Sunday, May 19, 2013

Resiliance

About a month ago, I found about 20 4x5 slide holders. Since until last fall, I hadn't done any developing since 2006 and therefore not taken any 4x5s since then; these were old so I wasn't expecting much when I developed them.

The negatives were very dense, probably their age and terrible miss-treatment; they'd been in a box in the woodshed since 2006 which means freezing in the winter and really hot in the summer.  But the wonder of digital technology is that one can fairly easily capture all the available information a negative has to offer, however traditionally 'unprintable'. These would have needed something like grade 6 paper - if such a thing existed;  the film base was so dark that there was very little difference in opacity between the highlights and the shadows. They looked as though they had been accidentally exposed to light. However, if you tell the scanner's AD converter how to interpret even these terrible negatives, it will generate a nice smooth 256 interval scale between the shadows and the highlights, regardless of the negative's overall density.

Here's one that was taken in about 2002.

  
The developer is Xtol (or EcoPro), 1+1.

Thursday, May 2, 2013

Three formats

Clips from the same scene in three formats; 12MP digital (the LX7), a 2,400dpi and a 1,800dpi scan both from the same RZ6x7  Arista 100 negative (significantly under-developed - the app is, or I am, doing something wrong) and a sheet of Arista 100 scanned at 1,200dpi from the SR45. The lenses were   Mamiya 50mm for the RZ67, a Schneider Super Angulon 90mm f8 for the large format. The LX7 was at 24mm equivalent.  

LX7


RZ67 @ 2400dpi
RZ67 @ 1,800dpi

SR45 @ 1200dpi
What I find slightly surprising is that the 1,200 dpi scan of the SR45 negative, which creates a file with fewer pixels than the higher res scan of the RZ67 neg, is more detailed. Further evidence comes from the files sizes. The high res RZ scan has 1.3 times as many pixels but generates a smaller file meaning that there are more regularities that the jpeg algorithm can compress.

width height px (M) file size (MB)
LX7 3776 2520 9.5 4.8
RZ67 @ 2,400 6280 4983 31.3 8.6
RZ67 @ 1,800 4710 3737 17.6 5.5
SR45 @ 1,200 5411 4300 23.3 9.1

(The saw cuts in the last image I find particularly satisfying). I also discovered that you can enter any scan resolution directly into the Epson dialog box; you don't have to use one of their presets in the drop-down. 1,800 dpi isn't one of the options, but the scanner did exactly what was asked when I put that resolution in... I should probably read the manual.