[sane-devel] ScanJet HP46x0

[m. allan noah]

On Wed, Jan 6, 2010 at 3:50 AM, Ille <tam.ille at free.fr> wrote:
> Hi,
> Happy New Year to every one!
>
> I did some work to get the hp4600/4670 working, based on the
> mustek_usb2 backend. So far, the communication with the scanner is not
> far from being completed, some registers are still not exactly
> identified, but they seem to be some kind of constant (resolution
> dependant although).

For many machines which have no documentation we have to send unknown registers.

> Now I'm trying to understand how gamma/color correction table is
> computed. The W$ driver send a table of 0x600 bytes (which I assume is
> Gamma/rgb correction),

That sounds like three channels (RGB) of gamma slopes. Is there a
pattern in this data?

> then another one whose size vary depending on
> resolution (I don't know what this is),

Almost certainly per-cell correction information, probably a black
offset and a gain, sometimes interlaced.

 then a third one who is 0x1000
> bytes (I think there are 4 times 0x250 bytes: values are presented in
> the first 0x250 bytes, padded with 0, then repeated in reverse order in
> the next 0x250 bytes, padded with 0, the first 0x500 bytes are then
> repeated).

Probably motor slope tables, positive and negative accelleration, both
forward and backward
Some models have many more specialized slopes for things like
backtracking or fast home.

All of these things are easy enough to test- reduce the numbers, and
see what breaks :)

allan
-- 
"The truth is an offense, but not a sin"