What is the truth about Phase One camera profiling?
I recently got a Sony A7RII, after using a Sony A7R for three years.
Somehow, I cannot get satisfying results in CO1, colors never look really right, and towards highlights, images look harsh and tend to look overexposed while the rest of the image really needs the exposure.
So after some attempts to create my own profile, without the wanted result, I began to get suspicious about the Sony A7RII profiles, as I had some issues with copied profiles brought over to new camera models with Pentax in the past.
And it does indeed appear, that the two Sony A7RII profiles, "generic" and "standard", are exact copies of the Sony A7R profiles "generic" and "standard", because switching between these profiles does not change look, color or tonality one slightest bit.
What does this really implicate?
Clearly, the profile for the Sony A7RII has not been made using the Sony A7RII camera, but the Sony A7R camera, which features a totally different sensor.
What are the actual implications on this ommitance of actual profiling, and simply using a previous profile?
Does it mean that Phase one has had word from Sony, that they have calibrated the A7RII to deliver the exact same output as the A7R? Surely not.
The Sony A7RII is a high end full frame camera, and yet it seems that it is not worth profiling individually?
Is there other software around that does justice to the Sony A7RII output?
Chris
Somehow, I cannot get satisfying results in CO1, colors never look really right, and towards highlights, images look harsh and tend to look overexposed while the rest of the image really needs the exposure.
So after some attempts to create my own profile, without the wanted result, I began to get suspicious about the Sony A7RII profiles, as I had some issues with copied profiles brought over to new camera models with Pentax in the past.
And it does indeed appear, that the two Sony A7RII profiles, "generic" and "standard", are exact copies of the Sony A7R profiles "generic" and "standard", because switching between these profiles does not change look, color or tonality one slightest bit.
What does this really implicate?
Clearly, the profile for the Sony A7RII has not been made using the Sony A7RII camera, but the Sony A7R camera, which features a totally different sensor.
What are the actual implications on this ommitance of actual profiling, and simply using a previous profile?
Does it mean that Phase one has had word from Sony, that they have calibrated the A7RII to deliver the exact same output as the A7R? Surely not.
The Sony A7RII is a high end full frame camera, and yet it seems that it is not worth profiling individually?
Is there other software around that does justice to the Sony A7RII output?
Chris
0
-
[quote="ChrisM" wrote:
I think from the two screen shots posted above it is truly glaringly obvious
you were alleging that with A7R there were no big issues (with P1 OEM profiles for A7R and C1 default settings) and with A7R2 there are big issues (with P1 OEM profiles for A7R2 and C1 default settings) - so you are challenged to show that with a pair of raw files of the same scene and equal exposure, not even your files... so ?[quote="ChrisM" wrote:
I am not going to make an effort to prove anything
😂
may be it is simply because you can't ...[quote="ChrisM" wrote:
And if anyone seeing the two screenshots posted above, doesn't understand what I am talking about
we do - a stream of consciousness, a flow of adjectives, no effort to prove your point in a serious manner0 -
Just to keep this thread to the point 😄 :
I took the effort to (visually) compare the Sony A7R and A7RII profiles, and found that the generic profiles are not totally the same: in the blue/purple region they differ somewhat (for the good). The standard profiles however, appear identical.
Also curve implementation seems to be identical, and whether it is Sony's raw format (I use uncompressed), CO1's curves or the limitation of the new sensor: where I could almost always get a good distribution of tonal values with the A7R from a single shot, even in high contrast outdoors, with only some shadows/midtones adjustment and some highlight correction, with the A7RII it seems that I have to start bracketing or heavily use all sorts of edits regarding base curve, shadows/midtones slider, curves, highlights slider, just to get it to look like a decent shot when contrast is medium or high.
PS camera raw offers a better starting point because it starts of with better shadows and less highlight clipping (see the screenshots below of default settings), and is by nature easier to use for getting a natural look out of outdoors images from the Sony A7RII. It is possible in CO1 with e.g. the settings in my former post, but it is hard to keep it looking natural and balanced. When it gets so critical because of the obvious limitations of a sensor, then it has proven that only a linear curve starting point can prevent images being ruined by high contrast and clipping in outdoor daylight.
I now really wonder what this "wide dynamic range" of the Sony A7RII is, that is praised in reviews. It certainly has not turned out to be an outdoor daylight camera. At least not with CO1.
Chris
http://www.abnovembris.nl/PSvsCO.jpg0 -
It's really difficult to tell from typical screen shots or processed jpgs, etc. but I would say that the C1 image has more detail in the dark areas (wooden bench in the shadow area for example) and possibly in the light areas too but I can;t see the image comparison on the right which is more relevant to to discussing the white end of the DR.
However, in the end it comes down to personal choice for preferences.
I usually find I am compressing some DR away even with the outputs from my ancient bodies.
My eyes, looking around a scene "live", auto-adjust from area to area but to see a smaller version of the whole scene in a single view I seem to prefer the adjustment to be applied to the whole in advance - which normally means some "flattening" of available DR. Much the same as would be delivered by a printer for example. However I tend to compress across the range, not just at the ends of the range, depending on the light available for the image when taken.
That may not be everyone's preference of course.
Grant0 -
[quote="SFA" wrote:
It's really difficult to tell from typical screen shots or processed jpgs, etc. but I would say that the C1 image has more detail in the dark areas (wooden bench in the shadow area for example) and possibly in the light areas too but I can;t see the image comparison on the right which is more relevant to to discussing the white end of the DR.
However, in the end it comes down to personal choice for preferences.
I usually find I am compressing some DR away even with the outputs from my ancient bodies.
My eyes, looking around a scene "live", auto-adjust from area to area but to see a smaller version of the whole scene in a single view I seem to prefer the adjustment to be applied to the whole in advance - which normally means some "flattening" of available DR. Much the same as would be delivered by a printer for example. However I tend to compress across the range, not just at the ends of the range, depending on the light available for the image when taken.
That may not be everyone's preference of course.
Grant
I know it is awkward judging from screenshots, but still, looking at the shadow areas in the CO1 snippet for a while, and I quickly see (even on my laptop screen), that something is not right: the shadows clip into below 0 black harshly. That is less so in the PS sample. I'm sure everyone has set in the sun and looked at the shadows: they always gradually transit into black. In the CO1 sample, it looks as if somebody has painted over the shadows with black ink..
Of course taste comes into it, but you have to have a workable starting point, and remember: it's easier to compress a tonal range, than to decompress it.
I'm running into unexpected limits regarding shooting the A7RII in broad daylight, and having dragged the midtone (brightness) and shadows slider around to extremes, I can see that the detail is still there, and noise is kept low. But the starting point is more often than not littered with blue and red overlays, if you know what I mean. Perhaps the Sony can not handle the contrast of the Zeiss f2 lenses, especially the Zeiss 25/2 is really terrible outdoors, where I liked it on the Sony A7R.
So...
Because I noticed the striking similarity between the Sony A7R and the Sony A7RII profiles (especially the " standard" profile is visually totally identical), I raised doubt about the quality of the A7RII profiling, since in the camera color profile, there is also a certain amount of contrast applied.
But if that is not the case, than the Sony A7RII is a body that needs extensive manipulation of the midtone and shadows areas ór bracketing like the Canon bodies do.
Chris0 -
[quote="ChrisM" wrote:
Clearly, the profile for the Sony A7RII has not been made using the Sony A7RII camera, but the Sony A7R camera, which features a totally different sensor.
Unfortunately, I have a similar problem with the Pentax K-1.
Phase One prides itself of creating individual support for each supported camera. On 5 April 2017, Ulf Liljegren from PhaseOne wrote"All cameras are individually optimized by a special/unique Phase One technology process."
A blog states that"...camera support starts when Phase One receives a new model camera at their headquarters in Copenhagen, Denmark. There is no way around that; no camera, no support. ... Next, data collection starts. Images are taken both in a controlled lab environment as well as outdoors. "
However, the camera profile in C1 for the Pentax K-1 appears to be simply the standard Pentax profile that is embedded in out-of-camera Pentax DNG files. I see a number of issues with this profile and while one could argue that this is Pentax's fault and not Phase One's, I expected Phase One to provide an alternative, i.e., a high-quality profile that they designed.
I know that I can tweak a profile with the C1 colour editor and/or that I can use software to create my own profiles. However, creating a quality profile that will be useful with a variety of illuminants is not an easy task. I have bought C9 and C10 (upgrade price only) and believe it is not unreasonable to expect C1's promise of providing individual support for each supported camera model to be fulfilled.
As Chris pointed out there have been issues with Pentax camera profiles in the past (with the K-5 requiring an update that came in the form of a V2 version), and the K-3 initially having to work with a plain K-5 profile. I don't mind these issues and highly commend Phase One for eventually providing great profiles for these camera models.
