1 - 7 of 7 Posts

#### slb

·
##### Registered
Joined
·
1,495 Posts
Discussion Starter · ·
When doing a luminance/gray scale sweep, it appears that Colorfacts always assumes that the black level uses setup (7.5IRE). I would think that this throws off the gamma curve calculations if no setup is used. Is there a way to change the software to use 0IRE? Or, must I calibrate the black level using setup to get an accurate gamma calculation, and then change it back to 0IRE afterward?

Regards,

Steve

#### PencilGeek

·
##### Registered
Joined
·
292 Posts

Quote:
Originally Posted by slb /forum/post/0

When doing a luminance/gray scale sweep, it appears that Colorfacts always assumes that the black level uses setup (7.5IRE). I would think that this throws off the gamma curve calculations if no setup is used. Is there a way to change the software to use 0IRE? Or, must I calibrate the black level using setup to get an accurate gamma calculation, and then change it back to 0IRE afterward?

Regards,

Steve

I'm probably going to step into another can of worms here, but it's been one of those types of days...so why not.

I don't know why the setup would affect the overall gamma curve calculations. The gamma curve calculation should be done iteratively over all of the measured points.

Let's say we're doing an 11-point computation (0, 10, 20, ..., 80, 90, 100). Whether there's a 7.5IRE setup or not, there's still 11 points being measured. To calculate the measured gamma curve, one doesn't need to know if there's a setup or not; they only need to know the maximum luminence (measurement at IRE-100), the current luminence (value measured at current IRE setting), and the current IRE setting (0, 10, 20, ... 80, 90, 100) being measured.

If the target luminence is given by the following forumula:

Tl = Lmax * ((Ti / 100) ^ Tg)

Where:

Tl = Target Luminence

Lmax = Luminence Max (IRE-100 reading)

Ti = Target IRE (IRE-90, IRE-80, etc.)

Tg = Target Gamma

Then the measured gamma can be calculated as follows:

Gamma = Log(Tl/Lmax) / Log(Ti/100)

The "Ti/100" in both equations is really setup-agnostic...if you think about it...because it's really just a ratio of current IRE divided by maximum IRE.

Here's some actual numbers from a spreadsheet I whipped together.

255

100 235 55 55

90 213 42.26384093 42 2.559436666

80 191 31.48383712 31 2.569404213

70 169 22.54798772 22.5 2.505973271

60 147 15.33701405 15 2.543496105

50 125 9.722718241 9.75 2.495957495

40 103 5.565608682 5.5 2.512941595

30 81 2.71122666 2.7 2.503446425

20 59 0.98386991 1 2.489896102

10 37 0.173925271 0.2 2.439332694

0 16 0 0 #NUM!

0

Column-2: RGB value with setup

Column-3: Calculated lumienence (based on gamma=2.5)

Column-4: My own made-up luminence values

Column-5: Calculated Gamma

Notice how column-5 is approximately 2.5 -- and how it doesn't matter on the setup values in Column-2?

Sorry, it looks like the formatting of the table is all screwed up. You could probably cut-paste into a spreadsheet to make it more readable.

#### slb

·
##### Registered
Joined
·
1,495 Posts
Discussion Starter · ·
Okay, if the calculation is done from point to point, you may be correct. My thinking was that if you use setup, then a 10IRE signal is 2.5% drive, but if you don't use setup, then 10IRE is 10% drive. I thought this might cause an error in the calculations if the software assumes setup is being used when it is not.

Thanks,

Steve

#### Gregg Loewen

·
##### Registered
Joined
·
964 Posts
7.0 will include a luminance calculator. You will enter the peak light output reading, and target gamma, then it will calculate the targets for the # of points that you select. This will be an excellent tool for gamma shaping. (sharp PJs, Lumagen scalers, Some Runco PJs etc etc)

#### PencilGeek

·
##### Registered
Joined
·
292 Posts

Quote:
Originally Posted by Gregg Loewen /forum/post/0

7.0 will include a luminance calculator. You will enter the peak light output reading, and target gamma, then it will calculate the targets for the # of points that you select. This will be an excellent tool for gamma shaping. (sharp PJs, Lumagen scalers, Some Runco PJs etc etc)

Well, that actually answers a nagging question that I had about 6.0 -- whether or not the grayscale wizard was already gamma-corrected. If you're adding a luminence calculator, then the answer is an obvious: no.

