Non-destructive edit applications and future viability

[jpringle]

So all these cool, non-destructive apps are starting to appear - even those that will handle JPGs (Lightview Lightzone, Capture NX, Lightroom). I'm excited by the possibility, but am also concerned about the future viability of images processed by such apps.  Even if a "neutral" file format were used (such as DNG), I doubt that the processing instructions from, say, Capture NX would be honored or even recognized by Lightroom or Lightview Lightzone.  So that means once you start making adjustments in that application, those (processed versions of) images are tied to it.

I suppose you could argue that doing edits in photoshop and saving as PSD is the same thing, but I expect some variant of PS to be around for quite some time (or have clear migration paths) as Adobe is a giant in the business.  I have no such faith in the smaller companies (it's a cutthroat market).

What thoughts do folks out there have on this topic?

--Jason

[pvonk]

There will be two categories of photo software: the image editing apps and metadata writers that write edit code into such formats at DNGs, and the viewers that display the edited image files as part of their workflow (like DAM software).  Of course there will be the apps that straddle both fields.

I agree with you Jason, the giants like Adobe will call the shots on formats and the smaller companies will continuously scramble to keep up.  That's one reason I didn't look at Aperture too closely (aside from the original price) - Apple is not an image processing company like Adobe.  Of course I don't consider Apple small.  The really small companies will struggle, but I'm sure there will be those that carve out a sustained niche market with clever applications.

- Pierre

[johnbeardy]

Jason

You raise a very good point and I also agree with you. I'd predict three categories though - the "One Ring..." programs that lock your editing instructions in their database and black interfaces, those that write it openly in the XMP metadata, and those that hide it away in the raw file and misuse US or EC copyright law to protect our work as if it's their proprietary data. I guess you can see which of the three I prefer and think is more consistent with DAM principles.

It's the same as my analogy of DAM with serial monogamy. Assume the relationship is not going to be for life and you need a prenup that guarantees you can take your metadata with you when you leave - then you're going to be much more comfortable with starting off life together. Except with raw processors we are going to be a touch more polygamous. Are we really going to have to reset all our WB's or white point / exposure every time we shack up together with our next favoured processor, or just have a passionate 30 day trial that shortens those hotel room nights but ultimately leads nowhere?

As for smaller companies, one model is to exploit open standards for recording editing instructions, then focus on their own strengths. For instance, I think it would really help Lightzone to become a great editor of DNG files and save their editing instructions into the DNG metadata (their RT version hides them in tifs). I've encouraged them to do so and given them contacts with people at Adobe who could resolve any difficulties they may have.

Anyway I've got a date with NX tomorrow, maybe doing a review of it soon, and portability of editing metadata is something I think many reviewers will overlook. As you think, its absence will hurt us in the long run.

John

[jpringle]

John -
I agree with your points, but (not knowing too much about the DNG spec) I would suspect that many of the processing instructions generated by these apps would only make sense to that app unless that "way of thinking about the image processing" became standard across image processing apps.  For example, how would one represent the effect of NX's control points in some portable manner?  However, I'd think things like whitepoint, overall exposure, etc should be generally portable across applications.

I'd settle for some intermediary "universal" format that the results of a set of instructions could be translated to, but typically wouldn't be used by an application otherwise.  Something like you could take a tweaked image and translate the app-specific instructions into an equivalent set of layers, filters and masks within a TIFF (for example).

I suppose another view one could take is the "do I really care" attitude - most photos will either be in a "virgin" state where it doesn't matter, or in an adjusted state where you probably only care about the resulting image and could save that as a TIFF/JPG/somefuturecooltype when ready to retire your application of choice.  If you were to reprocess that image years later you'd probably want to start over from scratch anyway with your new wiz-bang application.

Anyway I've got a date with NX tomorrow, maybe doing a review of it soon, and portability of editing metadata is something I think many reviewers will overlook. As you think, its absence will hurt us in the long run.

