[sane-devel] FW: Sane Scanimages Too Dark

[Stef]

On 05/01/2013 01:45, Michael Watson wrote:
> Message
> I may have answered one of my questions after researching the sane 
> pages.  The "curve tool" is a transfer function graph.
> http://www.scantips.com/curve2.html
> Good stuff.
> R/
> Michael
> -----Original Message-----
> *From:* Michael Watson [mailto:mwatson33 at comcast.net]
> *Sent:* Friday, January 04, 2013 5:28 PM
> *To:* 'sane-devel at lists.alioth.debian.org'
> *Subject:* Sane Scanimages Too Dark
>
> Request insights from the brain trust.  Sane scanimage files are too 
> dark while windows' scans (Canon drivers/applications) are fine?
>
> I have an archlinuxarm box and an Lide210 scanner.  Utilizing 
> SaneTwain on a windows box the network scans were too dark. Similarly, 
> scanimage (local) scans were also too dark.  Trying to compensate the 
> SaneTwain gamma was experimented with.  What worked well was a 
> straight line curve that maxed out at the top while 3/4 of the way to 
> the right (0,0 --> 191,65535 --> 65535, 65535)and not the top right 
> corner ((0,0 --> 65535, 65535).
>
> With this information the scanimage (local) was adjusted with 
> gamma4scanimage 1.75 0 181 255 65535, which works well.
>
> I do not understand why the upper limit is at approximately 191 and 
> not 255.
>
> From the sane debug it appears the gamma is initially set to 1.7, 
> which produces dark images because the initial gamma table is 
> equivalent to gamma4scanimage 1.7 0 255 255 65535?
>
> [genesys] sanei_genesys_create_gamma_table: size = 256, maximum = 
> 65535, gamma_max = 65535, gamma = 1.7
>
> Also, SaneTwain will send the gamma table to the scanner when 
> initially set.  If it is saved (ini file) and SaneTwain is started up, 
> the custom gamma table checkbox (custom-gamma=yes/no) has no visible 
> impact on the scanned image - too dark.
>
> Any insights appreciated.
>
> R/
> Michael
>
>
>
>

     Hello,

     I have recently fixed an issue with gamma table handling in the 
genesys backend. This should improve the darkness issue you see. To test 
it you'll have to compile the latest source code available from SANE's 
git repository.

Regards,
     Stef
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/sane-devel/attachments/20130106/757ac3b0/attachment.html>