Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Adobe DNG Profile Editor vs X-Rite Passport Editor  (Read 15090 times)
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« on: November 09, 2009, 08:44:50 AM »
ReplyReply

Micheal's review of the X-rite Passport editor informed us of an interesting new product and was thus useful. However, a complete review compares the new product to existing alternatives, giving the reader guidance which he may choose to heed or ignore. However, the concluding remarks in Micheal's review are not all that helpful:

Quote
How good are these profiles created by Colorchecker Passport? Are they as good as the standard Dual Illuminant ones that Adobe creates and includes with camera Raw and Lightroom?

Likely not, since Adobe uses calibrated light sources and a much larger set of colour patches to make their profiles. But, when compared side by side I find the Passport DNG profiles to be very good, and a couple of colour gurus that I've spoken with are similarly impressed.

In summary, I now regard the xrite Colorchecker Passport as a must-have – a standard part of my field kit. It's a simple as that.

An alternative to using the Adobe supplied profiles is to tweak them with the Adobe DNG editor, which does not make a profile de novo but merely refines existing profiles. Since the Adobe provided camera profiles are made using sophisticated techniques with many patches, one would think that it would be advantageous to tweak such a profile rather than making one anew using only the patches on a Colorchecker. Of course, one could use the Colorchecker Passport with the DNG Profile Editor as well as the x-rite editor. Has anyone done comparisons? One could measure color accuracy objectively using Delta Es or merely use subjective comparison with important memory colors such as human flesh tone, blue sky, or foliage.
Logged
sandymc
Full Member
***
Offline Offline

Posts: 235


« Reply #1 on: November 09, 2009, 09:06:57 AM »
ReplyReply

Quote from: bjanes
An alternative to using the Adobe supplied profiles is to tweak them with the Adobe DNG editor, which does not make a profile de novo but merely refines existing profiles.

Actually, if you use a color patch chart in conjunction with the Adobe profile editor, it pretty much creates a new profile; comparing what's in the original profile to the new one shows that only the color matrixes are preserved. Which I'd think would also be the case for X-Rite. On that basis, I'd think that there would not be much difference at all between the DNG profile editor and the X-Rite product. Of course the X-Rite has the advantage of greater convenience.

Sandy
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #2 on: November 09, 2009, 09:51:19 AM »
ReplyReply

Quote from: sandymc
Actually, if you use a color patch chart in conjunction with the Adobe profile editor, it pretty much creates a new profile; comparing what's in the original profile to the new one shows that only the color matrixes are preserved. Which I'd think would also be the case for X-Rite. On that basis, I'd think that there would not be much difference at all between the DNG profile editor and the X-Rite product. Of course the X-Rite has the advantage of greater convenience.

Sandy
The color matrices are a very critical part of the profile. As I understand things, the DNG profile editor tweaks the matrix profile with lookup tables to correct individual colors. This impression was verified by Eric Chan.

Therefore, your statement that there might not be much difference between the two editors may not be correct.
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8027



WWW
« Reply #3 on: November 09, 2009, 12:09:04 PM »
ReplyReply

Quote from: bjanes
Micheal's review of the X-rite Passport editor informed us of an interesting new product and was thus useful. However, a complete review compares the new product to existing alternatives, giving the reader guidance which he may choose to heed or ignore. However, the concluding remarks in Micheal's review are not all that helpful:

Well at least based on the testing I did with the product and a Canon 5DMII, Michael’s off base, I preferred the Passport profiles to the custom profiles I also built using the Adobe DNG profile editor. It was hardly night and day, but the differences were visible. YMMV (as may Michaels).

And custom DNG profiles using the Adobe product were IMHO, better than those suppled (canned profiles) which you’d kind of expect.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 743


« Reply #4 on: November 09, 2009, 12:31:13 PM »
ReplyReply

Quote from: digitaldog
Well at least based on the testing I did with the product and a Canon 5DMII, Michael’s off base, I preferred the Passport profiles to the custom profiles I also built using the Adobe DNG profile editor. It was hardly night and day, but the differences were visible. YMMV (as may Michaels).

And custom DNG profiles using the Adobe product were IMHO, better than those suppled (canned profiles) which you’d kind of expect.

Andrew, was it your (beta) version that was able to make ICC profiles ? can you, please, post 2 profiles (dcp and icc) made from the same target shot  ? does not matter how good or accurate they are - I just want to see at least what we lost once XRite disabled ICC/ICM generation in the final version of the software... unless posting such examples again violates some NDAs, etc... thank you.
Logged
sandymc
Full Member
***
Offline Offline

Posts: 235


« Reply #5 on: November 09, 2009, 01:17:23 PM »
ReplyReply

Quote from: bjanes
As I understand things, the DNG profile editor tweaks the matrix profile with lookup tables to correct individual colors.

I'm a bit lost - so far as I understand, they both do that.

What's your understanding of the difference?

Sandy
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8027



WWW
« Reply #6 on: November 09, 2009, 01:29:40 PM »
ReplyReply

Quote from: deja
Andrew, was it your (beta) version that was able to make ICC profiles ? can you, please, post 2 profiles (dcp and icc) made from the same target shot  ? does not matter how good or accurate they are - I just want to see at least what we lost once XRite disabled ICC/ICM generation in the final version of the software... unless posting such examples again violates some NDAs, etc... thank you.

I can’t really discuss the beta and ICC profiles due to NDAs.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 743


« Reply #7 on: November 09, 2009, 03:56:08 PM »
ReplyReply

Quote from: digitaldog
I can’t really discuss the beta and ICC profiles due to NDAs.
even the resulting icc profile covered ? damn it !
Logged
seanmcfoto
Full Member
***
Offline Offline

Posts: 176


« Reply #8 on: November 09, 2009, 06:27:45 PM »
ReplyReply

I don't have a Passport, but I do have a mini Color Checker. I was able to download the X-Rite software by creating an account on their website, which worked with the Color Checker. Maybe that's an option?
Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 8027



WWW
« Reply #9 on: November 09, 2009, 08:21:43 PM »
ReplyReply

Quote from: seanmcfoto
I don't have a Passport, but I do have a mini Color Checker. I was able to download the X-Rite software by creating an account on their website, which worked with the Color Checker. Maybe that's an option?

You don’t need anything more really. X-Rite decided to make the software free since you already purchased a Macbeth color checker. They could have (not should have, sorry) made software that only worked with a new target you had to buy. If you have a 24 patch color checker, the Adobe and X-Rite solutions don’t cost you a dime.
« Last Edit: November 09, 2009, 08:22:28 PM by digitaldog » Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
neil snape
Sr. Member
****
Offline Offline

Posts: 1432


WWW
« Reply #10 on: November 13, 2009, 02:01:03 PM »
ReplyReply

Both are going about correcting colors by modifying the color matrix. The Passport is fully automated, easily transported, easily learned without any understanding of manually editing color patch values.

It is a great move on X-Rite's part to have the software work on the regular checker as well. It doesn't have the nice extras the Passport chart does but the profile part should be pretty much the same.

I did make my custom camera calibrations in ACR/ with various charts but they were not as reliable nor easy as the passport.

For example this afternoon I had to shoot some products, with mixed variable daylight and flash. It was so easy to throw the Passport in the picture which picked up all the surrounding reflections and made accurate color a snap. That is what it is about, how you get there now has more than one way.
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #11 on: November 13, 2009, 05:09:01 PM »
ReplyReply

Quote from: neil snape
I did make my custom camera calibrations in ACR/ with various charts but they were not as reliable nor easy as the passport.

For example this afternoon I had to shoot some products, with mixed variable daylight and flash. It was so easy to throw the Passport in the picture which picked up all the surrounding reflections and made accurate color a snap. That is what it is about, how you get there now has more than one way.
You don't say if you made your previous calibrations with Bruce Fraser's manual method, an automated script (such as the Fors script) or the DNG editor. If you think that Passport profile is more reliable, can you post some data? The Digitaldog preferred the DNG editor profiles. Michael was uncertain, but assumed that the Adobe profiles would be more accurate.

I posted some testing that I did in another thread. I did not find the Passport profile superior nor did Erik Kaffehr (who also posted in the referenced thread).

The ability to include the Passport in the scene is an advantage. Is the miniature Colorchecker in Passport mounted on durable plastic or merely cardboard? Can it stand up to field use?
Logged
Alan Goldhammer
Sr. Member
****
Online Online

Posts: 1573


WWW
« Reply #12 on: November 13, 2009, 07:29:12 PM »
ReplyReply

Quote from: bjanes
The ability to include the Passport in the scene is an advantage. Is the miniature Colorchecker in Passport mounted on durable plastic or merely cardboard? Can it stand up to field use?
Everything is in a hard plastic clam shell case.  Go to the website and see Seth Resnick's video on its use:  http://www.xritephoto.com/  I've had it out in the field for a month now and it is extremely useful.
Logged

neil snape
Sr. Member
****
Offline Offline

Posts: 1432


WWW
« Reply #13 on: November 14, 2009, 02:23:51 AM »
ReplyReply

Quote from: bjanes
You don't say if you made your previous calibrations with Bruce Fraser's manual method, an automated script (such as the Fors script) or the DNG editor. If you think that Passport profile is more reliable, can you post some data? The Digitaldog preferred the DNG editor profiles. Michael was uncertain, but assumed that the Adobe profiles would be more accurate.

