Page 1 of 1

GANE vs USM & Descreen for grain and grain aliasing?

PostPosted: Tue Sep 20, 2005 7:43 am
by MicroSF6
if my goal is to reduce grain and grain aliasing, is it better for me to use the GANE filter or the USM & Descreen filter? i have studied the silverfast tutorial on GANE, and i have studied the ian lyons tutorial on USM & Descreen, but i am unclear as to which filter is better for my needs. i am usually scanning kodak portra nc 160, in the 6x7 cm format.

thank you!

mark

PostPosted: Fri Sep 23, 2005 4:04 pm
by LSI_Noack
Dear Mark

actually, I suggest you want to to use multi-sampling first when scanning transparencies (4x to 8x). Descreening is intended to eliminate artifacts that occur when scanning screened / printed images.

GANE actually is the tool intended for want you want to achieve, but cannot be used in combination with the unsharp masking tool.

Hope I could point you in the right direction, best regards
Sonny Noack
- Manager Technical Support, LaserSoft Imaging AG -

PostPosted: Sat Sep 24, 2005 8:38 pm
by MicroSF6
thank you, sonny noack.

your write "Descreening is intended to eliminate artifacts that occur when scanning screened / printed images."

but ian lyons has a tutorial recommending the use of descreening & USM for transparencie as a way of addressing the grain aliasing problem. i think he wrote it in 2002. is it no longer recommended?

you write "actually, I suggest you want to to use multi-sampling first when scanning transparencies (4x to 8x)."

why only 4x to 8x? why not 16x? i would like to be able to do this, and i can initiate a scan in 16x or 8x multiple scan mode, but doing so results in an extremely distorted ("stretched") and merely partial scan image, a severey cropped and distorted version of a very small section of the much larger film image that i select for the scan. i am using the microtek artixscan 120tf with mac osx 10.3.9 with 2 gigs of RAM on an apple G5 1.8 GHz dual-processor tower (PLENTY of RAM and processing power). i have had this problem with the last 2 or 3 versions of silverfast ai 6. i was told by your representatives via e-mail 2 or 3 versions ago that lasersoft would "fix it in the next version." i have now updated to silverfast ai version 6.4.2r5d(MicroIT8). the problem is still not fixed: when i scan in 48-bit mode, non-raw (not 48-bit HDR but just 48-bit), i can multiple scan at 1x, 2x, 4x, and 8x, but not 16x. I now need to be doing my scanning in raw mode (48-bit HDR color), and in that mode I cannot multiple scan 16x or 8x without getting the same kind of extreme distortion and arbitrary severe cropping of the image. please note: this does not happen with scans of 35 mm film; it does happen when I scan 6 x 7 cm film.

this is extremely frustrating and disappointing. I need to be able to do raw scans in 16x Multiple Scan mode. would upgrading to AI studio solve this problem? could you encourage the bug-fixers there at lasersoft to fix this problem for me and notify me when it it fixed? thank you!

sincerely,

mark