mjh Posted September 10, 2007 Share #61 Posted September 10, 2007 Advertisement (gone after registration) It might be quite a bit more than twice as long - the camera would have to apply all vignetting corrections to 16-bits files as well, how long that would take with the limited computing capability af a digital camera is anybodies guess. But as far as I can tell, they are applying the corrections to the 16-bit values, the compression to 8 bits being the last step in the internal processing. Trying to apply any kind of correction to the 8 bit values is a bad idea. Quote Link to post Share on other sites More sharing options...
Advertisement Posted September 10, 2007 Posted September 10, 2007 Hi mjh, Take a look here LFI exposed?. I'm sure you'll find what you were looking for!
sandymc Posted September 10, 2007 Share #62 Posted September 10, 2007 The result is a 8-bit format requiring less storage space than a 12-bit format, but giving superior results in the critical shadow areas, similar to a 14-bit format. Well, not really - the table below is the number of levels per EV zone, for uncompressed 12 and 14-bit data, Nikon NEFs, and the M8 DNG. The only place the M8 is better than 12-bit data is in zone 0, which is really black. Note BTW, the kink in the NEF compression - from black up to about EV6.5, Nikon preserve as much data as they get out of the A/D converter. After that, they start compressing. Sandy 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! Quote 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/32890-lfi-exposed/?do=findComment&comment=350466'>More sharing options...
Julius Bjornsson Posted September 10, 2007 Share #63 Posted September 10, 2007 Could you please explain the table. Your 14 bit values only go up to half the values in a 14 bit space which should go up to 16384 and the Leica column I simply do not understand at all, does it mean that the Leica DNG only has 74 differing luminosities? I must be misunderstanding something here, please explain. JKB Quote Link to post Share on other sites More sharing options...
scott kirkpatrick Posted September 10, 2007 Share #64 Posted September 10, 2007 The M8 works best with coded lenses so that it can do its internal magic on the files before writing the DNGs. It would take more computing power, for sure, to do all those corrections at a higher bit level. I think that when Leica first realized the IR problem, they must have thanked their lucky stars they had both coding and 8-bit RAW files in place. Indeed. In the discussion of "what Leica must be doing" that i read when CornerFix was being developed it was concluded that writing time constraints had forced the color vignetting corrections to be applied in the form of modified encoding tables, that depended on how far from the center of the image each pixel is. You'd think that all these factors would just be parameters in some formula that is evaluated pixel by pixel, but that would take many seconds per frame and thus the correction must be pre-computed. Then the problem becomes finding enough space in the firmware for all these tables. Tables which map raw data at any resolution into an 8-bit space require 256 Bytes each, and you can have hundreds of them. Tables which map 12-16bit raw data into a 14-bit space require 32KB apiece, and that gets big very quickly. The reason this direction, which is actually rather clumsy, is needed is that Leica uses available microprocessor and controller chips rather than designing powerful ASICs for image processing. ASICs are probably too expensive for the market size that Leica can hope for. And they take time to develop and more time to change as the needs evolve. scott Quote Link to post Share on other sites More sharing options...
sandymc Posted September 10, 2007 Share #65 Posted September 10, 2007 Could you please explain the table. Your 14 bit values only go up to half the values in a 14 bit space which should go up to 16384 and the Leica column I simply do not understand at all, does it mean that the Leica DNG only has 74 differing luminosities? I must be misunderstanding something here, please explain. JKB Julius, the table shows the number of levels per EV step (or zone), so, for the 14-bit table, the total of all the steps will be 16834, not any individual step Sandy Quote Link to post Share on other sites More sharing options...
Peter Branch Posted September 10, 2007 Share #66 Posted September 10, 2007 .......... The result is a 8-bit format requiring less storage space than a 12-bit format, but giving superior results in the critical shadow areas, similar to a 14-bit format............ A 16-bit format would give you still better resolution of the highlights that aren’t blown, but you don’t need that. OK. All of this I understand – but – it does not address the issue. I and it seems others with plenty of experience, find that manipulating the highlights in Leica M8 DNG files, despite their richness of highlight data, results in significantly more problems than manipulating the shadows. The shadow recovery is unquestionably very good. Indeed we have had reports on this forum of people deliberately underexposing by up to -2EV to force the highlight data into the region where it is possible to control the result. So if “16 Bit” is not the answer to this dilemma what is? BTW I'm not persuaded that the shadows are more critical than the highlights perhaps that is the problem. Quote Link to post Share on other sites More sharing options...
rosuna Posted September 10, 2007 Share #67 Posted September 10, 2007 Advertisement (gone after registration) Indeed. In the discussion of "what Leica must be doing" that i read when CornerFix was being developed it was concluded that writing time constraints had forced the color vignetting corrections to be applied in the form of modified encoding tables, that depended on how far from the center of the image each pixel is. You'd think that all these factors would just be parameters in some formula that is evaluated pixel by pixel, but that would take many seconds per frame and thus the correction must be pre-computed. Then the problem becomes finding enough space in the firmware for all these tables. Tables which map raw data at any resolution into an 8-bit space require 256 Bytes each, and you can have hundreds of them. Tables which map 12-16bit raw data into a 14-bit space require 32KB apiece, and that gets big very quickly. The reason this direction, which is actually rather clumsy, is needed is that Leica uses available microprocessor and controller chips rather than designing powerful ASICs for image processing. ASICs are probably too expensive for the market size that Leica can hope for. And they take time to develop and more time to change as the needs evolve. scott This is a very interesting and complete explanation. Thanks. The Leica M8 is a very simple camera. It doesn't have autofocus or a complex metering system. It doesn't have to handle 10 frames per second. Etc. I suppose a simpler electronic board can do the job. Maybe is a problem of processor's power and/or internal memory. Maybe Leica can update these specifications easily... Quote Link to post Share on other sites More sharing options...
mjh Posted September 10, 2007 Share #68 Posted September 10, 2007 The shadow recovery is unquestionably very good. Indeed we have had reports on this forum of people deliberately underexposing by up to -2EV to force the highlight data into the region where it is possible to control the result. But doesn’t that simply mean that the deliberate “under-exposure” is dragging the highlights back within the sensor’s dynamic range? So that it is not so much a matter of under-exposure but of exposing for the highlights? Because under-exposure doesn’t improve the rendition of highlights that would have been resolved even without under-exposure; rather it gets worse (see Sandy’s table above). Quote Link to post Share on other sites More sharing options...
mjh Posted September 10, 2007 Share #69 Posted September 10, 2007 For illustration, here is the comparison between the linear 8 and 12 bit formats and Leica’s compressed 8-bit format that I had used in my LFI 2/2007 article. 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! Quote 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/32890-lfi-exposed/?do=findComment&comment=350704'>More sharing options...
ChrisC Posted September 10, 2007 Share #70 Posted September 10, 2007 .......BTW I'm not persuaded that the shadows are more critical than the highlights perhaps that is the problem. Likewise. Me too. My name is Chris: I have the same affliction. ........... Chris Quote Link to post Share on other sites More sharing options...
rosuna Posted September 11, 2007 Share #71 Posted September 11, 2007 About the price of flash memory cards... The Online Photographer: The Cost of Flash Memory R. 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.