I posted some testing that I did in another thread. I did not find the Passport profile superior nor did Erik Kaffehr (who also posted in the referenced thread).

The ability to include the Passport in the scene is an advantage. Is the miniature Colorchecker in Passport mounted on durable plastic or merely cardboard? Can it stand up to field use?



Did I say it was more accurate?

It is more reliable, and for me that is what counts. Yes I used both Bruce Fraser's recommendations and Eric Chan's. In fact Eric helped me with the use of the SG Color Checker which is probably the root of the variability.  Many others have tried with the SG, it has proven not reliable.

I simply do visual comparisons on calibrated monitors, the HP Dream Color being the more reliable as the extended gamut covers the Color Checker patches very well.
Using the Adobe Profile Editor introduces more variability , thus your results will differ depending on your input. If you are careful there is no reason why the results cannot be just as good as PassPort, which is why I said which ever way you get there is up to you, yet the PassPort is the easier route, faster, more fun. That said if the chart is not photographed correctly , the results will be less than optimal in any case too.
Logged
sandymc
Full Member
***
Offline Offline

Posts: 235


« Reply #14 on: November 14, 2009, 03:05:20 AM »
ReplyReply

Quote from: neil snape
Both are going about correcting colors by modifying the color matrix.

Not the case for the DNG profile editor. The matrixes are unchanged; what is written is a new HueSatMap table.

Sandy
Logged
neil snape
Sr. Member
****
Offline Offline

Posts: 1432


WWW
« Reply #15 on: November 14, 2009, 03:18:24 AM »
ReplyReply

Quote from: sandymc
Not the case for the DNG profile editor. The matrixes are unchanged; what is written is a new HueSatMap table.

Sandy
Is not the HSL adding a power curve on the matrix?
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #16 on: November 14, 2009, 07:02:42 AM »
ReplyReply

Quote from: neil snape
Did I say it was more accurate?

It is more reliable, and for me that is what counts. Yes I used both Bruce Fraser's recommendations and Eric Chan's. In fact Eric helped me with the use of the SG Color Checker which is probably the root of the variability.  Many others have tried with the SG, it has proven not reliable.

I simply do visual comparisons on calibrated monitors, the HP Dream Color being the more reliable as the extended gamut covers the Color Checker patches very well.
Using the Adobe Profile Editor introduces more variability , thus your results will differ depending on your input. If you are careful there is no reason why the results cannot be just as good as PassPort, which is why I said which ever way you get there is up to you, yet the PassPort is the easier route, faster, more fun. That said if the chart is not photographed correctly , the results will be less than optimal in any case too.

I'm not certain how you are using the terms reliability and accuracy . In statistics, reliability is getting the same results on repeated measurements and is essentially the same as precision. Accuracy is how close the measured value is to the actual or true value, and the results may vary from measurement to measurement. In practice, one wants both accuracy and precision (reliability). A reliable method may give the same incorrect results on repeated measurements.

Are you just looking at the results on your calibrated monitor, or are you actually making measurements (i.e. how the rendered patches differ numerically from what they should be)? If you are just looking at the results, then you are not being that scientific. Imatest allows both methods. Shown below are my results for profiles made with Passport and the DNG profile editor. I would advise opening both images on screen and placing them side by side. The DNG profile has more accurate flesh tones (patch 2) and foliage tones (patch 4). The DNG profile editor gives the user the opportunity to further tweak the profile. I did not test reliability by repeating the measurements multiple times.

Of course, in many instances one wants pleasing color, not accurate color. Back in film days, landscape photographers switched from the more accurate Kodachrome to the more saturated Velvia. However, for skin tones Kodachrome or Astia was preferable. For landscapes with my D3 I often use the Adobe supplied Camera Landscape profile since it pumps up saturation, especially the greens. I could use the DNG profile editor to make my own landscape profile, but have found no need to do so. With the Passport, this would not be possible.

