AVS › AVS Forum › Display Devices › Display Calibration › eeColor Processor - CalMAN - 3D LUTs
New Posts  All Forums:Forum Nav:

eeColor Processor - CalMAN - 3D LUTs

post #1 of 297
3/12/13 at 11:35am
Thread Starter 
At present, both the eeColor Processor and the SpectraCal ColorBox are supported in CalMAN Ultimate along with the ability to save the completed LUTs on your hard drive.

CalMAN Enthusiast supports only the SpectraCal ColorBox but Derek has stated that the eeBox will be supported in a future Enthusiast update.

The differences between the two processors:
1. Faceplate
2. Provided OSDs
3. Firmware

What they do:
1. Hold and process 6 separate 3D LUTs
2. 1 HDMI In, 1 HDMI Out
3. Pass side-by-side and top-bottom 3D

What they don't do:
1. Power on/off is not available with the remote control and must be operated with the power switch on the back of the Box.
2. No USB 3.0 available for Uploading LUTs - USB 2.0 OK
3. Does not pass frame packing 3D

It is possible to use the eeBox with the present version of Enthusiast. I'm doing some screen captures today that will demonstrated the procedures - will provide in post #2 later
Edited by buzzard767 - 3/13/13 at 5:29pm
Reply
Reply

AVS Top Picks

post #2 of 297
3/12/13 at 11:36am
Thread Starter 
eeColor Processor workflow in CalMAN Enthusiast

Select, with the ee remote control, the LUT (1-6) you want to calibrate prior to starting CalMAN

Hooking Up



Select the LUT number matching that which you selected on the eeBox prior to starting CM

DO NOT do a DDC reset



Select Video



Select these Black and White Level Targets



If using CM internal patterns do not select PC level



Unlike the Lumagen Radiance, do not run the RGB Balance module. Skip to the Calibrate 3D Cube LUT module



Select the size Cube and check the Limit Calibration to Reference White box



Run the calibration
Edited by buzzard767 - 3/12/13 at 6:01pm
Reply
Reply
post #3 of 297
3/12/13 at 11:46am
Glad to hear we can use the eecolor box now with Enthusiast. I want to compare the Mini with the eecolor box.

Looking forward to your next post #2..

Thanks,

ss
Reply
Reply
post #4 of 297
3/12/13 at 12:00pm
Quote:
Originally Posted by buzzard767 View Post

It is possible to use the eeBox with the present version of Enthusiast. I'm doing some screen captures today that will demonstrated the procedures - will provide in post #2 later

You may have found an oversight in our licensing that will be fixed. We should not allow a eeColor to be treated as a ColorBox. We do a firmware/serial number check to see if it is a eeColor or a ColorBox. In any case the LUT update for the eeColor and ColorBox will be different once the next release of CalMAN 5 is out. The eeColor procedure will provide you with a LUT file you then use TruVue tools to upload. The ColorBox we write directly to memory.

But in either case once the next CalMAN 5 Enthusiasts update is out you can create and use LUT’s in either version.
Reply
Reply
post #5 of 297
3/12/13 at 12:10pm
What sizes of LUT's will be supported by calman enthusiast?
Reply
Reply
post #6 of 297
3/12/13 at 12:52pm
Quote:
Originally Posted by visca blaugrana View Post

What sizes of LUT's will be supported by calman enthusiast?

You'll be able to run any of our combinations of calibration sizes.

The output you'll want to use is the native 64(65) point LUT. The EE software likes a 65 point file, but only uses 0-1008, any triplet that includes 1023, is excluded from the LUT data.
Reply
Reply
post #7 of 297
3/12/13 at 2:16pm
Thread Starter 
Quote:
Originally Posted by derekjsmith View Post

You may have found an oversight in our licensing that will be fixed. We should not allow a eeColor to be treated as a ColorBox. We do a firmware/serial number check to see if it is a eeColor or a ColorBox. In any case the LUT update for the eeColor and ColorBox will be different once the next release of CalMAN 5 is out. The eeColor procedure will provide you with a LUT file you then use TruVue tools to upload. The ColorBox we write directly to memory.

