eeColor Processor - CalMAN - 3D LUTs - Page 2

If the EE wants a 65 point file, we'll generate a 65 point file with accurate data throughout.

I believe you, but how it can be uploaded using eeColor Application for user with CalMAN Enthousiast Licence if it's 64-Point TXT File? eeColor Application requests only 65...

This is not true, if you look at all the LUTs provided with the eeColor application (TruVue) they are all 64^3. So if CalMan writes 64^3 files the TruVue app will download them to the box, there is no point in writing 65^3 files.

The plan is enthusiast will look exactly the same, except that it will have a shorter list of supported export formats.

Sorry but this is not correct...

Installing the eeColor Application you see 64-Point LUT's but if you open them the 1.00000 1.00000 1.00000 (100%) Line is missing.

This is not a true 64-Point LUT File, no valid 64-Point cube (True 64-Point LUT has data for 100% also). You need to upload a 65-Point to eeColor see 64-Point, last line is removed from software. Got it now?

Do you have eeColor and you have upload 64-Point LUT from CalMAN that worked for you?

If you try to export eeColor 65-Point Cube from CalMAN, then a 65_16_CUBE.TXT file is created and if you open that file you will see that the 1.00000 1.00000 1.00000 (100%) Line is missing, so it's a 64-Point LUT for eeColor Use only, and that is the correct way.

It looks like there was a little confusion, internally.
The formatting code for the EE file format was stripping out the last entry. So the 65 point setting was creating a 64 point file, the 64 point setting was creating a 63 point file. That's why the "64 point"(63 in reality) wasn't working correctly.

We just had a meeting about this, and it will be simplified in the future.

If you want to know how much data is in the file, you need to look at the number of lines, not the last output triplet. The last output triplet simply means the input values are output at 1023,1023,1023 It doesn't reveal what the input triplets are. Proper EE files have 262,144 lines (64^3).

Hey Buzz. Out of curiousity, how long does it take to do a eeColor calibration through Calman? The Radiance 5x5x5 calibration takes me about 35 minutes. Just wondering if it adds a huge amount of time for a 65x65x65 calibration.

It all will depend on your generator and meter used. So for CalMAN to generate a 17x9 to be used with a 64^3 that will be about 3000-3500 reads. So if your meter and generator combination is about 1 second it would be about an hour. But if your meter and generator combination is 3 seconds then it will be around 3 hours.

Also we have found the timing issue Ted and Buzz were having , so this will also be much faster in general.
Edited by derekjsmith - 3/13/13 at 12:23pm

Thanks for the clarification, I have verified with my ee and it's application software that it will accept either 64^3 or 65^3 lines and had assumed it just throws out any line with a value of 1.0, is that correct? or must you have exactly the correct number of lines? Also related, what are the exact RGB values being modified by line, is it:

R G B
line 1 0 0 0
line 2 0 16 0
line 3 0 32 0
...
line 64 0 1008 0

or...

line 1 0 0 0
line 2 0 15 0
line 3 0 31 0
...
line 64 0 1007 0

Thanks! That helps. I would like to think about what I would need to calibrate the eeColor. I really don't have a super expensive meter, nor would I want to spend that much money. I'm using the common combination of an i1pro and Chroma 5. The i1pro seems pretty good for gamut calibration, but it gets pretty loose at lower light levels (like under 40 IRE). My 25% primaries are usually all over the place (up to 5-6 DeltaE sometimes). My Chroma 5 gets all wonky when I try to do gamut and it doesn't work at all, but it does low-light greyscale pretty well. So what I do now, is profile my Chroma5 against the i1pro for greyscale and then use the i1pro for the 125 point Radiance CMS.

I can't remember how many reads were required for the 125-point Radiance CMS. I think I remember someting around 353 reads for a 35 minute process (correct me if I'm wrong). If this is correct, then I could be running this thing for 9-10 hours to do a calibration on the eeColor (based on your 3000-3500 reads). Is there a better alternative for us hobbyists? Would the OEM i1Display or C6 meters read gamut okay and within acceptable speeds for a home calibrations? Or would the i1pro still be a better meter for the gamut?

