Luminous Landscape Forum

Raw & Post Processing, Printing => Adobe Lightroom Q&A => Topic started by: simplify on May 13, 2010, 10:05:04 pm

Title: RGB numbers?
Post by: simplify on May 13, 2010, 10:05:04 pm
Lightroom displays color percentages under the histogram.  Is there a way to change to percentage to rgb numbers (0-255)

There must be a way to change it.

Thanks,
JP
Title: RGB numbers?
Post by: Schewe on May 13, 2010, 10:10:04 pm
Quote from: simplify
There must be a way to change it.

No, there isn't since the actual numbers (0-256) are undefined since there is no specified output color space inside of Develop. So 0-100% is the only relevant method of displaying color info.
Title: RGB numbers?
Post by: simplify on May 13, 2010, 10:31:25 pm
Thanks for the info.

It would be very nice if could approximate prophoto rgb numbers.  Its hard to understand percentages after working with rgb numbers for so long.

Thanks again,
JP Jespersen
Title: RGB numbers?
Post by: JeffKohn on May 13, 2010, 10:41:47 pm
0-100% actually seems more intuitive to me, assuming that it's perceptually uniform like L* in LAB. Easier to understand than gamma 2.2 RGB numbers IMHO, where I can never remember what value middle gray is. ProPhoto is even worse, because it's gamma 1.8 - or should we use use gamma 1.0 since that's what LR uses internally? See the problem?
Title: RGB numbers?
Post by: simplify on May 13, 2010, 10:53:29 pm
I work with RGB numbers all day long every day.  So I feel much more in tune to those numbers.
Title: RGB numbers?
Post by: digitaldog on May 14, 2010, 09:34:07 am
Quote from: simplify
I work with RGB numbers all day long every day.  So I feel much more in tune to those numbers.

The percentages actually make a lot more sense but seem odd to people used to 0-255 (much as the metric system seems odd to us Americans who use the Standard system even though the metric system makes far more sense).

Funny too that people who work in CMYK love a percentage system, based on ink but not the actual documents encoding. We still have a 0-255 scale behind these percentages but they have to use percentages because that’s traditionally what has been done in prepress forever.

It would be interesting if Photoshop’s info palette would toggle to RGB percentages like LR, problem is, the color space used in LR isn’t used anywhere else and they wouldn’t match. So like anyone using a new scale, its best not to try a literal translation here, just understand what zero and 100% represent, and that when all three values are identical, you have a neutral. Since the LR underlying color space is only used in LR, trying to decipher it into something in a 0-255 scale later in Photoshop will just make your head explode.
Title: RGB numbers?
Post by: Guillermo Luijk on May 14, 2010, 10:40:23 am
Quote from: digitaldog
(much as the metric system seems odd to us Americans who use the Standard system even though the metric system makes far more sense).
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?
Title: RGB numbers?
Post by: digitaldog on May 14, 2010, 10:49:33 am
Quote from: Guillermo Luijk
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?

I have no idea. My guess would be that isn’t happening. The educational system in the US doesn’t seem to be improving.
Title: RGB numbers?
Post by: bjanes on May 14, 2010, 11:40:15 am
Quote from: simplify
Thanks for the info.

It would be very nice if could approximate prophoto rgb numbers.  Its hard to understand percentages after working with rgb numbers for so long.

Thanks again,
JP Jespersen
The percentage output values in Lightroom are merely normalized values. For example, a value of 50% would be 127.5 in 8 bit notation. The normalized values are in the range of 0..1.0 and the 8 bit values are in the range of 0..255.

If you want the ProPhotoRGB values of the Lightroom readout you have to do a tone curve conversion, since Lightroom uses an sRGB tone curve with ProPhoto chromaticities. sRGB uses a linear segment in the shadows, and an exponent of 2.4 plus an offset for the other tone values, but can be approximated by a gamma of 2.2 (see Wikipedia  (http://en.wikipedia.org/wiki/SRGB)for the exact formula).

For example, mid gray (L* = 0.5) has a gamma 2.2 normalized pixel value of 0.4635 or 118.8 in 8 bit notation. To convert to the gamma 1.8 value, one would first convert to a linear space (0.4635^2.2 = 0.1842). Then convert to gamma 1.8 (0.1842^(1/1.8) = 0.3907). For 8 bit notation, multiply 0.3907 by 255 to get the gamma 1.8 pixel value of 100.

Here is a chart giving pixel values in 0.3 stop decrements over the range of densities in a Stouffer wedge target showing sRGB values, gamma 2.2, and gamma 1.8.

[attachment=21972:Gammas.gif]
Title: RGB numbers?
Post by: Greg Barnett on May 14, 2010, 12:01:47 pm
Quote from: Guillermo Luijk
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?

Somewhat subject dependent, I'm seeing both in use in our classrooms.
Title: RGB numbers?
Post by: digitaldog on May 14, 2010, 12:08:00 pm
Quote from: bjanes
If you want the ProPhotoRGB values of the Lightroom readout you have to do a tone curve conversion, since Lightroom uses an sRGB tone curve with ProPhoto chromaticities. sRGB uses a linear segment in the shadows, and an exponent of 2.4 plus an offset for the other tone values, but can be approximated by a gamma of 2.2 (see Wikipedia  (http://en.wikipedia.org/wiki/SRGB)for the exact formula).

The other “problem” is that the values provided, and the color space and TRC used to actually process the data (and the data you get out when you select a color space) are all different. So working by the numbers here has limits (as does all working by the numbers techniques IMHO).
Title: RGB numbers?
Post by: bjanes on May 14, 2010, 05:12:23 pm
Quote from: digitaldog
The other “problem” is that the values provided, and the color space and TRC used to actually process the data (and the data you get out when you select a color space) are all different. So working by the numbers here has limits (as does all working by the numbers techniques IMHO).
I'm not an advocate of working by the numbers either. However, you can use Photoshop to define and save a custom color space which we will call Melissa using the ProphotoRGB primaries and white point but a gamma of 2.2. If you export the Lightroom rendering to this color space, the LR info values will be very close to those of the file exported in this custom space, except for the deep shadows, where the Lightroom values will be lower due to the linear segment of the sRGB tone curve that Lightroom is using.

Here is what I got using rendered values from the neutral patches of a MacBeth Color Checker:

[attachment=21976:MelissaTest.gif]
Title: RGB numbers?
Post by: digitaldog on May 14, 2010, 05:38:07 pm
Quote from: bjanes
I'm not an advocate of working by the numbers either. However, you can use Photoshop to define and save a custom color space which we will call Melissa using the ProphotoRGB primaries and white point but a gamma of 2.2.

Yes, I’ve done this. But the raw processing is Melissa using a 1.0 TRC. So you’ve got a readout in LR that isn’t the same as what is happening under the hood, then you have an encoding color space that (at least in my case) is ProPhoto with its original 1.8 TRC. Its a bloody mess. What do you show? What do you translate? In the end, it seems easier to forget the whole thing and just view the percentages in LR and forget Photoshop number translation until you actually get there, use its scale based on the actual data you are working with.
Title: RGB numbers?
Post by: bjanes on May 14, 2010, 10:16:09 pm
Quote from: digitaldog
Yes, I’ve done this. But the raw processing is Melissa using a 1.0 TRC. So you’ve got a readout in LR that isn’t the same as what is happening under the hood, then you have an encoding color space that (at least in my case) is ProPhoto with its original 1.8 TRC.
The purpose of the preview histogram and readout is to estimate the output values in the selected output space, not what is going on under the hood, so to speak. The linear values have everything scrunched up on the left side of the histogram and mid gray would be 18%. Such information is not that useful.

Quote from: digitaldog
Its a bloody mess. What do you show? What do you translate? In the end, it seems easier to forget the whole thing and just view the percentages in LR and forget Photoshop number translation until you actually get there, use its scale based on the actual data you are working with.
I agree that the preview pixel values in Lightroom are a bloody mess. I don't know why anyone would want to know the values in this bastardized space. The solution that Thomas Knoll uses in ACR where the user can select an output space and see the output values for that space is more logical and useful. ACR uses the same linear internal working space that Lightroom uses, but no one is interested in what those values are. As with Lightroom, the ACR image is not rendered into the output space until you hit done and the image is dumped into Photoshop.

I have no problem in reporting normalized pixel values-- it makes more sense than using 0..255. However, normalized pixel values in a bastardized space that no one uses are not that helpful.

If you have to wait until you export the image from Lightroom to ProPhotoRGB to see what the actual pixel values are, what good is the preview? It is nearly as ridiculous as Nancy Pelosi saying that they would have to pass the health care bill so they could find out what was in it.
Title: RGB numbers?
Post by: Wayne Fox on May 15, 2010, 12:46:24 am
Quote from: Guillermo Luijk
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?
I think most students are exposed to it but since it isn't used  it's irrelevant, so only the small percentage involved in technical and scientific worlds, or perhaps those heavily involved in international business become proficient.

The problem with Americans converting is we are challenged to think in pure metric.  We have to translate it to what we are used to it ... Not efficient and no motivation.  The only way to really make a switch would almost be cold turkey, but think about in reverse and you get some idea of how hard it would be.

I read a report a long time ago which studied ways to make it happen, with no easy solution, and costs in the 100's of billions.  They even believed their would be some significant physical injuries and deaths if a forced conversion were to occur.  Seemed a stretch, but there is no doubt it would be costly and difficult.

(sorry to wander off topic)

Title: RGB numbers?
Post by: bjanes on May 15, 2010, 09:45:49 am
Quote from: Guillermo Luijk
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?
For the last past half century or more in the USA, metric units have been extensively used in science and medicine and they are now creeping into everyday life. In photography, we measure the focal length of our lenses in millimeters and resolution in lp/mm, but photographic paper is still measured in inches. Our wine comes in 750 ml bottles, but beer is usually sold in 12 ounce containers. Mechanics have to have a set of metric wrenches, since metric nuts and bolts are widely used in American machinery and universally so with imports.

Wikipedia  (http://en.wikipedia.org/wiki/Metrication_in_the_United_States)has an informative article on metrication in the United States, and the article mentions the failure of a NASA Mars orbiter because one contractor used English units while others used metric units, and someone forgot to convert between pound force seconds instead of newton seconds.

To answer your question, the metric system is taught in our schools, but those who do not use metric units forget much of what was taught. One reason that the metric system has not caught on in the USA is that Americans tend to be traditional and do not like to be told what to do by the government.

Not mentioned in this thread is that the metric system has been superseded by the SI (Système International) nomenclature. American physicians have been measuring blood glucose concentration in milligrams per deciliter for decades, but now some journals are reporting it in the SI units of millimoles per liter. Under SI, if we photographers were still doing chemical photography, we would have to go to the camera store and tell the sales clerk to give us 1 mole of sodium thiosulfate!

Regards,

Bill
Title: RGB numbers?
Post by: digitaldog on May 15, 2010, 11:16:27 am
Quote from: bjanes
The solution that Thomas Knoll uses in ACR where the user can select an output space and see the output values for that space is more logical and useful.

True but far more difficult to do in LR since we put off export (or open in Application) for a future date and depending on the multitude of encoding options for the presets, they will all be different. In ACR, you have the one workflow option and when you click Done or Open, you get what you get. I might export from LR in sRGB for a JEPG, ProPhoto for a master and for that matter, I have the option to select any RGB profile on my system. ACR only provides 4. So getting feedback for numbers and a histogram in Develop, where the output in Print, or Web will be different makes this a difficult prospect.

At the very least, I’d like to option click on the Histogram in Develop and Library, then select an encoding color space recognizing that what I may end up with based on a preset will not necessarily be what I see now. I think that is better than showing me numbers I’ll never get (Melissa RGB).
Title: RGB numbers?
Post by: bjanes on May 15, 2010, 01:27:36 pm
Quote from: digitaldog
True but far more difficult to do in LR since we put off export (or open in Application) for a future date and depending on the multitude of encoding options for the presets, they will all be different. In ACR, you have the one workflow option and when you click Done or Open, you get what you get. I might export from LR in sRGB for a JEPG, ProPhoto for a master and for that matter, I have the option to select any RGB profile on my system. ACR only provides 4. So getting feedback for numbers and a histogram in Develop, where the output in Print, or Web will be different makes this a difficult prospect.

At the very least, I’d like to option click on the Histogram in Develop and Library, then select an encoding color space recognizing that what I may end up with based on a preset will not necessarily be what I see now. I think that is better than showing me numbers I’ll never get (Melissa RGB).

