[sane-devel] Canon IX-4015 not working
Greetings, I cannot see from here (I am on vacation) which scanner model this is, but it may be a case for the canon-scsi backend. I can have a look at the code again, but you will have to wait for 2 weeks. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\...\ Tel. +49 (0)221 470-4543, Fax -4900 ___L|L|__|_\ http://www.uni-koeln.de/deiters/
[sane-devel] Possible bug in canon backend or scanimage
If xsane works correctly with the scanner, the problem should not be in the Canon backend. Kind regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\...\ Tel. +49 (0)221 470-4543, Fax -4900 ___L|L|__|_\ http://www.uni-koeln.de/deiters/
[sane-devel] Art project
Is it really necessary to remodel the scanner? Alternatively, you might use a lens/mirror system to project an image of your large object onto a standard A4 flatbed scanner. You would have to provide additional lighting for the object, and you would have to to something to protect your scanner from stray light, but it should be possible. You will, of course, not achieve a greater image resolution. Best regards, Ulrich Deiters
[sane-devel] What exactly does scanning software do?
Hi, what the scanner adjustments do depends very much on the scanner. The selection of colour/grayscale/lineart affects the sampling and scanning speed on practically all scanners. Adjustments of brightness, contrast, colour hues are used to fine-tune the signal amplifiers on many scanners, but there are some where the adjustments are done by software after the scan. If your scanner produces 3x8 bit RGB data only, you can expect that it does its adjustments by hardware, for otherwise you would lose too much quality (e.g., a too dim picture using pixel values between 0 and 63 can be "lit up" by software to use the full range, 0 to 255, but then you would still have merely 64 shades of colour instead of 255). If your scanner produces 3x12 or 3x16 bit RGB data, it is possible that the signal amplifiers work with fixed settings, and corrections are done by software. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\...\ Tel. +49 (0)221 470-4543, Fax -4900 ___L|L|__|_\ http://www.uni-koeln.de/deiters/
[sane-devel] Possible help with Canon FS2720U
Hi, the FS2710S (S standing for SCSI) is supported, but I do not know the state of the USB backend for the FS2720U. But unless you have got a slide scanner with a magazine feeder, you will need *months* two digitize your collection. Even if you do not have to do complicated colour adjustments, I should think that 2 hours per 50 slides is already quite a good speed. You might want to consider taking your collection to a professional scanning service. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] CVS Problem
Hi, I think the "pserver" access has been turned off for quite a while. Set your CVS shell to "ssh" and try "ext": export CVS_RSH=ssh cvs -z3 -d "ext:alioth..." This is explained in more detail on the SANE homepage. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] ambiguous post-increment in canon-sane.c
Interesting! The incriminated C code looked perfectly innocent to me (if the incrementing operation is carried out after all assignments, as it should, there is no ambiguity). But if it makes you compiler happy, I will use your substitution code. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] Help with Canon 2700F
I did test runs with the new canon.conf with my 2710F and everything worked OK. I suspect there is a hardware error that causes the sense_handler to take action. Perhaps it is a minor problem that can be ignored? Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] Help with Canon2700F
I played a little bit with my configuration files. In my /usr/etc/sane.d/dll.conf there is a line "#canon" which I left as it is. In /usr/etc/sane.d/canon.conf I tried several combinations. The scanner is recognized as canon:/dev/sg0 if canon.conf contains the line "/dev/sg0" or "scsi". "dev/scanner" works only if I create a symlink sg0 --> scanner (understandably). The canon.conf of the contribution contains: #canon.conf /dev/scanner #/dev/sg0 I suggest changing that to #canon.conf scsi Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] Help with Canon 2700F
I have the same problem since I switched to Linux. Now I made sure that canon.conf contains /dev/sg0 and /dev/scanner AND I created a symbolic link: ln -s /dev/sg0 /dev/scanner. Now "scanimage -L" sees my Canon scanner. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] sane-find-scanner ...
The undetected scanner is a Canoscan 2710S, an SCSI slide scanner. It used to function properly with my previous computer (HP-UX). It does not matter whether I run SANE as user or superuser. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] sane-find-scanner ...
