Luke_Miller Posted March 18, 2015 Share #281 Posted March 18, 2015 Advertisement (gone after registration) I don't have a golden eye either. As a long time Nikon shooter I was oblivious to Nikon's transition from CCD to CMOS sensors. I just appreciated the improved ISO range. That said - I continue to enjoy the images from my M9 and have not acquired an M-240. Nothing against the M-240, but the M9 meets my needs. Link to post Share on other sites More sharing options...
Advertisement Posted March 18, 2015 Posted March 18, 2015 Hi Luke_Miller, Take a look here CCD vs CMOS: Can you tell which is which?{merged}. I'm sure you'll find what you were looking for!
CheshireCat Posted March 19, 2015 Share #282 Posted March 19, 2015 Not to the people who can not hear it. Exactly. Link to post Share on other sites More sharing options...
Rick Posted March 20, 2015 Share #283 Posted March 20, 2015 This is quite possibly, by far, the biggest BS thread I've ever read here. Rick Link to post Share on other sites More sharing options...
lct Posted March 20, 2015 Share #284 Posted March 20, 2015 Could be worst. How to mimic a Lomo with an MM for instance with the conclusion that both are the same camera. Link to post Share on other sites More sharing options...
thighslapper Posted March 20, 2015 Share #285 Posted March 20, 2015 This is quite possibly, by far, the biggest BS thread I've ever read here. Rick ..... forgotten some of yours Rick ?? Link to post Share on other sites More sharing options...
Rick Posted March 20, 2015 Share #286 Posted March 20, 2015 ..... forgotten some of yours Rick ?? Moi? Link to post Share on other sites More sharing options...
t024484 Posted March 21, 2015 Share #287 Posted March 21, 2015 Advertisement (gone after registration) I scaled the differences *127 before storing back into the DNG file, then used LR sliders to bring out the image. These are from the M9. L1000441 by fiftyonepointsix, on Flickr The data being discarded by DNG8 compression, scaled for "visualization purposes only" using LR sliders. Lots of high-frequency, noise and signal. L441DFSC by fiftyonepointsix, on Flickr And a recomputation of the prior image. L439DFSC by fiftyonepointsix, on Flickr "The Daughter is in the Details". Sometimes she can be a Devil, but not here... Roy Lichtenstein comes to mind when looking at these from a few feet away. It's interesting to read some of the discussions regarding the use of DNG8 when the M9 first came out. Lenshacker, I understand that you have compressed/expanded an uncompressed DNG file. What I do not understand is that you substracted the two images instead of replacing each pixel value with the percentage in illumination between the compressed/expanded pixel value and the uncompressed value, because relative difference is what makes it visible or not, but not the absolute difference. And then displaying this file in LR, you should leave the sliders in exactly the same position as with the original image. In your case it looks as if an enormous amount of information has been lost, even magnified by using the LR sliders, but it is not posible to use this to tell how effective the compression algorithm is. The "relative difference" image will give you a completely different and more valid view on what compression/expansion does to a file. Hans Link to post Share on other sites More sharing options...
Lenshacker Posted March 22, 2015 Share #288 Posted March 22, 2015 I stated that the image could be reconstructed from the data thrown away by the compression scheme. dataloss_vs_pixelvalue Here is a plot of bin size vs pixel value. At the high end, the bin size is a count of 127- "almost" 7 bits are being thrown away. The compression scheme increases peak-to-peak noise: noise that occurs on threshold boundaries push the pixel to the next higher step. The scheme reduces average noise, once in the step, is not seen. The M8 was plagued by color noise at ISO2500, which disappears when uncompressed RAW files are used. The scale could have been better distributed. Values below the black-level are pretty much wasted. As all of this is done with look-up-tables, they could have been more efficiently distributed. I have a routine that computes spatial-gradient and some other scene metrics. I'll add it to my list of things to call and compare original with truncated images. For the original question, I got one "guess" wrong looking at frequency content. Apparently, the use of compressed M9 files set it off. Right now, I'm learning how to write DNG files for the custom monochrome conversion using yellow filters with the color camera. Link to post Share on other sites More sharing options...
Lenshacker Posted March 22, 2015 Share #289 Posted March 22, 2015 A "bit" off topic.... TIFF tags A nice, concise table of tags for reading and creating DNG and TIFF files. Another example of the compression loss, this one shows the color noise issue of the compression scheme. Original: L1015636 by fiftyonepointsix, on Flickr L15636DF by fiftyonepointsix, on Flickr For a company that prides itself, and charges a premium, for lenses that have good micro-contrast: this is a really bad way to compress data. Lossless compression using running difference would have been much better. Link to post Share on other sites More sharing options...
t024484 Posted March 25, 2015 Share #290 Posted March 25, 2015 A "bit" off topic.... For a company that prides itself, and charges a premium, for lenses that have good micro-contrast: this is a really bad way to compress data. Lossless compression using running difference would have been much better. I completely disagree with your conclusions. To substantiate your thesis there is only one way to prove it. Show us an uncompressed picture together with the same picture after being compressed/expanded and tell us where the differences can be found. You will have a hard time in finding any differences at all. What you have shown so far does not prove all that Leica's compression scheme is inferior. On the contrary, it is a very clever system that is hard to improve. Hans Link to post Share on other sites More sharing options...
CheshireCat Posted March 25, 2015 Share #291 Posted March 25, 2015 What you have shown so far does not prove all that Leica's compression scheme is inferior. On the contrary, it is a very clever system that is hard to improve. It is easy to debunk this myth. In the M240, Leica finally replaced the DNG8 lossy compression with lossless DNG compression. The average file size is the same and in some cases even lower, but no information is lost. I call this a big improvement. It was not even hard to improve it. This "new" lossless compression scheme had been around since 1993 and used for years by most competitors. Link to post Share on other sites More sharing options...