I hope that some day, the same will be done for the Pentax K-1 (that appears to enjoy much better than expected sales).
I have lodged a respective support case in August 2016. I realise that the Pentax user base is not large and that Phase One needs to prioritise tasks. On the other hand, I'm not asking for anything that Phase One did not promise themselves.0 -
Class A.
I'm not sure that Phase would have much to offer for a previous process embedded in a DNG file.
I might be wrong in my understanding but I get the impression that to do much with DNG's starting at the source level the source RAW data need to be available to the Phase process. Once processed to DNG (in camera) that might no longer be the case?
Grant0 -
[quote="ChrisM" wrote:
[
So...
Because I noticed the striking similarity between the Sony A7R and the Sony A7RII profiles (especially the " standard" profile is visually totally identical), I raised doubt about the quality of the A7RII profiling, since in the camera color profile, there is also a certain amount of contrast applied.
But if that is not the case, than the Sony A7RII is a body that needs extensive manipulation of the midtone and shadows areas ór bracketing like the Canon bodies do.
Chris
Chris,
If I recall correctly the "Standard" profile offered is the in camera profile for those cameras that include such a thing.
If so that would suggest that Sony use the same profile for both bodies.
That said I would not suggest that you should take my word for that. I think there was a post that related to the subject some months back. But if it cannot be found I would suggest raising a Support case to ask the question.
On the other hand looking at your samples (I have now seen the lighter samples on the right) I'm not really seeing huge differences in the results and one might consider the differences that do seem to be there as marginal - some potentially better for a purpose on one interpretation, other better on the other but overall I can't see either being a significantly better starting point for processing based on the evidence available, although I suspect I could get a little more out of the C1 interpretation eventually.
Grant0 -
[quote="SFA" wrote:
Class A.
I'm not sure that Phase would have much to offer for a previous process embedded in a DNG file.
I might be wrong in my understanding but I get the impression that to do much with DNG's starting at the source level the source RAW data need to be available to the Phase process. Once processed to DNG (in camera) that might no longer be the case?
Grant
No Grant,
The color profile in the Pentax DNG is not applied yet, the DNG still contains all the raw data. The raw converter (like Adobe camera raw) can give you the option to use the embedded profile, but should still also give you their own profile, like Adobe does.
Phase oen may have extracted the color profile from a Pentax K1 dng and converted it to an .icc profile and call it Pentax K1 generic, but I don't have a K1, and I have my hands full with my own suspicion of "conspiracy" on the Sony A7rII 😉
Chris0 -
I got around comparing the Sony A7R and A7RII profiles, since raising the suspicion of "not up to standard" profiling may have me chased out of the Phase one community with tar and feathers 🤓 🤓 🤓
So here is a direct comparison of both the A7R & A7RII "generic" profiles and the A7R & A7RII "standard" profiles.
Both profiles are overlayed, one is displayed in wire frame, the other as a semi transparent color drawing.
First the A7R and A7RII standard profiles. As is visible, both are 100% identical":
http://www.abnovembris.nl/Sony%20A7R%20A7RII%20standard.jpg
So no doubts whatsoever: both profiles are one and the same. Meaning what? That Sony made the sensor output of the A7RII 100% identical to that of the A7R? Or were differences too small to justify individual profiling of the A7RII?
Then the A7R and A7RII generic profiles. Two images here, as they show two things: the Sony A7RII profile is indeed also based on the A7R profile, because 80% of the gamut is also identical between both profiles, and individual profiling of the A7RII should at least show some differences.
My take is, that the A7RII generic profile is the A7R generic profile with a modification to slightly shrink the gamut in the blue/ purple region ( The A7R has the larger gamut), as I already saw when visually comparing both profiles:
http://www.abnovembris.nl/Sony%20A7R%20A7RII%20generic.jpg
http://www.abnovembris.nl/Sony%20A7R%20A7RII%20generic_2.jpg
Again, I can only state that with CO1, the advantage over other raw converters can be superior color. But again it shows, that that advantage is only reached when the camera is individually profiled.
My experience with the revised Pentax K3 profile as well as with the Sony A7R profile was one of superior color in CO1, so there was not even a doubt about using CO1 as default raw converter.