But in either case once the next CalMAN 5 Enthusiasts update is out you can create and use LUT’s in either version.

Roger that on oversight. I can run the same flow with equal results using either an eeBox or my SpectraCal box. In any case, I've worked out a workflow that is producing acceptable LUTs and will be putting it in post #2. I can't get to it until later as all the graphics are on my work computer and it's running a LUT right now. Once the eeBox is supported I'm sure everything will not only become easier but also faster and maybe even more accurate but the dEs are well beneath 3 as is.
Reply
Reply
post #8 of 297
3/12/13 at 6:02pm
Thread Starter 
eeBox Enthusiast workflow now inserted in post #2 above
Reply
Reply
post #9 of 297
3/12/13 at 6:14pm
Thread Starter 
Quote:
Originally Posted by sotti View Post

You'll be able to run any of our combinations of calibration sizes.

The output you'll want to use is the native 64(65) point LUT. The EE software likes a 65 point file, but only uses 0-1008, any triplet that includes 1023, is excluded from the LUT data.

The ee software not only "likes" a 65x65x65 LUT, the ee software demands it and will reject 64x64x64. This will (has to) be available on the next CalMAN release supporting the eeBox as there is no sense saving anything other than a 65x3 LUT on the hard drive for uploading to the eeBox, now or later.

Also, for TV use the peak LUT value that is active is 940. There is no active picture above that in TV legal.
Reply
Reply
post #10 of 297
3/12/13 at 6:48pm
Quote:
Originally Posted by sotti View Post

Quote:
Originally Posted by visca blaugrana View Post

What sizes of LUT's will be supported by calman enthusiast?

You'll be able to run any of our combinations of calibration sizes.

The output you'll want to use is the native 64(65) point LUT. The EE software likes a 65 point file, but only uses 0-1008, any triplet that includes 1023, is excluded from the LUT data.

The Data's from the 64-Point LUT (262.144) you are creating is adjustable?, i mean from the last point (64) or the LUT....

You mean that from the 65-Point LUT (274.625) the difference of 12.481 more data (vs. the 64-Point) is adjustable but eeColor has that range fixed? and any change of that final 12.481 lines of data has no impact to the Output LUT?
Edited by ConnecTEDDD - 3/12/13 at 7:16pm
Reply
Reply
post #11 of 297
3/12/13 at 7:14pm
Quote:
Originally Posted by ConnecTEDDD View Post

The Data's from the 64-Point LUT (262.144) you are creating is adjustable?, i mean from the last point (64) or the LUT....

You mean that from the 65-Point LUT (274.625) the difference of 12.481 more data (vs. the 64-Point) is adjustable but eeColor has that range fixed? and any change of that final 12.481 lines of data has no impact to the Output LUT?

That's correct, if you actually measure the changes between 251 and 255 in digital 8bit land you'll see that at 251 it's calibrated and at 255 it's native. In ten bit land (the way the box works) there are entries for 0,16,32,...976,992,1008, and then that's it. so the 65/64 LUT is the same spacing, just the 64 lacks all the points with 1023 in the triplet.

The EE itself doesn't physically have memory locations for them. The memory map has each LUT spaced 1MB apart with fixed addressing. The byte order they use packs the 10bit RGB data into a single 32 bit double word and then writes that using exactly 1MB of space to store the data.

There is literally no place for the EE to store that extra information. This isn't a CalMAN decision, it's a limitation of the hardware. Which is why the box is not appropriate for PC levels.
Edited by sotti - 3/12/13 at 7:19pm
Reply
Reply
post #12 of 297
3/12/13 at 7:17pm
Quote:
Originally Posted by sotti View Post

Quote:
Originally Posted by ConnecTEDDD View Post

The Data's from the 64-Point LUT (262.144) you are creating is adjustable?, i mean from the last point (64) or the LUT....

You mean that from the 65-Point LUT (274.625) the difference of 12.481 more data (vs. the 64-Point) is adjustable but eeColor has that range fixed? and any change of that final 12.481 lines of data has no impact to the Output LUT?

That's correct, if you actually measure the changes between 251 and 255 in digital 8bit land you'll see that at 251 it's calibrated and at 255 it's native.

