[sane-devel] Reflecta DigiDia 5000 not found

[Reinhard Kotucha]

Hi,
first of all, thank you very much for sane.

I've bought a slide scanner (Reflecta DigiDia 5000) from a friend
which is classified as "Untested". However, its successor (DigiDia
6000) is supported and I doubt that there are significant differences.

It's detected by sane-find-scanner but not by scanimage -L .

I'm running Debian 9 on an AMD64 machine. It provides

sane 1.0.14-12
sane-utils 1.0.25-4.1
libusb-1.0-0

I created a file

/etc/udev/rules.d/40-libsane.rules

containing the line

ATTRS{idVendor}=="05e3", ATTRS{idProduct}=="0142", ENV{libsane_matched}="yes"

The file

/etc/sane.d/pieusb.conf

contains the line

usb 0x05e3 0x0142

and everything else is commented out.

# dmesg
usb 2-3: new high-speed USB device number 2 using ehci-pci
usb 2-3: New USB device found, idVendor=05e3, idProduct=0142
usb 2-3: New USB device strings: Mfr=0, Product=0, SerialNumber=0

The last line is quite strange, I've never seen something like that
before.

# sane-find-scanner
found USB scanner (vendor=0x05e3, product=0x0142) at libusb:002:002

# lsusb
Bus 002 Device 002: ID 05e3:0142 Genesys Logic, Inc. Multiple Slides Scanner-3600

# scanimage -L
No scanners were identified.

Whenever I've run scanimage -L , dmesg sais

usb 2-3: USB disconnect, device number 2

and, as expected, sane-find-scanner and lsusb dont't find the scanner
anymore. This _only_ happens after I've run scanimage -L .

Furthermore,

# strace -eopen scanimage -L 2>str
# grep pie str
open("/usr/lib/x86_64-linux-gnu/sane/libsane-pie.so.1", O_RDONLY) = 36
open("/usr/lib/x86_64-linux-gnu/sane/libsane-pie.so.1", O_RDONLY|O_CLOEXEC) = 36
open("./pie.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
open("/etc/sane.d/pie.conf", O_RDONLY) = 36

It seems wrong to me that libsane-pie.so.1 is found but not
libsane-pieusb.so.1, as I would expect.

When I add pieusb to /etc/sane.d/ddl.conf I get

# grep pie str
open("/usr/lib/x86_64-linux-gnu/sane/libsane-pieusb.so.1", O_RDONLY) = 36
open("/usr/lib/x86_64-linux-gnu/sane/libsane-pieusb.so.1", O_RDONLY|O_CLOEXEC) = 36
open("./pieusb.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
open("/etc/sane.d/pieusb.conf", O_RDONLY) = 39
open("/usr/lib/x86_64-linux-gnu/sane/libsane-pie.so.1", O_RDONLY) = 39
open("/usr/lib/x86_64-linux-gnu/sane/libsane-pie.so.1", O_RDONLY|O_CLOEXEC) = 39
open("./pie.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
open("/etc/sane.d/pie.conf", O_RDONLY) = 39

This looks better because libsane-pieusb.so.1 is read but scanimage -L
still fails. Is it on purpose that pieusb is not in dll.conf by default?

Though I don't believe that anything regarding Reflecta scanners was
changed recently, I temporarily installed the current git version of
sane-backends. It didn't solve my problem either but introduced a new
room of freedom. Entries in pieusb.conf now have the form

usb <vendor-id> <device-id> <model-nr> <has-slide-transport>

and I do not know how to determine the <model-nr> of my scanner.
Therefore I switched back to the version provided by Debian but can
activate the current git version at any time if desired.

I'm absolutely clueless now. What am I doing wrong?

Though my scanner is classified as "Untested" I'm quite convinced that
I can use the driver provided by Klaus Kämpf for the successor because
both devices return "idVendor=05e3, idProduct=0142" and thus seem to
be compatible.

Please let me know if you need more information.

