Differences between ACR and DPP

[Manfred M]

Colin - you oversimply things delightfully. First of all, the camera software is not "free", it comes as part of the camera kit, so you are paying for it. Second of all, I highly suspect that the lens profiles (at least for the lenses that communicate with the camera) are in the software / firmware. I come to that conclusion in how the 5D Mk III handles the colour fringe correction. If you have a modern lens that the camera can read, the fringes are reduced, if not, you have to manually enter the profile into the lens to get that functionality.

Unfortunately, Canon, Nikon, Adobe, etc. do not share their "secret sauce" with members of the photographic community, so we can only speculate what they have done in the background. Personally, I ususally use ACR because it is usually "good enough" for what I do and more convenient from a workflow content.

But, as I have stated before a commercial photographer showed me some of the issues he was having with ACR and (being a Nikon shooter) demonstrated how Capture NX2 gave him much more accurate results. Ray is one of the most experienced photographers I know and spent a number of years as head of photography for Lee Valley Tools, whose iconic catalogues, have become collector's items because of the beautiful photography.

http://www.leevalley.com/en/home/OnlineCatalogs.aspx

http://www.photographyottawa.com/Lee...ey%20Tools.htm

Ray has a very specific workflow of getting images into Photoshop, that I have documented elsewhere in this thread.

All I am saying is that not everyone would agree with your assessment of these tools. Again, you are right, that the standard profiles that the camera / firmware / software produce do have their limitations. I would also argue that your approach has limitations as well; two data points provide you with a linear interpolation in a non-linear environment.

Let me sum up by saying I have seen the work of two different photographers (you and Ray) who have taken a totally different approach in how they import RAW images into Photoshop. Both seem to be convinced that their workflow is best, and have the images to prove it...

[Colin Southern]

Colin - you oversimply things delightfully. First of all, the camera software is not "free", it comes as part of the camera kit, so you are paying for it. Second of all, I highly suspect that the lens profiles (at least for the lenses that communicate with the camera) are in the software / firmware. I come to that conclusion in how the 5D Mk III handles the colour fringe correction. If you have a modern lens that the camera can read, the fringes are reduced, if not, you have to manually enter the profile into the lens to get that functionality.

Unfortunately, Canon, Nikon, Adobe, etc. do not share their "secret sauce" with members of the photographic community, so we can only speculate what they have done in the background. Personally, I ususally use ACR because it is usually "good enough" for what I do and more convenient from a workflow content.

But, as I have stated before a commercial photographer showed me some of the issues he was having with ACR and (being a Nikon shooter) demonstrated how Capture NX2 gave him much more accurate results. Ray is one of the most experienced photographers I know and spent a number of years as head of photography for Lee Valley Tools, whose iconic catalogues, have become collector's items because of the beautiful photography.

http://www.leevalley.com/en/home/OnlineCatalogs.aspx

http://www.photographyottawa.com/Lee...ey%20Tools.htm

Ray has a very specific workflow of getting images into Photoshop, that I have documented elsewhere in this thread.

All I am saying is that not everyone would agree with your assessment of these tools. Again, you are right, that the standard profiles that the camera / firmware / software produce do have their limitations. I would also argue that your approach has limitations as well; two data points provide you with a linear interpolation in a non-linear environment.

Let me sum up by saying I have seen the work of two different photographers (you and Ray) who have taken a totally different approach in how they import RAW images into Photoshop. Both seem to be convinced that their workflow is best, and have the images to prove it...

I give up.

[Glenn NK]

My camera is set to Adobe RGB and Neutral profile Colin.The file is set to RAW and temperature to custom 5600K, I do not know how to change the profile.

Just found out that ACR is fixed on Adobe Standard Profile; which is something else I haven't touched.

This image looked perfectly alright to me until a yellow cast was pointed out; I'm not very good at seeing colour.

I really need to have an accurate colour, because I cannot easily see a colour cast as illustrated here:

Therefore I think I have to trust people to point out colour imperfections.

Steve:

I use Lightroom, but I think it works similarly to ACR.

Generally, I try out several settings: Camera Faithful, Camera Neutral, Landscape, Portrait, and the others which I don't recall right now (my photo computer is off). I find that some images are better with Camera Faithful, some with Camera Neutral, and etc.

The camera profile is my starting point. I've thought about the process of creating camera profiles as Colin described, but haven't bothered to put in the effort.

One thing that one should be aware of is that there are instances (particularly with indoor lighting) where there is more than one light source, and these sources often have different temperatures. In these cases, getting the correct white balance is hopeless it seems. I took some images on a cruise in the dining room, and the first thing I noticed was the variable colour of the lighting - and the images just could not be corrected because of this.

Glenn

[Manfred M]

ACR and Lightroom use the same engine, so results that are the same are what one would expect (assuming the same settings, of course).

I've always found the mixed light settings during "magic hour" interesting, because from a colour balance standpoint you get so many different light / colour temperature inputs, there is no "right answer". The image is a shot I took a few weeks ago at sunset in Venice. I didn't have my tripod along, so I cranked up the ISO and handheld for this scene on the Grand Canal right beside the railway station. I did a custom white balance with white card and then took some shots. The tungsten lights have their typical yellow cast and the fluorescents shoe their distinctive green cast. Lights that pass through coloured glass play different "games".



