[filmscanners] Re: Noise in Polaroid SS 4000 scans

["Arthur Entlich" <artistic-1@shaw.ca>]

Harry has sent me a full res sample to analyze, and the problem is
definitely grain.

As mentioned before, with negatives, scanner CCD noise appears mainly in
the darkest areas of the neg or the lightest areas of the positive
inverted image.  The "defect" on his image is throughout the image, but
more visible in the middle and lighter areas on the neg.  However, even
assuming there was something very wrong with the scanner either
physically or electronically, there is another clincher.

Noise is totally random from pixel to pixel, that's why its called
"noise" in that it does not relate to the image itself in any manner
other than that we know CCDs tend to show it more as they are "gasping
for information" (in darker areas).

However, when zooming in on Harry's sample, it is very obvious that
these "artifacts" are of numerous variable pixels in size in all
directions, and all involve groups of pixels, and each pixel tends to
slightly change in density going toward a more defined "edge".  This is
exactly what grain appears as.  Since the scanner uses three lines of
CCD sensors, one pixel in height, each line would have little to no
relationship to the next in terms noise distribution.  Instead, every
artifact I can see in the scan has "buddy pixels doing the same basic
thing to either side, above, below and diagonally, in "clump".

Verdict: GRAIN

Now, I am not saying there isn't some grain aliasing taking place, as
sampling errors creep in with all scanners, less so with higher res
ones, obviously.  But that's math at work, and nothing can be done about
it.

So, not noise, but grain.

My recommendation: either poorly exposed, poorly processed or poorly
manufactured film, or a film type inappropriate for scanning.

Try exposing a bit more, a different processor, and if that doesn't
help, a different film type.

Art


Harry@vdKrogt.nl wrote:

> OK, I scanned some negatives Fuji Super G 100 I had. The scans were more
> clear, no comparison with the Fuji NPS, I think. So it must be the
> combination of the Film and development by the lab or the way the
> scanner reacts to the grain in the film, or both. In both cases I must
> look for another 160 ASA film (and I will have other NPS 160 developped
> by other labs, and see if that makes a difference).
>
> Thanks for the answers,
>
> Harry
>
>
> -----Oorspronkelijk bericht-----
> Van: filmscanners_owner@halftone.co.uk
> [mailto:filmscanners_owner@halftone.co.uk] Namens Alessandro Pardi
> Verzonden: woensdag 11 juni 2003 11:20
> Aan: Harry@vdKrogt.nl
> Onderwerp: [filmscanners] RE: Noise in Polaroid SS 4000 scans
>
>
> Hi Harry,
>
> to rule out the possibility that the problem is the film you can shoot a
> test roll of a good 100ASA film: Kodak Supra, Fuji Reala or Superia
> should all be so far from the grain (assuming it's grain) you see that
> it would be immediately clear whether the problem is in the film or in
> the scanner. In the meantime, a major difference between noise and grain
> is that noise is
> random: so if you use the multiple pass option of Vuescan noise should
> be greatly reduced, whereas grain would remain substantially unchanged.
>
> Alessandro
>
>
>>-----Original Message-----
>>From: Harry@vdKrogt.nl [mailto:Harry@vdKrogt.nl]
>>Sent: Wednesday, June 11, 2003 11:11
>>To: alessandro.pardi@inferentiadnm.com
>>Subject: [filmscanners] RE: Noise in Polaroid SS 4000 scans
>>
>>
>
> <snip>
>
>>I am a bit surprised to see that the opinions are so different, grain
>>versus noise. What can I do to make a more definite test? Make scans
>>at 4000 dpi and put them on the page? That would mean even more
>>bandwith.
>>
>>Harry
>>
>>
>
> ------------------------------------------------------------------------
> ----------------
> Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe
> filmscanners' or 'unsubscribe filmscanners_digest' (as appropriate) in
> the message title or body
>
>
>


----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body