JVC Autocalibrage avec la i1D3 !!!!

[jacko]

Merci à JJ-34 pour le lien !

post180217587.html#p180217587

Tout se trouve sur le lien suivant :

https://autocal.banjobyte.com/?lang=en

Merci Steven l'auteur du DLL.

[jacko]

Ce qui suit est tiré su site de Steven.

Dernière version du DLL

[jacko]

Description

This project shall provide a DLL file which can be integrated into the Autocal software provided by JVC. This makes it possible to use the i1 Display Pro (i1d3) sensor for the calibration process.

The whole thing has been tested so far with:

* on my JVC X5000 with Projector Calibration Software 7
* on a NX9 with Projector Calibration Software 11

If you have tested it successfully you can write me with which projector it worked. Just write a mail to autocal@banjobyte.com with the subject "Test success Autocal".

If you find errors or have suggestions, you can also send them to the above mentioned mail address, with a little luck I'll put that in

Attention!

I'm not responsible for anything that happens with your PC, projector, sensor or anything else. Everything happens at your own risk!

[jacko]

Support
You like this project and therefore want to support me? Of course this would motivate me a lot to develop further. Just use my PayPal-Me Link. For the amount you can orientate yourself on common calibration solutions and the amounts you have invested in hardware so far No seriously, I would be happy about every Euro.

[jacko]

Download
Versions
v0.34 (2020-03-26)
Configuration file for setup reading mode and times.

A settings.ini is now created in the directory Documents\autocal. Here you can define the measurement times and the measurement mode. Available modes are:

0 -> AIO
1 -> CRT
2 -> LCD
3 -> Burst
Times can be defined for Y measurements greater than 1, less than 1.0, less than 0.1 and less than 0.01.

v0.33 (2020-03-25)
Reduce integration time for low light to 2 seconds

v0.32 (2020-03-19)
Support sensors without AIO mode

v0.31 (2020-03-15)
Quick installation guide

Simply extract the file into the JVC installation folder. It is best to rename the dccmtr.dll before and unzip it. Additionally you need the i1d3SDK.dll. You will find it if you have installed the x-rite software in the directory "C:\Program Files (x86)\X-Rite\Devices\i1d3". Copy them into the JVC directory with the rest. After that you should be able to check with the i1test.exe if you can access your i1D3. If this works you can try the calibration process. Make sure that you align the sensor towards the screen and select the Spyder sensor in the software as normal. Don't take it too seriously with the distance display and brightness, that's not really true with the i1D3. It probably simply shows too little light here, but it doesn't matter. The correction matrix is created after the first use or after running the i1test.exe in the "My Documents\autocal" folder.

[jacko]

Whats happening?
In the JVC software directory (c:\program files\Projector Calibration Software ...) there is a file dccmtr.dll. This file normally addresses the Spyder sensor for the autocal process. This file is renamed or backed up somewhere else and then replaced with the new file. After that the autocal process is started as usual in the software and the Spyder sensor is selected again. But now the Spyder sensor is no longer addressed, instead the i1d3 is used. But this leads to the fact that everything is displayed a little bit wrong when the sensor is aligned. Here, it was proved to be very useful to align the sensor to the screen, as it get wrong readings by aligning it in the direction of the projector because it is to bright. The software now probably shows it is too dark, you can simply ignore that. Then go on with autocal like it is always done. Before you start the whole process, it is worthwhile to align the sensor with another software (e.g. HCFR) to measure maximum light. This way there is more reserve for the dark test fields.

Some more information about Autocal in general (applies to the X-series, for the N-series I don't know for sure yet)

All information is without guarantee of correctness and completeness

If you have any corrections or more hints, please let me know

Gamma calibration

The gamma correction shows the test fields completely uncorrected. Thus, here, it is measured how much light arrives on the screen if I control the LCD-panel with a value of X. Thereby, the control is done in a range of values from 0 to 4095. This is now accomplished with different values and from these, a curve is interpolated. The brightness is scaled to a range from 0 to 1. This curve is then used to create a LUT (Look-up-table) which the projector can use to look up: If I want to achieve an actual brightness of e.g. 50% with the red panel, what value should I use to control the panel. In the table the value is e.g. 3127, so with this calculated LUT he can calculate exactly with how many volts he has to control the display panel in order to generate a desired real brightness (which he actually wants to have because of the image information). In theory, this can also compensate the gamma drift as long as the panel can darken in different degrees.

Now the important thing when calibrating: This gamma LUT is available in the projector per:

2D / 3D Input
CMDrive on / off (only for 2D input)
High / low Lamp Power
CineFilter on / off
Filter Lamp Power 2D/3D CMDrive
CineFilter off low Lamp Power 2D CMDrive off
CineFilter off low Lamp Power 2D CMDrive on
CineFilter off low Lamp Power 3D CMDrive doesn't matter
CineFilter off hoher Lamp Power 2D CMDrive off
CineFilter off high Lamp Power 2D CMDrive on
CineFilter off high Lamp Power 3D CMDrive doesn't matter
CineFilter on low Lamp Power 2D CMDrive off
CineFilter on low Lamp Power 2D CMDrive on
CineFilter on low Lamp Power 3D CMDrive doesn't matter
CineFilter on high Lamp Power 2D CMDrive off
CineFilter on high Lamp Power 2D CMDrive on
CineFilter on high Lamp Power 3D CMDrive doesn't matter
There are therefore a total of 12 memory locations that can be calibrated. Simply calibrate the slots that you really use.

Color Calibration

The color calibration is done by displaying 4 patterns. White, red, green and blue. Everything again without further correction directly with the panel with 88% brightness. White is not used very much here, the three colors are important. From these, a 3x3 matrix is calculated. This matrix describes the maximum possible colour space of the projector and is needed in order to display the colour to be displayed correctly in the projector colour space. From this matrix and the white point (which is programmed to D65) a vector with 3 values is calculated. This vector is responsible that the calibrated projector is not as bright as it could be. The values here are actually always smaller than 1 and are multiplied by the colors to be displayed. The reason is that the panels can't display brighter than 100% but have a different native white point. So they have to be dimmed partially to get the other white point.

Now to the important thing when calibrating: This matrix is available for the following settings

aperture with following boundaries: -12, -8, -4 and 0 (the greatest last value in each case)
Lamp Power
CineFilter
Aperture Lamp Power CineFilter
-15 to -12 low off
-11 to -8 low off
-7 to -4 low off
-3 to 0 low off
-15 to -12 high off
-11 to -8 high off
-7 to -4 high off
-3 to 0 high off
-15 to -12 low on
-11 to -8 low on
-7 to -4 low on
-3 to 0 low on
-15 to -12 high on
-11 to -8 high on
-7 to -4 high on
-3 to 0 high on
So here are 16 memory locations that can be calibrated. The difficult thing here is the aperture, because it changes constantly between all values with automatic aperture enabled.

[jacko]

Versions
v0.4
ID Title State Created at
9 Spyder 5 (bypass) über alte DLL unterstützen open 2020-03-14
8 Klein K10-A unterstützung open 2020-03-14
7 Auswahl zwischen verschiedenen Sensoren bei Initialisierung hinzufügen open 2020-03-14
Maybe for later
ID Title State Created at
6 i1d3SDK durch eigenen Code ersetzen open 2020-03-13
5 Testprogramm erweitern, evtl. GUI open 2020-03-12
4 Matrix / CCSS Korrekturen für die original DLL (Spyder Sensor) open 2020-03-12
3 CCSS-Korrektur ermöglichen open 2020-03-12

[jacko]

Who am I
Who am I, anyway? That is a good and valid question. My name is Steven, I am currently 30 years old and I have been a full-time software developer in the SAP area since 2008. My degree is "Diplom Wirtschaftsinformatiker (FH)". You can say that I am a developer with full passion and have been doing this since I was a child. About 16 years ago I started developing with C/C++ and over the years I got to know many other languages like C#, PHP, Ruby, Python, Java, Assembler, ABAP and JavaScript (order unknown). Nevertheless C++ is still my favourite language but mostly I do command line tools and less GUI and this mostly under Linux. I like to make websites in Ruby on Rails, the performance is annoying, but you get results really fast when programming.

So why does this website look so shitty? Because I didn't want to make it pretty yet / because I'm not a designer / because I simply created it to have a project page at all

[le shad]

Si une âme charitable et francophone pouvait traduire cela, ce serait au top..

Après oui un coup de Google traduction pourrait suffire mais bon c’est quand même pas pareil..

Merci

[jacko]

Le but du thread est de suivre les évolutions de ce projet qui va grandement améliorer la performance de l'Autocal pour JVC.

Bravo à Steven !

[jacko]

Si une âme charitable et francophone pouvait traduire cela, ce serait au top..

Après oui un coup de Google traduction pourrait suffire mais bon c’est quand même pas pareil..

Merci

C'est l'idée, mais ça ne se fera pas en 5 minutes

[yoyoman]

Testé et approuvé avec un jvc x7500 et auteur de la dll récompensé.

C’est vraiment une bonne nouvelle !


Envoyé de mon iPhone en utilisant Tapatalk

[jeff51]

Avez-vous testé aussi l'autocal HDR?

[padhc]

Et du coup le résultat est aussi bon pour les couleurs, au point de rendre l'étape de calibration manuelle qui suivait l'autocal avec la Spyder inutile ?

[le shad]

Apparemment oui, à confirmer par ceux qui ont effectué une calibration..