M does now. Really, we are not talking about difficulty here. It is a godod thing Leica does not have a division in the nuclear plants field. There is nothing in a pseudo number generator preventing it from repeating elements. And nothing that makes it unique across multiple cameras. It is harder than that. A function that takes a monotonically increasing parameter and some unique salt like a random bit appended to the serial number. I think this is what Leica did but they didn't do a good enough job. The good news for Leica is that they don't need a mathematically provably perfect algorithm. They just need something like 2^20 (about a million) unique numbers per camera and then they just distribute the bits of the salt through the rest of the bits of the unique id. Link to post Share on other sites More sharing options...
faxao Posted April 14, 2016 Share #36 Posted April 14, 2016 Maybe I am missing something but I have never had this problem with Nikon and/or Fuji cameras. Why is that? Link to post Share on other sites More sharing options...
Exodies Posted April 14, 2016 Share #37 Posted April 14, 2016 Different firmware Link to post Share on other sites More sharing options...
CheshireCat Posted April 15, 2016 Share #38 Posted April 15, 2016 There is nothing in a pseudo number generator preventing it from repeating elements. And nothing that makes it unique across multiple cameras. Just use a 64 bit random number, and you will not have a significant chance of hitting a duplicate ID amongst all men who ever lived and ever will live on earth, shooting continuously with their past and future Leica cameras, out of 18,446,744,073,709,551,616 possible random IDs. And by the way, just using a string like: "sequential_shot_counter" and then appending the "camera_serial_number" will give a universal unique ID across multiple cameras without even knowing what a random number is. There is no excuse. Really. How Leica failed on that one is quite worrying. Link to post Share on other sites More sharing options...
marchyman Posted April 15, 2016 Share #39 Posted April 15, 2016 Computer systems vendors have been screwing up random number generation for over 50 years. Why should you expect a camera vendor to do any better? I'll bet somewhere in the world there is code still using RANDU, "one of the most ill-conceived random number generators ever designed" according to Knuth. Link to post Share on other sites More sharing options...
CheshireCat Posted April 15, 2016 Share #40 Posted April 15, 2016 I'll bet somewhere in the world there is code still using RANDU, "one of the most ill-conceived random number generators ever designed" according to Knuth. No, seriously, even RANDU would work great for this purpose. Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.