When will Lightroom take over PhotoShop?

[Digiteyesed]

Lightroom will never substitute Photoshop, they´re two completely different beasts.
[a href=\"index.php?act=findpost&pid=125306\"][{POST_SNAPBACK}][/a]

I downloaded the trial, looked for layers, couldn't find them, deleted it and upgraded to CS3.

[NikosR]

While I understand and accept that the two programs are developed under different design paradigms, people should not overstress the differences between 'pixel editing' and so called 'parametric editing'. Doing so only causes confusion, is counter-productive and, if I may say so, seems to me to serve purposes other than educating and enlightening people.

After all, in both cases pixels are being edited, aren't they?. There's is a fundamendal difference in that in PS case pixels are edited 'on-the-spot' while in LR's case change instructions are stored as metadata and applied only when an RGB image is saved or exported but, in the end of the day, there's pixel editing going on in both cases.

This has as an implication that there is fundamentally nothing which can be done in so called 'pixel editors' that can't be implemented in a 'parametric' program in a way consistent with its underlying design paradigm. Please note I used the term 'underlying design paradigm' and not 'underlying product positioning decisions'.

People often wrongly bring up, as an example of the different design paradigms,  the difference between 'global' and 'local' changes but this is not a fundamental difference between 'pixel-editors' and 'parametric editors' as programs like Capture NX (a 'parametric' program which can gracefully handle local pixel changes, masking and layer-like blending functionality without distorting its design paradigm) show.

In doing so, NX silently introduces the concept of 'sequenciality' (steps) on top of its parametric nature. It could introduce this concept more elegantly, IMHO, but that means nothing with regard to  the soundness of the underlying 'parametric' principle.

Introducing layers, as we know them from PS, in LR would break away from its design paradigm, but introducing the capability for local changes, masking and layer-like blending, would not.

I find 'parametric' editing a much more elegant way of doing things, and, in fact, I believe that if PS was invented now rather than many years ago it would be 'parametric' as well without compromising its pixel-editing functionality.

It all depends on design decisions heavily influenced, amongst other things, by marketing, product positioning requirements and legacy burdens.

[Schewe]

After all, in both cases pixels are being edited, aren't they?. There's is a fundamendal difference in that in PS case pixels are edited 'on-the-spot' while in LR's case change instructions are stored as metadata and applied only when an RGB image is saved or exported but, in the end of the day, there's pixel editing going on in both cases.
[a href=\"index.php?act=findpost&pid=125342\"][{POST_SNAPBACK}][/a]

Nope...with Photoshop when you open the image, make an adjustment and save the image, those changes are resaved in the file-forever changing those pixels. In Lightroom open an image into Develop and make changes then close Lightroom. Have the pixels in the image changed? Nope. That is a huge conceptual difference.

If you want to work on 100 images in Photoshop, you have to open, change and then save those 100 images (well except now that Camera Raw 4 and Bridge can apply parametric changes to tiffs & jpgs now). In Lightroom adjust a single image and push all those changes to the other 99 images. You didn't have to open, change and save them...you only altered the paramters.

The parameters WILL be applied to an image only upon rendering in Lightroom. That's what makes parametric editing fundimentally different than pixel editing in Photoshop. Sure, Lightroom and Camera Raw share the same approach...they even share the same pipeline. And that is completely different than Photoshop's approach...

[NikosR]

Nope...with Photoshop when you open the image, make an adjustment and save the image, those changes are resaved in the file-forever changing those pixels. In Lightroom open an image into Develop and make changes then close Lightroom. Have the pixels in the image changed? Nope. That is a huge conceptual difference.

If you want to work on 100 images in Photoshop, you have to open, change and then save those 100 images (well except now that Camera Raw 4 and Bridge can apply parametric changes to tiffs & jpgs now). In Lightroom adjust a single image and push all those changes to the other 99 images. You didn't have to open, change and save them...you only altered the paramters.

The parameters WILL be applied to an image only upon rendering in Lightroom. That's what makes parametric editing fundimentally different than pixel editing in Photoshop. Sure, Lightroom and Camera Raw share the same approach...they even share the same pipeline. And that is completely different than Photoshop's approach...
[a href=\"index.php?act=findpost&pid=125348\"][{POST_SNAPBACK}][/a]

I agree, but I fail to see what I have said differently. May be I should have used just the word 'export' or 'render' rather than 'save or export' in my orginal post, but I had in mind also other 'parametric' editors.

