A good long while back I did a comparison of RAW and JPEG as image formats. The conclusion then was that
RAW has suddenly become useful for some cases - when one needs to capture a high-dynamic range scene and can do so by exposing for highlights, can shoot at base ISO, and can't or won't bracket. To that I can add that RAW is very useful for astrophotography, and when the white balance is difficult or you expect to require different white balance settings for different parts of the image.
I still stand by my estimate back then that
this is quite a rare event. But the Raw vs. Jpeg comparison was done over three years ago - and rare as the event may be, over three years you do accumulate a couple of RAW files. About 2,000 to be exact. (Compare that to about 40,000 JPEGs.) Challengers like Google's WebP[a] and now BPG[b] notwithstanding, I expect JPEG to be "good enough" for a good many years more, and readable for a great many years beyond that.
The same thing can't be said for Nikon's raw NEF files. Being Nikon's proprietary file format, they contain stuff only Nikon knows how to interpret and should Nikon go belly up... ...well, hope you interpreted what you needed to interpret by then. So I started looking into converting my NEF files into Adobe's Digital Negative format - DNG.
DNG is sort of an "open raw format". It is supposed to be a superset of all RAW formats, being able to store everything that every RAW format contains in a standard way. The American Society of Media Photographers[d] recommends using DNG for archiving, and the Library of Congress lists DNG as a preferred alternative to RAW[e].
I'm generally suspicious of "supreme data formats". Since they rarely are ahead of the curve when defining data items, they tend to have almost but not perfect ability to absorb all the data of other formats: 99% gets converted OK, but the last 1% is lost. So I set out to see just what DNG would drop on the floor. As a sample, I used a RAW photo of my toy camel.
I started with the out-of-camera NEF.
Then I wrote GPS data to the NEF and to an XMP sidecar file.
I wrote a description and a Microsoft face tag to the XMP sidecar file.
DNG with embedded original file
I extracted the NEF from the DNG with embedded original file.
3. Summary of Results
The image data was handled without any issues. The DNG, when opened in Adobe Camera Raw, yielded exactly the same image data as the original NEF.
The "embed original image data" option also does exactly what it says on the box. The original NEF could be extracted with no changes whatsoever. The sidecar XMP could not be extracted, however.
The metadata was shuffled around a bit. The XMP data was merged with the NEF data and put into the DNG file, with the NEF data having priority in case of data being defined in both. The DNG converter then added some fields of its own. Both the description and the face tags were converted perfectly, something I verified by having
exiftool extract XMP from the DNG.
For the raw data, see the diff viewer below.
Well, gee, I dunno... Having the file format endorsed by the Library of Congress is a good thing, I suppose - but then they also recommend JPEG2000, which makes me a bit nervous as that is one of those "state of the art better than JPEG" formats that failed to go anywhere. Most likely, I'll go for DNG with embedded NEF. If something goes wrong, and I find out that I should have used a different converter, or a different setting, or something else, I'll at least be able to get my NEFs back. Yes, my DNGs will be twice as lovely, but then again I don't shoot much RAW so in the grand scheme of things it won't make that much of a difference.
5. Diff Viewer