Rick, that looks more like some kind of dithering pattern to me... it's too regular to be grain. Look at the parallel lines on that green patch. I think you're right about it just being older technology...
Rick, that looks more like some kind of dithering pattern to me... it's too regular to be grain. Look at the parallel lines on that green patch. I think you're right about it just being older technology...
Wow. That is very interesting. It looks very different than what I consider grain aliasing in my scans. However, that could be because I'm not seeing what I think I am
or the scanners just produce a very different high-frequency response. Does this scanner have different resolutions in the horizontal and vertical directions because the pattern is so polarized?
I use a Howtek 4500. I usually scan current E6 films at 4000dpi with an aperture of 13 microns (2000spi). This filters the highest frequencies and then over samples so I know I optimally resolve everything up to 2000dpi and minimize aliasing. That produces a very nice smooth scan and as best I can tell picks up everything if I'm shooting at f22 or smaller. At f16 and larger there is sometimes more detail to be had and If I decide it's worth it I'll scan at 4000dpi with a 6 micron aperture (4000spi) and in this case there is no oversampling and no smoothing of the data at the sampling frequency. The resulting scans do have a bit more detail (provided the transparency is really good enough) but much more noise. The interesting thing is that the increase in noise is not just at the sampling frequency but also bleeds out into slightly lower frequencies. Then if I use noise reduction and attenuate the "grain noise" I'm still left with the lower frequency noise. This noise looks just like grain noise only larger and at lower amplitude. It looks randomly distributed and isotropic. The ironic result is that the higher frequency scan can end up looking like it has larger grains.
Originally Posted by rdenney
Here is a scan of a small piece of a negative that reveals what I believe is grain aliasing:
The image is shown at 100% (scanned pixel = screen pixel), and the 400-pixel width covers about 0.15" of negative. The scan density was 2720 sensors/inch at the full optical resolution of the scanner, which was an Acer Scanwit 2720. The image was made 8 or 9 years ago, so the scanner technology is pretty ancient. I replaced that scanner with a Minolta Multi II, and then again by a Nikon 8000, and have not had this sort of a problem since.
Rick "not positive that this is aliasing, but reasonably sure" Denney
Could well be. That was a cheapie film scanner when there was no such thing.
Rick "too good to be true" Denney
Rick: that looks more like periodic noise to me, although there is an aliasing effect in halftone reproductions called "wriggley worms" which looks similar. The fact that the worms move together over the whole height of your section makes it more likely to be noise though.
Grain aliasing can be unambiguously identified if you have a good enough scanner to reduce the sampling density well below the grain size. In that case the apparent grain shrinks as you increase the sampling rate. Kodachrome 200 on a Nikon 5000 scanner shows it nicely. My LF and MF film on an Epson desktop is grainier than the film viewed through a microscope, but that could just as well be noise and posterisation, and I can't do anything about it anyway.
Sounds like you are on autotrack. I more often scan at 13, and let the dpi be 4000. Works great...
Lenny
That's what I said. What does it have to do with autotrack? Yes it does work great.
Must have mis-read that... nevermind.
Lenny
OK
What a beautiful photograph!
Same here, 13, 4000dpi. However E6 is a lot easier to focus than negs.
And with Ektar 100 I can't even see any grain, just noise.
Posting Permissions
Reply With Quote
Bookmarks