[attachment=17918:090904_M...t_colors.png]

[attachment=17919:DNG.png]
« Last Edit: November 14, 2009, 07:05:11 AM by bjanes » Logged
neil snape
Sr. Member
****
Offline Offline

Posts: 1432


WWW
« Reply #17 on: November 14, 2009, 07:52:27 AM »
ReplyReply

Yes I did say any comparisons I do are on screen, not measured. Accuracy would be defined by measured values for which I don't need to do. I don't believe any of these profile editors are made for accurate colour repro from scene rendering anyway, rather targeting a closer precision for output.

Reliability comes from the simple notions of varying light and reproductions of the colour I see of thins like make up the all important key for what I do.
Problem is with Canon raw images the reds are way out of reality compared to what your eye sees. With the Beta1+2 shipping calibrations in LightRoom there was a huge improvement in all areas including the reds. The advantage of the DNG editor or Passport is the ability to customise your camera and light to bring the output closer to your expectations.
While I found the DNG edits to perform very well for that condition, I find the calibration/profile to not work as well in varying light, which perhaps is the same in Passport, but as I said, you simply and easily shoot a card for each set up and it's pretty much an automatic and the new profile makes the set up reliable.  Now you could do so in DNG editor for each and every set up and why wouldn't the results be as good?

Hope this shows a little into where I see making a workflow more reliable, which is closer to what I said in that both will get you there , which you use is up to you.

PS. I didn't work on the beta/alpha of this chart, I did have some knowledge of it's existence, and it's works, but I'm not at all an expert in the math, numbers etc .

I am pretty sure that Adobe and X-Rite worked on this together, hence my assumption that the underpinnings lead to a similar result or goal inside the frameworks.


I just looked at the two images above. Mind you the profiles were stripped out> but if I tag them as sRGB, I see the same type of stuff between PassPort, DNG, Beta1+2 profiles etc.

The PAssport has better blue green, cyans, The DNG edit profile closer reds pinks and orange showing a more true (less saturated) colours closer to my PassPort chart which is different of course to my other charts....
« Last Edit: November 14, 2009, 08:05:45 AM by neil snape » Logged
sandymc
Full Member
***
Offline Offline

Posts: 235


« Reply #18 on: November 14, 2009, 09:04:02 AM »
ReplyReply

Quote from: neil snape
Is not the HSL adding a power curve on the matrix?

Don't know what you mean by power curve in this context, sorry.

Sandy
Logged
TheSuede
Newbie
*
Offline Offline

Posts: 16


« Reply #19 on: November 14, 2009, 11:22:18 AM »
ReplyReply

I'd say both are better than the "Adobe Standard" - which should come as no surprise. AS is geared towards maximum usability under many different conditions, and can therefore not be as "sharp" in the huecorrections as a "one-condition-only" profile. I do however prefer the Adobe Profile Editor, as the passport software very often does exactly what the examples that bjanes showed above; oversaturate and undercorrect. Hues are slightly more off (than the Adobe solution), and to saturated. Do also note the difference in exposure curve, I find the Adobe more natural, and the Passport to contrasty.

But it has also been mentioned that accuracy isn't always the ultimate goal, "visually pleasing" is also a factor to be added into the equation. But also here, I find the Adobe solution preferable - but that's highly personal preference based, not a case of "right" or "wrong"! Others may find they like output from the passport software more.

What I check in a profile is mostly that similar hues are placed at CORRECT DISTANCES from eachother, after this comes colour saturation per hue in "equal-saturation" slices of the gamut, then brightness per hue. I find this prioritization order to give the most freedom in post-processing choices (some would call it "processability") and usability/flexibility. Only in controlled shoots (mostly product/studio work) is it worthwhile to aim for very "sharp" or "accurate" correction. The sharper you correct, the worse the "worst case" errors get when your lighting conditions deviate from the conditions the sample shot was taken under.

There will be more products available for automatically tuning both icc and dcp/dng profiles in the near future (one launch that I know of should be around February). People outside the "product/corporate" sector are starting to realize that they spend thousands of dollars on colour accuracy in screen/print but next to no though at all of what they put IN to the workflow. Most people that I know (or have heard from) are very unanimous in the feeling that after they started to use calibrated input into their workflows they have actually LOWERED their workload in PP - not increased it.

Thank you for your work with the dcp/dng "hue-twist" question btw, Sandy - I've learned a lot from it!
Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad