jaapv Posted September 4, 2014 Share #361 Posted September 4, 2014 Advertisement (gone after registration) Mine has not: a spot check yields 24MB (M) vs 18.4MB (M9).Strangely enough my cameras return the same figures... Link to post Share on other sites More sharing options...
Advertisement Posted September 4, 2014 Posted September 4, 2014 Hi jaapv, Take a look here The New LEICA M-P: Discreet, Faster, Harder. I'm sure you'll find what you were looking for!
pico Posted September 4, 2014 Share #362 Posted September 4, 2014 Using lossy compression on a picture will make it more difficult to process the image because you're bound to introduce more artefacts or because the desired data has gone. These are, admittedly, borderline cases, but I strongly subscribe to the doctrine that "you never know". With respect, the difference between a compressed image and a non-compressed is zero. Compressed simply obviates the unused bits of the so-called non-compressed image. Link to post Share on other sites More sharing options...
CheshireCat Posted September 4, 2014 Share #363 Posted September 4, 2014 Thank you, I do know quite well what we're talking about.The EXIF says the M9 picture is uncompressed, as it should. Technically, after the lossy phase that destroys sensor information (14 bits to 8 bits), the file is not entropy-encoded. Therefore the EXIF info will tag your file as "8 Bits Per Sample, Uncompressed". I do not "import pictures into the computer" and I most certainly do not let any software do it that damages the goods. I merely copy them from the SD card to one of the HDDs. Well, your M9 DNG compression is actually damaging the goods. If you knew how to use software correctly, your pictures would be perfectly safe, they would contain the original raw sensor data, and still be about 18 MB in size. Link to post Share on other sites More sharing options...
CheshireCat Posted September 4, 2014 Share #364 Posted September 4, 2014 Nonsense. For the first four seconds of continuous shooting both cameras have the same speed. After that, the M-P will take more frames per second than the M. "More frames per second" is called "faster". The difference will grow asymptotically smaller, of course, after an additional four seconds. Are you in Leica's marketing ? The M-P is not any faster than the M. It will just shoot at the same M speed for 4 more seconds, and only if the M buffer is full (which won't happen to most users). Link to post Share on other sites More sharing options...
CheshireCat Posted September 4, 2014 Share #365 Posted September 4, 2014 Strangely enough my cameras return the same figures... From the Leica M9 user guide: "File Size: DNG™: compressed 18 MB, uncompressed 36 MB, JPEG: approx. 2–10 MB" Therefore you are also talking about lossy DNG, or you are correctly using your software Link to post Share on other sites More sharing options...
jaapv Posted September 5, 2014 Share #366 Posted September 5, 2014 With respect, the difference between a compressed image and a non-compressed is zero. Compressed simply obviates the unused bits of the so-called non-compressed image. Umm... The compression on the M9 does not do that. It applies the same type of compression as the M8, which is rather intelligent, but not lossless. The M240 employs true lossless compression. Link to post Share on other sites More sharing options...
CheshireCat Posted September 5, 2014 Share #367 Posted September 5, 2014 (edited) Advertisement (gone after registration) With respect, the difference between a compressed image and a non-compressed is zero. Compressed simply obviates the unused bits of the so-called non-compressed image. Nope. There are lossy compressions and lossless compressions. The M9 DNG "compression" is only removing information (actually color shades) from your raw files. They look similar but will be prone to posterization in postprocessing. The M Typ 240 DNG compression is instead using Lossless-JPEG. Lossless-JPEG is a totally different algorithm than classic JPEG and no information is lost. You may find this thread interesting: http://www.l-camera-forum.com/leica-forum/leica-m9-forum/278070-new-m-dng-compression-3.html Edited September 5, 2014 by CheshireCat Link to post Share on other sites More sharing options...
pop Posted September 5, 2014 Share #368 Posted September 5, 2014 I wonder about exactly what information is lost. Can anyone demonstrate the bits thrown out that diminish the outcome? I don't know which is the compression function in the M9. Actually, I can't even recall if the M9 does a lossy compression since I never intented using it. I do know, however, that the M8 reduced the color depth to a non linear function which made nasty gaps in the brightness histogram when trying to restore an image with very little contrast or exposure. In theory, that is. I don't think anyone here was able to actually produce that effect. Still, as I said before, you never know. :-) Anyway, that's off topic. Link to post Share on other sites More sharing options...
jaapv Posted September 5, 2014 Share #369 Posted September 5, 2014 (edited) Nope. There are lossy compressions and lossless compressions. The M9 DNG "compression" is only removing information (actually color shades) from your raw files. They look similar but will be prone to posterization in postprocessing. The M Typ 240 DNG compression is instead using Lossless-JPEG. Lossless-JPEG is a totally different algorithm than classic JPEG and no information is lost. You may find this thread interesting: http://www.l-camera-forum.com/leica-forum/leica-m9-forum/278070-new-m-dng-compression-3.html The compression in the M9 does not lead to posterisation as it only loses significant information in the lighter areas, which will not really be affected visibly. It makes the camera behave like a 12 bit one instead of a 16 (14) bit one. See the extensive discussions on this forum on the M8 in 2007 and the relevant article in LFI at the time. Basically the camera writes 8 bit pointers to a 16 bit LUT. The M9 additionally writes the black point at different levels depending on the ISO. Actually the M8 compression is less destructive. You might care to read for instance this thread: http://www.l-camera-forum.com/leica-forum/leica-m9-forum/101279-compressed-versus-uncompressed.html#post1067232 Edited September 5, 2014 by jaapv 1 Link to post Share on other sites More sharing options...