Now two months with the Sony A7RII has me continously struggling to get "superior" color out of CO1. In fact, I find I am liking DxO and PS Camera Raw better, and try to get the same output out of CO1, but need endless tweaking going nowhere.
So I raise doubt about the A7RII profiling and what appears to be the case: I think the above images answer that question.
To be clear: it is not only about color or color fidelity: these are becoming less and less important it seems, as "creative color editing" takes over. To my eyes, it is all about the combination of natural color with good tonality. The colors I get out of the Sony A7RII images in CO1 just don't fit nicely with the tonality (luminousity), and so the colors may be "ok", but the images lack the "life" of the A7R images. This is especially true for outdoor images, where there is no controlled environment and tonality is key (in tandem with natural color) to convincing images.
Perhaps something similar is the case with the Pentax K1.
Chris0 -
[quote="SFA" wrote:
I'm not sure that Phase would have much to offer for a previous process embedded in a DNG file.
Thanks for your response.
To clarify: C1 gives me the standard (subpar) Pentax colours for PEF (Pentax RAW) files and for DNG files (that Pentax cameras can optionally produce in camera) alike.
I'd understand if Phase One did not support the application of their own engineered camera profiles to DNG files (whether the latter are the result of conversions or come out of camera). However, a PEF file should definitely be converted using Phase One's own camera profile per default.
As Chris pointed out, DNG files, as created by Pentax cameras, do contain the full original raw data. They just also happen to contain an embedded camera profile which a RAW converter may or may not use.
I configured my K-1 to produce PEF files because I figured that C1 would provide the best support for this file type.
My problem is different to Chris's as I don't suffer from the application of a profile that was engineered for a different camera. The camera profile used by C1 for the K-1 is genuinely suitable for the K-1 as it appears to be just Pentax's original profile. However, it suffers from a number of problems for which I can only blame Pentax but yet wish I would not be facing them when using C1 (with its ambition to provide its own high-quality profiles).
I share with Chris that both our camera models have not received the dedicated camera profile engineering that Phase One appears to promise. I'm still hoping that there may be a Phase One engineered camera profile for the K-1 in the future.
I am trying to make more Pentax shooters enthusiastic about C1 through forum posts, in part because I think C1 is great in many ways, and in part because I am hoping that with higher Pentax user numbers, support for Pentax will become better. I hope that Phase One realises that if they offered a high quality camera profile for the K-1 as an alternative, more Pentax shooters would be immediately drawn to C1 when they experiment with a trial version.
As it stands, C1 has to convince through its interface/functionality but provides no advantage in terms of out-of-the-box appearance of the images. Adobe's "Standard" profile may be of questionable quality (they have been known to use twisted hues) but at least it gives first time experimenters an alternative and many seem to like it. I understand, Adobe has even started to support different camera profiles (landscape, portrait, ...) for Pentax recently.
Personally, I cannot see myself returning to Lightroom, but for others Lightroom is the default choice and it is regrettable that C1 does not have a killer camera profile for the K-1 that could entice users to choose it over Lightroom.0 -
[quote="ChrisM" wrote:
So here is a direct comparison of both the A7R & A7RII "generic" profiles and the A7R & A7RII "standard" profiles.
unfortunately you totally don't understand that you are not really comparing color transforms guided by camera profiles using this tool ... as an exercise "compare" C1 profiles from few modern Canon dSLRs between themselves, then few modern Nikon dSLRs ... then repeat the statement like "because 80% of the gamut is also identical between both profiles, and individual profiling of the ... substitute Canon/Nikon model ... should at least show some differences."0 -
[quote="ChrisM" wrote:
well as with the Sony A7R profile was one of superior color in CO1
and you again is not able to show on a pair of raw files of identical scene (imaging resource) how A7R with C1 profiles for it is better than A7R2 with C1 profiles for it ... because you can't 😂0 -
Hello deejjjaaaa
Sorry, I don't understand your posts. They are gobbledegook (see dictionary definition) to me.
Chris is doing a sterling job (in the absence of any P1 input) and should be encouraged to continue.
My only question is, why aren't other A7R2 users chiming in. Is Chris's experience a one-off? Perhaps he can borrow another A7R2 for a bit of DIY testing?
For what it's worth I have virtually given up using C1 colour tones with my Canon cameras (5D2 and 5Ds) because I have to fiddle about to do justice to my mind's eye image. (Yes, I know, personal preferences and perhaps slightly wonky eyesight.)
Peter0 -
[quote="Peter" wrote:
Chris is doing a sterling job
Chris does not understand basics and Chris is a chicken refusing to do a simple thing - take 2 raw files of the same scene from I-R and put the money where his mouth is (that C1 A7R profiles with A7R raw files were way better than C1 A7R2 profiles with A7R2 files)... because that is the simple way to proof his allegations (that P1 screwed A7R2 profiles), but why do you think he can't do this for days 🙄 ... instead he tries to show us differences between C1 and other converters... those again mostly addressed by using a different curve or creating your own, even with OEM profiles[quote="Peter" wrote:
My only question is, why aren't other A7R2 users chiming in. Is Chris's experience a one-off? Perhaps he can borrow another A7R2 for a bit of DIY testing?
I am A7R2 user since its release ... I have no issues (except that I naturally do not use "film standard" curve, that's given) and when I do not like how C1 OEM profiles render "colors" I simply use my own profiles created with dcamprof (or now with LrPD)...0 -
[quote="deejjjaaaa" wrote:
[quote="ChrisM" wrote:
So here is a direct comparison of both the A7R & A7RII "generic" profiles and the A7R & A7RII "standard" profiles.
unfortunately you totally don't understand that you are not really comparing color transforms guided by camera profiles using this tool ... as an exercise "compare" C1 profiles from few modern Canon dSLRs between themselves, then few modern Nikon dSLRs ... then repeat the statement like "because 80% of the gamut is also identical between both profiles, and individual profiling of the ... substitute Canon/Nikon model ... should at least show some differences."
Well, enlighten me then. 😄
I am always eager to learn.
But also explain to me the following: the camera raw color space, as defined by the chromaticities of its primary colors, is unique for each camera (or should be)? Why on earth should a new camera profile for raw color transforms, LUT et all. feature an exactly identical raw camera color space? How does an identical raw color space make its way 1:1 into a new profile?
Look for instance at the profiles for the Sony A7: standard and generic have very different raw color spaces (more or less undermining your statement that all modern camera color spaces are created equal).
Anyway: your eyes eventually tell the "true" story: a few posts back, before I opened these profiles to take a look at their raw color spaces, I already said that the standard profiles of A7R and A7RII look 100% identical, but the generic profiles yielded different results in the blue/ purple region.
So I took a look at the raw color spaces, and this supports what I already saw: the raw color space has been modified in the blue/purple area.
So tell me how a new profiling job ends with an identical raw camera color space, like with the standard profiles of both A7RII and A7R, or like with 80% of the gamut of the generic profiles? You are saying that only the LUT is created new and the raw color space is copied over? To put it in perspective: with the Pentax K3, Phase one álso shifted one of the primaries (in the red area) of the Pentax K5 color space, and did so considerably, giving a similar type of "edited" raw color space that was leading to subtantially oversaturated reds in the K3 images. That was corrected in the new K3 profile, that featured a very different raw color space.
Chris0 -
[quote="deejjjaaaa" wrote:
Chris does not understand basics and Chris is a chicken refusing to do a simple thing
Thank you, you are too kind 😕[quote="deejjjaaaa" wrote:
I am A7R2 user since its release ... I have no issues (except that I naturally do not use "film standard" curve, that's given) and when I do not like how C1 OEM profiles render "colors" I simply use my own profiles created with dcamprof (or now with LrPD)...
The standard film curve is not a "bad" curve at all. I use it with the Pentax K3 and used it on 80% of my A7R images.
It is just with the Sony A7RII that I get up to severe issues with blackening of the shadows and washing out of highlights.
It may well be that the Sony A7RII has, although for post processing a wide dynamic range, for single shot dynamic range a limited ability to capture outdoor scenes. In that case, just to be clear, Phase one is not really to blame. If it is the best the Sony A7RII can do, than so be it, and I will have to accept heavy editing. However, there may well be a profiling weakness contributing to the inability of the A7RII to handle broad daylight photography.
Chris0 -
[quote="deejjjaaaa" wrote:
[quote="ChrisM" wrote:
well as with the Sony A7R profile was one of superior color in CO1
and you again is not able to show on a pair of raw files of identical scene (imaging resource) how A7R with C1 profiles for it is better than A7R2 with C1 profiles for it ... because you can't 😂
This is about high contrast outdoors shooting in broad daylight. "In studio" controlled shooting won't be an issue.
I get the message about how you rely on numbers and proof presentation. I however rely on a pair of trained eyes.
To put it all in perspective: with a lot of curves work, the A7RII shots can look very good in CO1, it is just that the starting point leaves a lot to be desired, more so than with any other camera I used in almost 10 years of CO1 use.
Chris0 -
Hi Folks,
This discussion is a bit too technical for me. I Encoutered a similar problem using A Canon G3X.
Using the default cameraprofile for the camera (fil standard) the flower (I noticed when taking the picture that is was bright red came out a bit yellowish.
That picture is shown on the left. Then I created a new camera profile with lumariver software and used it (also film standard) on this picture. You can see that on the right. Apart from this profile both settings are equal. The embedded preview from the camera is very much closer to the right one.
https://c1.staticflickr.com/5/4287/35461774751_9e2c5e0e79_c.jpgC1 by Germ Wind, on Flickr
I would tend to agree that some profiles are not 100% spot-on.
Regards from Holland
Germ Wind0 -
[quote="Germie" wrote:
Using the default cameraprofile for the camera (fil standard) the flower (I noticed when taking the picture that is was bright red came out a bit yellowish.
because P1 does not make repro-like profiles for your (and other dSLR) cameras - they have a certain idea what colors are good and they (even if mechanically) make LUTs to do such work and sometimes this is because profile creator(s) want(s) to "clamp" highly saturated "colors" within a certain gamut upon color transform into a proper color space (whatever it is) - sometimes saturated "reds" go "orange"...
but this discussion is not about colors though - this discussion is about allegations from the topic starter that
1) P1 simply repeated the profiles from A7R for A7R2
and
2) somehow those allegedly repeated profiles results in "glaringly" (or whatever was the word) worse conversion when A7R2 raw files are used with A7R2 profiles from P1 vs A7R raw files are used with A7R profiles from P1...
as a proof topic starter suggested comparisons between C1 and non C1 conversions and gamut comparison where "20%" difference in volume (80-20) or whatever was claimed as a proof that profile was simply repeated by P1 ... did I miss any new proofs since my last posting here ? did he finally bothered to take A7R and A7R2 raw files ?[quote="Germie" wrote:
That picture is shown on the left. Then I created a new camera profile with lumariver software and used it
the way to go... dcamprof or LrPD is what doctor prescribed0 -
[quote="ChrisM" wrote:
The standard film curve is not a "bad" curve at all. I use it with the Pentax K3 and used it on 80% of my A7R images.
It is just with the Sony A7RII that I get up to severe issues with blackening of the shadows and washing out of highlights.
and where are the comparison on a pair of raw files from I-R website ? one from A7R , one from A7R2 ... so that we can see that raw files with the same scene and equal within 1/6 EV exposure are so glaringly different when "film standard" curve is used and the rest of parameters are the same ?0 -
[quote="ChrisM" wrote:
Well, enlighten me then. 😄
consider just 2 raw triplets : R1G1B1 and R2G2B2 before color transform ... let one profile map it to 2 coordinates in whatever color space and then 2nd profile to map just the opposite way... good enough ? that is just to show that while identical profiles will absolutely identical "volume" of the resulting output that alone proves nothing about different profiles at all...0 -
[quote="ChrisM" wrote:
But also explain to me the following: the camera raw color space
input devices (= raw data) do not have color space, so they don't have gamuts... color 101
color space and gamut appears only upon color transform from raw data numbers (that are __not__ coordinates in any proper color spaces) to some proper color space and you can have totally different color transforms for the same input device ... so mapping of raw data to proper color space is quite arbitrary process0 -
[quote="deejjjaaaa" wrote:
[quote="ChrisM" wrote:
But also explain to me the following: the camera raw color space
input devices (= raw data) do not have color space, so they don't have gamuts... color 101
color space and gamut appears only upon color transform from raw data numbers (that are __not__ coordinates in any proper color spaces) to some proper color space and you can have totally different color transforms for the same input device ... so mapping of raw data to proper color space is quite arbitrary process
A camera as input device may not have a defined color space, but it surely has a color response, i.e. an ability to capture color data, that at a later stage will be mapped to a color space. So a native camera color space will at maximum extend to the limits of the camera to record color data, also limited by noise levels. Are you saying that this color response is not measured and defined into a gamut per individual camera, linked to the LUT's, but that the gamuts incorporated in phase one camera profiles are arbitrarily assigned, merely to be able to enable a workflow to output images?
I assumed that the color gamuts shown in profile inspectors are based on the LUT's and that a phase one camera profile with a modified LUT will automatically show up as a differently shaped color gamut. But if you say that a differently shaped color gamut in a phase one camera profile says absolutely nothing about the LUT's in a that profile, but that the gamuts shown, are in fact arbitrary color spaces, then ok.
Chris
Chris0 -
[quote="ChrisM" wrote:
A camera as input device may not have a defined color space
it is not "may" ... it does not[quote="ChrisM" wrote:
, but it surely has a color response, i.e. an ability to capture color data
it does not capture "color" it captures "light"... color (coordinates in a proper color space) is the result of color transform executed by software and guided for example by the content of camera profile (and also in many case by the code in software itself - C1 is a fine example) ... what you refer above by "color response" probably is spectral sensitivity of CFA filters, IR/UV cut filter, lens mounted, etc that simply tell how much of a particular spectrum goes through to the sensel[quote="ChrisM" wrote:
So a native camera color space
input devices do not have color space... based on sensor architecture, firmware, etc the sensel "data" recorded as digital numbers (DN) in raw file can be anywhere from 0 to whatever max integer implementation allows... and those DNs are not coordinates in any color space.[quote="ChrisM" wrote:
but that the gamuts incorporated in phase one camera profiles are arbitrarily assigned, merely to be able to enable a workflow to output images?
you can (if you wish so) create a profile that will guide color transform through LUT effectively for example well within sRGB gamut or for example well above AdobeRGB gamut... it is a design choice... PS: and in C1 camera profiles are applied at the very end of its pipeline... after almost all those sliders in UI (things like Color Editor of course operates on LUT in camera profile itself)...0 -
[quote="ChrisM" wrote:
show up as a differently shaped color gamut.
it will, but what conclusions can you draw from that ?[quote="ChrisM" wrote:
but that the gamuts shown, are in fact arbitrary color spaces
not arbitrary icc/icm camera profiles have PCS ( cielab for C1 , the color transform guided by the data in that container shall result in coordinates in PCS color space and raw converter code then further in its pipeline may/shall execute further color transform from that PCS color space to output color space like for example ProPhoto RGB or whatever for export, etc ) - gamuts shown are based on the resulting output of color transform encoded in profile, but why are you making any conclusions about whether P1 profiled camera or copied from elsewhere from the shapes beats me... even what you consider a "small difference" does not prove anything ... and on top of that we really don't know how much different (if different at all) for example CFA and IR/UV cut filters are between A7R camera and A7R2 camera - there is of course a chance that P1 discovered through test of otherwise that no material difference is and then simply still did some changes because may be the way they create their "look" evolves and it resulted in "slightly different" LUTs... shall we say in this case that P1 did not individually profile camera ? if I discover through test that I do not need to mod LUTs and leave them be - does it mean that I did not profile the camera ? the challenge that you did not answer still stands - you did not show that things suddenly went wrong in C1 with A7R2 raw files and A7R2 camera profiles vs A7R raw files and A7R camera profiles... we read your statements, but no proof shown (comparing output from other raw converters with C1 output does not prove that with A7R everything was better so much, no)0 -
[quote="deejjjaaaa" wrote:
[quote="ChrisM" wrote:
show up as a differently shaped color gamut.
it will, but what conclusions can you draw from that ?[quote="ChrisM" wrote:
but that the gamuts shown, are in fact arbitrary color spaces
not arbitrary icc/icm camera profiles have PCS ( cielab for C1 , the color transform guided by the data in that container shall result in coordinates in PCS color space
Ok. Then at least we don't disagree on that.
Again, thanks for your input and insights, I'm sure it keeps the discussion relevant. Also regarding what you are saying below. A dedicated profile is a puritan's wish, but if color filter and other sensor stack engineering is similar up to the point where you would not notice the difference, and thus any profile for that sensor stack may be as good (or bad) as another, then the good quality of the A7R profiles might be a reason to reuse them. What makes me weary though with CO1, is their implementation of color profiles, and especially the nature of the curve in the color profile itsef.
However, Phase one was kind enough to respond, and tells me that they do not try to adjust exposure or contrast in their .icc camera profiles, but that I should be looking at the base curve. Since it is mostly about overbrightened highlights washing out and shadows darkening and losing detail, I understand that the impact of the base curve on this behavior is far greater than the mapping in the color profile itself.
