Calibrating Adobe Camera RAW
in Olympus E-System , Friday, September 03, 2004
Following the less than perfect support for the E-1 in PhaseOne CaptureOne, I decided to go back to ACR, and, armed with Bruce Fraser's Real World Adobe Camera Raw, and his earlier Creative Pro article, I tried recalibration ACR for my E-1.
I photographed a 24 patch Macbeth Color Checker in ambient light, downloaded the ProPhoto reference file from Bruce's website, and got going.
The basic results were as shown here:
If you can't read the image, the numbers are:
Green Hue +8 Green Saturation +6
Blue Hue +20 Blue Saturation -15
Red Hue -20 Red Saturation +31
Be aware that you must first use the Adjust tab settings as Bruce describes to reach a neutral white balance. I'm not sure how "portable" these results are - I'd be interested to hear.
Generally speaking the results are not bad at all - far better than my first attempts. I reached them after 3 iterations, and was in the ballpark from the 1st. The biggest problem is that the E-1 reds are very hot - this I believe is typical of all DSLRs, but the E-1's sensor definitely has a different kind of response to others. Not worse, not better, just different. We know it is different in design to, say, the Sony CCD used by Nikon and others, so no big surprise there.
You have to basically try to achieve a good relative balance between the complementary colours on each main R/G/B patch - in other words, sort out the R/B balance for G, R/G for B, etc. I think you have to live with high R everywhere - bringing R down to the "correct" value badly distorts the overall balance. but letting it ride high doesn't seem to matter too much. Getting a good balance is the main thing - global corrections can be made, as ever, on an image by image basis in the Adjust tab.
Two things are worth emphasising: first, before starting, play around with the calibration sliders to understand what they do. The thing to get in mind is that each colour slider affects the other two colours, not itself - hence Red Saturation affects Blue & Green together (increase Red and the other two go down) and Red Hue affects the balance of Blue & Green. As Bruce says, start off with getting Green Saturation right using the Adjust Saturation slider, then balance the Red & Blue in the Green patch using the Green H/S sliders, then move on to Blue, and finally Red. It works, once you get used to it.
The other thing is that Bruce Fraser needs a technical editor. The two versions of his instructions (see above) are not totally consistent. It isn't so much that they are contradictory, more that they both tell 80% of the story. Together they make up the whole. A few other things let him down, for example he advises that there should be no specular highlights in the frame when you shoot the checker, for example a metal clip. Good advice, but what do we see in fig 3-25 ? Yep. a metal support with a specular highlight.
I think that the book has been artificially extended to make it "long enough". There is far too much repetition. But there is also a lot of value, especially on the file browser. Possibly this is repeated in Real World Photoshop CS, I don't know, I haven't read it. But it should be.... Personally I would have preferred to see a more in depth (and far better laid out) section on calibration - for example a cut'n'paste from Creative Pro (!) and a bit less fluff. Maybe this would not have satisfied his editors - technical publishers seem to rate value by weight these days. In that case maybe a self-published, downloadable PDF would have been better (in the style of Digital Outback's guides).
So what is the bottom line on RAW converters ?
Adobe ACR - very deep potential, advanced highlight recovery, direct access to "gurus" (and the guy who wrote it) on Adobe forums, workflow advantages if you're using Photoshop CS anyway, free if you're using Photoshop anyway.... but a Macbeth color check is vital and is not cheap. And time is money, and it takes time to get it right. Nevertheless, you learn a lot of useful information going through the calibration process, which can be applied elsewhere. The Photoshop file browser is highly configurable, and basically is as good as (or better than) anything else apart from the lack of a histogram display. I'd expect it to improve...
Capture One - very nice software, very good workflow, very good browser, a few quirks to get used to, but the pre-conversion crop & rotate, and the background processing are killer features. If you don't mind some issues with reds in extreme cases, have faith that they will fix these in the long run, don't mind being outside the Photoshop workflow, don't mind somewhat uneven technical support, and can tolerate the price, then this the winner at present.
Olympus Studio - some nice concepts, excellent browsing and selection options, good downloader, camera control, "intelligent" lens distortion correction, but horrible, badly documented and limited RAW converter interface, and abysmal performance. On the other hand many people (including me) believe it gives the best absolute results. And only Olympus know exactly how their sensor works. It is also very prone to crashing on my system. I'm not sure about Olympus Viewer, but anyway who need that (at least) for firmware upgrades.
