[Q] Software control of LCD backlight?

[Alex Deucher]

On 8/10/07, Hal V. Engel <hvengel at astound.net> wrote:
> On Friday 10 August 2007 07:26, Xavier Bestel wrote:
> > On Fri, 2007-08-10 at 09:05 -0500, Pat Kane wrote:
> > > Is it possible to vary the intensity of the backlight on
> > > modern LCD panels (e.g. Dell 17'') from software
> > > such as the X server?
> >
> > Yes, there is http://ddccontrol.sourceforge.net/ which should do it.
> > This kind of thing should be handled by RandR imho, but apparently the
> > monitors don't follow a common spec.
> >
> >       Xav
>
> The "spec" is really two specifications from VESA called DDC/CI (the
> communication protocol) and the Monitor Control Command Set (the op codes
> that can be passed to the device).  So there are a number of things like
> monitor luminosity, RGB channel gains... that are part of VESA DDC/CI and the
> Monitor Control Command Set specifications.  Therefore if you have a monitor
> that supports DDC/CI along with a video card and driver that support it an
> application like ddccontrol should work for those things that are in the VESA
> specifications and are supported by the monitor (IE. much of the basic
> stuff).
>
> Apparently it is OK for a monitor to support either a subset or a superset of
> the the VESA specifications.  Many monitors do support supersets and some
> vendors even go as far as implementing some of the basic controls as part of
> a superset and not implementing those controls in the standard way because
> they believe that they get a proprietary advantage from doing this.
> Unfortunately none of these vendors have published the interface
> specifications for these devices and the only way to figure out these
> extensions is by probing the device to get a list of extensions and then
> trying each extension to see what it does.
>
> So far the only platform that provides a native DDC/CI API is Vista.  How much
> of the VESA specification is exposed in the Vista API is something I don't
> know since the specifications are not available to the public without paying
> a substantial fee ($350/document and you need two documents to fully define
> the standard).   But apparently the Vista developers felt that what they
> exposed was enough to be useful and that they could figure out how to make
> this work with most monitors that support DDC/CI.  I don't know much about
> how this works on Vista but I suspect that Microsoft expects vendors of
> DDC/CI capable monitors to include hooks of some sort in the
> monitor "driver".
>
> It is apparent that the situation with DDC/CI is not a good one.  It is also
> apparent that most of the monitor vendors don't seem to mind the current
> situation since the DDC/CI specification has been through several revision
> cycles and so far the problems have not been rectified.
>
> To make things even worse there is also the USB HID Monitor specification
> which does the same basic thing but using a USB connection to the monitor. At
> least in this case there is a publicly available specification for the
> communications protocol.  Please see  the link  here
> http://www.usb.org/developers/hidpage/ for details.  This specification is
> supported by some vendors such as the Eizo and Apple among others.  This
> specification appears to use the same op codes as specified by VESA in the
> Monitor Control Command Set.
>
> At some point most monitors will stop providing front panel controls and
> require a software interface for making these adjustments.  Therefore in the
> long run this functionality does belong somewhere in the X server and I guess
> that RandR is as good a place as any.  But there will be significant issues
> getting a good implementation of this in place.  Personnaly I would like to
> see this sooner rather than later.

I don't know anything much about the ddc/ci spec, but it would be
pretty easy to expose the i2c interface for each output in randr.
What's we'd do wtih that interface depends on what's needed for
ddc/ci.

Alex