Much more interesting than this discussion why sane-find-scanner works would be a hint what can be done if scanimage -L does not detect the scanner. I have this problem under Linux with an SCSI scanner (i.e., sane-find-scanner sees it, but scanimage not). Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] The xsane problem
I suspect a problem with your file permissions. Do you have writing permissions for /tmp, or whereever your computer puts temporary files? Try to do a scan as superuser: If this works, it is certainly a permissions problem. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] Epson 3590 Scanning issues
I have had a similar problem once. In my case it had been caused my an improper mixing of library versions. If you install SANE from a Linux CD and then update manually to the latest version, it is likely that some links to libraries point the wrong way. It is best to remove the old version completely before installing a new one. regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] Backend driver for Canon CanoScan 4200f
The problem is (probably) not so much the programming, but knowing the machine codes. Unless someone does a good job of bus traffic sniffing, it is not possible to do much. Regards, Ulrich Deiters -- Prof. Dr. Ulrich K. Deiters __ Institut f. Physikalische Chemie \ Luxemburger Str. 116, D-50939 Koeln Universitaet zu Koeln /\/\__. \ Tel. +49 (0)221 470-4543, Fax -4900 L|L|__|\ http://www.uni-koeln.de/deiters/
[sane-devel] sane-canon: Endian inverted data when scanning 48bbp raw mode pnm
Hi, I think I am the one who is responsible for the byte switch problem. Indeed, I am using a big-endian machine. However, I thought that the PNM standard specifies the byte order. Please advise me if this is wrong. Best regards, Ulrich Deiters
[sane-devel] Epson 3490
Using a SANE version compiled from CVS in parallel with an older version installed with the operating system can be tricky, because some programs will load libsane-XXX.so libraries, whereas others seem to look for libsane-XXX.so.1 first. Both of them are links to the "real things", libsane-XXX.so.1.0.16 . It may be a good idea to check the /usr/lib and /usr/lib/sane directories for SANE-related entries and to make sure that all links point to files of the same version. Regards, Ulrich Deiters
[sane-devel] off-topic 25 mm scanner
Genuine slide scanners usually have - or should have - additional features like focus control (you need that if you have framed slides). I am not sure that such a feature is built into an add-on to a flatbed scanner, unless it is a really expensive model. I think the resolution of commercial fine-grain slide film is about 5000 dpi. There may be higher resolutions possible, but you have to ask the CIA or the KGB about them. :-) But these 5000 dpi can only be realized if you write on the film with a specialized optical system. If you use a lens that is not made of calcium fluoride, if you are not restricting yourself to taking pictures at infinite distance, and if you are not using monochromatic light, you are most likely down to 3000 dpi. If you use a zoom lens, 2500 dpi is more likely. My advice would be to look for a used slide scanner (Canon 2710F) is OK, for instance). Or photograph your slides with a digital camera! Regards, Ulrich Deiters
[sane-devel] Epson Perfection 3590
I just downloaded the CVS-version of the snapscan backend and tested it.
It works (for the flatbed part of the scanner)!
The only thing that needs correction is the list of resolutions, which
at present is
Options specific to device `snapscan:libusb:004:002':
Scan Mode:
--resolution
auto||150|200|240|266|300|350|360|400|600|720|800|1200|2400|3200|0|10|50|
75|100|150|200|300|400|600|1200|2400|0|0|0|0|0|10|50|75|100|150|200|300|
400|600|800|1600|0|0|0|0|0|10|50|75dpi [300]
Sets the resolution of the scanned image.
This is probably not what was intended. :-)
I did this test after an upgrade to SuSE Linux 10.0. Perhaps somebody
would like to know what was necessary to get the scanner going (again):
1. Updated SuSE 9.3 --> 10.0.
2. Updated my SANE backend files from CVS.
3. Ran ./configure --prefix=/usr again (because of some change in the
gcc libraries.
4. make
5. make install (as root)
6. chmod 666 /proc/bus/usb/004/* (so that ordinary users can access the
scanner)
7. cd /usr/lib/sane
rm *.1 (because these links all point to the old version)
ld *.so | xargs -i ln -s {}.1.0.16 {}.1
Step 7 should be superfluous, but evidently something makes my linker
prefer the *.so.1 libraries over the *.so ones. Perhaps this is a problem
of the configure script?