Regards,
Reinhard
--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Klaus Kaempf]

Post by Reinhard Kotucha
Hi,
first of all, thank you very much for sane.
I've bought a slide scanner (Reflecta DigiDia 5000) from a friend
which is classified as "Untested". However, its successor (DigiDia
6000) is supported and I doubt that there are significant differences.

Moin Reinhard,

welcome to the wonderful world of Reflecta/PIE interfaces. ;-)

Please run
export SANE_DEBUG_PIEUSB=11
in a shell and then start xsane from there.

This should report the model number. Add it to pieusb.conf and you
should be all set.

Post by Reinhard Kotucha
I created a file
/etc/udev/rules.d/40-libsane.rules
containing the line
ATTRS{idVendor}=="05e3", ATTRS{idProduct}=="0142", ENV{libsane_matched}="yes"

Please add 'GROUP="lp",' to this list to give you full access to the
usb device.

Post by Reinhard Kotucha
The file
/etc/sane.d/pieusb.conf
contains the line
usb 0x05e3 0x0142
and everything else is commented out.

That's the old format and insufficient for your device.

Post by Reinhard Kotucha
This looks better because libsane-pieusb.so.1 is read but scanimage -L
still fails. Is it on purpose that pieusb is not in dll.conf by default?

Your version of 'sane' is too old.

Post by Reinhard Kotucha
Though I don't believe that anything regarding Reflecta scanners was
changed recently, I temporarily installed the current git version of
sane-backends.

Awesome, you're all set. The 'model number' is the only missing part.

Post by Reinhard Kotucha
It didn't solve my problem either but introduced a new room of freedom.
Entries in pieusb.conf now have the form
usb <vendor-id> <device-id> <model-nr> <has-slide-transport>
and I do not know how to determine the <model-nr> of my scanner.
Therefore I switched back to the version provided by Debian but can
activate the current git version at any time if desired.
I'm absolutely clueless now. What am I doing wrong?

Nothing. You just happen to be the first to use this scanner.


Klaus

[Reinhard Kotucha]

Post by Klaus Kaempf
Please run
export SANE_DEBUG_PIEUSB=11
in a shell and then start xsane from there.
This should report the model number. Add it to pieusb.conf and you
should be all set.

Hi Klaus,
thank you very much! Yes, it returns

[pieusb] model .......................: 0x002f

This is the same value as for the DigitDia 4000.

It works now but the scanner is not detected unless 'pieusb' is added
to dll.conf .

Post by Klaus Kaempf
Please add 'GROUP="lp",' to this list to give you full access to the
usb device.

Done. Can now run it as a mortal user.

The next problem to solve is to find a suitable ICM file for this
scanner. Colors are still too dark and incorrect and I doubt that I
can determine the appropriate values manually. I played with
Gamma/brightness/contrast but the colors are still wrong and I
therefore assume that these values have to be determined separately
for each color.

Iv'e got two DVDs, "CyberView X" and "Adobe Photoshop Elements 6". On
the "CyberView X" DVD I found a file "Generic Positive.ft" but I fear
that the file format isn't understood be sane. See an excerpt of this
file below.

How did you solve this problem? Do you have an appropriate ICM file
for your Reflecta 6000? I'm not sure whether what I bought is still
complete.

Another thing I have not found out yet is how to make use the slide
transport supported by the scanner. When I set SANE_DEBUG_PIEUSB=13
and activate "Advance slide", I see

[pieusb] set advcane [#10] to 1

(there is a typo BTW, s/advcane/advance/) but the same slide is
scanned again when I acquire a new preview. I have abt. 7200 slides
and I hope that I can digitize them without too much manual surgery.

The new format of pieusb.conf is

# usb <vendor-id> <device-id> <model-nr> <has-slide-transport>

When I set <has-slide-transport> to 0x01, xsane complains about
IO-errors[1]. On the other hand we have

# Autodetect (built-in)
# Reflecta DigitDia 6000 Multiple Slide Scanner
# usb 0x05e3 0x0142 0x3a 0x01

So I assume that 0x01 should be the correct value for my scanner too.