In the long run we'll have to see what Olympus and Phase One come up with. But if they wait too long, ACR has a very good chance become the defacto standard.
If you can't read the image, the numbers are:
Green Hue +8 Green Saturation +6
Blue Hue +20 Blue Saturation -15
Red Hue -20 Red Saturation +31
Be aware that you must first use the Adjust tab settings as Bruce describes to reach a neutral white balance. I'm not sure how "portable" these results are - I'd be interested to hear.
Generally speaking the results are not bad at all - far better than my first attempts. I reached them after 3 iterations, and was in the ballpark from the 1st. The biggest problem is that the E-1 reds are very hot - this I believe is typical of all DSLRs, but the E-1's sensor definitely has a different kind of response to others. Not worse, not better, just different. We know it is different in design to, say, the Sony CCD used by Nikon and others, so no big surprise there.
You have to basically try to achieve a good relative balance between the complementary colours on each main R/G/B patch - in other words, sort out the R/B balance for G, R/G for B, etc. I think you have to live with high R everywhere - bringing R down to the "correct" value badly distorts the overall balance. but letting it ride high doesn't seem to matter too much. Getting a good balance is the main thing - global corrections can be made, as ever, on an image by image basis in the Adjust tab.
Two things are worth emphasising: first, before starting, play around with the calibration sliders to understand what they do. The thing to get in mind is that each colour slider affects the other two colours, not itself - hence Red Saturation affects Blue & Green together (increase Red and the other two go down) and Red Hue affects the balance of Blue & Green. As Bruce says, start off with getting Green Saturation right using the Adjust Saturation slider, then balance the Red & Blue in the Green patch using the Green H/S sliders, then move on to Blue, and finally Red. It works, once you get used to it.
The other thing is that Bruce Fraser needs a technical editor. The two versions of his instructions (see above) are not totally consistent. It isn't so much that they are contradictory, more that they both tell 80% of the story. Together they make up the whole. A few other things let him down, for example he advises that there should be no specular highlights in the frame when you shoot the checker, for example a metal clip. Good advice, but what do we see in fig 3-25 ? Yep. a metal support with a specular highlight.
I think that the book has been artificially extended to make it "long enough". There is far too much repetition. But there is also a lot of value, especially on the file browser. Possibly this is repeated in Real World Photoshop CS, I don't know, I haven't read it. But it should be.... Personally I would have preferred to see a more in depth (and far better laid out) section on calibration - for example a cut'n'paste from Creative Pro (!) and a bit less fluff. Maybe this would not have satisfied his editors - technical publishers seem to rate value by weight these days. In that case maybe a self-published, downloadable PDF would have been better (in the style of Digital Outback's guides).
So what is the bottom line on RAW converters ?
Adobe ACR - very deep potential, advanced highlight recovery, direct access to "gurus" (and the guy who wrote it) on Adobe forums, workflow advantages if you're using Photoshop CS anyway, free if you're using Photoshop anyway.... but a Macbeth color check is vital and is not cheap. And time is money, and it takes time to get it right. Nevertheless, you learn a lot of useful information going through the calibration process, which can be applied elsewhere. The Photoshop file browser is highly configurable, and basically is as good as (or better than) anything else apart from the lack of a histogram display. I'd expect it to improve...
Capture One - very nice software, very good workflow, very good browser, a few quirks to get used to, but the pre-conversion crop & rotate, and the background processing are killer features. If you don't mind some issues with reds in extreme cases, have faith that they will fix these in the long run, don't mind being outside the Photoshop workflow, don't mind somewhat uneven technical support, and can tolerate the price, then this the winner at present.
Olympus Studio - some nice concepts, excellent browsing and selection options, good downloader, camera control, "intelligent" lens distortion correction, but horrible, badly documented and limited RAW converter interface, and abysmal performance. On the other hand many people (including me) believe it gives the best absolute results. And only Olympus know exactly how their sensor works. It is also very prone to crashing on my system. I'm not sure about Olympus Viewer, but anyway who need that (at least) for firmware upgrades.
In the long run we'll have to see what Olympus and Phase One come up with. But if they wait too long, ACR has a very good chance become the defacto standard.