Regards,
Ulrich Deiters
[sane-devel] Epson 3590
Here are the latest results for my 3590:
As Bouke Jan wrote, the firmware files are the same; I did
a diff on a hexdump and could not find any difference.
I tried more resolutions. The results are funny:
- 200 dpi is OK
- 150 dpi works, too
- 100 dpi works, but the scanned area was displaced in y direction
- 75 dpi works, but scans a larger y distance than expected
- 50 dpi is OK,
- values under 50 are rounded to 50
It seems that the y controls enjoy some intellectual freedom
with this scanner.
About the preview bit: the old code was
source = 0x00;
if (pss->preview) {
source |= 0x80; /* no high quality in preview */
}
else {
source |= 0x40; /* no preview */
}
The new version is
source = 0x40; /* always set no-preview bit */
if (pss->preview)
source |= 0x80; /* no high quality in preview */
The film scanning unit needs more work. If I set the "transparency
mode" in xsane, the film scanner lamp turn on, but subsequent
scans move the flatbed unit, not the film unit.
Best regards,
Ulrich
[sane-devel] Epson 3590
I set the no-preview bit (source |= 0x40;) in snapscan-scsi.c after line 922, and the problem vanished, i.e., the scanner does a quick, undistorted scan of the whole scan area without hanging itself. Regards, Ulrich
[sane-devel] Epson 3590
> I don't quite understand that. Are you saying that preview is slower
> than a normal scan at 200 DPI? (200 DPI should be the lowest
> resolution for the 3590.) I was under the impression that any
> frontend would use the lowest available resolution as the default
> resolution for previewing.
The scanner is slow in preview mode and makes a different noise
(probably because it is doing very small y steps). Right now
the preview mode is not the same as the normal 200 dpi mode.
But this is not surprising: Somewhere around line 870 in snapscan-scsi.c
you inform the scanner that it is in preview mode:
source = 0x0;
if (pss->preview) {
source |= 0x80; /* no high quality in preview */
}
else {
source |= 0x40; /* no preview */
}
I suspect that the scanner takes that as an excuse to get lazy
(with respect to the y motion). :-)
I will download the latest CVS version during the weekend and
see if the problem still persists.
Regards,
Ulrich
[sane-devel] Epson 3590
I manually clobbered all files in my installation that contained the word "sane" or were in "sane" directories, and re-installed the back- and frontends, but the problem with the preview persists. But as the scanner is slower in preview mode than in normal scan mode at low resolution, it might be a good idea to map the preview option to some low-res option in the backend. Another (although theoretical) problem with the Epson/Snapscan backend: The backend reads the path of the firmware file from the snapscan.conf file. What happens if there is more than one scanner? Regards, Ulrich Deiters
[sane-devel] canon-sane.c:2113 SSIZE_MAX...
Thank you for checking the files! I put an "#indef SSIZE_MAX..." definition into canon.h . This should take care of the problem, but won't hurt if one of the sanei header files is modified. Regards, Ulrich Deiters
[sane-devel] canon-sane.c:2113: error: `SSIZE_MAX' undeclared
I think there may be some trouble: Someone used an old backend version, made a change in the section for the FB1200, and overwrote the one in CVS. This means that all developments from the previous year (improved sense handler, focus control for the slide scanners, autoexposure, ...) have been overwritten. I have to consult the Canon-SCSI team members about the situation. Please do not make any changes to the CVS version; I will probably reset the Canon backend this evening (German time) when I am home. Regards, Ulrich
[sane-devel] canon-sane.c:2113: error: `SSIZE_MAX' undeclared
Hi, are you talking about the Canon SCSI backend? I do not have any references to a variable SSIZE_MAX in my program files. Regards, Ulrich Deiters
[sane-devel] Problem reading HP ScanJet 6200C/6250C
My HP 6200C is only recognized properly on the SCSI bus if its SCSI address is 0. Regards, Ulrich Deiters
[sane-devel] film scanner in linux?
