November 27, 2014, 12:39:07 AM

Author Topic: PP for realistic look - is DPP the best?  (Read 4694 times)

SwissBear

  • Rebel SL1
  • ***
  • Posts: 89
    • View Profile
Re: PP for realistic look - is DPP the best?
« Reply #15 on: April 17, 2013, 06:46:28 PM »
Not a long time ago there was the dilemma before pressing the shutter release. That was in the anlogue age, where you had to choose the right roll of film before the real action. ISO, temperature, color style and so on.
I am very happy these times are over - as a student, film was too expensive and too slow for learning.

Now, all these decisions must be made afterwards, and as many have pointed out, something as "right" or "real" is rather hard to get.
(Some presets that try to imitate all these films would be nice ;) )

But anyway, any decent RAW converter should be able to give you the results you wish, but clearly DPP, being from the camera-dev, has the ability to render a RAW similar to the jpg you can view on the camera.

But that is, again, not "realistic" or anything.
EOS 50D, 600D - 24-105L - 40f/2.8 - 85f/1.8 - 580ii - Lensbaby Muse, Composer with Plastic, Double Glass, Edge 80
Tokina 11-16f/2.8 - Sigma 150mm OS Macro
Flickr

canon rumors FORUM

Re: PP for realistic look - is DPP the best?
« Reply #15 on: April 17, 2013, 06:46:28 PM »

TrumpetPower!

  • 1D Mark IV
  • ******
  • Posts: 951
    • View Profile
Re: PP for realistic look - is DPP the best?
« Reply #16 on: April 17, 2013, 08:08:21 PM »
(Some presets that try to imitate all these films would be nice ;) )

Raw Photo Processor does exactly that. It's not something that interests me -- RPP's superb colorimetric accuracy is why I use it -- but I understand that it does a very, very good job at emulating all sorts of favorite film stocks.

Cheers,

b&

sandymandy

  • Canon 7D MK II
  • *****
  • Posts: 606
    • View Profile
Re: PP for realistic look - is DPP the best?
« Reply #17 on: April 17, 2013, 10:07:56 PM »
(Some presets that try to imitate all these films would be nice ;) )


There you go: http://vsco.co/film

:)

Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
Re: PP for realistic look - is DPP the best?
« Reply #18 on: April 17, 2013, 10:49:11 PM »
You are basically out of luck. Images taken from cameras and printed or displayed on monitors simply cannot match what we see. Theoretically, they could, but this requires color filters in the Bayer sensor, and pixels on your monitor which well match the human vision.

What the camera records is a 3D projection of an image with infinite dimensional spectrum. Your eyes see a different projection.  There are still better and worse RAW converters, of course, but you can never make it even close to perfect.

The worst of all is that you may see two objects as being of same color in one light, and different in another (metamers). So the object you want to shoot has a color depending on the lighting, really. I am not talking just about WB.

TrumpetPower!

  • 1D Mark IV
  • ******
  • Posts: 951
    • View Profile
Re: PP for realistic look - is DPP the best?
« Reply #19 on: April 17, 2013, 11:13:41 PM »
You are basically out of luck. Images taken from cameras and printed or displayed on monitors simply cannot match what we see. Theoretically, they could, but this requires color filters in the Bayer sensor, and pixels on your monitor which well match the human vision.

While it is true that no camera actually satisfies the Luther Condition, it is also true that all modern cameras come pretty close. They're all plenty good enough for all but the most exotic multispectral work...and, even then, you can still use them for multispectral imaging by shooting multiple frames through a series of carefully-selected (but standard off-the-shelf) filters and feeding the results to the right kind of software.

Quote
There are still better and worse RAW converters, of course, but you can never make it even close to perfect.

Utterly false. I regularly make giclée prints using a 5DIII and an iPF8100 that the artists have to look carefully at to be able to distinguish copy from original. Perfect? No -- of course not. But damned close.

Quote
The worst of all is that you may see two objects as being of same color in one light, and different in another (metamers). So the object you want to shoot has a color depending on the lighting, really. I am not talking just about WB.

You're referring to metamerismic failures, and those are rarely significant in practice. It's almost never of any concern in general photographic scenes. When it comes to fine art reproduction, the artist generally knows up front that the pigments exhibit significant amounts of metamerismic shifts and will know that nothing but that actual pigment will function the same way. Then it's a simple choice: either you shoot the artwork under your standard studio conditions and profile to D50 as normal for a good all-purpose choice; you shoot under the known intended viewing conditions and profile to those conditions for a print only intended to be displayed in said conditions; or you shoot as normal and profile to D50 as normal and hand the print back to the artist for hand-application of the special pigment. The same applies, of course, to iridescent, metallic, fluorescent, and other special colorants / materials / whatever.