Taking a look at the linear curve, it is clear how much the base curve film standard (and also extra shadow) impacts the image: it adds exposure and contrast up to a degree where you could say that it defines the look of an image. I never warmed to HDR or other tone mapping software, because with a modern digital sensor the curves tool and a highlight repair tool are all you should need. Neither film "standard" nor "extra shadow" give me the post processing ease with the A7RII, that they gave me with the A7R and I am still looking (as I have this week off from work) if anything can provide that ease. I would say I join you in the wish for CO1 to provide, either in their software or outside, a dedicated way to edit the base curve directly, and save it as your own.
[quote="deejjjaaaa" wrote:
and raw converter code then further in its pipeline may/shall execute further color transform from that PCS color space to output color space like for example ProPhoto RGB or whatever for export, etc ) - gamuts shown are based on the resulting output of color transform encoded in profile, but why are you making any conclusions about whether P1 profiled camera or copied from elsewhere from the shapes beats me... even what you consider a "small difference" does not prove anything ... and on top of that we really don't know how much different (if different at all) for example CFA and IR/UV cut filters are between A7R camera and A7R2 camera - there is of course a chance that P1 discovered through test of otherwise that no material difference is and then simply still did some changes because may be the way they create their "look" evolves and it resulted in "slightly different" LUTs... shall we say in this case that P1 did not individually profile camera ? if I discover through test that I do not need to mod LUTs and leave them be - does it mean that I did not profile the camera ? the challenge that you did not answer still stands - you did not show that things suddenly went wrong in C1 with A7R2 raw files and A7R2 camera profiles vs A7R raw files and A7R camera profiles... we read your statements, but no proof shown (comparing output from other raw converters with C1 output does not prove that with A7R everything was better so much, no)0 -
I am retiring from this thread... peace. 0 -
For over-brightness it appear that in C1 10 they change default contrast 'zero' values and same (?) Film Standard (and two others except so-called Linear) curves produce now this too-bright highlights and too-dark shadows with the same raw data. But Shadow and Highlights recovery tools 'works' much better now, sort of;-]
...
I shot A7r and occasionally edited some files from A7r-II - it usable only with Linear Curve both.0 -
[quote="NN174596UL" wrote:
For over-brightness it appear that in C1 10 they change default contrast 'zero' values and same (?) Film Standard (and two others except so-called Linear) curves produce now this too-bright highlights and too-dark shadows with the same raw data. But Shadow and Highlights recovery tools 'works' much better now, sort of;-]
...
I shot A7r and occasionally edited some files from A7r-II - it usable only with Linear Curve both.
You may well be right about a somewhat higher "baseline" contrast, although revisiting images from e.g. Pentax also seems to indicate that fairly recent cameras suffer progressively from this issue.
It is always far easier to post process starting from a relatively neutral image, instead of having to "undo" additional contrast in the default rendering. For me somehow the Sony A7RII is the point where it has become simply awkward to edit ftom the default rendering with auto curve (film standard both shadow clipping and highlight overbrightening), but also with "extra shadow" where shadow darkening is less, but highlight overbrightening still present.
As an additional experiment, I tried linear input curve, with the added "RGB contrast curve" preset in the curve tool panel, and auto adjusting exposure, and compard that to a variant with the standard film curve and auto adjusting exposure, to see if both settings would render similar contrast, as you might expect (assuming the RGB contrast curve preset is based on the same principle as the standard film curve), but indeed the film standard curve variant suffers from an added unnatural look due to blacker shadows and more overbrightened highlights.
The thing however with the approach of using linear input curve and occupying the curves tool with a (tweaked) rgb contrast curve, is that the curves tool becomes inaccesible to creative curve editing presets, and working on extra layers does not allow for loading curve presets.
And anyhow, I remain of the opinion that a default setting should be there as an optimal starting point for further post processing. That is hardly the case with CO1 and the Sony A7RII.
To add something positive to all the criticism, I fully agree with the stellar new tools in CO1 v10 regarding shadow lifting, highlight recovery, and my favorite: the new brightness tool. They excell at retaining a natural balance color wise and overall look. But first, you have to undo the side effects of the rather awkward default rendering.
Chris0 -
For me - with A7r i prefer to set Linear Curve and Luminosity contrast curve leaving RGB for creative edits and advanced colour correction.
I also use home-brewed (not by me) icc profile, it's relatively soft but in some cased C1s Standard is also used.0
Please sign in to leave a comment.
Comments
60 comments