> Hello, and sorry if this is slightly off-topic. I've been thinking > about buying a dedicated slide/negative scanner, but none of the models > I've seen in the market seem to be supported by SANE. Could anyone give > me recommendations about which film scanners work well under Linux? If "in the market" includes also older or second-hand models, thre are some SANE-supported film scanners. The Canon FS2710S does work (2700 dpi, SCSI interface); if you need higher resolutions or IR channel, you should look for Minolta or Nikon. Regards, Ulrich Deiters
[sane-devel] Problems with Canonscan 2700f
dear Tobias, I have just sent an updated (still experimental) Canon backend to your home address. I am not sure, however, that your case is a backend problem. Regards, Ulrich Deiters
[sane-devel] Canon 2700 slide scanner
The trouble spot in the Canon SCSI backend has been located, and we have an experimental backend that seems to support both the 2700 and the 2710 now. But there is still some testing and cleaning-up to be done, although that should not take too long. Regards, Ulrich Deiters
[sane-devel] No device available
Are you sure that the SCSI ID of the scanner is different from any other ID on the bus? Is the bus properly terminated? Is the SCSI cable of permissible length? If the SCSI controller cannot initialize its bus, it might declare itself as offline at boot time. Regards, Ulrich Deiters
[sane-devel] Problems with Canon 2700F
Hi, I just want to let you know that I am listening. What are the Canon backend versions we are talking about (the version number at the top of canon.c)? Regards, Ulrich Deiters
[sane-devel] Problems with scsi filmscanner canon2700
I am sorry for the problem, but it appears that the introduction of focus controls for the Canon FS2710 somehow disrupted the focus controls for the FS2700. Unfortunately, I do not have an FS2700, and so I cannot test the backend for this scanner. If you say that sane version 1.0.13 worked for you, this gives me a clue where to start hunting for the bug. But please don't expect a quick answer. As a workaround you can copy the relevant files (canon.c, canon-sane.c, canon-scsi.c, canon.h) from the 1.0.13 version to your actual backend directory and recompile. That should help, provided that the bug is where we think it is. Best regards, Ulrich Deiters
[sane-devel] Re: Fwd: Canon 2700f problem
> My question is, why and since when > did the focus stop working? I've tried vuescan, which focusses quite > alright, so the focus mechanism must work. I've tried the latest > backend (0.14), and it has the same issues. I wish I knew! I only have a 2710S and could not run any tests on the older 2700. I think the first error reports came in after I had put in focus support for the 2710S. My suggestion is that you retrieve older Canon backends from CVS (or I can mail them to you); you compile them and try them out on your scanner until we know which version worked. I know that this is tedious, but it seems to be the only way to find out what exactly went wrong. Regards, Ulrich Deiters
[sane-devel] Canon IS-12 Scanner head (for canon BJC 50,55 & 80 printers)
Hi Tim, what exactly is this IS-12 - an add-on device for a printer? How is it controlled by the computer - SCSI, USB or parallel port? Perhaps one of the existing backends can be used. Regards, Ulrich Deiters
[sane-devel] Canon FS4000: image data is sensor data, not RGB?
Do you receive 3 * 2 bytes per pixel? The FS2710S also produces 2 bytes, of which 12 bit are significant, in the order RRGGBB. The Canon-SCSI backend contains some portions that convert these raw data to the usual 8 bit representation. Perhaps this code can be modified to fit your case. Regards, Ulrich Deiters
[sane-devel] Canon FS4000 support
I once had an opportunity to look at an SCSI protocol of
a FS4000. The protocol differs significantly from those
of the SCSI flatbed scanners and the older film scanners.
Moreover, there appear several SCSI commands which my
handbook lists as "proprietary/undocumented". Reverse
engineering will therefore require much experimental
work.
