sandymc Posted January 21, 2017 Share #1 Posted January 21, 2017 Advertisement (gone after registration) For those interested in the gory technical details, my usual "new Leica camera" analysis is up. Although in this case there's quite a lot that's (as yet) unknown. Leica M10 raw file (DNG) analysis http://chromasoft.blogspot.com/2017/01/leica-m10-raw-file-dng-analysis.html Sandy 12 Quote Link to post Share on other sites More sharing options...
Advertisement Posted January 21, 2017 Posted January 21, 2017 Hi sandymc, Take a look here Leica M10 raw file (DNG) analysis. I'm sure you'll find what you were looking for!
Tobers Posted January 21, 2017 Share #2 Posted January 21, 2017 (edited) Nice on Sandy. My DNGs were imported onto my PC (not a Mac) hard drive using Lightroom. The files I linked to on this forum were taken from my hard drive, rather direct from my SD card. I wonder if this explains the bits of Lightroom fragments in the DNG files (though I'm not sure if my files showed this). The EXIF is interesting and would explain the inability of Lightroom to show the aperture the M10 shot was taken at. It appears that Leica may have messed up the aperture calculation. It would be good if anyone with an M10 could confirm if they are seeing the aperture displayed correctly in Lightroom. Edit: just re-read your post and see that the camera I used didn't show these fragments. Edit edit: The M10 does allow you to rate images using up to 5 stars in camera, which is transferred to LR. Maybe that's a reason why the LR info is in there. Edited January 21, 2017 by Tobers 1 Quote Link to post Share on other sites More sharing options...
digitalfx Posted January 21, 2017 Share #3 Posted January 21, 2017 (edited) Thanks for sharing Sandy. Would you like some DNG's straight off the SD card? Also I wonder if some of these images were from earlier FW. The camera I received has v1.0.1.0 and I noticed difference in the default camera settings vs all the reviews. Edited January 21, 2017 by digitalfx Quote Link to post Share on other sites More sharing options...
sandymc Posted January 21, 2017 Author Share #4 Posted January 21, 2017 Thanks for sharing Sandy. Would you like some DNG's straight off the SD card? Also I wonder if some of these images were from earlier FW. The camera I received has v1.0.1.0 and I noticed difference in the default camera settings vs all the reviews. Thanks for the offer. What I think would interesting would be if you post a DNG from an SD card, then also transfer the same image via WIFI, and then post that. Also perhaps an image that you rated in camera vs an image you haven't. Comparing might allow us to home in on what's going on exactly. Quote Link to post Share on other sites More sharing options...
wlaidlaw Posted January 21, 2017 Share #5 Posted January 21, 2017 Sandy, Many thanks for your very knowledgable observations. It seems common practice for Leica to issue a firmware update within a few weeks of a camera's announcement then it is many months until the next update. Hopefully if it is the same case with the M10, for the sake of M10 users, of which I will not be one, I hope Leica update what looks like somewhat sloppy coding of the DNG and sidecar file information, written by the camera. Wilson 1 Quote Link to post Share on other sites More sharing options...
ramarren Posted January 21, 2017 Share #6 Posted January 21, 2017 Nice on Sandy. My DNGs were imported onto my PC (not a Mac) hard drive using Lightroom. The files I linked to on this forum were taken from my hard drive, rather direct from my SD card. I wonder if this explains the bits of Lightroom fragments in the DNG files (though I'm not sure if my files showed this). The EXIF is interesting and would explain the inability of Lightroom to show the aperture the M10 shot was taken at. It appears that Leica may have messed up the aperture calculation. It would be good if anyone with an M10 could confirm if they are seeing the aperture displayed correctly in Lightroom. Edit: just re-read your post and see that the camera I used didn't show these fragments. Edit edit: The M10 does allow you to rate images using up to 5 stars in camera, which is transferred to LR. Maybe that's a reason why the LR info is in there. I thought I read somewhere that the embedding of Aperture into EXIF was intentionally disabled since it wasn't all that accurate. 2 Quote Link to post Share on other sites More sharing options...
digitalfx Posted January 21, 2017 Share #7 Posted January 21, 2017 (edited) Advertisement (gone after registration) I thought I read somewhere that the embedding of Aperture into EXIF was intentionally disabled since it wasn't all that accurate. You are correct, it was deliberately omitted on the M10. But I don't understand why. It wasn't perfect data, but it was still useful...far better than nothing. Edited January 21, 2017 by digitalfx 5 Quote Link to post Share on other sites More sharing options...