Thanks again and Merry Xmas,
Reinhard


[1] "Io, the Greek goddess of input and output"
Donald E. Knuth, The METAFONTbook


File "GAMMA/Generic Positive.ft"
------------------------------------------------------------------
# Version 1.4;
# Film term file for all reversal film;
# Created using Q60 unit;
# Film term created by B. Lin;

FilmClass REV;
FilmMfgr Generic;
FilmName Generic Positive;
FilmLevel 4 4 4;
CreationDate 07102003;
FilmBalance 140 140 140;
FilmLut
0.00 0.00 0.00
7.50 7.50 6.00
15.06 15.11 12.20
22.67 22.83 18.60
30.33 30.67 25.20
38.06 38.61 32.00
[...lines omitted...]
4093.22 4093.18 4093.85
4094.11 4094.09 4094.43
4095.00 4095.00 4095.00;
FilmMatR 1.4234 -0.3230 0.0061 -29.7088;
FilmMatG 0.0600 0.9677 0.0719 -31.3147;
FilmMatB 0.1562 -0.5104 1.4530 -28.9059;
------------------------------------------------------------------
--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Klaus Kaempf]

Hi Reinhard,

Post by Reinhard Kotucha
It works now but the scanner is not detected unless 'pieusb' is added
to dll.conf .

yes, that's to be expected.

Post by Reinhard Kotucha

Post by Klaus Kaempf
Please add 'GROUP="lp",' to this list to give you full access to the
usb device.

Done. Can now run it as a mortal user.

:-)

Post by Reinhard Kotucha
The next problem to solve is to find a suitable ICM file for this
scanner. Colors are still too dark and incorrect

Happened to all of my slides too. I used gimp to normalize colors and
had spectacular results.

I don't know if SANE has any color-correction build in.

Post by Reinhard Kotucha
Another thing I have not found out yet is how to make use the slide
transport supported by the scanner.
When I set <has-slide-transport> to 0x01, xsane complains about
IO-errors[1].

Hmm, 0x01 is the right value. Send me the full log via private mail.


Klaus

[Julian Hauser]

Hello,

I created medium profiles in xsane to get the colours right for various negative films (on my Reflecta 10T). I guess the same should work for slide film too.

The process is a bit fickle and takes a while, but once the profile is created it works without any problems: http://www.xsane.org/doc/sane-xsane-medium-definition-doc.html

I didn't bother creating a scanner profile because, at least for negatives, the corrections are going to be different for every brand/line of film anyway.

Best,
Julian

Post by Klaus Kaempf
Hi Reinhard,

Post by Reinhard Kotucha
It works now but the scanner is not detected unless 'pieusb' is added
to dll.conf .

yes, that's to be expected.

Post by Reinhard Kotucha

Post by Klaus Kaempf
Please add 'GROUP="lp",' to this list to give you full access to the
usb device.

Done. Can now run it as a mortal user.

:-)

Post by Reinhard Kotucha
The next problem to solve is to find a suitable ICM file for this
scanner. Colors are still too dark and incorrect

Happened to all of my slides too. I used gimp to normalize colors and
had spectacular results.
I don't know if SANE has any color-correction build in.

Post by Reinhard Kotucha
Another thing I have not found out yet is how to make use the slide
transport supported by the scanner.
When I set <has-slide-transport> to 0x01, xsane complains about
IO-errors[1].

Hmm, 0x01 is the right value. Send me the full log via private mail.
Klaus

[Klaus Kaempf]

Post by Klaus Kaempf
Hmm, 0x01 is the right value. Send me the full log via private mail.

Reinhard,

thanks for the logs.

Please check out https://github.com/kkaempf/sane-backends.git. The
'kkaempf' branch is one commit on top of current sane-backends#master.