And who knows - perhaps Canon built some unconventional
commands into the scanner ("eject light bulb", "burn
the slide") to discourage people like us? :-)
Regards,
Ulrich Deiters
[sane-devel] scanning negatives
Right now the only viable solution for negatives with yellow filter layers is to scan a portion of the film strip *between* two photos, play with the colour filters of xsane or whatever scanner software you are using until you have obtained pure white, and use this filter setting for your photos. Regards, Ulrich Deiters
[sane-devel] compilation problem under HP-UX
> Our configure script/Makefiles don't set any optimisation at all as > far as I can see. Is that done by autoconf or libtool? The file README.hp-ux tells how to compile under HP-UX using the Hewlett-Packard compiler, and here an optimization option is specified. I think it is good technique to turn optimization on for RISC machines, hence the example line should be kept; it would be good, however, to warn against higher level optimizations: ... For the HP compiler, invoke configure like this: CC=cc CFLAGS="-Ae -O" ./configure Higher optimization levels than -O (equivalent to +O2) may fail to compile correctly. ... Best regards, Ulrich Deiters
[sane-devel] compilation problems under HP-UX
The main problem is solved: The loader error occurs at optimization level +O3 only (which includes function inlining). Level +O2 (equivalent to -O) is safe. However, the HP compiler reports a few errors: - sanei/sanei_pp.c, line 125: "#warning" is not a permissible preprocessor statement. As mentioned before, this line can (and should be) removed without creating troubles. - backend/snapscan.c, line 86: The HP compiler does not know what an __attribute__ might be. - backend/hp5400_internal.c, line 469ff: Some complaint about incompatible pointer types. Perhaps the backend maintainers could look into this. Best regards, Ulrich Deiters
[sane-devel] compilation under HP-UX
I just noted that the SANE backends do no longer compile under HP-UX (using the HP C compiler "cc -Ae"). One problem is a line #warning ... in sanei_pp.c, which is not understood by the cc preprocessor. If it is removed, the compilation proceeds, but than the loader fails with a rather cryptic error message. Does someone know what has been changed since version 1.0.8 that might have affected the building of the shared libraries? Regards, Ulrich Deiters
[sane-devel] What's a good slide scanner?
"Solarized scans"? Be happy, some people work hard to achieve this effect. :-) Please try to run the scanner under XSane and use the colour filters of that program. If you scan in raw colour mode, you will have to set gamma to 2-2.5 either in XSane or in whatever program you use to process your image data. Regards, Ulrich Deiters
[sane-devel] CanoScan 2700F
If the latest Sane version crashes *after* doing the prescan and *before* doing the main scan, turn off the autofocus capability for the FS2700 in canon.c. This is a known problem, but I could not take care of it because I do not have a FS2700. But if you see this bug, you have already made a lot of progress. :-) Regards, Ulrich Deiters
[sane-devel] OT: fast usb document scanner
Why don't you use a fax device? Ulrich Deiters
[sane-devel] What's a good slide scanner
Hi Sigurd, as the focus control of the FS2700S had been working in an earlier Canon-SCSI backend version, I am confident that we can make it work again. However, I just came back from a) vacation and b) a scientific conference, and c) found my mailbox stuffed with 500 e-mails. I will come back to the problem as soon as possible. Regards, Ulrich Deiters
[sane-devel] What's a good slide scanner?
The Canon FS2710S is supported (SCSI scanner with up to 2700 dpi resolution, 3 x 12 bit per pixel). It takes film strips and framed slides up to 3 mm thickness. Regards, Ulrich Deiters
[sane-devel] program to convert SANE format to JPEG
xsane can produce JPEG directly. scanimage produces PNM formats (PPM for colour pictures). You can convert PPM to JPEG with - cjpeg (part of the PBMPLUS package) - xv - ImageMagick - ... and of course some picture manipulation programs like the GIMP or Pixel!FX Regards, Ulrich Deiters
[sane-devel] Problems with the Canoscan 2700F
Could you let me have a debug printout (after setting SANE_DEBUG_CANON=255), so that I can locate the problem? Regards, Ulrich Deiters
[sane-devel] Affordable film scanner
> I have the same scanner. But I think I'm just using 3x8 bit when > scanning - every time I tried 3x12bit all I got was a "solarized" image. > So how should I proceed to 3x12bit scanning? Get the latest Canon backend. The "solarization effect" occurs when the FS2710 is not recognized and has a 8-bit gamma table sent to it by the backend. The scanner seems to consider that as an insult. :-) Regards, Ulrich Deiters
[sane-devel] Affordable film scanner
The Canon FS27XX scanners are cases where you need a large gamma to obtain good results (2.0-2.5). Right now the backend collects 12-bit data and converts them to 8 bit with a 4096 byte gamma table (or exports them as 16-bit data). This usually works well; but I admit that there are situations where a few more bits of depth would be welcome. Even if your scanner has got exposure controls, the question remains whether it scans with fixed settings and manipulates the raw digital data, or whether it really adjusts the gain and zero of the amplifiers. This is probably the point where some companies try to save money; good analog circuitry is expensive. Regards, Ulrich Deiters
[sane-devel] Affordable film scanner
> Also backends truncating to 8bit for output are *very* nasty when it > comes to film scanners Why? First of all, truncation is an option, not a must. Second, you usually want to end up with JPEGs, and that implies 3x8 bit encoding. Regards, Ulrich Deiters
[sane-devel] Affordable film scanner
Some recent film scanners appear not to have exposure controls. As far as I know, the Canon FS2710S is such a case. But it scans 3x12 bits, and so it is OK to do the gamma correction in the backend and then truncate to 3x8 bit. I think this is just the modern trend in computer pheripherals: Build them as simple as dumb as possible, and provide a fancy driver to compensate the faults. Regards, Ulrich Deiters
[sane-devel] Re: Affordable film scanner supported by SANE
If the backend provides focus control, the frontends make it available to the user. For some film scanners focus control is already supported by the SANE programs. Regards, Ulrich Deiters
[sane-devel] Blue cast ...