jonoslack Posted January 21, 2017 Share #8 Posted January 21, 2017 You are correct, it was deliberately omitted on the M10. But I don't understand why. It wasn't perfect data, but it was still useful...far better than nothing. I was sad to see it go as well . . And maybe we could petition to get it back, but I think they got fed up of people complaining that it wasn't accurate (hardly surprising I guess). 3 Quote Link to post Share on other sites More sharing options...
jonoslack Posted January 21, 2017 Share #9 Posted January 21, 2017 Sandy, Many thanks for your very knowledgable observations. It seems common practice for Leica to issue a firmware update within a few weeks of a camera's announcement then it is many months until the next update. Hopefully if it is the same case with the M10, for the sake of M10 users, of which I will not be one, I hope Leica update what looks like somewhat sloppy coding of the DNG and sidecar file information, written by the camera. Wilson HI There Wilson I hope you're well . . I've got some insight into this . . If you're making the cameras, it's fairly important that they all go out with the same firmware, and you need that to be tried and tested, so when you start manufactuing production models you fix the firmware for the first shipments because it would be a real balls-ache to have to unpack the cameras and update the firmware - much easier to put the new firmware up on the website in a couple of weeks, when you're sure it's good. 5 Quote Link to post Share on other sites More sharing options...
LocalHero1953 Posted January 21, 2017 Share #10 Posted January 21, 2017 (edited) I was sad to see it go as well . . And maybe we could petition to get it back, but I think they got fed up of people complaining that it wasn't accurate (hardly surprising I guess). Agreed - the data was imperfect, but still useful when you don't have a clue otherwise. Edited January 21, 2017 by LocalHero1953 Quote Link to post Share on other sites More sharing options...
scott kirkpatrick Posted January 21, 2017 Share #11 Posted January 21, 2017 About not even trying to save the estimated aperture -- who said that? Was it official or an observation that can be immediately made once pictures with EXIF show up? Anyway, using the Adobe dng_validate program (which Sandy explains on his blog) you find deep in the file an entry that lists the focal length and what looks like an aperture, but the value is complete garbage, as if the output program wrote the things it was expected to write, but the camera forgot to put any valid number in the location from which that value would normally be taken. Any of us who has written code has made this mistake at some point. As a result, there is a warning code in every file I have looked at about the invalid value. That's why I wondered if leaving out the aperture was an official statement. It looks as if it just got lost somewhere. scott Quote Link to post Share on other sites More sharing options...
jonoslack Posted January 21, 2017 Share #12 Posted January 21, 2017 About not even trying to save the estimated aperture -- who said that? Was it official or an observation that can be immediately made once pictures with EXIF show up? Anyway, using the Adobe dng_validate program (which Sandy explains on his blog) you find deep in the file an entry that lists the focal length and what looks like an aperture, but the value is complete garbage, as if the output program wrote the things it was expected to write, but the camera forgot to put any valid number in the location from which that value would normally be taken. Any of us who has written code has made this mistake at some point. As a result, there is a warning code in every file I have looked at about the invalid value. That's why I wondered if leaving out the aperture was an official statement. It looks as if it just got lost somewhere. scott Hi ScottThe 'assessed' aperture is important, and is used by the camera for lens corrections etc. Displaying it to the user is something they've decided not to do (I think because they've been hassled that it isn't completely accurate) So perhaps what is showing in the DNG is something the camera is using without any intention of the end user seeing it. 4 Quote Link to post Share on other sites More sharing options...
wlaidlaw Posted January 21, 2017 Share #13 Posted January 21, 2017 Jono, I assumed that the interpolated aperture had to be used for something, otherwise why put an external CdS sensor on the digital M's. I don't expect the aperture to be accurate and once you appreciate that it is interpolated, you know it can only be an approximation but nevertheless, often useful to have. I bet Leica engineers wish they could use a time machine and go back to add an aperture position lever to the M mount in 1952, when it was designed. Wilson Quote Link to post Share on other sites More sharing options...
scott kirkpatrick Posted January 21, 2017 Share #14 Posted January 21, 2017 Hi Scott The 'assessed' aperture is important, and is used by the camera for lens corrections etc. Displaying it to the user is something they've decided not to do (I think because they've been hassled that it isn't completely accurate) So perhaps what is showing in the DNG is something the camera is using without any intention of the end user seeing it. So "they" really did decide to leave it out? I imagine this must have been a subject of private conversation. Anyway the number that appears where the estimated aperture perhaps should have been seems to be a random 3 or 4-digit integer. Examples are 248, 973 etc. Not even something useful for internal processing. I think leaving it out has caused an annoying but non-fatal bug. The things that the camera would do with the estimated aperture would include adjusting the strength of vignetting corrections. It would be nice for those to show up in EXIF, but I don't think that is the practice. scott 1 Quote Link to post Share on other sites More sharing options...