This version gets rid of the 'flags' column in pieusb.conf and
evaluates the inquiry data instead.

Compile with 'make', install with 'sudo make install' and then run
xsane.

Good luck !

Klaus

[Reinhard Kotucha]

Post by Klaus Kaempf

Post by Klaus Kaempf
Hmm, 0x01 is the right value. Send me the full log via private mail.

Reinhard,
thanks for the logs.
Please check out https://github.com/kkaempf/sane-backends.git. The
'kkaempf' branch is one commit on top of current sane-backends#master.
This version gets rid of the 'flags' column in pieusb.conf and
evaluates the inquiry data instead.

Thank you Klaus. Works fine now after adding pieusb to dll.conf.

Regards,
Reinhard
--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Klaus Kaempf]

Post by Reinhard Kotucha
Thank you Klaus. Works fine now after adding pieusb to dll.conf.

Awesome, thanks for testing !

Does xsane offer the 'advance slide' checkbox ?
Does it advance the slide after the scan ?

Thanks,

Klaus

[Reinhard Kotucha]

Post by Klaus Kaempf

Post by Reinhard Kotucha
Thank you Klaus. Works fine now after adding pieusb to dll.conf.

Awesome, thanks for testing !
Does xsane offer the 'advance slide' checkbox ?

Hi Klaus,
yes, there is an 'advance slide' checkbox.

Post by Klaus Kaempf
Does it advance the slide after the scan ?

It advances the slide once. A preview of the first slide is displayed
and the second slide is pulled in and nothing else happens after that.
The last lines of the log are:

[pieusb] sane_control_option()
[pieusb] get br-y [#16]
[pieusb] get br-y [#16] val=2379677
[pieusb] sane_get_parameters
[pieusb] sane_get_parameters from option values
[pieusb] resolution 300.000000
[pieusb] width x height: 22.971008 x 34.706558
[pieusb] colors: 3
[pieusb] sane_get_parameters(): SANE parameters
[pieusb] format = 1
[pieusb] last_frame = 1
[pieusb] bytes_per_line = 813
[pieusb] pixels_per_line = 271
[pieusb] lines = 409
[pieusb] depth = 8

but the program is still running.

Regards,
Reinhard
--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Klaus Kaempf]

Post by Reinhard Kotucha
yes, there is an 'advance slide' checkbox.

Post by Klaus Kaempf
Does it advance the slide after the scan ?

It advances the slide once. A preview of the first slide is displayed
and the second slide is pulled in and nothing else happens after that.

If you choose "Acquire preview", it does not advance to the next slide.

A normal scan will advance to the next slide.

You can increase the number of slide in xsane's main window. If the
'advance slide' checkbox is marked, it will scan multiple slides.

Hth,

Klaus

[Reinhard Kotucha]

Post by Klaus Kaempf

Post by Reinhard Kotucha

Post by Klaus Kaempf
Does it advance the slide after the scan ?

It advances the slide once. A preview of the first slide is
displayed and the second slide is pulled in and nothing else
happens after that.

If you choose "Acquire preview", it does not advance to the next slide.
A normal scan will advance to the next slide.
You can increase the number of slide in xsane's main window. If the
'advance slide' checkbox is marked, it will scan multiple slides.

Sorry for the noise. Yes it works as expected now. I'm just quite
unfamiliar with xsane.

Thank you very much again,
Reinhard
--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Reinhard Kotucha]

Post by Klaus Kaempf

Post by Reinhard Kotucha
The next problem to solve is to find a suitable ICM file for this
scanner. Colors are still too dark and incorrect

Happened to all of my slides too. I used gimp to normalize colors and
had spectacular results.

Hi Klaus, Julian, and Dave and thank you very much for your
suggestions.

As I mentioned in a previous mail, there was a file on the CD called
"Generic Positive.ft". The file format is neither understood by sane
nor by me. The good thing is that it's pure ASCII.

