[GRLUG] LINUX and NOT LINUX

[Michael Mol]

On Tue, Sep 15, 2009 at 2:05 PM, Bob Kline <bob.kline at gmail.com> wrote:
> On Tue, Sep 15, 2009 at 1:28 PM, Michael Mol <mikemol at gmail.com> wrote:
>> > Looks to me like this is a big enough job
>> > that it should only be done once, preferably
>> > by the people doing the reverse engineering.
>>
>> No, because there are manufacturing level differences that are
>> important if you want precision.
>
> I suspect for most people those
> differences are quite minimal.  i.e.,
> the sensors are solid state, and
> probably pretty uniform to start with.
>
> That said, the little package for windoz
> that comes with the unit does have a
> calibration step when you turn on the printer.
> It uses the white surface of the cover, which might
> be where the uniformity you refer to
> really comes in. ( For this reason one
> doesn't want any material in the scanner
> when they start..... )
>
> I just looked at my HP unit, and there
> too, the white cover could be used.  If
> one knew how that calibration program
> works,

I can see three possibilities.  Either it assumes a linear response
curve per channel from pitch black to full bright(Stupid; The curve
isn't going to be linear.), it takes the two values of black and
bright and extrapolates based on an estimated curve (also stupid,
because variations in the detector and support circuitry will lead to
variances in the curve), or it's capable of varying the brightness of
the illumination in controlled, precalibrated steps. (Extremely
unlikely.)

In short, if it works, it works badly.

> it would be nice to have under
> Linux.  That way one only has to assume
> they know the white cover material is
> in place.
>>
>> > It take equipment the typical putterer is not
>> > likely to even have access to, much less own.
>>
>> Reverse-engineering USB traffic between Windows and a USB device?
>> Windows VM guest on a Linux host, using the tools that exist on Linux
>> for sniffing USB traffic.  No special hardware required.
>>
> No, the photometric equipment and
> charts referred to.  Less of on issue on
> output,  you nevertheless need to have
> some concept of when you've arrived.
> Maybe color charts.  Things like color
> accuracy span the gamut from your
> device to room light to whether you're
> color blind in any way.  Real precision
> is an art.

In output, yes. In input, not so much. My day job involves
observation, measurement and analysis with optics, albeit in a
different area of the spectrum.

> Go to a paint store for a color
> today and they just might have you
> stand in a booth where you first adjust
> the lighting, to give you some idea what
> will happen after you buy it and splatter
> it on your walls.

Read up on IT8 targets, and get your hands on one of agreeable quality.

http://www.computer-darkroom.com/it8cal/it8_page_1.htm

>
> Anyway, this is all good stuff to know
> about.  Even though my Canon FB630Ui
> is ten years old now, and likely long out
> of production, I will at least interact with
> the people doing the real work and see if
> they're interested in tuning up support
> for it.

Interaction is always a good thing.

-- 
:wq