Page 1 of 1

16-bit Pixel Type

Posted: Tue Aug 28, 2012 4:23 am
by impsnldavid
Certain types of LED fixtures can operate in 8-bit or 16-bit modes, where when operating in 16-bit, each property (red, blue, green, white, intensity, etc.) occupies two DMX channels, a coarse and a fine. Whilst obviously when mapping from an 8-bit video source using 16-bit control is pointless, it would still be handy to have to option to specify this as a pixel type. There are lots of situations where it is impractical to change the operating mode of the fixture (there may not be time, the fixture may be rigged and inaccessible, or it may be being used by other systems) so the ability to address fixtures operating in this mode from MadMapper is needed.

Re: 16-bit Pixel Type

Posted: Tue Aug 28, 2012 5:36 pm
by franz
good remark.
can you name any fixture that uses 16bit mode for RGB/Luminosity ?
(it would be handy to get one for testing)

at the moment, you can still use an offset in-between luma values, so that only the MSB (most significant bit) will be sent through artnet, hence reducing your 16bit value to half precision.

Re: 16-bit Pixel Type

Posted: Thu Aug 30, 2012 5:41 am
by impsnldavid
The PixelRange SmartLine fixtures have a 16-bit mode, http://www.pixelrange.com/pixelrange/index.php?id=3.

Re: 16-bit Pixel Type

Posted: Thu Oct 18, 2012 6:45 am
by firionicable
Does this works best with garage Cube?

Re: 16-bit Pixel Type

Posted: Tue Aug 29, 2017 7:41 am
by CaseyJScalf
I am curious how to interface with the pan and tilt of certain moving head lights.

I understand there are only 8 bits of color coming in. 256 steps does not fit nicely into 360+ degrees of rotation.

It looks a little jerky when controlling although very handy. Although other lighting software controls them butter smooth.

How can I smooth out the pan and tilt? Use a 16 bit pixel type? But still, it is only coming in at 8 bits.

How can this be resolved so I can run it all from 1 program, a la MadMapper?

Re: 16-bit Pixel Type

Posted: Mon Oct 09, 2017 7:30 pm
by mad-matt
If you use materials on fixtures you are working with a floating point source (not a RGB 8 bits per component source, like images, movies or Syphon textures - they are 8 bpc in most cases but MadMapper supports other formats) - so you can have more than 8 bits precision in your source !

Re: 16-bit Pixel Type

Posted: Tue Nov 27, 2018 5:09 am
by CaseyJScalf
Do I need to setup or configure the MM fixture or light fixture itself in any special way to control them using such floating point values?

Re: 16-bit Pixel Type

Posted: Tue Nov 27, 2018 5:11 am
by CaseyJScalf
I would second this. I use those low-mid range moving head lights and they look a little jerky with 8-bit only. I would be great to have a way to achieve the 16-bit movement (pan, tilt, etc) using software and not having to set all the fixtures by hand, etc.

Thank you MM!

Re: 16-bit Pixel Type

Posted: Tue Nov 27, 2018 5:24 pm
by mad-matt
Do I need to setup or configure the MM fixture or light fixture itself in any special way to control them using such floating point values?
No. Inside a GPU shader we're always working with floating point.
Cheers!

Re: 16-bit Pixel Type

Posted: Wed Dec 05, 2018 7:25 am
by CaseyJScalf
Thank you.

I understand that the floating point nature of the shaders exists.

However, these fixtures usually use a fine and a coarse adjustment for say a moving head light.

How do I setup the fixture? Do both write the luminance value to 8-bit, 1 channel apart?

I guess I should break one out and experiment.

Re: 16-bit Pixel Type

Posted: Wed Dec 05, 2018 10:12 am
by mad-matt
In MadMapper fixture editor you can make 16 bits channels for red, green, blue, luminosity, or channels controlled with slider etc.
So the 16 bits channel getting the RED value of the shader can be used for a a 16 bits PAN control, the green for tilt etc. If you want to control the pan & tilt with a material shader & the color with another material you can split your fixture in two and set some channels to "unused" in each fixture definition. Then you patch both at the same address. A bit tricky but it opens many possibilities