Chinese LED moving head clone colour calibration

Hi Darren

I have nothing specific to answer with, but I recall from previous discussions that the "wrong mixing" of LED and other devices is inherent in the Eos family. 

(ETC: correct me if I'm wrong)  ETC farmed out the contract to someone else to provide the colour mixing programming (IE blue at 80%, red at 40% yellow at 22% on a Seledor will equal R80 for the seledor).  I think it is a known weakness that the preprogrammed colours don't match up to the Rosco values by eye.

The work-around (long and tedious as it is) is to make your own colour palettes matched to a known piece of gel.  Do a few colours a show, and keep a Base SHow File, and eventually you'll get a library of Marsha's favourite colours.

Good Luck

Andrew Riter

Thanks Andrew. Hoping I will hear something from Anne @ ETC soon as she mentioned a couple years ago that they were trying to develop some sort of colour editing tool for the Ion/Eos line.

Cheers

Darren

And here she is.  I'm not sure that 'wrong mixing" inherent in the design of Eos is exactly the correct way to put it.  Color matching is a complex issue.  Mike Wood wrote an excellent article on this topic several months ago.

http://www.mikewoodconsulting.com/articles/Protocol%20Spring%202012%20-%20Color%20Pickers.pdf    

Mike is far more articulate than I in this regard and this article is most certainly worth a read.   However, there is a comment to be made about expectation management and that Eos uses the same core color match data as several other main-line manufacturers.  

First things first.   There are both calibrated and uncalibrated fixtures in the library.  Color calibration requires physically bringing a unit into a color lab --- with the thousands of fixtures in the library, and the struggle just to keep up getting profiles built for the rapid proliferation of devices, this is not always possible.  We try to get the mainstream, higher-end fixtures calibrated.   If you see a black gamet line in the color picker, you know the fixture has been calibrated.  If not calibrated - we use a generic model that should at least get you in the ballpark.    But for this discussion, assume the fixture has been calibrated.

If you are working with a dichroic mixing system, there are many many things that impact the resultant mixed color of the unit:

The lens coatings, the reflector coatings, the mechanics of the mixing system, the dichroic coatings themselves,  the color temperature of the lamp (which is not exact from the box, and shifts over time), the CRI of the lamp (even two brand new lamps from the same batch are not guaranteed to have the same color temperature and CRI).  

LED fixtures can have huge binning problems.  Higher end fixtures will often have calibration built in to compensate for this.   More cost effective fixtures do not have this and often have very little bin control.  LED fixtures can also be subject to color shifts as the units heat and cool.      Matt Kerr or Jim Uphoff from the fixtures team will know more about the Selador color matching process, and have asked them to comment.  

120V will render color differently than 240V.   Even voltage differences in the same venue can be enough to impact certain colors, dependent on the color you are trying to mix.

It is also worth noting that currently the color engine is set to match color to a 3200 source.   

Add to all of this that we ALL see color differently.  If you bring two designers and two programmers into a room, ask them to mix two identical fixtures with calibrated lamps (to assure the same color temperature and same CRI) to R80, you will end up with two different sets of output values.  

Now, having said all of that - this is an ongoing area of experimentation for us.   We have improved the color matching several times in the past couple of years when issues were identified.   2.0.1 introduced new algorithms to provide greater precision.   We have several other things in work as well.  

However, as you can tell.... this will NEVER be perfect.   And it isn't always going to be the fault of the color engine in the desk. But we are always working on ways to make it better and to provide tools for transitioning through color space.  As those tools are introduced, you'll read about them here first.  

Hope that helps!!

a

Hi Anne,

Thank you so much for your quick reply.

i knew it was a complex set of information to deal with and try to match colours, but wow.  I really appreciate the info.

So here is the next problem then...if I go with Andrew's suggestion and start building colour palettes that can be saved as a template file and merged into the active show, the fixtures will always have to be on the same channel, as palettes are channel related data (if I'm not mistaken) unlike fixture specific info (personalities).

i guess this is possible as long as I edit each palette to alter the channels on each production. Or am I missing something that would make this process easier?

Thanks in advance.

Darren

Interesting stuff, isn't it?  :-)   It is always recommended to put data into palettes/presets when possible.  It makes maintenance of the show much easier.   Color palettes are channel specific, but it is easy to move channel data around as needed.  So you could start with a basic patch, build your palettes and when you start each show, move the channels around as you need to for that particular show - using the "move to entire show" function in patch.  

You might also try using by-type palettes, using the source data for them from higher numbered channels that might not otherwise be used.    However, I would suspect that due to the afore mentioned binning issues on certain fixtures, this might not be worth it for you.  A quick way to determine that is to set them all to a couple of specific colors from the gel picker and see if they all match - even if it isn't a color you are going to want to use later - it's just a test. If you end up having to tweak them individually to get the same color from them all, you will probably be better off going with plan A.   But there is a good description in the manual of building and maintaining by type palettes - and providing discrete data for channels when needed - in the manual should you decide to go that route.

Hope that is useful!

a 

Interesting stuff, isn't it?  :-)   It is always recommended to put data into palettes/presets when possible.  It makes maintenance of the show much easier.   Color palettes are channel specific, but it is easy to move channel data around as needed.  So you could start with a basic patch, build your palettes and when you start each show, move the channels around as you need to for that particular show - using the "move to entire show" function in patch.  

You might also try using by-type palettes, using the source data for them from higher numbered channels that might not otherwise be used.    However, I would suspect that due to the afore mentioned binning issues on certain fixtures, this might not be worth it for you.  A quick way to determine that is to set them all to a couple of specific colors from the gel picker and see if they all match - even if it isn't a color you are going to want to use later - it's just a test. If you end up having to tweak them individually to get the same color from them all, you will probably be better off going with plan A.   But there is a good description in the manual of building and maintaining by type palettes - and providing discrete data for channels when needed - in the manual should you decide to go that route.

Hope that is useful!

a