Re: PING: ALL YOU FILM LUDDITES !

On Mar 24, 1:01 pm, "Scott W" <biph...@xxxxxxxxxxx> wrote:

This is a tricky area, film scanners tend to over sample which in one
way is good because you don't get aliasing in a film scan. At least
you don't get aliasing from the image detail, grain aliasing is a
different subject. But DSLRs tend to under sample, they want to get
all the detail they can from the pixels they have. This means a
digital camera will capture more detail on a pixel pre pixel basis but
at the expense of possible aliasing.

Good point!


Well I don't believe you need to have a tiff file, if you look at some
of my images you will see they are very clean and they are jpeg. But
if you really want to post a 35MB tiff let me know and I will set you
up with a FTP account and a directory to upload to.

jpg compression works mostly by throwing away detail, Scott.
It *must* lose detail, or else it's not working properly!
:-)
Yeah, no worries: send me an address and I'll upload some
tiff images for you to have a go at analysing. the email address
above in yahoo is good to send stuff to.



But just to show what a jpeg can do this image has the same number of
pixels that a 4000 ppi scan of 35mm film would have.http://www.sewcon.com/largephotos/church_small.jpg

yeah, but number of pixels doesn't translate into detail.
jpg works by "aliasing" - also known as the "jpg artifacts"
similar image areas. It doesn't matter how many pixels
it's got, if it's a jpg it has already lost detail. check
what happens to the tiles in the roof as it progresses
to the left. And how the tiles under the tower
are less detailed than the tower walls themselves: that's
the jpg compression stufing things up.

down. Film scans don't compress as well BTW due to the higher noise
in the scan, you get the same effect if you shoot at very high ISO
setting with a digital.

Exactly. Run the film scan through a noise reducer
and then compress it with jpg: big difference!

.

Relevant Pages

  • Re: Value (and values) 35mm
    ... pixels in width and then up sample to one that is 5400 pixels wide. ... to upsize the 9x6 I already optimized for printing with much better ... x 533px *.jpg all look the same, except for some jpg compression ... Scott always bashes film. ...
    (rec.photo.equipment.35mm)
  • Re: When is wide not wide?
    ... > actual image size in pixels is the same. ... > effect on the transmission. ... It is indeed a lower quality image then. ... >>>films, allowing me to see the film as was intended. ...
    (uk.tech.digital-tv)
  • Re: PING: ALL YOU FILM LUDDITES !
    ... lines/mm at contrast ratio of 1.6:1 - barely perceptible, 140 lines/mm at a contrast ratio of 1000:1, and the MTF curve is down to 25% at about 60 lines/inch. ... Being generous using the 1000:1 figure, gives a resolution of 3556 lines/inch. ... Of those 3,504 pixels, only half as many register green light, ... beaten by a 35mm SLR using fine grain side film. ...
    (rec.photo.equipment.35mm)
  • Re: PING: ALL YOU FILM LUDDITES !
    ... MTF curve is down to 25% at about 60 lines/inch. ... Of those 3,504 pixels, only half as many register green light, ... beaten by a 35mm SLR using fine grain side film. ... Every pixel of a digital image made from scanning film gives complete ...
    (rec.photo.equipment.35mm)
  • Re: I didnt like The Queen in 3D
    ... It turns out that the typical grain size in 100 ASA colour film is 3 ... That gives a grain density of under 350 grains per mm. ... To digitise a film without loss of resolution, ... suitable non-aliasing algorithm to 350 pixels per mm. ...
    (uk.legal)