Good Day Flexers!
Today, I'd like to share a result from the processing thru SilkyPix 5. It is
not a concern or anything buggy, just a result I usually get with all the
images taken with the beloved Fujifilm X20, a small compact reminding me of the
Leica III serie (I love the design).
Of course most of you prefer another converter, and therefore, I'll issue below
a link to the Raw file so you can draw your own conclusion. Please yourself
with a 18.3 Mega byte file: http://1drv.ms/21eMa9R
Before you start clicking, for whoever just want to see an image, reasonnable
in size, here: http://1drv.ms/1U3Nufa
Back to the post, all stuff is large, it is a "modem burner" of a subject
The image developped "normally" with the options one can expect, such as:
Here is an amplified of the image view with a grid forming, I have placed red
lines to demonstrate: http://1drv.ms/21eKahJ
The developped image is : http://1drv.ms/21eMYeE
As you may see, this grid appears in some areas of the image, in some parts, it
is not so visible.
Why this could be?
In fact, if I change the "unsharp option" in the development, if I remove it, I
get this image (amplified):
As you can see, the 'grid' pattern has almost disappeared.
After some research, I learned that thee unsharp mask is an algorithm applied
after the development is done. It must be true as the processing time with no
"unsharp" is less on my computer.
In the development, there is a choice for normal or pure détails:
Using Pure details and no unsharp leads to lesser grid effect, however, the
image looks a bit smeared: http://1drv.ms/21eKVYi
In some way, it looks like the original image JPG, out the camera:
The difference? the camera takes 5 seconds to process and Silkypix a good
When Pure detail is applied as well as Original unsharp is chosen, the grid is
All together, the conclusion of such is that, it is really counter productive
to change too many parameters, they cancel each other. But this does not come
as a surprise. Complex process of correction is usually not reproductible from
image to image and time waste anyway.
Although the latest update of the firmware of the X20 has really improved the
images, better colors. a few bugs out in the Man Machine Interface, I strongly
believe that the 'grid' effect is a signature of this 4x3 X Trans Sensor
(strangely, this effect does not exist in the 2x3 APS Sensor of the X E2, also
a X Trans of second generation....).
And it needs to remove the unsharp mask when developped at full size, no big
In small size such as 900x1200, the unsharp mask 'for web' option does not
imrove a lot.
Interestingly, with the Digilux 3or the various Panasonic LX serie compacts I
had, it did improve.
The pattern of the pixels are reminicent of the chemical film when they were
observed with a loupe. To me, it is a signature as well. When the unsharp mask
parameters are over set (choice 'for L size printing') the pattern looks like
'pointillism', this method developped by Van Gogh and others. It is somehow
nice. I'll explore it one day...
And if you want to see the files produced (same photo, various parameters),
look through the folder: http://1drv.ms/21eMw0d
Ah, some of you are going to be picky, with reason, showing a portrait of a
lady with so much details does not do her justice. My photograpphic subjects
are often more or less from my universe....
From : Xavier F. BILLE
mail : hot_billexf@xxxxxxxxxxx
Maisons Alfort - France