Main Site Documentation

Cobra AmbieLight


I would like to show you my new Ambie Light Clone Projekt.
A ghi cobra uses a camera to get the tv picture and then analyze the average picture color. this color is send to led strip of 128 led’s. no laptop or other hardware is needed.

Here is a example video :

Tell me what you think or is someone making another ambielight clone ?



RRRRRR! You stole one of my great ideas! :wink: The problem with every other AmbieLight solution is they only work when the computer is the controller. Using a camera makes it work with any source. Now you just need to work on breaking down the sampling so you can get differing colors on different parts of the edges instead of every LED showing the same color. Looking good!

BTW, for the video tags you only need to put “m2Gg9_XzCqY” between the tags.


@ af6767

Nice one! :clap:


@ Architect:

@ Ianlee74:
Different parts of the edges is a camera problem. I cant get the whole picture of the tv, this is only possible if the camera is nearly 1 m in front of the tv. No waf :smiley:



Would it be possibly to create a hdmi sniffer to get the average color instead of a camera?


@ wouter
That is what i am thinking about.


Thats totally cool!

What if you mount your cam to the ceiling, so you can see the entire picture and than separate it?

@ wouter I think a hdmi sniffer is not possible due to handshakes, keys and protection stuff… (


Wow man! This is the first “image processing” NETMF project EVER! Nicely done. Simply yet awesome.


My plan was to mount the camera on the wall opposite the TV. This obviously could be a problem depending on the layout of your room.

Yep, many have tried the HDMI sniffing route before without success. If you can figure that one out you’ll have geek cred at the infamous level. :wink:


So where is the code? :wink:


Code? Heck, I want a to see a full WIKI write up on this one :slight_smile:

What camera did you use?


I used a Logitech C910, because all other cameras are ugly :slight_smile:
Positioning the camera is the main and only problem in this solution.
You have no option to position the camera and then zoom perfectly the tv screen.



Another point, maybe someone knows a trick, giving the bitmap array to rlp is impossible in performance sight. so i have to get the average color in manage code. invokeex is useless too.
i thinking of a memory area, that is availeble for both, managed an native code. Or the camera writtes directly to a native memory area. Someone have a hint ?



@ Weiti: in the same article:

So it is possible, because the master key is out somewhere. You’re just not allowed to create a device that uses it.

Maybe interresting device: