Phocus Reproduction Mode Not Usually Helpful Outdoors?

[BradP]

I'm struggling how best to use custom camera profiles in my new Phocus workflow. 

As background, I am predominately an outdoors, non-studio, mostly landscape photographer.   I recently bought my first Hasselblad (X1D).  Since then, I've become convinced that Phocus is a better place to process RAW files from this camera than Adobe's CR/LR.  The colors are better, the highlight recovery seems better (unless there's exceptionally large DR in the files), lens profiles exist (and appear likely better when Adobe eventually implements them), among other reasons.

What's confusing me is that during the last five years, the very first step in my Canon and Sony post processing workflow has been to apply a Colorchecker Passport profile correction as the very first thing I did in my workflow.  The colors out of both those cameras seemed improved by that, and because a color profile can affect all the RGB channels, it was a logical place to start.  With Phocus and its Reproduction Module, I'm not seeing the same improvement.  In fact, through rereading the manual and what little comments I can find on the internet, trying many different permutations of the settings, everything I do generally makes the files slightly LESS pleasing. 

I'm becoming convinced that either:

(1) The X1D and Phocus factory color profiles are so well tuned that a ColorChecker Passport color correction is no longer worthwhile unless truly critical color correction is needed, like for artwork or documents.  I'd love this to be the correct answer and it's apparently possible.  But my past experience makes me skeptical.   The way I'm accustomed to using the ColorChecker is to create different standard profiles for different lighting/filter conditions.  For example, I typically have been used to creating profiles with the following general parameters and applying the color correction to the scene:  Camera used, lens used, filter used (if any), general lighting condition (e.g., shade, sunlight, tungsten).  I'm actually hoping that creating general profiles like these are no longer necessary because of the color accuracy of the system (except maybe for some of my filters - I suppose I'll have to test all of them, but for the most part they're pretty neutral Singh Ray). 

(2) I'm using the Reproduction Module incorrectly.  I'm almost ready to rule this out as I've been over the instructions so many times and tried it out so many different ways, one of them should have worked.  But there could be a set of instructions and a process I haven't followed.  Is there anything other than the pretty cryptic Phocus manual or announcement releases that describe how to use color profiles in general landscape work?

And/Or

(3) The Reproduction Module and functionality is designed primarily for color critical artwork/document work in a low dynamic range environments using more linear color response curves, not for general use.  I think this may be true generally, although I'm guessing the standard "Response" selection in Reproduction is designed to work the way I would otherwise be accustomed to working.  It may be just that outdoor lighting conditions, coupled with the camera's color accuracy, make my old way of dealing with profiles actually less accurate.

Hope this makes sense.  Any thoughts on the above are appreciated.

[NickT]

(1) The X1D and Phocus factory color profiles are so well tuned that a ColorChecker Passport color correction is no longer worthwhile unless truly critical color correction is needed, like for artwork or documents. 

Brad this would be my take.

I was involved in some very early work on Hasselblad's "Natural colour" approach and it struck me then, and does now, as a very very good solution. I think in a scientific environment one could do better with reproduction mode and input profiles but I'm not that guy.

[BradP]

Thanks Nick.  Your opinion means a lot.  If this really holds up as the right answer, I've saved tons of time and H could have charged me even more for the camera (don't tell anyone!)

[BradP]

Thanks Hassilistic!

I just read your post on the ColorChecker Digital SG (which I've been eyeballing, but it seems difficult to shoot outdoors because of the semigloss), and I've been interested in getting that for some time.  Do you find it that target difficult to shoot outdoors?  I've read it requires a black umbrella and neutral background, so I haven't tried it yet.  Interested in your thoughts.

I also went back through Karl's tutorial, which I saw before, to confirm I was doing everything how he did it.  Karl did everything using the reproduction module on his video, which is how I've approached this so far.  The methodology is as follows:

1. First Karl neutralized the shot of the ColorChecker using the white balance neutralizer tool.

2. This step number 2 is mine, not Karl's, and is based on my reading of the manual and other posts.   In the reproduction module, I changed the color working space to Hasselblad L*RGB, which I've read (but like most things in Phocus am not certain) is a unique Hasselblad defined colorspace similar in gamut to ProPhoto RGB.  Karl left the default Working Space set to Hasselblad RGB (the only other choice there), which I've read might be a smaller colorspace (similar to, but different from, Adobe RGB).  I'll use the larger color space in my workflow even though it eats up storage.

3. Then Karl changed the "Response" curve to "Reproduction".  Here is where I'm having problems.  When I change the response curve to "reproduction" for outdoor shots generally, everything seems to go too flat, like pastels.  If I leave "Response" set to "Standard", then that resulting image suggests to my eyes that it might be the correct setting for applying the profile.  In the shots that I've taken so far, "Standard" does appear to mildly affect the histogram and color rendering, but surprisingly not as much as I'm used to with other camera systems.

4. Then Karl accessed the color calibration function again in the Reproduction Module by clicking on "Edit" in the "Color Calibration" drop down menu.  This appears to provide access to the same menu that can be accessed through the Windows>Color Calibration menu item on top, as you point out.  Using the + sign, one can then create a new color calibration profile (which after created also appears in the Windows>Color Calibration menu on top. 