Cool! - it would be great to get more feedback on NX. Much of the forum traffic on it has been (IMO) too narrowly focused on it as a RAW converter - the whole editing approach (control points) seems really slick and a huge timesaver for 90% of the types of adjustments one wants to make.

--Jason

[peterkrogh]

DNG solves most of these problems: either now or in the near future.

Different applications will create editing instructions stored in proprietary syntax, so don't expect the Instructions to make sense to other applications.  But you can also store a rendered version (soon, hopefully to be multiple rendered versions) inside the DNG file.
This should let you move through applications, porting your previous renderings with you.

The biggest mistake Nikon is making with Capture NX is that you will be married even more to the NEF file, which gets changes without notice or warning or documentation.  I think they could have a huge winner on their hands (assuming the software works right) if they would only store the adjusted files in DNG.  This will particularly important for people who want to use the software with images from other cameras.

Peter

[johnbeardy]

"I suppose another view one could take is the "do I really care" attitude ... you'd probably want to start over from scratch anyway with your new wiz-bang application."

Yes, probably true. As well as starting from scratch with new applications years later, one returns to images with new moods and more editing skills - or greater expectations of the possible. I'm not sure where one draws the line though - for instance a high ISO shooter might want default noise settings to be portable. DNG already does this, but the NEF would be much more acceptable if Capture NX held its settings like the adjusted White Point as unencrypted XMP - it would then be possible with today's technology to script that over to ACR. So it's back to my prenup argument - providing an exit door means people are more likely to buy into the program in the first place.

John

[johnbeardy]

As for the demo of Nikon Capture NX today, I'll just say the guy was demonstrating it on a laptop he'd only had for 3 days and had only seen the program for the first time yesterday. So he was reading his Powerpoints word by word, didn't know where his demo images were, and didn't know much about what the software could do. What's more, the other attendees had already contrived to spend nearly an hour asking questions about the already-obsolete Nikon View Pro. All I can say is NX's selection-based raw adjustments looked pretty damn slick but we only saw them for 5 minutes. When he went back to Powerpoint, I left for home.

John

[Marc Sabatella]

While it is true that DNG addresses the issue nicely for RAW files, at least when using Bridge/ACR or other RAW processing programs that can preserve the adjustments, I personally am excited by the possibility of non-destructive *JPEG* editing.  Of course, I realize that the edit instructions are going to be stored in some proprietary format and will be useless from an interoperability standpoint.  I fully expect I'd have to "export" my edits and create new JPEG (or other format) file before migrating.  In fact, I'd probably go ahead and create a full-resoultion, minimally-or-lesslessly-compressed "proof" immediately upon doing my initial editing (equivalent to the batch-oriented ACR adjustment that one might otherwise do in the initial image processing workflow) as part of my regular workflow and save it as a derivative.  This would be my "pretty good print", and I'd hope to use version control or some kludge using other metadata to keep it associated with the original, so I really was reaping most of the workflow advantages one would otherwise get only by shooting RAW.

The advantage of the non-destructive nature of editing on the original would be that if then decide to create a more carefully edited "master", I could do so from the original file and have the edits I had already made in creating my proof already be applied as a starting point.  I could keep going back to the original each time I wanted to make any more customized derivative, and have all the processing applied creating the master there by default.  I could save the derivatives (including the master) as JPEG's and not worry about generation loss from the compression.  In short, it would basically give me much of the same DAM capabilities with JPEG that one currently gets only with RAW and specifically with DNG.

I know, at some point I'll give up this reluctance to shoot RAW.  If Pentax starts compressing their RAW format so it isn't 3-4 times larger than their highest resolution JPEG, that might be the clincher.  But as someone who really wouldn't be taking advantage of most of the image-quality benefits of shooting RAW most of the time (I would probably be using the "As Shot" adjustments the vast majority of the time), non-destructive JPEG editing would be quite welcome.

[peterkrogh]

Marc,
Expect DNG to be able to store adjustments to JPG and TIFF files in the future.
Peter