pico Posted January 21, 2017 Share #15 Posted January 21, 2017 (edited) Jono, I assumed that the interpolated aperture had to be used for something, otherwise why put an external CdS sensor on the digital M's. I don't expect the aperture to be accurate and once you appreciate that it is interpolated, you know it can only be an approximation but nevertheless, often useful to have. I bet Leica engineers wish they could use a time machine and go back to add an aperture position lever to the M mount in 1952, when it was designed. I'm going to see if anything important happens to the image after I tape over the second sensor. Scott, if the entry is random, might it be from not putting nulls and instead just taking whatever was in the binary image? I cringe to think it might be so. Edited January 21, 2017 by pico 1 Quote Link to post Share on other sites More sharing options...
mjh Posted January 21, 2017 Share #16 Posted January 21, 2017 So "they" really did decide to leave it out? I imagine this must have been a subject of private conversation. Anyway the number that appears where the estimated aperture perhaps should have been seems to be a random 3 or 4-digit integer. Examples are 248, 973 etc. Not even something useful for internal processing. I think leaving it out has caused an annoying but non-fatal bug. Where do you see those random numbers? In my files the aperture is simply not specified; the tags used for specifying the aperture aren’t present. The inclusion of the guesstimated aperture info in the Exif data created by the M (Typ 240) was officially a bug that has been rectified with the M10. But then, different people have different views of whether something should be regarded a bug or a feature. 2 Quote Link to post Share on other sites More sharing options...
scott kirkpatrick Posted January 21, 2017 Share #17 Posted January 21, 2017 (edited) I'm going to see if anything important happens to the image after I tape over the second sensor. Scott, if the entry is random, might it be from not putting nulls and instead just taking whatever was in the binary image? I cringe to think it might be so. I've used that experiment. A finger over the "blue dot" works just fine if you are systematic about it. It should make the camera think that the lens is at its smallest (wide open) aperture. With the older firmwares, this lets you see in the Exif what value for that aperture is in the tables. Sometimes there were surprises. That's my hunch -- an uninitiated variable. Not so hard to do if the code that writes the file is large and written by somebody else. scott Edited January 21, 2017 by scott kirkpatrick 1 Quote Link to post Share on other sites More sharing options...
scott kirkpatrick Posted January 21, 2017 Share #18 Posted January 21, 2017 (edited) Where do you see those random numbers? In my files the aperture is simply not specified; the tags used for specifying the aperture aren’t present. The inclusion of the guesstimated aperture info in the Exif data created by the M (Typ 240) was officially a bug that has been rectified with the M10. But then, different people have different views of whether something should be regarded a bug or a feature. Hi Michael. If you run Adobe's dng_validate on one of the as-shipped cameras' files you find something like this: LensSpecificationExif: 50.0 mm f/511.0 LensMakeExif: "Leica Camera AG" LensModelExif: "Apo-Summicron-M 1:2/50 ASPH." followed a few lines further down by this (non-fatal) error message *** Warning: Possible MaxApertureValue conflict with LensInfo *** The number 511 in this example is some other equally unlikely value in each of the other files I have looked at. scott Edited January 21, 2017 by scott kirkpatrick Quote Link to post Share on other sites More sharing options...
jonoslack Posted January 21, 2017 Share #19 Posted January 21, 2017 Indeed Michael (and thanks for chipping in) The distinction between a bug and a feature is one of the most important considerations in my working life (and possibly yours as well). Clearly the estimated aperture is important for the lens corrections - but I do understand why they're fed up with users complaining it's inaccurate! Quote Link to post Share on other sites More sharing options...
waterlenz Posted January 21, 2017 Share #20 Posted January 21, 2017 I've used that experiment. A finger over the "blue dot" works just fine if you are systematic about it. It should make the camera think that the lens is at its smallest (wide open) aperture. With the older firmwares, this lets you see in the Exif what value for that aperture is in the tables. Sometimes there were surprises. That's my hunch -- an uninitiated variable. Not so hard to do if the code that writes the file is large and written by somebody else. scott Will covering the external light sensor mess up the in-camera lens corrections? 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.