OpenRAW

Essential data is documented

What could be a more open format than an ASCII text file? Yet it is obviously possible to write private data in an ASCII text file, encrypted if desired. So what?

What matters about DNG is what information is open, (and so not private and encrypted), and the significance of anything that might be private and even encrypted. This is what makes DNG different from other raw file formats, and gives confidence that our digital images will be preserved with DNG.

The DNG data held in openly documented format comprises the raw image data plus essential metadata needed for processing that raw image data now and in future. It won't be necessary in future to have separate knowledge of the camera model - essential information is held in openly documented form within each DNG file. Armed with the DNG specification, it will be possible to render our images without needing code and/or extra data for each of 100s of camera models.

The DNGPrivateData isn't necessary to render our images. (For example, ACR and Lightroom need only what is in the openly documented part, not what is in DNGPrivateData). That is why DNG still satisfies the objectives of OpenRAW.

Some of the information currently held in DNGPrivateData should ideally have been defined in other specifications, such as JEITA's EXIF specification, but those other organisations haven't caught up with the what camera manufacturers want to tell us about. So it is held in DNGPrivateData rather than lose it, the equivalent of Makernote in EXIF data. Examples include "Firmware", "Lens", "OwnerName", etc. If those other organisations bring their specifications up to date, this information too will become openly documented. (DNG exploits other specifications where they exist - it doesn't attempt to redefine standards unnecessarily).

Other information in DNGPrivateData are the camera manufacturers' "added value". These are such things as "Curves", "Picture Styles", etc. (They could also include things that are irrelevant for rendering images, such as diagnostic information). It is unrealistic to expect camera manufacturers not to add value to their raw files - and many photographers want such features. What we can hope is that camera manufacturers document this information in future, and this is one of things OpenRAW is trying to achieve. If camera manufacturers do so, then those last undocumented contents of DNG files will become documented. For this "added value", DNG is no worse than other formats, and typically is better, because DNGPrivateData has to conform to certain rules to ensure that it doesn't get damaged when DNG files are being handled.

Whether or not OpenRAW succeeds in causing the camera manufacturers to document their raw file formats, DNG will always be better than those other formats for purposes of digital image preservation. Until OpenRAW gets results, DNG files are far better documented than all alternatives, including all essential raw image data and metadata. If/when OpenRAW gets results, those last remaining fields in DNG files will suddenly become documented, and such criticisms will cease. And in both cases, DNG solves the major problem of the proliferation of formats that is just as much a problem as the fact that they are undocumented.

Barry, are you still under the impression that DNG is an open RAW format?

Unfortunately and opposite to your headline above, it MISSES ESSENTIAL information or hides it in private maker notes.

Somewhere you wrote that its not Adobe's fault that they have to wrap private maker notes also in their DNG format, but since they made the marketing decision to allow 'private' information inside their file format, they create a situation which as a bad as the situation camera makers created.

OpenRAW is not in competition with DNG, but even if DNG becomes a standard today, this will not solve the problem of already legacy RAW files. However, with an open documentation, Adobe has the chance to make DNG truly 'open'.

Juergen
--
http://www.openraw.org/

Yes, essential DNG is open

DNG is an open RAW format for essential image data, (the raw image data plus essential metadata). It enables images to be rendered via the open parts. This is shown by several products, (some from Adobe, some not), that can render DNG images without needing specific camera code or data built in. (How else could they do this?) Therefore, it satisfies the objective "Digital Image Preservation Through Open Documentation". It is better than all current alternatives.

DNG allows some non-essential undocumented data to be included. In that it is no worse than current alternatives. And if OpenRAW causes that data to be documented, that last criticism of DNG will be gone. In which case, it will clearly be better than openly documented alternatives, because it will be a common raw format that future software products will be motivated to support.

The decision to have the DNG Converter wrap up the Makernote as DNGPrivateData where it is well behaved, (Canon, Nikon, Pentax, etc), means that it is preserved in case it becomes understood in future. It would have been a mistake NOT to have DNGPrivateData, partly because there would be nowhere to hold the Makernote for the future, and partly because it is unrealistic to expect camera manufacturers not to want to "add value". We want them to use DNG in-camera, as 3 more cameras did in announcements in September 2006.

In some cases, where Adobe know what is in the Makernote, they add the details to the XMP metadata in DNG files, and that is openly documented. For example, Firmware, FlashCompensation, Lens, LensSerialNumber, etc. This offers scope for extracting DNGPrivateData information and turning it into XMP metadata in future.

If OpenRAW doesn't succeed, DNG solves most of the problem of legacy files, as long as there is a DNG Converter for them. If OpenRAW succeeds, it will solve the rest of the problem too.

"DNG without OpenRAW" is much better than nothing. "DNG plus OpenRAW" is ideal. "OpenRAW without DNG" won't exist, because DNG is already being used and will benefit from anything OpenRAW gains, so it isn't worth discussing. They are not competing, and DNG will continue to exist anyway, so OpenRAW should be building DNG centrally into its strategy.

well the thing is i have tried to understand the DNG specification.
though i am "not so experienced", i have noticed one thing-- yes the essential data fields are there. but you can always have related informations written in the private field which will give the converter accurate info to deal with the essential data. without which the quality of the image can be compromised.

i am pretty bad at explaining things. so i beg my pardon.
but anybody please say that i am wrong.
(because i deal only in DNG :( @#$@##@$)

Here is another problem stated http://openraw.org/node/1541