Rethink text and lineart scanning, especially for fax use

SimpleScan stores "text" mode scans as 2 bit per color grayscale images, whereas XSane stores them as real lineart (1 bit per pixel).

This can lead to all kinds of unwanted effects:

In PDF viewers, 2 bpc "grayscale" images appear light gray instead of black and white. Users tend to adjust the scan brightness setting and wonder why it doesn’t help. (See simplescan-scan.pdf—brightness was turned to full "dark".)
Fax software, like Roger Router, usually expects "lineart" (1 bpc) with 200x200 dpi resolution. Grayscale images (like SimpleScan’s) get rasterized and usually produce illegible output (see screenshot comparison).
OCR software sometimes also expects lineart (although many work with 8 bpc grayscale images, too).
2 bits per color/pixel is a rather unusual format and may not be handled well by all applications and printer drivers.

Screenshot showing faxed document from both XSane and SimpleScan side-by-side:

For verification, I enclose the following files:

original.pdf (which I printed out, scanned and faxed using Roger Router)
original-faxed.pdf (directly faxed without previous scan, for quality check)
simplescan-scan.pdf (300 dpi "text" mode scan done with SimpleScan)
simplescan-faxed.pdf (above faxed using Roger Router, hi-res mode)
xsane-scan.pdf (200 dpi "lineart" scan done with XSane)
xsane-faxed.pdf (above faxed using Roger Router, hi-res mode)

You can use pdfimages (from the poppler-utils package) to view the PDF image properties:

$ pdfimages -list simplescan-scan.pdf
page   num  type   width height color comp bpc  enc interp  object ID x-ppi y-ppi size ratio
--------------------------------------------------------------------------------------------
   1     0 image    2480  3507  gray    1   2  image  no         4  0   300   300 52.8K 2.5%

$ pdfimages -list xsane-scan.pdf
page   num  type   width height color comp bpc  enc interp  object ID x-ppi y-ppi size ratio
--------------------------------------------------------------------------------------------
   1     0 image    1632  2335  gray    1   1  image  no   [inline]     200   201 16.0K 3.4%

and you’ll see that SimpleScan stores "text" mode scans as 2 bit per color grayscale images, whereas XSane stores them as real lineart (1 bit per pixel).

Since SimpleScan is the most useful scanning application for the casual user as well as office users (XSane is clearly superior but much too complicated), I’d strongly suggest to either

rethink the decision to use 2 bpc for "text" mode and revert back to 1 bpc, or
have three scanning modes like "lineart", "grayscale", and "colour" that would produce 1 bpc black-and-white, 8 bpc grayscale, and 8 bpc RGB images, respectively.

IMO, SimpleScan would benefit greatly from offering real lineart scanning and being able to produce fax-able documents (especially since we now have 200 dpi resolution, see also #223 (closed)).

In some legislatures (like Germany), sending a fax is required for the document to be legally valid (as opposed to sending an email with a PDF attached). So there is still a need for this archaic technology. ;-)