ARGB vs PPRGB in Workflow

<p>Hi Francisco,<br>

I'm not familiar with LR so I don't fully understand when you say "because they don't reassign a camera profile once in LR".<br>

I assume that in ACR, when we see the default for camera profile ("Standard") it corresponds to the camera we're using. Is this correct or you consider as necessary to follow any other procedures, please?</p>

<p>Antonio,</p>

<p>This is an issue only with the DNG files exported from DxO optics Pro. When you import them in LR or ACR, they will look with muddy colors. If you then go to the camera profile section and select the one you prefer from your camera, the brigth colors will be restored. (it defaults to an embeded or adobe, don't remember now). For example, I use Nikon and I can select between standard, vivid, landscape, neutral, portrait, etc. </p>

<p>You could even apply a custom profile if you have a ColorChecker. The X-Rite software that creates the DNG profiles will accept the DNG files created by DxO optics pro.</p>

<blockquote>

<p>As a side note, playing around with gamut plots, in my experience the majority of images fall inside the gamut of AdobeRGB</p>

</blockquote>

<p>Raws processed in say an Adobe converter and encoded in ProPhoto? Because I can see lots of my images that show saturation clipping (any color in the histogram) when set to Adobe RGB (1998) in ACR that don’t clip when ProPhoto is set. You don’t need any 3D gamut tools (they are great), just set the rendering as you desire in ACR, toggle the workflow options link from sRGB (ugh) to Adobe RGB (1998) then ProPhoto and watch the histogram update. Of course, much depends on what you shoot (the gamut of the scene). A Bride in a white wedding dress will likely fall in sRGB. A sunset, a field of colorful flowers, a lot of man made objects that are painted? Not so much. </p>

<p>Francisco<br>

I think you're on to something here and I want to understand it. Are you saying that way to preserve ProPhoto colour space when transferring images from DxO to LR3 must be:<br>

1) Export from DxO as DNG <br>

2) Import DNG into LR3<br>

3) After opening the DNG file in LR3, re-apply the camera profile for whatever camera was used to take the image to that DNG file.<br>

It is step #3 that I am not certain about. AND when you do that which set of lens optical corrections is in effect - the one that you applied in DxO or LR3's?<br>

PS: One advantage of this workflow (transferring DNG's instead of TIFFs) is that the DNG is about half the size of the TIFF (for CR2=25MB DNG=70MB vs TIFF=140MB).</p>

<p>Andrew: Agreed, it depends on the gamut of the scenes.</p>

<p>David:<br>

You have the correct steps. About your concerns on step 3: It is only the DNG profile what I'm referring to, I haven't considered optical corrections.</p>

<p>IMHO: The only reason to use this workflow is if you prefer to do the optical corrections in Dx0 or you don't have LR 3 / ACR 6 (Earlier versions of LR / ACR don't have the optical corrections). I see no point in using this workflow if you then want to use the optical corrections in LR/ACR. What will you do in DxO then?</p>

<p>How do you derive a 3D gamut plot from a tiff image whose pixels are already baked in and defined by 0-255 RGB encoding without using a hardware measuring instrument. This defies logic. Last I checked, for me to get a 3D gamut plot of my display I have to calibrate and profile the response of its colors using a colorimeter.</p>

<p>Is this all a guesstimation by the 3D gamut tool software making comparisons on the rate of clipping between converting from source RGB numbers to output spaces of varying sizes and shapes (AdobeRGB or ProPhotoRGB) according to the rate and amount of spikes that show up in a histogram which represents output in ACR/LR?</p>

<p>From my understanding of Raw files which don't have a color space and must be given a source space to define its preview that being linear ProPhotoRGB in ACR/LR, I still don't see how you can get a 3D gamut plot from a tiff with baked in pixels whose colors are now limited by the display's gamut without the use of a hardware measuring device.</p>

<p>It sounds like sloppy color science going on under the hood with a lot of assumptions being made to derive information that looks cool and technical but can't be used or proven to be useful by the photographer or a color scientist for that matter.</p>

<blockquote>

<p>How do you derive a 3D gamut plot from a tiff image whose pixels are already baked in and defined by 0-255 RGB encoding without using a hardware measuring instrument. This defies logic.</p>

</blockquote>

<p>ColorThink Pro can do it. And you can even build a color list from that data (to say create a target). Nothing illogical about it. </p>

<blockquote>

<p>From my understanding of Raw files which don't have a color space and must be given a source space to define its preview that being linear ProPhotoRGB in ACR/LR, I still don't see how you can get a 3D gamut plot from a tiff with baked in pixels whose colors are now limited by the display's gamut without the use of a hardware measuring device.</p>

</blockquote>

<p>In terms of the RGB filters and the spectral response (if known), the data does have the ability to have a color space definition. That’s nothing special. Its when an assumption of the color space has to be made. Everyone (raw converter) does it differently. This assumption and the processing color space (ProPhoto 1.0 TRC in LR) are not the same. </p>

<p>Hello All:<br>

In the workflow I describe (Camera CR2 > DxO DNG > LR3) , I think DxO becomes in effect the camera and the DNG file that DxO outputs incorporates whatever optical corrections were dialed in in DxO.<br>

I have tried an experiment to confirm this and found that:<br />1) Applying camera lens correction in LR3 distorts the image. Conclusion: the DxO lens corrections are embodied in the DNG.<br />2) Camera calibration must be enabled in LR3. I got the best results using Process=2010 and Profile=Embedded. Conclusion: To make full use of the PPRGB data and the colour adjusments made in Dx0, LR3 must be set to reference whatever is contained in the DNG file.<br />3) The DNG files are 50% smaller than TIFFs for same image: Conclusion: DNG is more efficient.<br>

To my eye on my monitor, the that appears in LR3 when the DNG file is used is appreciably better than one based on a TIFF of that same image.</p>

<p> </p>

<p>Tim:<br /> To derive a 3D gamut plot you just need a set of points with 3 coordinates (RGB, Lab, XYZ). If you have a Tiff file, you have that set of points. Another way is to measure an image with a hardware instrument as you mention, but you will first compile the set of points and then derive the 3D plot from it.</p>

<p>As Andrew pointed out (thank you) you could do this with ColorThink Pro. I used Argyll CMS + Cortona3D because they are free open source tools, it would be easier to use ColorThink Pro</p>

<blockquote>

<p>Conclusion: the DxO lens corrections are embodied in the DNG.</p>

</blockquote>

<p>David:<br /> You are correct, maybe I didn't understand your previous post. DxO lens corrections are embodied in the DNG. That is the reason why those DNGs are linear and not RAW. It is required to do first the demosaicing and interpolation to apply the optical corrections. Sorry for not making this clear before.</p>