For once, we are in basic agreement. The various options regarding color spaces and readouts for LR were discussed by Mark Hamburg, Thomas Knoll, Bruce Fraser and others at some length in the Lightroom Podcast #8. Mr. Knoll suggested that LR should offer the user to choice of an output space as is done in ACR, but Mr. Hamburg wanted to keep things simple and not offer those choices. The choice of ProPhotoRGB chromaticity for LR was a given, since that is what ACR/LR use internally, but Mark chose to report out the numbers and base the histograms with an sRGB TRC, perhaps because that was what most inexperienced photographers were used to. Bruce suggested that LR should do the right thing and merely use ProPhotoRGB. IMHO, that would have made the most sense.

The end result of Mark's choice was that by simplifying things for the novice, he made things more difficult for the more experienced user who wants to have a better idea of what is actually going on. When you export from LR to a narrow space, you have no control over clipping and do not even have an indication that it is taking place. Some sort of perceptual rendering for the export would be useful if a suitable algorithm could be developed.

Although ACR offers only 4 output spaces, one can save in ProPhotoRGB and then convert to any space in Photoshop and get exactly the same results as if one had output to that space in the first place. If Ver 4 profiles were available, one might even be able to perform perceptual rendering. What are the possibilities here? I admit that I do not understand the Ver 4 process.
Title: RGB numbers?
Post by: digitaldog on May 15, 2010, 01:37:44 pm
Quote from: bjanes
For once, we are in basic agreement.

Actually I think its the 2nd time....
Title: RGB numbers?
Post by: bjanes on May 15, 2010, 01:58:36 pm
Quote from: digitaldog
Actually I think its the 2nd time....
 

Any words of wisdom on Ver 4?
Title: RGB numbers?
Post by: David Hufford on May 16, 2010, 12:37:12 am
Quote from: Guillermo Luijk
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?
I studied that in grade school, then quickly forgot most since it is generally irrelevant to people living in the US. We hardly ever use it. It's not really something for university-level study anyway as the conversion is simple. But after more than a decade of living in a country that uses the metric system, I still find the Celcius temp reading less intuitive and a more general, less nuanced---if that's the word---way of defining the temperature.
Title: RGB numbers?
Post by: DarkPenguin on May 16, 2010, 02:01:12 am
Why not just mentally multiplying the percentages by 2.5 in your head until you're used to them?

Quote from: Guillermo Luijk
Offtopic question: just curious, are American students now learning at school/university inches or cm? miles or Km? Celsius or Farenheit?

Back in grade school history's greatest monster (Jimmy Carter) had us learning it.  Then Ronnie Raygun got into power and put the kibosh on that and we went back to rods and hogs heads for measurements.
Title: RGB numbers?
Post by: Peter Stacey on May 18, 2010, 03:56:24 am
Quote from: DarkPenguin
Why not just mentally multiplying the percentages by 2.5 in your head until you're used to them?

For all the reasons earlier in the thread, that approach won't help much.

To show it graphically, the following image shows the percent values in Lightroom 3 Beta 2, followed by the same image converted to ProphotoRGB, AdobeRGB and sRGB in Photoshop.

Multiplying the percent values by 2.5 doesn't give the same values in the different output profiles:

(http://www.peterstacey.com/images/test/LightroomPercent-to-RGB-example.jpg)

For the specific values, the following table might show it more clearly than the image, including the 2.5 multiplier of the percent values:

(http://www.peterstacey.com/images/test/table.jpg)

The target image can be downloaded here:

target.psd (http://www.peterstacey.com/images/test/target.psd)
Title: RGB numbers?
Post by: digitaldog on May 18, 2010, 09:30:13 am
Quote from: Peter Stacey
The target image can be downloaded here:

Thanks for taking the time to build that. Most useful as a teaching aid for this topic!
Title: RGB numbers?
Post by: eliedinur on May 18, 2010, 10:00:31 am
Quote from: digitaldog
Thanks for taking the time to build that. Most useful as a teaching aid for this topic!

d.d. - Are the LR numbers ProPhoto corrected to gamma 2.2 instead of 1.8?
Title: RGB numbers?
Post by: digitaldog on May 18, 2010, 10:02:46 am
Quote from: elied
d.d. - Are the LR numbers ProPhoto corrected to gamma 2.2 instead of 1.8?

In LR, the numbers use the actual sRGB 2.2 TRC.