The different light sources just paint the scene with interesting light.

[Glenn NK]

I've always found the mixed light settings during "magic hour" interesting, because from a colour balance standpoint you get so many different light / colour temperature inputs, there is no "right answer". The image is a shot I took a few weeks ago at sunset in Venice. I didn't have my tripod along, so I cranked up the ISO and handheld for this scene on the Grand Canal right beside the railway station. I did a custom white balance with white card and then took some shots. The tungsten lights have their typical yellow cast and the fluorescents shoe their distinctive green cast. Lights that pass through coloured glass play different "games".

The different light sources just paint the scene with interesting light.

The light temperatures at this time of day (actually at high noon too) can be quite different, but when one critically looks at the scene, the colour differences are quite apparent. Good graphic artists have long recognized this and put a blue tint in shadow areas (for example).

On the other hand, in a room with two or more light sources with different temperatures, the difference in colour cast isn't quite as pleasing, particularly if people's faces are lit with differing light conditions.

My personal take on DPP (Canon) is as follows:

1) it MAY produce better colour than ACR/Lightroom (I will need much more convincing however than anyone can provide in this thread to convince me - and I'm likely in the majority),

2) ACR/Lightroom is getting better with each issue - which I doubt DPP is doing. Iin spite of Canon "knowing" exactly what their sensors produce, there is scant evidence that Canon is better at this than Adobe, for example. I see little evidence in forums that many critical photogs actually use the camera supplied software.

3) DPP is a PITA to use compared to ACR/Lightroom. If a beginning photog asked my advice, I'd suggest not even putting DPP on the computer (the software for DL is useful however). I removed DPP from my machine a couple of years ago. I saw no advantage, but many disadvantages.

Glenn

[arith]

Hi Glenn; DPP chooses the profile as shot I think. When I first got the camera I read the handbook only as far as setting RAW and custom white balance with the neutral profile, which I thought mean't do absolutely nothing, reasonable I thought since the output was to be RAW.

Therefore DPP was giving the more accurate white balance, since ACR updated itself to give choices of profile, proving there isn't a right white balance.

I've absolutely no interest in profiling, no interest in reading how, in fact even the thought of doing it is so mind numbingly painful I would rather chew my arm off. Don't know how to do it, don't want to know how to do it, which is incidently why I don't use Av or P or DEP although I've tried Tv for fast moving critters and even here I would have been better off without it. I don't even use ETTL if I can help it.

I'm suspicious that the only way to profile is with special equipment anyway.

The only thing I want is the certanty that the grey in the image is the same as on the card; it gives me a starting point so that when I'm tone mapping the colours don't stray too far from what I thought was reality.

There is nothing more devastating than putting images up on the web to see someone say I can detect a colour cast; when I worked around a grey card in ACR.

I like DPP because it gives an easier quicker reproduction which is camera and lens specific, but until now preferred ACR because it went straight into photoshop.

[Glenn NK]

Steve:

I believe you are right about DPP choosing the profile from the "as shot" RAW file.

As for accuracy, there is a possible glitch that could occur: the camera may not choose the correct white balance, so your best bet is probably to use the white balance tool. Many scenes have lighting that varies (from open sun to heavy shade for example), so the camera can easily be fooled and get the balance wrong.

Incidentally, I wouldn't put too much faith in critiques about "colour cast" from the public on a forum. I often see comments like "the image is too blue", when in fact the image was made bluer for impact and mood. There are many "techie slaves" to the "perfect white balance" that don't have an iota of creativity and simply don't understand the effect of colour on mood.

Glenn

[John Wiegley]

I have become concerned by the difference in white balance in Adobe Camera Raw v6.6 and Digital Photo Professional v3.9.3.0.

I have frequently had a yellow cast reported and now I'm unsure of the correct balance; but think DPP wins.

I'm surprised no one mentioned this, but just using a color picker shows that many (though not all) of the pixels in the top gray card are neutral (#d0d0d0), while almost none of the bottom ones are (most are #c0c2c1).

I would say then that at least in the JPEGs you uploaded, the top is closer to being correct with respect to the gray card being represented as neutral gray.

John

[arith]

That is a big surprise John.

[John Wiegley]

That is a big surprise John.

You know, since there is color noise in both cards, it was likely just luck-of-the-draw as to which pixel you ended up clicking in both cases. I'd try zooming to 400%, applying some chromatic noise reduction, and then pick a color temperature which makes most of the pixels even out around #d0d0d0 (i.e., 18% gray).

I can't see that _any_ picture development solution, whether ACR, LR, DPP, Photoshop, or any other, would get this wrong. Gray is gray, it's just that sometimes what we click on isn't what we think we're clicking on, due to hidden noise in the image. I often find myself clicking repeatedly within a supposedly white area until it looks like the color cast is gone.

[Colin Southern]

You know, since there is color noise in both cards, it was likely just luck-of-the-draw as to which pixel you ended up clicking in both cases.

ACR samples a 5 x 5 pixel area up to 100% zoom, and then uses 5 x 5 SCREEN pixels after that.