coz Posted September 9, 2009 Share #1 Posted September 9, 2009 Advertisement (gone after registration) I see DPreview's review showing 14bit capture as opposed to Leica's brochure of 16 bit? Any thoughts? Link to post Share on other sites More sharing options...
Advertisement Posted September 9, 2009 Posted September 9, 2009 Hi coz, Take a look here 16 or 14 bit?. I'm sure you'll find what you were looking for!
nugat Posted September 9, 2009 Share #2 Posted September 9, 2009 Frankly, Leica claim of 16 bit is hard to believe. Must be a typo. Link to post Share on other sites More sharing options...
Jamie Roberts Posted September 9, 2009 Share #3 Posted September 9, 2009 Looking at the file size and the DMR's in the past, I do believe that the uncompressed RAW file is written as a full 16bpp file, and it wouldn't surprise me if the AD converter was 16bpp as well. That doesn't mean the AD converter is going to use all those bits. So it might be effectively 14bpp, as it is with the current Nikons and Canons. On the other hand, Leica is the only one using bit depth for compression (that I've seen), so perhaps the M9 is inheriting technology from the S2... Link to post Share on other sites More sharing options...
georg Posted September 9, 2009 Share #4 Posted September 9, 2009 16bit files from a 14bit ADC because of 12 stops DR - just like other "16bit"-cameras. Link to post Share on other sites More sharing options...
luigi bertolotti Posted September 9, 2009 Share #5 Posted September 9, 2009 16bit files from a 14bit ADC because of 12 stops DR - just like other "16bit"-cameras. Me too think that the "16 bit" statement has to be interpreted like that - and is a no-problem by sure. Link to post Share on other sites More sharing options...
scott kirkpatrick Posted September 9, 2009 Share #6 Posted September 9, 2009 Leica claimed for a while that the M8 had 16 bit pixels. Look to Kodak to answer that question. The Kodak Imaging Systems Products pages have a press release posted for their new KAF18500 chip, the one used in the M9, but it says little that wasn't in today's press conference. There should soon be a full technical spec (several pages of pdf) that will contain the true potential dynamic range of the chip and the true IR sensitivity of the new cover glass and microlens overlay structure. scott Link to post Share on other sites More sharing options...
sandymc Posted September 9, 2009 Share #7 Posted September 9, 2009 Advertisement (gone after registration) Have looked in the DNG - 14 bits Sandy Link to post Share on other sites More sharing options...
Jamie Roberts Posted September 9, 2009 Share #8 Posted September 9, 2009 Have looked in the DNG - 14 bits Sandy ah well. Have you seen an uncompressed one yet? Would that matter? Link to post Share on other sites More sharing options...
sandymc Posted September 9, 2009 Share #9 Posted September 9, 2009 ah well. Have you seen an uncompressed one yet? Would that matter? No, haven't seen an uncompressed one. Unlikely to matter - different data widths depending on compression would be quite a bizarre thing to do, and confuse the %^&^ out of every raw developer except for Adobe - but possible it could make a difference. We'll see!! Somebody out there got an uncompressed DNG? Sandy Link to post Share on other sites More sharing options...
carstenw Posted September 9, 2009 Share #10 Posted September 9, 2009 The statement in the literature is: "choice of uncompressed or slightly compressed (by non-linear reduction of color depth)." That at least theoretically leaves the possibility that the uncompressed version might be 16 bits all the way through... Link to post Share on other sites More sharing options...
luigi bertolotti Posted September 9, 2009 Share #11 Posted September 9, 2009 The statement in the literature is: "choice of uncompressed or slightly compressed (by non-linear reduction of color depth)." That at least theoretically leaves the possibility that the uncompressed version might be 16 bits all the way through... But I seem to have red DNG uncompressed=36 MB, DNG compressed = 18 MB : this makes me think that the DNG compressed is a sort of "M8 emulation mode" with similar lookup table to have 8 bit (and obviously, 18 MP=18 MB with 8 bit, as in M8) Link to post Share on other sites More sharing options...
mjh Posted September 9, 2009 Share #12 Posted September 9, 2009 Given the 6.8 µm pixel pitch, shot noise alone ensures that a 16 bit ADC wouldn’t be worth the effort. But as 14 bit values will be processed as 16 bit words it would take some effort to get rid of those two superfluous bits prior to writing the raw data to the card. That would explain why uncompressed raw files take up 18 MP x 2 byte = 36 MB. Applying some kind of lossless LZW compression would also remove insignificant bits, but the M9, lacking the Maestro processor, may not be up to this task – just like the M8 before. Link to post Share on other sites More sharing options...
sandymc Posted September 10, 2009 Share #13 Posted September 10, 2009 That at least theoretically leaves the possibility that the uncompressed version might be 16 bits all the way through... I've gotten my hands on an uncompressed file and (as expected) it is 14-bit, I'm afraid Sandy Link to post Share on other sites More sharing options...
scott kirkpatrick Posted September 10, 2009 Share #14 Posted September 10, 2009 I wouldn't make too many assumptions based on the fact that the pixel pitch in the M9 is the same as in the M8. For some more information on the chip (KAF-18500), see the Kodak blog entry. The claim is that the changes in the chip beyond making it bigger include redesigning the pixel cell. scott Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.