Hello, I just converted my raws to the DNG format. Because I prefer to browse my images in XnView and not in a heavy application like Lightroom, the only reason for my move were the embedded preview JPEGs that (in DNG) can contain the adjustments I've done. My XnView is set to display the embedded 1:1 preview image, so after the conversion I can now browse 'pictures' instead of 'negatives'. Fantastic.
It would be extremely useful to have such an "adjusted-preview" feature for JPEGs as well. Unfortunately, the JPEG thumbnails can't be used for this, and if a JPEG is converted to DNG, the resulting file is approx. 7x larger than the original JPEG. The original JPG-compressed file is thrown away and replaced by a compressed TIFF in the container, making the whole idea of converting JPEGs quite absurd. My Fuji F30 unfortunately never supported RAW...
I propose to create a new image format - let's call it JNG - that would be a simple container that would contain the original JPEG and *could optionally* contain an adjusted JPEG as well. The user would then set his image browser to prefer to display the adjusted JPEG (if available) similarly to choosing between rendering actual RAWs or displaying JPEG previews for RAW files.
As opposed to DNGs 7x size, the proposed JNG would only be max ~2x larger, wouldn't force you to convert your original JPG to a compressed TIFF, and for non-adjusted files, it would actually be the same size as the original JPEG (1x)!
What do you think?
An alternative would be to update the DNG standard to allow complete compressed JPEGs files (i.e. including metadata, for easy extraction) to act as 'raw' data in a DNG. This would be quite complex, because the spec would e.g. have to resolve the duplicate metadata issue (DNG EXIF would have to point to or carbon-copy the JPEG EXIF etc.). Creating "JNG" spec from scratch would be incomparably simpler in that regard. A guy like Phil Harvey could have it done in a day (and avoid stupidity
http://www.sno.phy.queensu.ca/~phil/exiftool/standards.htmlhttp://www.sno.phy.queensu.ca/~phil/exiftool/commentary.html