There are standard "colour targets" available from photo suppliers. The one I am using is called a Q70 card (I think) and is from Kodak. We have been using the Pixel!FX scanner software (by Mentalix), and it provides a calibration routine against that colour target. It might be a good idea to build a similar calibration routine into the SANE frontends. Scanner lamps do age, and the standard calibration against a black+white portion of the scanner casing does not get rid of discolorations. Regards, Ulrich Deiters
[sane-devel] how to scan a multipage document with no adf?
The TIFF standard supports multiple images in one file, but a) not every TIFF reader can handle this. First make sure that your image viewer (or that of the mail recipient) can handle multi-image TIFF. b) the typical TIFF compression schemes (Packbits, LZW) are good for lineart only. For colour photos you need TIFF JPEG, but again, this cannot be handled by all viewers. PDF can handle multi-page documents with images, but the compression of colour graphics is not always good. See one of the previous discussion threads of this forum on how to produce PDFs. In short: Much depends on whether you want to scan colour photos or monochrome blueprints. Regards, Ulrich Deiters
[sane-devel] How to combine scanned images together and print them as a file
For a large number of pictures one might consider writing a script using tools of the PBMPLUS package: - use djpeg to convert JPEG to PNM format, - use pnmcat to combine two or more pictures you want to show on the same page (if this is required), - use pnmtops to make PostScript files - and finally ps2pdf to make PDF. Alternatively, you can use xv to view JPEGs and to print them to PostScript files; xv as a graphical interface that gives you some control over the placement of the picture on the page. Regards, Ulrich Deiters
[sane-devel] Canon-SCSI
The Canon SCSI backend now supports the FB1200S flatbed scanner. The crashes with Mandrake Linux and the problems with the preview of xscanimage mentioned in previous postings have been resolved. The revised backend has been committed to CVS. Ulrich Deiters
Fwd: Re: [sane-devel] Film scanning
May I see the debug file? Please mail it to my address directly, not to the SANE list. Ulrich Deiters
[sane-devel] Canon IX4025
The Canon-SCSI backend has not changed much from sane-1.0.9 to 1.0.10. Chances are quite good that your scanner will work with either version. Ulrich Deiters
[sane-devel] Canon IX4025
The Canon-SCSI backend lists the IX-4025 among the recognized scanners. Please give it a try. Ulrich Deiters
[sane-devel] Canon-SCSI backend
I just made a few minor changes (fixed an uninitialized variable, inserted checks for return status of some subroutines). Sincerely yours, Ulrich Deiters
[sane-devel] [patch] Fix up error checking in canon backend
The patch looks reasonable. I will test it this evening. Ulrich Deiters
[sane-devel] libsane-dll.a - where is it?
I had the broken link in my installation directory too after compiling sane-backends-1.0.8 without any special patches or flags. I simply deleted it. Another thing: The script sane-config, which is copied to the $prefix/bin directory, should have execution permissions. Kind regards, Ulrich Deiters
[sane-devel] more than 8 bits?