I also fail to see where your comment negates what I have said. The fundamental difference (and it is a big one, I agree) is in when / how 'pixel editing' is performed, thus this should have no effect, in principle, on the editing capabilities of a 'parametric' vs. a 'pixel-editing' program. I remind, this what all this thread is all about.

I guess what I'm trying to say is that LR could very well grow to be what PS is (at least for photographers) without compromising its parametric nature. It all depends on what Adobe decide to do.

[dennypage]

Nope...with Photoshop when you open the image, make an adjustment and save the image, those changes are resaved in the file-forever changing those pixels. In Lightroom open an image into Develop and make changes then close Lightroom. Have the pixels in the image changed? Nope. That is a huge conceptual difference.
[a href=\"index.php?act=findpost&pid=125348\"][{POST_SNAPBACK}][/a]

I think that it's easy to overstate the difference between the two approaches.

On a conceptual basis, the differences are minor.  Both PS and LR are step by step image editors.  Both record an ordered set of the steps taken, and both render the result at each step when it is performed.  With both systems, you can go backward and forward in the steps, and each time the result at that given step is rendered and displayed.

The difference between the systems doesn't affect things until you "save" the image.  LR persists the steps taken (discarding the rendered results), while PS persists the final rendered result (discarding the steps taken).  This is a difference, but it is not a huge difference.

There is nothing conceptually in the LR approach that precludes editing of individual pixels.


This having all been said, the PS/LR code bases could certainly be structured such that that appeared to be a huge conceptual difference.  This is often the case, especially in the case of an very old code base vs. a very new code base.  But concepts in the code are not the same as concepts to the user.  

[NikosR]

Just to add that describing PS kind of programs as 'pixel-editors' and contrasting them to LR kind of programs or 'parametric editors' is somewhat misleading and gives the wrong impression of their differences. It is not the case of a  'pixel editor' vs. 'parametric editor'.

Both are pixel editing programs in their nature (as opposed to, for example, vector processing programs).

If one wants to term LR-like operation as 'parametric' and contrast this to PS-like operation, then one has to invent another term for PS-like operation rather than use the term 'pixel-editor'. Maybe 'direct editor' or something to that effect would be appropriate.

The fundamental differences are important, but, from the user's point of view, limited mainly to workflow considerations and backtracking / change management. Not to the amount / level of pixel changes supported. This is just a feature-set decision.

And BTW, how would one define a program like Capture NX? A 'parametric editor with local change capability'?


PS. To draw a parallel, this reminds me of the old Computer Science discussions of 'procedural' vs 'functional' or 'declarative' programming languages.

[DavidW]

Thanks, David. I'll definitely start an expedition into Photoshop CS3 rather sooner than later.

I know Martin Evening's Photoshop books are pretty good. Any others you could recommend?
[a href=\"index.php?act=findpost&pid=125290\"][{POST_SNAPBACK}][/a]

I'm self-taught on Photoshop based on experience with other programs, online tutorials, forums and the commonality across the Creative Suite (there's a fair amount of commonality between Illustrator and Photoshop in the way that certain tools, key combinations and the like work).

If you liked Martin Evening's Lightroom book, then you could try his Photoshop book. I've also heard good things about Scott Kelby's books. Maybe you could go to a good bookstore and look through a few Photoshop books, just to get an idea of the style of them. There's sample chapters from some books online.


I think the differences between Lightroom and Photoshop will become clear when you start to work with Photoshop. As I said, I find using Lightroom as far as I can, then Photoshop when it's needed to be the best combination, and I think you'll find the same.

Having a common Camera Raw engine between Photoshop CS3 and Lightroom means that the settings can go backwards and forwards via XMP sidecars (or DNG files) - which is very helpful.


One of my tasks for today is to post-process a shoot from Tuesday, which was low light no flash work (typically 1/50s at f/3.5, ISO 1600). I'll use Lightroom 1.1 to take things as far as I can - including the new sharpening and Clarity features - then finish off with Noise Ninja and possibly PTLens in Photoshop, before importing the files back into Lightroom. Some have horrible lighting in one corner, which needs correcting - again, that will be done in Photoshop.

This is a perfect example of the synergy available. Lightroom is the richest environment for sorting, culling and ranking, also keywording and captioning. I can do the Develop work in Lightroom, then move to Photoshop for what Lightroom can't do.