pop Posted September 5, 2014 Share #370 Posted September 5, 2014 With respect, the difference between a compressed image and a non-compressed is zero. Compressed simply obviates the unused bits of the so-called non-compressed image. That, of course, depends a bit on the algorithm. RLE, Huffman encoding or LZW are well known examples of lossless compression procedures. On the other hand, there are indeed compression schemes which do discard significant data. Link to post Share on other sites More sharing options...
pop Posted September 5, 2014 Share #371 Posted September 5, 2014 The M-P is not any faster than the M. The M-P will - under certain circumstances - take less time to take the same number of frames than the M. Under no known circumstances will it take more time. Calling that "faster" is consistent with normal use of the term. Some users have commented here and elsewhere that they had experienced those circumstances with the M. The gain in speed, albeit limited, would be useful to those users under those circumstances. Link to post Share on other sites More sharing options...
CheshireCat Posted September 5, 2014 Share #372 Posted September 5, 2014 The compression in the M9 does not lead to posterisation as it only loses significant information in the lighter areas, which will not really be affected visibly. It makes the camera behave like a 12 bit one instead of a 16 (14) bit one. It may not be visible if your image is properly exposed, but still it really depends on what you need/want to do in postprocessing. Link to post Share on other sites More sharing options...
mjh Posted September 5, 2014 Share #373 Posted September 5, 2014 (edited) It may not be visible if your image is properly exposed, but still it really depends on what you need/want to do in postprocessing. There is a reason that despite these nagging suspicions, during all those years nobody has ever managed to demonstrate that it actually matters. And that reason is noise: the noise from the sensor acts as a dithering pattern effectively preventing any posterisation that otherwise could occur. Since you cannot get rid of the noise – except by employing algorithms that also eliminate posterisation – this does not really depend on what you do with the image data. The exception is the M Monochrom which consequently doesn’t offer a lossy compression option. The lossy compression algorithm of the M8 and M9 wouldn’t work that well with the M/M-P (Typ 240) either – there just isn’t enough noise. Edited September 5, 2014 by mjh 2 Link to post Share on other sites More sharing options...
mjh Posted September 5, 2014 Share #374 Posted September 5, 2014 I don't know which is the compression function in the M9. Actually, I can't even recall if the M9 does a lossy compression since I never intented using it. M8/M8.2: lossy compression by applying f(x) = sqrt(4x) to the value of each pixel M9/M-E (Typ 220): uncompressed or lossy compression (basically the same method as applied in the M8, apart from the black level issue mentioned by Jaap) M Monochrom: uncompressed M/M-P (Typ 240): uncompressed or lossless compression 4 Link to post Share on other sites More sharing options...
CheshireCat Posted September 5, 2014 Share #375 Posted September 5, 2014 The M-P will - under certain circumstances - take less time to take the same number of frames than the M. Under no known circumstances will it take more time. Calling that "faster" is consistent with normal use of the term. Now it makes sense. I totally missed the "under certain circumstances" sentence in the M-P home page (still can't find it, but I am sure it is written somewhere). Leica calls it "twice as fast at capturing shots in sequence", but it is clearly an understatement. Leica marketing should not be so modest ! They should call it "infinitely faster at capturing shots in sequence", because "under certain circumstances" the M locks up and needs to be power cycled, while the M-P may still be capturing frames at a blazing fast speed of 3.7 fps. 1 Link to post Share on other sites More sharing options...
mjh Posted September 5, 2014 Share #376 Posted September 5, 2014 Leica calls it "twice as fast at capturing shots in sequence" Yeah, the English translation is indeed misleading. The original German version states quite clearly: “Damit ist sie doppelt so gut für die Aufnahme schneller Bildsequenzen gerüstet wie die Leica M und bleibt dementsprechend zweimal so lange schussbereit.” 2 Link to post Share on other sites More sharing options...
CheshireCat Posted September 5, 2014 Share #377 Posted September 5, 2014 There is a reason that despite these nagging suspicions, during all those years nobody has ever managed to demonstrate that it actually matters. And that reason is noise: the noise from the sensor acts as a dithering pattern effectively preventing any posterisation that otherwise could occur. Interesting theory. But noise on highlights should be minimal and most of it removed by the binning operation. I will need to experiment with this. Sorry for the out of topic. Link to post Share on other sites More sharing options...
Paulus Posted September 5, 2014 Share #378 Posted September 5, 2014 They should call it "infinitely faster at capturing shots in sequence", because "under certain circumstances" the M locks up and needs to be power cycled, while the M-P may still be capturing frames at a blazing fast speed of 3.7 fps. You mean the Leica M-P wil not lock up in any circomstance? Link to post Share on other sites More sharing options...
pico Posted September 5, 2014 Share #379 Posted September 5, 2014 That, of course, depends a bit on the algorithm. RLE, Huffman encoding or LZW are well known examples of lossless compression procedures. On the other hand, there are indeed compression schemes which do discard significant data. You are correct, of course. I studied those algorithms back in the dark ages - in assembler! Fascinating stuff, even today. 1 Link to post Share on other sites More sharing options...
CheshireCat Posted September 5, 2014 Share #380 Posted September 5, 2014 You mean the Leica M-P wil not lock up in any circomstance? I really hope that with the new hardware, they worked around that bug. I guess they had to rework the motherboard because of this bug, and while they were at it, they updated the RAM. People upgrading to M-P will let us know if it's any better Link to post Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now