I'm okay waiting 2-3 hours, but 10 hours is pretty long. I'm running a Sony SXRD VW95ES.

Derek, I knew that eeColor has 16bit internal processing and support for 8-10-12bit input-output signal.

As you are saying, the 65/64 Point LUT is stored at 10bit space internally. So if I feed the eecolor with 12-bit signal then it's downscaled to 10bit to apply the correction and then upscaled to 12 bit to the output?

Not necessarily.

It's storing the data as 10 bit to fit it in the flash memory. But when it selects a LUT, it could easily be padding that data and processing the internal signal at a higher bit depth.

So... just to get an idea with a package sold by Spectracal...

If I'm using a Mini 3D as a pattern generator and a C6 as a meter, where does that fall? It's pretty quick with the 5x5x5 readings, still longer than 2 secs I think. But that is a different animal with the eeColor box as there's a device write process (to the Mini's calibration settings) that won't be done.

So... Mini 3D/C6/eeColor... a usual combo I would think... any idea of time?
Edited by erkq - 3/13/13 at 5:20pm

Success has a 1000 mothers, and fathers!

Here is a screen shot of how long it took Calman 5.1 to run a 64 point (see this post http://www.avsforum.com/t/1462960/eecolor-processor-calman-3d-luts#post_23071948) for the eecolor box.
I did not check low light handler. My meter is a C6 profiled by a I1pro 2.



ss
Edited by sillysally - 3/16/13 at 6:50am

nice ss

tell us more..., what's the difference in PQ between mini 3d and ee. How is ur display now with this lut...i guess all the calman results looks good.

dan

Correct me if I am wrong. but isn't 16^3 a total of 4096 points?

You have read only 1530, which with a system that reads each patch multiple times suggests a cube size well below 16^3

???

Can you post the post calibration 21-Step Grayscale + 5 Point Saturation Charts?

Dan, I need to view some source material. However I did run my S&M disc, the stepped grayscale is off plus it looks like eecolor is crushing the colors. So I have to figure out why that is happening and if its coming from the calibration or the eecolor box.
However the calibration seems to check out well running the calibration verification of Calmans.


This is a 64 piont LUT, see this post for details. http://www.avsforum.com/t/1462960/eecolor-processor-calman-3d-luts#post_23071948.

I am trying to figure out how to setup Calman 5.1 to run a eecolor box, I am coming from a Radiance Mini 3D with a max 5^3.

ss
Edited by sillysally - 3/16/13 at 6:51am

According to Buzz, using Calman 5.1 build 1130 I can't run a grayscale. see post (http://www.avsforum.com/t/1462960/eecolor-processor-calman-3d-luts#post_23071948)

So all I did was touch up the grayscale using the internal controls of my VT50 after running the cube (hate doing that because I can't touch up the cube after). And that is probably why my stepped grascale is a little off.

I tried to open the report that was saved in calman but it will not open, so I can't show anything except what I took a picture of.

ss

Capture a screen of 21- grayscale & 5-Step Saturation Charts if you have time

Here is a LUT i ran last night with my eecolor on my Kuro using the standard run(279 pts). I used a c6 profiled against an i1pro.



Tedd, can you post your results from Calman?

I had a software crash before saving, currenty I have only saved results from LightSpace, there

The LUT may be 64^3, but the measured points are a lot lower - you originally said 16^3, but I think it looks like actually 9^3.

^^^^
Yes you are correct.

ss

Yes I did as you suggested (Unity files) with my second run. I didn't do a DDC reset per your instructions.

I only did a 2 point (30/100) grayscale using my VT50's W/B controls. When I used my Mini for a 21 step grayscale ramp after the calibration, it was way off. Also I ran my B&S disc and it showed that the RGB/W are crushing.
Calman almost crashed at the end of this calibration before it download the settings to eecolor, and then calman crashed.

I am running a new calibration after I did a 10 point grayscale using the VT50's controls in ISF day mode, reloaded the ee unity files also.
The new calibration has about 576 points remaining, so when it is done I will report back.

Thanks Buzz,
ss

Ah, that's what I thought - so two and a quarter hours for approx 780 points, with 1530 reads, which is a standard time duration for an i1D3.