Meanwhile - to Nikos and denypage - I know what you mean. I tried to stick to the terminology already in use in the thread rather than introduce more confusion.

I agree that the easiest way to think of Lightroom is as saving an original file and a recipe, rather than the end product of what has been done to it. (If you understand Smart Filters and Adjustment Layers in Photoshop CS3 - that's really what we're talking about). This difference in approach explains, for instance, why the snapshots and undo history in Photoshop are lost once you close the file.


I do, however, see more of a conceptual bar to individual pixel editing in Lightroom than denypage does - because I don't see where there's room to build that functionality into the Lightroom program. Of course, the distinction could increasingly blur (if the cloning features in Lightroom became more general case), but the way Lightroom works with a file and a recipe - and without basic features of Photoshop such as layers - would mean a great deal of code had to be added to support many basic Photoshop features.

To my mind it may make more sense to evolve and refine Photoshop than to re-implement large chunks of it in Lightroom, especially when those who want Photoshop like features probably already own and use Photoshop.


In the end, I think this is a moot point. There's lots that people have already wished for that will keep the Lightroom team busy (the SDK, soft proofing - and much more besides) that hasn't made it into Lightroom 1.1. I certainly look forward to the future!



David

[NikosR]

I do, however, see more of a conceptual bar to individual pixel editing in Lightroom than denypage does - because I don't see where there's room to build that functionality into the Lightroom program. Of course, the distinction could increasingly blur (if the cloning features in Lightroom became more general case), but the way Lightroom works with a file and a recipe - and without basic features of Photoshop such as layers - would mean a great deal of code had to be added to support many basic Photoshop features.

[a href=\"index.php?act=findpost&pid=125385\"][{POST_SNAPBACK}][/a]

You don't need layers per se (as in the PS implementation we are all used to). But for some operations you might need to introduce the concept of serialization or 'sequencing' . Again, Capture NX, with all its idiosyncrasies and failings, is a good example of that.

Now, I have no clue if providing more advanced functionality would need a great deal of code and effort, or not. That is Adobe's problem and Adobe's decision. We are just talking here about if the LR design paradigm allows for more advanced editing features. denypage and I think it does, Jeff and you seem to be advocating that it doesn't for some reason inherent in its design paradigm which I fail to comprehend.

And there is no way anybody can convince me that the ability to do what you guys call 'pixel changes' and I prefer calling localised changes, is not a basic tool in a photographer's armory. Dodging, burning and masking capability, for example,  have been part of the photographer's toolset since the invention of photography and no tool worthy of a title invoking memories of the darkroom should be without those.

[The View]

The fundamental difference (and it is a big one, I agree) is in when / how 'pixel editing' is performed, thus this should have no effect, in principle, on the editing capabilities of a 'parametric' vs. a 'pixel-editing' program. I remind, this what all this thread is all about.

I guess what I'm trying to say is that LR could very well grow to be what PS is (at least for photographers) without compromising its parametric nature. It all depends on what Adobe decide to do.
[a href=\"index.php?act=findpost&pid=125351\"][{POST_SNAPBACK}][/a]

While I see the very different software layouts between Photoshop and Lightroom as they are today, I can see, from NikosR's and denypage's posts, that this doesn't necessarily have to stay that way.

Lightroom saves the pixel image and the changes as a recipe (great, pinpointing expression by DavidW), while Photoshop saves whole images change after change. (Yes, David, I guess when I have started working with Photoshop I will have a different outlook on this application. I see that going with Lightroom as far as possible and then doing specific tasks in Photoshop is a good method).

(On the other hand, coming fresh to a matter can result in "fresh" questions - with its double meaning of new, and also insolent, which can question values that have long been taken for granted, and put them under the light again).

To be fresh again:

I wonder if a software like Photoshop could also be thought as saving only the original pixel image, and then, additionally, only saving pixel changes as a recipe.

And would this bring these two approaches to editing an image closer together in terms of the processing language each of those applications "speak"?

[rck]

I wonder if a software like Photoshop could also be thought as saving only the original pixel image, and then, additionally, only saving pixel changes as a recipe.[a href=\"index.php?act=findpost&pid=125473\"][{POST_SNAPBACK}][/a]

Adjustment Layers? Layer Masks? Layers? Smart Filters? Smart Objects?