Lenshacker Posted March 25, 2015 Share #292 Posted March 25, 2015 I completely disagree with your conclusions. To substantiate your thesis there is only one way to prove it. Show us an uncompressed picture together with the same picture after being compressed/expanded and tell us where the differences can be found. You will have a hard time in finding any differences at all. What you have shown so far does not prove all that Leica's compression scheme is inferior. On the contrary, it is a very clever system that is hard to improve. Hans I really don't care if you agree or disagree. I stopped using it on the M8, and the high ISO performance improved dramatically. I do not use it on the M9. I know better. I coded up the running difference lossless algorithm in 1988 for the system that I worked on. Which image looks better? M8_3_F15_ISO2500 by fiftyonepointsix, on Flickr M8_3_DNG8_ISO2500 by fiftyonepointsix, on Flickr Same camera, same exposure, one is compressed DNG the other is not. Can you tell the difference? Link to post Share on other sites More sharing options...
dfarkas Posted March 26, 2015 Author Share #293 Posted March 26, 2015 I really don't care if you agree or disagree. I stopped using it on the M8, and the high ISO performance improved dramatically. I do not use it on the M9. I know better. I coded up the running difference lossless algorithm in 1988 for the system that I worked on. Which image looks better? M8_3_F15_ISO2500 by fiftyonepointsix, on Flickr M8_3_DNG8_ISO2500 by fiftyonepointsix, on Flickr Same camera, same exposure, one is compressed DNG the other is not. Can you tell the difference? The two images look to have an exposure and color difference. I darkened the compressed image slightly to roughly match. Of course, this is correcting form your JPG which is obviously limited, but I think the differences here are minimized. 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/241543-ccd-vs-cmos-can-you-tell-which-is-whichmerged/?do=findComment&comment=2787536'>More sharing options...
Lenshacker Posted March 26, 2015 Share #294 Posted March 26, 2015 The two images are the same exposure, the uncompressed image is shot at base ISO160 and underexposed 4 stops. LR4.4 used to bring the image back up. The DNG-8 ISO2500 algorithm first multiples the image by 16 (Left shift 4 bits) then compresses. You do not get the the full dynamic range to work with. The problem with the DNG-8 compression routine and High-ISO is that it magnifies peak-to-peak noise. M8 M8_4_F15_ISO5000 by fiftyonepointsix, on Flickr and M Monochrom, Monochrom4_F15_ISO5000 by fiftyonepointsix, on Flickr Both at the same exposure, with the M Monochrom at ISO5000 and the same shutter speed and F-Stop as metered on the Monochrom used on the M8. Uncompressed DNG at ISO160 with the M8 and pulled up in post. Default NR in LR. The M8 could have shown ISO5000 performance like this 8 years ago. The DNG-8 algorithm crippled the performance of the sensor. Link to post Share on other sites More sharing options...
Lenshacker Posted March 26, 2015 Share #295 Posted March 26, 2015 Noise Reduction Off, No Sharpening, Histograms as equal as possible pushing the LR "Curves" and Exposure. The DNG-8 shifts the pixels 4 bits, basically kills off the shadow detail as it does. I have not looked at the changes in the table as the ISO is changed on the camera, but will do so. 100% crop From Uncompressed DNG L1015454_Crop by fiftyonepointsix, on Flickr DNG-8 L1015453_crop by fiftyonepointsix, on Flickr DNG-8, full res Jpeg uploaded. L1015453 by fiftyonepointsix, on Flickr L1015454 by fiftyonepointsix, on Flickr I believe the DNG-8 increases peak-to-peak noise, but reduces average noise. The Peak-to-Peak is what stands out in High ISO images, my opinion only. DNG files available if anyone wants them. Link to post Share on other sites More sharing options...
t024484 Posted March 27, 2015 Share #296 Posted March 27, 2015 The DNG-8 shifts the pixels 4 bits, basically kills off the shadow detail as it does. ??? I believe the DNG-8 increases peak-to-peak noise, but reduces average noise. ??? I stopped using it on the M8, and the high ISO performance improved dramatically ??? It is all a mystery to me, but I do not need any more "prove",. Link to post Share on other sites More sharing options...
IWC Doppel Posted April 1, 2015 Share #297 Posted April 1, 2015 Hi The pictures look interesting regarding noise and more importantly avoiding it. I am probably being stupid but if I take high ISO shots on my M9-P at 2500 using compressed and uncompressed settings I can't see any difference ? Link to post Share on other sites More sharing options...
Lenshacker Posted April 1, 2015 Share #298 Posted April 1, 2015 Take a look at 100% crops, especially the shadow areas. I have not looked at compressed output from the M9 above base ISO. I don't know if they omitted the shift operation. I will take a look. I have not used compressed DNG on the M9 in over 4 years. The noise on the M9 is about 1/2 that of the M8. Link to post Share on other sites More sharing options...
IWC Doppel Posted April 1, 2015 Share #299 Posted April 1, 2015 Take a look at 100% crops, especially the shadow areas. I have not looked at compressed output from the M9 above base ISO. I don't know if they omitted the shift operation. I will take a look. I have not used compressed DNG on the M9 in over 4 years. The noise on the M9 is about 1/2 that of the M8. I did look at 100% crops on a 27" iMac screen that has been calibrated, I couldn't reliably see any difference Link to post Share on other sites More sharing options...
Lenshacker Posted April 2, 2015 Share #300 Posted April 2, 2015 Now I'm very curious whether it's the lower noise of the M9 or if the firmware has a new trick in it for suppressing noise when using DNG-8. Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.