But why just implement the feature as a calculator? That's useful...but not as useful as it could be. Why enter a reading at all?

It would be far more preferrential to have a wizard -- almost exactly like the grayscale wizard -- that took the max reading, adjusted the grayscale...then worked its way down the gamma curve. The wizard would take the target gamma, and number of gamma points as it's input, and do the rest. It would read the max luminence itself... Since you know what the luminence should be at the next reading, you could say "lower green" , then red/blue...and keep this process going down the gamma curve until finished -- at which point you'll want to work your way back up/down again to fine tune any changes that occurred during the intermediary process.

Even better yet...there should be a check-box in the grayscale wizard: "Adjust grayscale with gamma correction (yes/no)." Seamless integration into a feature users should already be using.

If I were a programmer writing this stuff and wanted to maximize its usefulness, that's what I'd do.

#### krasmuzik

·
##### Registered
Joined
·
11,213 Posts
Wizards are for the SpyderTVPRO customer - I want instruments & tools in ColorFacts so I can work the way I want - and absolutely do not want some software engineer thinking they know how to calibrate better than I. Ever use the automated calibration "feature" in a Runco VX2? It will leave you spinning as you circle the target with your continued lies about the sensor readings to force the circle to land. Luckily they ask you for sensor readings rather than read it themselves from the sensor so you have the opportunity to lie.

#### PencilGeek

·
##### Registered
Joined
·
292 Posts

Quote:
Originally Posted by krasmuzik /forum/post/0

Wizards are for the SpyderTVPRO customer - I want instruments & tools in ColorFacts so I can work the way I want - and absolutely do not want some software engineer thinking they know how to calibrate better than I. Ever use the automated calibration "feature" in a Runco VX2? It will leave you spinning as you circle the target with your continued lies about the sensor readings to force the circle to land. Luckily they ask you for sensor readings rather than read it themselves from the sensor so you have the opportunity to lie.

Remember, computers only do what they are programmed to do...they don't lie. They are inherently truth-tellers, dealing with 1's and 0's. It's worth mentioning these two situations with a "lying sensor" -- even though it doesn't sound like either of them apply to you. 1) If the sensor has some type of LCD telling you the values that don't match the data being returned on the USB/Serial connection, then that sounds like a bug in the sensor. Such a sensor would be very expensive; and being so expensive, I'd go straight to the manufacturer and insist this be fixed immediately. After all, the sensor is absolutely useless if it returns false data, so I don't see the manufacturer refusing such a request for such an expensive item. 2) The sensor really does return the correct data, but the software screws it up. This isn't the fault of the sensor, this is a programming error in the software, and should be fixed. If the sensor returns the correct data and is displayed on the user-interface correctly, but sends a different set of data to an application-programming-interface (API), then clearly there would be a MAJOR bug in the software.

But I think you're in a third situation with your sensor. I think you said that the software requires you to manually input data, but that the sensor gives you false data. If that's the case 1) how do you know the sensor returns false data? Then how do you know what data to lie about and input into the software? 2) why haven't you replaced it or had it fixed? I certainly wouldn't use such a sensor.

A well designed wizard interface would not be a burden to you, and would still allow you complete control over the outcome. I didn't suggest that the wizard would take control and auto-tune anything on your behalf (ala C3) -- according to what some pinhead software engineer (like myself) thinks is best. The wizard would only be there to make your job easier, more productive, more cost-effective, and therefore more profitable. If these are not things you value, then you still have the freedom to ignore the feature and continue using your own techniques.

On the other hand, if it did offer a C3 auto-Gamma, I wouldn't hesitate to use it. My inclination would be to trust the computer long before the subjective opinions of a human. But that doesn't mean I wouldn't cross-check the results to build confidence in the computer method. After it was done, I would verify the results and make any fine-tuning adjustments that seemed necessary. If it proved untrustworthy, then I wouldn't use it again.

1 - 7 of 7 Posts