Cheers,

b&

Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
Re: PP for realistic look - is DPP the best?
« Reply #20 on: April 18, 2013, 01:08:43 AM »
While it is true that no camera actually satisfies the Luther Condition, it is also true that all modern cameras come pretty close. They're all plenty good enough for all but the most exotic multispectral work...and, even then, you can still use them for multispectral imaging by shooting multiple frames through a series of carefully-selected (but standard off-the-shelf) filters and feeding the results to the right kind of software.
Right. As I said, you get a 3D projection. When you use filters, you increase the number of the dimensions.
Quote
Utterly false. I regularly make giclée prints using a 5DIII and an iPF8100 that the artists have to look carefully at to be able to distinguish copy from original. Perfect? No -- of course not. But damned close.
Then I wonder how come something that you can do in your office cannot be achieved by multinational companies worth billion of dollars? Why is that no two bodies have the same color rendering with whatever software? 
Quote

You're referring to metamerismic failures, and those are rarely significant in practice. It's almost never of any concern in general photographic scenes.
You could not be more wrong. Take a picture in daylight of something with many colors, including sensitive to the eye, like skin tones. Next, do it in artificial light. Use a gray card. The color will not match. Use a Nikon, and the colors will be different. Wait, how can this be true if "all modern cameras come pretty close"? Look at the DXO measurements. Different bodies have very different color filters. They even measure how close they are to the human vision, and the numbers are different. Nikon are known for troubles with skin tones, Canon - with blues. Why would that be true?
Quote
When it comes to fine art reproduction, the artist generally knows up front that the pigments exhibit significant amounts of metamerismic shifts and will know that nothing but that actual pigment will function the same way. Then it's a simple choice: either you shoot the artwork under your standard studio conditions and profile to D50 as normal for a good all-purpose choice; you shoot under the known intended viewing conditions and profile to those conditions for a print only intended to be displayed in said conditions; or you shoot as normal and profile to D50 as normal and hand the print back to the artist for hand-application of the special pigment. The same applies, of course, to iridescent, metallic, fluorescent, and other special colorants / materials / whatever.

You just confirmed what I said. The colors recorded by the camera change, and simple WB would not fix it. When you profile, you actually are trying to get the best approximation after the fact that you did have metamerismic shifts, and they were significant enough.

TrumpetPower!

  • 1D Mark IV
  • ******
  • Posts: 951
    • View Profile
Re: PP for realistic look - is DPP the best?
« Reply #21 on: April 18, 2013, 09:30:32 AM »
While it is true that no camera actually satisfies the Luther Condition, it is also true that all modern cameras come pretty close. They're all plenty good enough for all but the most exotic multispectral work...and, even then, you can still use them for multispectral imaging by shooting multiple frames through a series of carefully-selected (but standard off-the-shelf) filters and feeding the results to the right kind of software.
Right. As I said, you get a 3D projection. When you use filters, you increase the number of the dimensions.
Quote
Utterly false. I regularly make giclée prints using a 5DIII and an iPF8100 that the artists have to look carefully at to be able to distinguish copy from original. Perfect? No -- of course not. But damned close.
Then I wonder how come something that you can do in your office cannot be achieved by multinational companies worth billion of dollars? Why is that no two bodies have the same color rendering with whatever software? 

Because the popular raw development engines (including those built into the cameras that do the conversion to JPEG) aren't attempting colorimetric renditions. They're going for "pleasing" color, with everybody tuning their algorithms for their own special sauce.

What, you thought all those different picture styles were random flailings at trying for colorimetric accuracy?

Quote
Quote
You're referring to metamerismic failures, and those are rarely significant in practice. It's almost never of any concern in general photographic scenes.
You could not be more wrong. Take a picture in daylight of something with many colors, including sensitive to the eye, like skin tones. Next, do it in artificial light. Use a gray card. The color will not match.

If your gray card isn't matching, you've got the worst gray card ever manufactured.

But you also don't understand what the Luther Condition is, either. A camera that meets the Luther Condition will render metamerismic shifts the same way your eyes do. So the fact that the cameras display the same reaction to metamerism as your eyes just goes to further prove my point.

Cameras are not meant to be spectroscopic instruments. They're designed to model human vision, and they do a damned fine job of it.

