[sane-devel] Re: sane-devel digest, Vol 1 #826 - 6 msgs

[Gregory C. Johnson]

>Using lower bitdepth may still be faster than higher ones. So even if
>the size of the resulting image file is the same, lower depth may be
>useful if your scanner is faster with the lower depth.

True, though that's not the case here. (AFAIK, I've not time()ed the
scans, though I probably will soon.)  Does anyone know of an instance
where a scanner is noticibly slower on an 8-bit scan v. a 4-bit scan?

>> Does anyone know of a SCSI sniffer for W2k available for free/student
>> prices?  a tcpdump equivalent would be ideal.
>
>I haven't looked at the w2k SCSI architecture. In the win95 days I
>used to use the ASPI wrapper dll but I don't know if ASPi is used
>nowadays.
>
>http://www.meier-geinitz.de/sane/misc/develop.html

Thanks, I'll go the W95 route with aspispy if I can't get the
Patin-Couffin driver to log.  I've already struc out with about everything
else.

-Greg


BTW, pedantically speaking, shouldn't:

              *dst-- = (*src << 4) & 0xf0;
              *dst-- = (*src--) & 0xf0;
read as:

              *dst-- = ((*src << 4) & 0xf0) + 0x07;
              *dst-- = ((*src--)    & 0xf0) + 0x07;

The intensity drop from the zeroed bits is not great, but it is easy to
notice when flipping between 4/8 bit scans of the same thing.  If you
agree, this might deserve a note in the backend development docs.