RAW processing

[asharpe]

Hi folks,

It occurs to me that I don't know how much of my editing in PL is in "raw" mode with a raw image. With Photoshop, you use ACR, and then import into Photoshop. With PL, the image is already in PL. So, is all the editing I'm doing in PL already past the "raw" stage, and the image has already been converted (to perhaps tiff)? Do I need to use something like RawTherapee or UFraw to stay in raw mode for processing, or is that what PL is doing until I save the image into another format? I obviously would like the answer to be the latter, of course...

[bkh]

So, is all the editing I'm doing in PL already past the "raw" stage, and the image has already been converted (to perhaps tiff)? Do I need to use something like RawTherapee or UFraw to stay in raw mode for processing, or is that what PL is doing until I save the image into another format? I obviously would like the answer to be the latter, of course...

In PL (current version), you normally work with the raw data. The actual raw conversion takes place in the two adjustment layers which PL creates for you (one is the child layer of the adjustment layer, the other is above the background layer). You can set some conversion options in the Layer Attributes of the bg layer. If you save a .pld file, it will contain the raw data, and subsequent processing will also be like in a raw converter.

Modify the existing adjustment layers and/or add your own, and do your retouching on a new layer, and you'll have a workflow similar to other raw converters. Be sure not to change the background layer itself, otherwise it will be converted to a normal RGB layer.

(Most of what normal raw converters do is "normal" non-destructive image processing anyway – it takes place after the image has been converted to a normal RGB image in a standard colour space.)

Cheers

Burkhard.

[photoken]

I've always thought that it was kind of silly for Photoshop and PaintShopPro to have separate "RAW Editor" windows. In the case of PL, since it's full complement of powerful editing tools work on 16-bit images anyway, once the RAW data has been de-mosaicked (converted into a viewable image), you might as well simply use its full editor on RAW images.

I got into a raging argument about this over in the PSP user forum, but none of the "editing window" advocates were able to demonstrate any visible benefits from using a separate editor. Having said that, I've since discovered that there are two areas where it is beneficial to apply corrections as far upstream as possible: Chromatic Aberration (CA) correction, and highlight recovery/adjustment.

I'm a huge fan of RawTherapee, and with regard to CA correction, it has two CA correction tools: one works on the RAW data before demosaicing and the other works on the image after it has been demosaiced.

When the new version of PL is released, it will use the latest version of the "dcraw" RAW image tool to process the RAW data, and that will mean updated processing algorithms for various cameras. In any case, it's up to you to decide whether you need to add RT to your workflow -- try it and see if it's appropriate for your images and their ultimate usage.

[asharpe]

I'm a huge fan of RawTherapee, and with regard to CA correction...

I work in black and white, CA is not a problem.

The actual raw conversion takes place in the two adjustment layers which PL creates for you (one is the child layer of the adjustment layer, the other is above the background layer).

Hmm. I only see one adjustment layer named (Image) and the background layer when I open a raw file.

Be sure not to change the background layer itself, otherwise it will be converted to a normal RGB layer.

Ah. I have been spotting in the background layer; that's obviously not a good idea. But if I make a duplicate layer of the background, it is already converted to RGB, isn't it? So any further processing won't be in "raw" 'mode, will it? That seems to be a conundrum. Perhaps I should spot at the very end of my processing? But spotting after sharpening seems quite odd, too...

[bkh]

I

The actual raw conversion takes place in the two adjustment layers which PL creates for you (one is the child layer of the adjustment layer, the other is above the background layer).

Hmm. I only see one adjustment layer named (Image) and the background layer when I open a raw file.

You should see a small triangle (Mac) or [+] before the layer – like a group. If you click on the symbol (or use "Enter Gruop" from the menu), the child layer becomes visible. EDIT: if you have specified a camera profile in the colour management settings, then there is no such layer because you are working in the camera's colour space.

Ah. I have been spotting in the background layer; that's obviously not a good idea.

Not a very serious problem, either (you lose a few conversion options in the layer attributes). The main disadvantage, imo, is that PLD files get much larger that way.

But if I make a duplicate layer of the background, it is already converted to RGB, isn't it?

A duplicate layer is still a RAW layer (until you start editing it, that is) – but again, it becomes a normal RGB layer when edited.

For retouching, I just create a new (transparent) layer above the background layer (ussing the "New Layer" menu command). If you use the Copy Brush for retouching, make sure you set it to "merged copy"; the "Remove Brush" (my preferred tool) just works as expected. I use this technique for other images as well, because it's easy to undo changes (just erase them) and to do before-after comparisons. A transparent layer with just a few changes doesn't need much storage space, either.

Cheers

Burkhard.

[asharpe]

For retouching, I just create a new (transparent) layer above the background layer (ussing the "New Layer" menu command). If you use the Copy Brush for retouching, make sure you set it to "merged copy"; the "Remove Brush" (my preferred tool) just works as expected. I use this technique for other images as well, because it's easy to undo changes (just erase them) and to do before-after comparisons. A transparent layer with just a few changes doesn't need much storage space, either.

This is simply brilliant. I have no idea if I could have done this with Photoshop or not, but it doesn't matter anyway, since I no longer use it. But I have been using the remove and copy brushes already to do spotting, and the transparent layer over the background works perfectly. Thanks!