[sane-devel] Epson ES-50 errors and misbehavior

[Sean Greenslade]

On Wed, Aug 27, 2025 at 03:16:32PM -0400, Sparr Risher wrote:
> I am using sane 1.4.0 with my Epson ES-50 (aka J421B). I am experiencing a
> few problems, potentially some or all of which have the same root cause.
> I'm moderately proficient with troubleshooting and coding in C, but I know
> nothing about scanner protocols or how the sane codebase is organized. I am
> hoping someone here might be able to give me some pointers on where I
> should look first in digging into one or more causes for these issues:
> 
> In some frontends, including Gnome simple-scan, when I attempt to scan an
> additional page into a batch, I am told there is no document in the feeder
> on alternating attempts. First page scans fine (almost?) every time. First
> attempt to add a page fails. In scanimage, as soon as that attempt fails,
> the batch ends. In Gnome simple-scan, I can click the button to add a page
> to the batch again and the second attempt will succeed, then I click again
> and the third attempt fails, and again and the fourth attempt succeeds, etc.
> 
> Every frontend I've tried produces 8.5x14" images when I scan 8.5x11"
> pages. The unscanned area is solid white.
> 
> Every page I scan with scanimage produces additional output to stderr
> "Corrupt JPEG data: found marker 0xd9 instead of RST3"
> 
> When I run scanimage in -T -v mode it also ends its output with "FAIL
> Error: End of file reached"
> 
> Here is one session showing all of the problems except the last one, where
> I attempt to scan two 8.5x11 pages and get one 8.5x14 output image then a
> failure:
> 
> % scanimage -L
> device `epsonds:libusb:003:004' is a Epson ES-50 ESC/I-2
> 
> % scanimage -v -b --batch-prompt --format=pngScanning infinity pages,
> incrementing by 1, numbering from 1
> Place document no. 1 on the scanner.
> Press <RETURN> to continue.
> Press Ctrl + D (EOF) to terminate.
> 
> Scanning page 1
> Corrupt JPEG data: found marker 0xd9 instead of RST4
> scanimage: scanning image of size 1696x2795 pixels at 24 bits/pixel
> scanimage: acquiring RGB frame
> scanimage: min/max graylevel value = 0/255
> scanimage: read 14220960 bytes in total
> Scanned page 1. (scanner status = 5)
> Place document no. 2 on the scanner.
> Press <RETURN> to continue.
> Press Ctrl + D (EOF) to terminate.
> 
> Scanning page 2
> scanimage: sane_start: Document feeder out of documents
> Batch terminated, 1 page scanned

Hi, Sparr.

I happen to own the same Epson scanner. I've encountered most of the
quirks you mentioned here, though I've just worked around them with my
particular workflow.

The "found marker 0xd9" I believe is fairly harmless. I think this
happens because the stream of JPEG data from the scanner gets truncated
as soon as the end of the paper is detected. I never dug any deeper into
this issue as it didn't appear to hurt the output of the scans at all.

I get the same 8.5 x 14 inch output with white area as the excess. In
fact, with my ADF area fixes in place, I get a much larger (8.5 x 72 inch)
file. I only use this scanner as a receipt scanner, and I have a little
python script that trims the scans down to size automatically, so this
issue never bothered me enough to fix within SANE. I seem to recall that
the SANE APIs might make retroactively resizing the scan area a little
awkward to do, but I'm sure it's possible.

The "out of documents" error I believe is related to the sane_cancel bug
I fixed in my merge request here:

https://gitlab.com/sane-project/backends/-/merge_requests/753

I rebased it recently(ish), so hopefully it still builds. Give that a
try and see if it at least fixes the "out of documents" error.

--Sean