I first extracted the lookup table

FilmLut
0.00 0.00 0.00
7.50 7.50 6.00
[...lines omitted...]
4094.11 4094.09 4094.43
4095.00 4095.00 4095.00;

and plotted the curves. See

http://ms25.ddns.net/plot-lut.pdf

The first page contains a linear plot, the second page shows the same
plot scaled logarithmically. Please note that when scaled
logarithmically, the Gamma value is just the slope (dy/dx) of the
curve. As you can see, two different Gamma values are used.

I then loaded a scanned image into Gimp and adapted the curves
according to the linear plot. The results were amazing. I did not
expect to get such good results almost immediately. But the result
isn't perfect. It turned out that the lower left part of the curve is
extremely critical. Small changes make huge differences and there is
no way to get things correct if the only tool you can use is a mouse.

The next step is to write a script which applies the lookup table to
pnm files. The pnm file format is quite simple and I think that I can
write such a script within a reasonable amount of time.

I could probably improve the results if anybody tells me what these
values mean and how to apply them:

FilmMatR 1.4234 -0.3230 0.0061 -29.7088;
FilmMatG 0.0600 0.9677 0.0719 -31.3147;
FilmMatB 0.1562 -0.5104 1.4530 -28.9059;

I suppose that I can scan all my slides automatically now and can
apply the color profile of my scanner afterwards. Everything I can do
without manual interaction is fine for me (I've abt. 7200 slides) and
GUI tools are not very helpful at this stage. I'm sure that GUI tools
like GIMP are helpful in order to enhance individual images.

Regards,
Reinhard
--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Olaf Meeuwissen]

Hi,

Post by Reinhard Kotucha
[...]
I could probably improve the results if anybody tells me what these
FilmMatR 1.4234 -0.3230 0.0061 -29.7088;
FilmMatG 0.0600 0.9677 0.0719 -31.3147;
FilmMatB 0.1562 -0.5104 1.4530 -28.9059;

I'm not sure about the rightmost column but this is a color correction
matrix. You apply it on the incoming (or gamma corrected?) RGB values
via matrix multiplication like so

R' = 1.4234 * R + -0.3230 * G + 0.0061 * B + -29.7088
G' = 0.0600 * R + 0.9677 * G + 0.0719 * B + -31.3147
B' = 0.1562 * R + -0.5104 * G + 1.4530 * B + -28.9059

R, G, B are the values before correction, R', G' and B' the ones after.
You may have to clamp the results into the [0,255] range (assuming 8bit
samples).

Hope this helps,
--
Olaf Meeuwissen, LPIC-2 FSF Associate Member since 2004-01-27
GnuPG key: F84A2DD9/B3C0 2F47 EA19 64F4 9F13 F43E B8A4 A88A F84A 2DD9
Support Free Software https://my.fsf.org/donate
Join the Free Software Foundation https://my.fsf.org/join

[Dave Platt]

Post by Klaus Kaempf

Post by Reinhard Kotucha
The next problem to solve is to find a suitable ICM file for this
scanner. Colors are still too dark and incorrect

Happened to all of my slides too. I used gimp to normalize colors and
had spectacular results.
I don't know if SANE has any color-correction build in.

Some of the SANE frontends can attach an ICM profile to an image
during the scanning process. This will allow for a fairly automatic
correction during display, and GIMP can automatically transform the
colors to a standard color-space when it loads the image if you wish.

The DigitDia 5000 may have originally been sold with an individual
profile based on an IT8 target image on a Kodak slide. If this file isn't
available, or of you want to scan slide made on a different brand or
type of transparency, you will need to make a custom profile by
scanning a reference target. Argyll CMS can process the resulting scan,
using the target reference data provided by the maker of the target, and
create a good profile for that individual scanner when used with that
particular type of transparency film.

A similar process can be done for reflective (e.g. flatbed) scanners.

http://www.targets.coloraid.de is one possible source for targets.

This process takes a bit of time and money but it's well worth doing.