Re: Minolta Scan Multi Pro: An Initial Impression

[Kennedy McEwen <rkm@xxxxxxxxxxxxxxxxxxxx>]

In article <101676961801@wi.net>, david soderman <scapes@wi.net> writes
>Hello Kennedy,
>
>Indeed, I appreciate your help with my plight.  Regretfully, I must confess
>that your assistance is going "way over my head".  I'll comment on your
>quotes below:
>
>> Try scanning at the full resolution of the scanner and downsampling in
>> Photoshop then after some mild "softening" at the higher resolution.

>Yes, I know how to downsample in P.S.  What do you mean by "softening"?
>Scan 6x6 negs at 3200ppi (or 4800?) then apply some form of "blur" before
>downsampling...let's say down to 300 ppi?

More or less - but I am talking about getting to the resolution you 
require, not something like 300ppi.  In other words, the Minolta has a 
4800ppi optical resolution in one axis at least, so use that as your 
basis for scanning instead of 3200ppi.  Then apply a small amount of 
blur to this the amount of which depends on the effective scan 
resolution you are attempting to achieve.  For example, for an effective 
3200ppi resolution, a gaussian blur with a radius of around 0.7-1.0 
pixels should be enough to eliminate the image components which would 
alias.  Then downsample in PS to 3200ppi to get your effective final 
scan.  If your problem IS grain aliasing, then this should help to 
control it
>>
>> I don't know what algorithm is used in the Minolta to get 3200ppi from
>> the native 4800ppi in one axis and whatever the native resolution of the
>> other axis is (the specs seem to leave it undefined, referring ONLY to
>> the interpolated resolution - most unusual and, to my mind, very
>> suspicious.
>Don't know; can't tell ya.  Too technical for me.
>
> Perhaps there is more info in the operators' manual than on
>> the published specifications on their web site!).
>Nope.  'Taint so.
>
Sounds very fishy to me - very reminiscent of the "1200x600dpi" 
resolution figures that are quoted in some flatbeds.  Everyone knows 
that the 1200dpi figure isn't realistic because the low resolution axis 
requires x2 interpolation.

>  However. on the
>> Nikon, all scans at whatever resolution are made at the native
>> resolution of the scanner and downsampled from there - specifically to
>> control spatial aliasing and reduce noise.
>Grrrr!  Makes me wish that I'd kept the Nikon 8000ED after all.  ;-(

Damn good scanner, but I am sure you had good reason for choosing the 
Minolta and its not my intention to make you regret your decision - only 
help to resolve your problem.
>>
>> The object is to scan with minimum aliasing and then filter out the
>> spatial components which would alias before downsampling.
>This sounds really "techy" and clinically confusing to me.  Can you rephrase
>this in laymen's terms please?
>
I'll try - on the next message.  :-)
-- 
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers
-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.