Page 2 of 2

Re: Filter to improve some tif images

Posted: 2018-06-24T06:51:37-07:00
by snibgo
Jean-Pierre Coulon wrote:Then is there the right filter for me in grayscale?
Without seeing the original, I can't say.

The original may be printed at a resolution of 6000 dpi, but you are scanning at only 600 dpi. This gives antialiasing: gray pixels at the boundary of black and white.

If you want to save space, try saving the grayscale as JPEG, at different quality settings.

Re: Filter to improve some tif images

Posted: 2018-06-24T23:02:04-07:00
by Jean-Pierre Coulon
snibgo wrote: 2018-06-24T06:51:37-07:00
Jean-Pierre Coulon wrote:Then is there the right filter for me in grayscale?
Without seeing the original, I can't say.
You can obtain a greyscale example by replacing the header of my brief example with P2 50 15 255 and changing all 1 into 255's.
I could have named my topic: Filter to improve some jpg or png images
The original may be printed at a resolution of 6000 dpi, but you are scanning at only 600 dpi. This gives antialiasing: gray pixels at the boundary of black and white.
The jag comes from the irregularity of the paper and printing process
If you want to save space, try saving the grayscale as JPEG, at different quality settings.
I want a monochrome result.

The filter I need should be called -unjag or -dejag :)

Re: Filter to improve some tif images

Posted: 2018-06-25T01:04:58-07:00
by snibgo
Jean-Pierre Coulon wrote:You can obtain a greyscale example by replacing the header of my brief example with P2 50 15 255 and changing all 1 into 255's.
The result will still be black and white only. It won't replace data that has been lost.

How did you create the black and white version? Please show the exact command. If you dithered, this might account for some of the noise.

Re: Filter to improve some tif images

Posted: 2018-06-25T02:55:57-07:00
by Jean-Pierre Coulon
snibgo wrote: 2018-06-25T01:04:58-07:00
Jean-Pierre Coulon wrote:You can obtain a greyscale example by replacing the header of my brief example with P2 50 15 255 and changing all 1 into 255's.
The result will still be black and white only.
Sure. But some complain about my using tif or pbm. :)
It won't replace data that has been lost.
Please read again my request: "I would be happy if my 6th line was transformed into0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 "
I am not trying to replace lost data.
How did you create the black and white version? Please show the exact command. If you dithered, this might account for some of the noise.
Simply with a scanner at 600 dpi.