5. Then Karl selected the color calibration profile he just created in the Color Calibration tab in the Reproduction Module.

6.  Not really a step, but I noticed that Karl in the video left the "Input Profile" set in its default position, "Factory".  Interestingly, there is a choice in that drop down box for "Camera RGB", which I don't understand, other than selecting it seems to disturb the colors negatively.  I guess that this is part of a function that allows one somehow to profile their own particular camera sensor colors independently, but that's just my guess until someone corrects me. 

There's the steps I've been following so far.  The "Reproduction" mode in step number 3 seems to be the wrong choice to apply a color correction for general post processing, and I'm guessing the right choice would be "Standard".  If that's all right, then what I'm finding in three shooting sessions so far is that so long as one has the right white balance and tint selected (by shooting a grey card from time to time) the colors out of the camera and Phocus are quite accurate already, so much so that shooting a ColorChecker Passport has surprisingly little effect for Landscape use.  I've not formed a firm view on that and my experience teaches me to doubt that conclusion, thus my post . . ..  Regardless, I have little doubt that a ColorChecker chart would be helpful under more stable lighting conditions and I may need to do more experimentation shooting the card and measuring color values against the MacBeth RGB values.

[Alex]

Th SG target is quite difficult to capture well because being semigloss it suffers easily from reflections and flare which can lead to the creation of some very misleading target images from which pretty bad quality profiles get made.
In a lot of ways the less reflective standard Colorchecker will give better results out "in the field", because of being less prone to reflections and glare.

1. If you want to forge ahead with calibrating with SG outside then primarily pick a day when the characteristics of the sky are constant (cloud/sun), otherwise you have a nightmare on your hands with shifting colour temperatures.
2. Stage the target against a dark background to minimise camera glare.
3. Position it to face a dark background (not leaning up at the sky)(use a dark bedsheet/background paper on stands - about 3m away), to reduce glare and be large enough to block-out spectral reflections.
4. Set the camera up in front of this background perpendicular to the target with the longest focal length you have so as to further minimise glare.
5. Slightly defocus the lens (at the chosen aperture), to remove dust & particle contamination on the colour tiles.
6. Set Phocus (tethered), to Hasselblad RGB (doesn't really matter for internal profiling), Reproduction Mode Low Gain, and the Colour Readouts to LAB.
7. Shoot and neutralise your Grey card at approximately L=50
8. Set your camera to expose the White Tiles for L=94-96
9. Check the black tiles (they are L=5) so aim to get them below 20 in the photo - if not then check & remedy what's causing the reflections.
10. Create your profile using the date, weather and WB for naming.

Aim to create a profile with a Delta AE Average of <1.5 and a max of <4 (which is good working from a generic reference file)

After selecting your custom made profile in Color Calibration, you can then go back to putting the "Factory curve" under Response, reduce your exposure time back to normal (~-1/3EV) and you'll be good to go.

As an internal calibration this type of color correction is in place before (and therefore not effected by) the RAW processor's tonal response modifications.

Also like scene calibrations, it is embedded (put not applied) within the RAW file and so a portable, easy to share solution.

I am not sure as to what extent the calibration covers the various colour temperatures A - D65, as well as lighting deficiencies FL10 etc: Maybe someone from Hasselblad could chime in with what the ideal choice of color temperatures are for us to create calibrations at so as to cover the full range?

Best of luck,

Alex

[BradP]

Thanks for that Alex.  You have just convinced me to do a bit of research in hopes to avoid all that!

I am thinking now that with the out-of-the box colors with the X1D, I am going to experiment for a while anyway with not doing any color profiling unless I'm (1) reproducing color critical art, products or logos, or (2) shooting in non sunlight, non tungsten environments (e.g., LEDs, fluorescents or controlled studio lighting). In the latter case, I believe that I outlined the correct process to work with color profiling in Phocus above -- until I'm corrected anyway. 

For the last 5 years I've been operating with a working assumption that the colors I got with ColorChecker Passport and Adobe profiles were such that profiling with a ColorChecker Passport always was a worthwhile undertaking.  With the X1D and its natural color solution, I may be rejecting that assumption going forward for most of my work.   We'll see how it goes.  But if true, Yippee!!!

Here's a good old link I just read that, together with observing the results of the X1D, helped me reach that conclusion for those desiring to read further.  If you don't know them, pay attention to the names Shewe, Thomas, Chan and van der Wolf.  They are names I pay attention to anyway.  http://forum.luminous-landscape.com/index.php?topic=62539.msg504308#msg504308

[Alex]

My 0.02$

Hasselblad's out-of -the-box colour is very very very very well designed and one of the main reasons why one invests in their hardware. You're not working within LR's demanding working colour space so can leave a lot of the first aid equipment behind.

Reaching for profiling tools is really either for when you want to sort out some problematic lighting issues, you want colormetric accuracy (and hence why it is so critical to get the target data spot on), or you working within LR RAW file converter.
The thing is if you are going to go to that extent of accuracy then really one should get the individual target measured (which will bring a visual improvement), which leads one to go get software that'll do an even better job of improving.