A 12 bit scanner needs indeed a gamma table of 4096 entries. But it is not necessary that the frontend maintains such a table. It is sufficient to exchange 256 byte gamma tables and to expand them by interpolation in the backend. Regards, Ulrich Deiters
[sane-devel] Compilation error on Compaq Tru64
... perhaps it does not apply to HP-UX 11.X, but on 10.X it is generally a good idea to include the compiler option "-Ae". Ulrich Deiters
[sane-devel] Canon-SCSI backend
> Then, if gamma tables aren't supported (or at least aren't > supported well), they shouldn't be offered to the frontend, right ? Not necessarily: The backend emulates the contrast, brightness, ... gamma options. This is even necessary for the FS2710, because it requires a huge gamma correction of about 2.5 to begin with, and the automatic optimization of xsane does not work well if it has to go to such extremes. Best regards, Ulrich Deiters
[sane-devel] Canon-SCSI backend
I remember very well the troubles I had with my FS2710S when I tried to upload gamma tables! The effects on the scans were very interesting from the artist's point of view - a *very* modern artist, in fact. :-) Now that I know the gamma table format the scanner recognizes I will try to play again with this feature, but I have some doubts about the success: Canon apparently does not use the gamma tables; furthermore, the standard controls of SCSI scanners like contrast or brightness are not responding at all. I agree that it would be better to handle such things within the scanner. But, according to Canon, we are not supposed to run the scanner without their software, and for them it may make sense to keep the scanner as simple as possible, while handling all the finer details of scanning by the computer. Best regards, Ulrich Deiters
[sane-devel] SANE 1.0.8, Canoscan 300 on Adaptec AVA-1502p ISA SCSI -> scani
Debug code removed??? I certainly did not do this, and the recent 1.0.8 backend still contains the DBG() statements. Are you sure you have the correct version? Regards, Ulrich Deiters
[sane-devel] HP, ADF and PDF
Scan, save as PostScript (or convert to PostScript by means of GhostScript), concatenate the files, convert to PDF my means of ps2pdf (which is part of the GhostScript distribution). I hope this helps. Ulrich Deiters
[sane-devel] film scanner recommendation
I think the HP photo smart scanner is within our price range, but I cannot recommend it: It grips negative strips or slides with rubber transport rollers, so you can be sure that you will have transport problems once the rubber surface show some wear. Furthermore, you cannot handle slides in glass frames with this scanner. Some Canon film scanners are also supported by SANE, but you probably will not find a new one for <500 US$; the FS2710 is probably at 650-700 US$ at the moment. But perhaps you can find a good used one? Sincerely yours, Ulrich Deiters
[sane-devel] Canon backends
...try stthd0.pc.uni-koeln.de instead :-) ^ |
[sane-devel] Canon backends
> > Hi, > > On Thu, Apr 04, 2002 at 09:09:12AM +0200, Ulrich Deiters wrote: > > The Canon-SCSI backend 1.0.9 (available via anonymous ftp from > > stthd0.pc.uni-koeln.de) supports the FB1200S. > > Is this code intended to be included into SANE CVS? Does it supersede > the current canon backend? > > Bye, > Henning Yes, by all means! I thought that it had already been accepted. The new backend supports the FB620S flatbed scanner and the FS2710 film scanner. I have not followed up the development for parallel port and USB scanners. These are not supported by this backend. Best regards, Ulrich Deiters
[sane-devel] Canon backends
The Canon-SCSI backend 1.0.9 (available via anonymous ftp from stthd0.pc.uni-koeln.de) supports the FB1200S. If scanimage does not find the scanner, it is often a problem with the file permissions of the scanner device file. Can you run scanimage as root? Kind regards, Ulrich Deiters
[sane-devel] Binary (worm/virus?) mails claiming to come from
... off-topic, but important, so I don't mind discussing that on this forum. We (U. of Cologne) have been attacked by various means over the last months: - Someone abused an anonymous FTP server for swapping MP3 files. That was not harmful in itself, but used up bandwidth and caused "partition full" errors. It stopped when I replaced some MP3 files by symbolic links to /dev/zero and removed the writing permissions to the ftp directory. It seems that some people, or perhaps even robots, are scanning the Internet for free storage capacity. - We have *LOTS* of attempts to abuse our machines as mail relays. - Recently there are lots of e-mails with a cleverly concealed executable which overwrites the dialer information in modem-connected PCs with expensive telephone numbers. The executable is not really a worm or virus, but even so fairly widespread because the originators use automated scripts to generate thousands of e-mail accounts with wanadoo.fr, yahoo.com, gmx.de, etc., mail as much of their junk as possible, and close the account. Best regards, Ulrich Deiters
