[sane-devel] LiDE 90
pierre at pirsoft.dnsalias.org
Tue Dec 2 23:39:50 UTC 2008
guillaume.gastebois at free.fr schrieb:
> I'm back with my LiDE 90. No more significant result since month. I don't know where to find.
> Just a little question. I see in windows snoop that inversion bit is set in frontend, not in sane.
> Is it possible that this inversion has an effect on calibration and scanning (different full scall...) ?
In theory, there should be no difference, since the digital filtering
would essentially stay the same, when using inverted/non-inverted or
arbitrarily biased input, and the analog stages cannot be changed in
polarity via programming.
> I want to test with inversion, but I can't find in genesys backend code where to software reinvert datas.....
You can't find that code, because there is no code. For use with
inverted data from the adc, you'd need to change all calibration
procedures to handle that data, and add a stage into the filter chain.
The filter chain is the stuff around genesys_conv.c, you can probably
adapt one of the reorder functions. That, or processing the data
directly from the scanner to appear non-inverted to the rest of the
calibration logic. This would still need changes to at least the shading
calibration, since the coefficients are different for inverted input.
> Can someone help me ?
> I did some new snoop (without calibration phase ?) using usbsnoop and parsed it using parseusbsnoop_gl841.awk from John Dalton (thank you).
> They can be found at : http://ggastebois.free.fr/lide90_snoop/UsbSnoop_600_216x17.4_nocal.log and http://ggastebois.free.fr/lide90_snoop/UsbSnoop_600_216x17.4_nocal_parsed.log (be carefull with headake !)
More information about the sane-devel