If you are interested in the Reproduction mode then work with the Low Gain version as it's a good deal more accurate to what the Hasselblad is really capable of (especially with Highlights). As you are working with Landscapes and colour go have a look at Joseph Holmes's Chroma Variation working spaces (probably DCam3 or 4), working with these & luma curves built upon the Repro. Low Gain base will give you some pretty powerful control and results.

Alex

[BradP]

Thanks Alex.  No sooner than I got out of CCP have you made me think about getting into CCDSG.  At least I had eight hours of freedom.  I'm reading his pages now...

As I go through it, my first thoughts are that I can appreciate that his color spaces can be more accurate especially in the highlight/shadow areas, that they might do better in a RAW converter and Photoshop, and that those color spaces might be especially helpful using luminosity curves (as I do) to manage highlight/shadow areas in particular (as those areas typically need the most management IMO and also are most damaged by gamma).  However all the Photoshop plugins I regularly use to manage different things in post (e.g., Zerene, PTGUI, Topaz, In Focus) operate in ProPhoto 16 bit space at best and I have doubts whether I can use those plugins in my normal workflow with his color spaces.  If so, for me, that would likely neutralize the benefit.  I'm completely unfamiliar with using those gamuts, however, so could easily be wrong, so any thoughts you have on that would be appreciated. 

[DenisM]

Quick question re the Reproduction Low Gain....

Should one be capturing with Response/Reproduciton Low Gain already selected? (and giving the exposure the extra that this setting would permit. i.e. 1/3-1/2 stop)

Or does it matter what is selected in Response at the capture stage?

i.e. Can you simply bracket the exposure and the select Reproduction Low Gain at the Adjust stage for the file with the most exposure?

I'm guessing that it can be applied after capture, but I just want to be sure.

I take the points above about using this feature for repro work (which I do a lot of).

But, for what it's worth, in my DSLR photography I have taken to using the equivalent in Capture One for ALL photography. i.e. choosing Linear Response in the Base Characteristics/Curve tab, instead of the default Auto/Film Standard.

Simply put, it totally transforms the elasticity (for want of a better word) of my Canon 5D Mk IV files. No longer am I fighting against the inbuilt curve when processing. I get to manipulate exactly what was captured by the sensor, exactly as I wish.

To maximize what your (very expensive) sensor can do, I would recommend using a linear approach whenever possible. Why accept the software creator's idea of an average "look" as your starting point?

My $00.02 worth.

D.

Ps. Below is a link to a webinar on processing of landscape, using a linear approach, by a (ahem!) rival camera manufacturer. Just about everything in it can be applied to Hasselblad/Phocus. (For what it's worth, I don't like his point about starting with Auto Levels. I prefer to raise the level of the darkened Linear/Low Gain image by using Exposure. But, that can be subject dependent, as well as being a matter of taste.)

https://www.youtube.com/watch?v=tyx-sZB4kVQ&index=16&list=PLBZS3EGK3tQ_r65ab8XMT1gHmKo2_AZNg

[Alex]

Should one be capturing with Response/Reproduciton Low Gain already selected?

Nope it's just the (lack of inclusion of) the response curve and so you can assign it later on in post - however as it's the true indication of what is hitting the chip at said exposure, it's worth ascertaining what exposure compensation one should dial into the body's exposure meter for ones oft used subjectmatter  i.e. grey card

[DenisM]

Should one be capturing with Response/Reproduciton Low Gain already selected?

Nope it's just the (lack of inclusion of) the response curve and so you can assign it later on in post - however as it's the true indication of what is hitting the chip at said exposure, it's worth ascertaining what exposure compensation one should dial into the body's exposure meter for ones oft used subjectmatter  i.e. grey card

As I thought...

Many thanks.

D.

[Alex]

it's the true indication of what is hitting the chip at said exposure

(After ascertaining max poss exposure with Raw Digger)

Low Gain Repro / Linear is a real life saver when ones trying to get the most out of ones CCD sensor.

[DenisM]

it's the true indication of what is hitting the chip at said exposure

(After ascertaining max poss exposure with Raw Digger)

Low Gain Repro / Linear is a real life saver when ones trying to get the most out of ones CCD sensor

Which is precisely my point; why permit an inbuilt curve to influence the dynamic range of your sensor in any way?

D.

[Alex]

Because it's easier (less exposure time) and besides - it's got the look!
https://www.youtube.com/watch?v=_jCuroTbqBI

[BradP]

Ps. Below is a link to a webinar on processing of landscape, using a linear approach, by a (ahem!) rival camera manufacturer. Just about everything in it can be applied to Hasselblad/Phocus. (For what it's worth, I don't like his point about starting with Auto Levels. I prefer to raise the level of the darkened Linear/Low Gain image by using Exposure. But, that can be subject dependent, as well as being a matter of taste.)

https://www.youtube.com/watch?v=tyx-sZB4kVQ&index=16&list=PLBZS3EGK3tQ_r65ab8XMT1gHmKo2_AZNg

Interesting video. Thanks Denis. I had never thought to start out editing with a linear response curve, but when doing heavy editing in landscapes particularly that seems to make a lot of sense.