[sane-devel] Canon CLC10/CJ10 backend
The SCSI-Canon backends will not recognize your scanner model, but they should at least give it a minimal support. Are you sure that the permissions of your scanner device file are OK? Best regards, Ulrich Deiters
[sane-devel] picking a robust sheetfed scanner
I think the old (and still good) releases of pbmplus and xv still contain the GIF tools. If you do not mind wirting a UNIX script, you can make SANE store data in PNM format and then convert to GIF. There are scanner tools like Pixel!FX by Mentalix (www.mentalix.com) which handle GIF, but they are expensive. Best regards, Ulrich Deiters
[sane-devel] Sane is definately not Scanner Access Now Easy
> No, there is no /.bash_profile file. ... not /.bash_profile, but ~/.bash_profile or, more explicitly, $HOME/.bash_profile! If you don't have this file, create it in your home directory and enter your PATH definition there: PATH=$PATH:whereever_the_SANE_binaries_hide export PATH Be sure to log out and in to make the definition active. Sincerely yours, Ulrich Deiters
[sane-devel] Sane is definately not Scanner Access Now Easy
My man pages say that bash looks for a file ~/.bash_profile to obtain user-specific variables and settings. Do you have this file? My own workstation (HP-UX 10.20, with CDE desktop) offers a special trap for its users: There is a special startup file for the desktop, .dtprofile, and there is by default an instruction not to read the other profile files. I do not know if this applies to Linux/KDE, however. Sincerely yours, Ulrich Deiters
[sane-devel] GTK
It may be necessary to re-run configure with some flags; typically something like --with-gtk-dir=/wherever/I/dumped/it. Of course, this only works if you have gtk installed. :-) Regards, Ulrich Deiters
[sane-devel] Re: Problem compiling 1.0.7 on HPUX 10.20
I a on HP-UX, too. Unfortunately I have not found the time yet to upgrade to the latest sane and xsane versions. Did you (or the configure script) set the compiler flags -Aa -DHPUX_SOURCE (or better -Ae)? Regards, Ulrich Deiters
[sane-devel] Film holder control (and adf)
"Film scanner" would be more accurate in my opinion, because these devices usually can handle slides, negative strips, and spooled film. Is there a film scanner with automated frame selection? On the ones I know you have to manually position a tab on the film strip holder. Regards, Ulrich Deiters
[sane-devel] Which new backends ...
The latest version of the Canon-SCSI backend is 1.09. It supports the Canon FB620S and the FS2710 (the latter has still some features missing) and has been confirmed to run under Linux and HP-UX. So far I did not dare to tackle those 900 compiler warnings ... Best regards, Ulrich Deiters
[sane-devel] gain/offset tables
The HP 6200 has some "internal scan area", too. But this is probably common technique. Another question is whether the gain and offset tables are readable or even writeable for the user. Some modern scanners like the Canon FS2710 film scanner seem to keep the gain control private and let all fine tuning of brightness, contrast, etc. to the computer. Regards, Ulrich Deiters
[sane-devel] segmentation fault
Dear Brian, "C" is just the dumb standard (ASCII keyboard) - should always work! Of course, accented characters, umlauts etc. might be inaccessible. Usually wrong language settings do not make programs crash. But "LANG=C; export LANG" is probably worth a try. Greetings, Ulrich
[sane-devel] segmentation fault
Hi, what is the setting of your LANG environment variable? It should contain an existing language specifier (e.g., de_DE.iso88591) or simply "C". Regards, Ulrich Deiters
[sane-devel] Canon backend (FS2710 problems)
A slightly modified Canon backend, which should avoid the segmentation
faults reported in connection with film scanners, can be picked up
by anonymous FTP from
stthd0.pc.uni-koeln.de (cd to pub!)
Has someone tried the raw scanning mode for a FS2710 and obtained a
picture containing random data? If yes, please let me know and tell
me what kind of computer you are using; there might be some misunderstanding
about the proper byte order ("bigendian" vs. "smallendian") which can
be easily corrected. Naturally, I cannot check this on my machine,
but must wait for your feedback.
Ulrich Deiters