The EE itself doesn't physically have memory locations for them. The memory map has each LUT spaced 1MB apart with fixed addressing. The byte order they use packs the 10bit RGB data into a single 32 bit double word and then writes that using exactly 1MB of space to store the data.

There is literally no place for the EE to store that extra information. This isn't a CalMAN decision, it's a limitation of the hardware. Which is why the box is not appropriate for PC levels.

The Last Line of the 64-Point LUT is 0.98438, that means adjustable data range for eecolor is up to 251 (8bit) - 1004 (10bit), correct?
Edited by ConnecTEDDD - 3/12/13 at 7:22pm
Reply
Reply
post #13 of 297
3/12/13 at 7:23pm
Quote:
Originally Posted by buzzard767 View Post

Also, for TV use the peak LUT value that is active is 940. There is no active picture above that in TV legal.

There are a large number of ways that you can have values above 940.

Many YCC legal values convert to RGB values that have individual values > reference white. or content like live TV may not have been fully legalized, or they may be just using the head room for specular highlights.

Charles Poynton, Don Munsil (Spears and Munsil), and many others can confirm that the headroom should be maintained and correctly calibrated.


That said, for video content the difference between being correctly calibrated from 64-1008 or 64-1023 is minuscule, for PC level content that's an entirely different matter.
Reply
Reply
post #14 of 297
3/12/13 at 7:24pm
Quote:
Originally Posted by ConnecTEDDD View Post

The Last Line of the 64-Point LUT is 0.98438, that means adjustable data range for eecolor is up to 251 (8bit) - 1008 (10bit), correct?
Correct
Reply
Reply
post #15 of 297
3/12/13 at 8:32pm
Are the lookup (input) values adjustable? all the the demo files I've looked at have identical values. I also don't understand why lightspace scales the LUTs for video level output, why doesn't the box do this automatically when converting back to YCbCr?
Reply
Reply
post #16 of 297
3/12/13 at 8:37pm
Quote:
Originally Posted by zoyd View Post

Are the lookup (input) values adjustable? all the the demo files I've looked at have identical values.

The EE file supports two LUT for import, one for the calibrated white point and one for the native white point. This is due to the original goal of the box, producing perceptive LUTs that corrected flesh tones while leaving the edge of the gamut over saturated.

So when you look at the file, you see two sets of output values that are identical, the input values are inferred from their position in the file.
Reply
Reply
post #17 of 297
3/13/13 at 1:23am
Thanks Buzz,

I was wondering about the RGB Balance module and how that would fit into the calibration for eecolor box, now I know.

I should be getting the eecolor box you sold me today, so I will setup CalMAN Enthusiast as you have shown and report back tomorrow.

ss

btw, from what I see in post #2, the calibration writes directly to eecolor box, is that correct?
Edited by sillysally - 3/13/13 at 2:34am
Reply
Reply
post #18 of 297
3/13/13 at 3:07am
Quote:
Originally Posted by sotti View Post

The EE file supports two LUT for import, one for the calibrated white point and one for the native white point. This is due to the original goal of the box, producing perceptive LUTs that corrected flesh tones while leaving the edge of the gamut over saturated.

So when you look at the file, you see two sets of output values that are identical, the input values are inferred from their position in the file.

So you effectively can choose from 12 LUTs?


Does the box handle YCbCr input/output seamlessly or do you have to scale the output LUT? Any signs of banding? (sorry for all the questions but I'm interested in how it works smile.gif )
Reply
Reply
post #19 of 297
3/13/13 at 3:31am
Quote:
Originally Posted by buzzard767 View Post

The ee software not only "likes" a 65x65x65 LUT, the ee software demands it and will reject 64x64x64. This will (has to) be available on the next CalMAN release supporting the eeBox as there is no sense saving anything other than a 65x3 LUT on the hard drive for uploading to the eeBox, now or later.

I'm not sure about this, I downloaded the truvue application and all of the LUTs they have in there stop at bit 1008 (0.985337).
Reply
Reply
post #20 of 297
3/13/13 at 4:33am
Thread Starter 
Quote:
Originally Posted by sillysally View Post

Thanks Buzz,

I was wondering about the RGB Balance module and how that would fit into the calibration for eecolor box, now I know.

I should be getting the eecolor box you sold me today, so I will setup CalMAN Enthusiast as you have shown and report back tomorrow.

ss

btw, from what I see in post #2, the calibration writes directly to eecolor box, is that correct?

Yes, the present version of Enthusiast loads directly to the eeBox. LUTs are also written to the hard drive but they are the wrong format for uploading with the ee TruVue software. An upcoming CM version will include saving 65x LUTs to the hard drive which can archived and loaded anytime.


Quote:
Originally Posted by zoyd View Post

I'm not sure about this, I downloaded the truvue application and all of the LUTs they have in there stop at bit 1008 (0.985337).

Here's the top end of one of the ee canned LUTs and LightSpace produced LUT. They both load.



Ultimate provides the option to produce various LUTs, 65x and 64x included. The 65s load, the 64s don't.
Reply
Reply
post #21 of 297
3/13/13 at 4:43am
Thread Starter 
Here's the result of a LUT produced by CalMAN yesterday. The display is a Samsung C630 LCD with some non defeatable green in the backlight.





12Mar 630 Studio Report-3_12_2013.pdf 717k .pdf file
Reply
Reply
post #22 of 297
3/13/13 at 4:51am
Quote:
Originally Posted by buzzard767 View Post



Ultimate provides the option to produce various LUTs, 65x and 64x included. The 65s load, the 64s don't.

65-Point LightSpace LUT is using Data Points with 6 demical (Input-Output) while 64-Point LUT is using 5 demical Data's wink.gif
Reply
Reply
post #23 of 297
3/13/13 at 4:53am
Thanks Buzz - btw I was looking at the unity LUT from ee and they appear to be wrong, ending in 0.985337 (1008/1023) instead of 0.98438 (1008/1024)

last line from 3DLUT_Unity (from calibration folder)
Code:
0.985337 0.985337 0.985337 0.985337 0.985337 0.985337

last line from 3DLUT_1 (from General folder)
Code:
0.98438 0.98438 0.98438 0.98438 0.98437 0.98438
Reply
Reply
post #24 of 297
3/13/13 at 7:16am
Quote:
Originally Posted by ConnecTEDDD View Post

65-Point LightSpace LUT is using Data Points with 6 demical (Input-Output) while 64-Point LUT is using 5 demical Data's wink.gif

The number of decimal points does not matter past 5 because the internal LUT format is a 10 bit integer 0-1023. The TruVue tools take this normalized 0-1 floating point and scales it to 0-1023 as it is writing it to memory. You get no additional precision with more decimal places. The only thing you are doing is creating a larger file.
Reply
Reply
post #25 of 297
3/13/13 at 7:32am
Here is an example of the unity 3D LUT loaded for the eeColor used by the TruVue tools. Note that the last line is 262144 64^3 not 274625 65^3. Yes you can have values in the LUT file that go from 0-1 which will allow you to achieve 0-1023. But and this is what Joel has been talking about with a 64 your last point in the internal processor is always anchored to 1023 it is fixed. This is why the eeColor LUT files only go to 0.985337 * 1023 = 1008. Yes you can put a 1.00000 * 1023 = 1023 but all you have done is clipped the top values.

I know how the eeColor/ColorBox store LUT data in memory and how it is processed. We have been working with them for over two years and the very lowest level in memory. It does not matter what the LUT file format is or how many data points are in it or how many decimals. The internal LUT format is 10 bit RGB with 262144 entires.

Snapshot of the unity LUT file see the number of lines:


I will load up a snapshot of the LUT memory once I get into the office today. So you can see what it really is.
Reply
Reply
post #26 of 297
3/13/13 at 7:38am
Quote:
Originally Posted by buzzard767 View Post

Ultimate provides the option to produce various LUTs, 65x and 64x included. The 65s load, the 64s don't.

That is because when we create the file for the eeColor we always drop the last entry. So a 65 becomes 64 and a 64 would have become a 63. So yes you are correct a 64 will load and a 63 will not.