If you're not able to get quality colorimetric results from your gear, it's your problem. Your gear is quite well up to the task, I assure you. Granted, it's not something you can do with the software that comes in the box, and it's not something many people care about so the knowledge of how to do it isn't common. But it most definitely is possible with common off-the-shelf equipment.

b&

canon rumors FORUM

Re: PP for realistic look - is DPP the best?
« Reply #21 on: April 18, 2013, 09:30:32 AM »

Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
Re: PP for realistic look - is DPP the best?
« Reply #22 on: April 18, 2013, 09:44:58 AM »
Because the popular raw development engines (including those built into the cameras that do the conversion to JPEG) aren't attempting colorimetric renditions. They're going for "pleasing" color, with everybody tuning their algorithms for their own special sauce.

What, you thought all those different picture styles were random flailings at trying for colorimetric accuracy?

They are different optimizations with different parameters. You would think that each company would offer at least one "colormetric accurate" profile, right? Actually, they are trying.

Since you cannot have color accuracy, going for pleasant colors is the right thing to do.
Quote
If your gray card isn't matching, you've got the worst gray card ever manufactured.

You did not understand what I said. The gray card might be perfect, you camera's filters are not. Not to mention your monitor or the inks of your printer.
Quote
But you also don't understand what the Luther Condition is, either. A camera that meets the Luther Condition will render metamerismic shifts the same way your eyes do. So the fact that the cameras display the same reaction to metamerism as your eyes just goes to further prove my point.
Too bad such cameras do not exist, proving my point. Canon's red filter, for example, is much close to the human one than Nikon's. etc. 5D's blue filter is much cleaner than 7D's, one, etc.
Quote
Cameras are not meant to be spectroscopic instruments. They're designed to model human vision, and they do a damned fine job of it.

If you're not able to get quality colorimetric results from your gear, it's your problem. Your gear is quite well up to the task, I assure you. Granted, it's not something you can do with the software that comes in the box, and it's not something many people care about so the knowledge of how to do it isn't common. But it most definitely is possible with common off-the-shelf equipment.

Sure. Go to IR, download their RAWs and try to match the colors of different cameras. Good luck. And that is not even close to reality, where lighting changes.

Why don't you do the experiment I suggested: same scene with different lighting? If you are right, a simple tweak of the WB would match the colors perfectly.

TrumpetPower!

  • 1D Mark IV
  • ******
  • Posts: 951
    • View Profile
Re: PP for realistic look - is DPP the best?
« Reply #23 on: April 18, 2013, 10:14:56 AM »
Because the popular raw development engines (including those built into the cameras that do the conversion to JPEG) aren't attempting colorimetric renditions. They're going for "pleasing" color, with everybody tuning their algorithms for their own special sauce.

What, you thought all those different picture styles were random flailings at trying for colorimetric accuracy?

They are different optimizations with different parameters. You would think that each company would offer at least one "colormetric accurate" profile, right?j

Worng. Very, very worng.

Nobody wants colorimetric accuracy save for those doing fine art reproduction, and those of us doing that kind of work don't need any hand-holding from the manufacturers.

Quote
Actually, they are trying.

ORLY? They are?

Show me just one canned picture style / whatever from just one manufacturer that doesn't have an S-curve applied after gamma adjustments.

Dude, they're not even pretending to try. Because that's not what their customers want.

Quote
Since you cannot have color accuracy, going for pleasant colors is the right thing to do.

No. You can have color accuracy, if you know what you're doing. I do that all the time. As I've already written, the artists whose work I reproduce have to look carefully and study the print in a side-by-side comparison to be able to spot the differences.

Going for "pleasing" colors is the right thing to do for the major manufacturers because that's what their customers are demanding.

Quote
Sure. Go to IR, download their RAWs and try to match the colors of different cameras.

Why on Earth would I waste my time on such a fool's errand?

That's not how colorimetric work is done. That you think that that's how it's done more than amply demonstrates that you have no clue about what you're talking about.

Have any camera you like delivered to my studio along with any samples you like of artwork that doesn't contain fluorescent / iridescent / metallic / etc. pigments and, for my standard fees, I'll return the camera and artwork along with a print that's a very close colorimetric match within the limits of the iPF8100's (rather large) gamut. If the prints will be evaluated in non-standard viewing conditions, I'll need a proper spectrophotometric reading of those conditions, such as what can be obtained with an i1 Pro in ambient measurement mode. The measurement must be in GCATS format and the instrument must be in the same position as the print will be viewed in.

Cheers,

b&

Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
Re: PP for realistic look - is DPP the best?
« Reply #24 on: April 18, 2013, 10:38:05 PM »
The "discussion" degenerated too much. I am done.

canon rumors FORUM

Re: PP for realistic look - is DPP the best?
« Reply #24 on: April 18, 2013, 10:38:05 PM »