Code clip from the file formattter for the eeColor format:
Code:
for(int b=0;b<pointsPerSide-1;b++)
{
        for(int r=0;r<pointsPerSide-1;r++)
        {
                for(int g=0;g<pointsPerSide-1;g++)
                {
                        double red,green,blue;
                        BOOL dirty;

                        if(!getPoint(LUT_POINTS[r],LUT_POINTS[g],LUT_POINTS[b],&red,&green,&blue,&dirty))
                        {
                                red = LUT_POINTS[r];
                                green = LUT_POINTS[g];
                                blue = LUT_POINTS[b];
                        }

                        char line[256];
                        line[0] = NULL;
                        DWORD len = 0;

                        len = sprintf_s(line,256,"%.08f %.08f %.08f %.08f %.08f %.08f\n",red/double(maxValue),green/double(maxValue),blue/double(maxValue),red/double(maxValue),green/double(maxValue),blue/double(maxValue));

                }
        }
}



How much misinformation are we going to have to clean up after LS. Why is this being debated I have all the low level docs, have been working with this hardware for over two years and know exactly how it stores and uses its internal memory.

Easy test take any of the LUT files created for the eeColor by TruVue, CalMAN or LightSpace. Open it up and count the lines you will find 262144 lines so 64^3.
Edited by derekjsmith - 3/13/13 at 7:54am
Reply
Reply
post #27 of 297
3/13/13 at 7:46am
Point: This is one of the big differences between LightSpace and SpectraCal. We get our hands dirty and write LUTs directly to the hardware, Davio, Pluto, ColorBox, FSI, etc... In doing so we have to get very low level docs on memory formats etc... but are under NDA in most cases. But with these low level docs we then know exactly how the hardware is using its LUTs and what the limits are.

I don’t want to turn this thread into a LS vs. CalMAN like what happened in the LS thread. So stop posting but LS said BS here.

I WILL answer any question about how the eeColor/ColorBox work. How CalMAN creates its LUT file for the eeColor or uses the direct memory in the ColorBox.
Reply
Reply
post #28 of 297
3/13/13 at 7:54am
Thread Starter 
Quote:
Originally Posted by derekjsmith View Post

That is because when we create the file for the eeColor we always drop the last entry. So a 65 becomes 64 and a 64 would have become a 63. So yes you are correct a 64 will load and a 63 will not.

How much misinformation are we going to have to clean up after LS. Why is this being debated I have all the low level docs, have been working with this hardware for over two years and know exactly how it stores and uses its internal memory.

Easy test take any of the LUT files created for the eeColor by TruVue, CalMAN or LightSpace. Open it up and count the lines you will find 262144 lines so 64^3.

Can you tell I'm not a techie? lololol

I think you should leave LS out of this so we can keep it CM only. But, for the record, Steve Shaw knows exactly what is going on, works closely with not only Entertainment Experience but several other hardware manufacturers as well, and has found a few problems with their equipment. You might be seeing some changes coming before too long. smile.gif
Reply
Reply
post #29 of 297
3/13/13 at 8:49am
Derek, when you will add support for eeColor 65-Point LUT, what exactly you will do? Add the 65th line to your 64 Point LUT untoutched like 1.00000 1.00000 1.00000 and the eeColor Appplication will discard that line to it's calculations and using similar way the total additional 12481 lines of LUT Data?
Reply
Reply
post #30 of 297
3/13/13 at 10:08am
Quote:
Originally Posted by ConnecTEDDD View Post

Derek, when you will add support for eeColor 65-Point LUT, what exactly you will do? Add the 65th line to your 64 Point LUT untoutched like 1.00000 1.00000 1.00000 and the eeColor Appplication will discard that line to it's calculations and using similar way the total additional 12481 lines of LUT Data?

The eeColor does NOT have a 65 point LUT. It only has 262144 entires in it's LUT. I have all the low levels docs I can show you. Open any of the LUT files and count the lines it is 262144. When we add support for the eeColor LUT file it will be 64^3 just like what we do now.
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration

AVS Top Picks

AVS › AVS Forum › Display Devices › Display Calibration › eeColor Processor - CalMAN - 3D LUTs