Help with poor scan from negative

Discussion in 'Digital Cameras' started by Yi-Zen Chu; Yiren Qu, Jul 13, 2003.

  1. Hi everyone,

    I recently got 7 rolls of negatives scanned by Color Lab Photo. The
    scans turned out to be very noisy / grainy. A sample scan is as follows.

    The small version (scaled down using Batch Thumbs) is here: (200KB)

    The original version is here: (3.1MB)

    The strange thing is that I got 2 4R prints from the image files and
    they look good.

    1) Why does this happen - i.e. why does the image look poor on the
    screen but print well?

    2) What can I do to make the image look good on the screen? I hope to
    make a webpage out of the images I scanned. I do have photoshop,
    although I do not know how to use it.

    Thanks for the help!

    Yi-Zen Chu; Yiren Qu, Jul 13, 2003
    1. Advertisements

  2. Color Lab Photo is a local lab here in Singapore, in case anyone is

    Yi-Zen Chu; Yiren Qu, Jul 13, 2003
    1. Advertisements

  3. Yi-Zen Chu; Yiren Qu

    Comrade Guest

    it is much easier to make a print from a negative, then to scan a negative
    on most scanners, so, just scan your print

    if you want the resolution of film, and you are going to bracket, there is
    no reason not to just use slide, unless you need prints, or you need
    critical protrait flesh

    to scan a slide, the color engineering of the scanner just has to convert
    the spectral transmission of the film, to CIELAB, and is a very easy
    algorithm, and so it is the same with a print, except there are
    reflectances not transmittances

    now, with a negative, the scanner must interpret the intent of the
    negative, because a negative is designed for the purpose of making a
    print, so, to get the intended look the negative is designed for, from the
    scanner, it must consider the design factors of the negative itself, the
    printer assumptions used in that design, and the print paper
    characteristics, used in that design, negative and print, is a system,
    with printers in the middle of it

    and you can "repurpose" a negative, from its design, by mathematically
    adjusting the considerations of such items, like paper/printer, in the
    scanner color path, or ICC profile, so that you can get a "different"
    look, that what the negative entails, IN FACT, you can even correct for
    deficiencies in the chemical system in achieving it's design, AND, you can
    even remove all design, and revert the negative system, to a system of
    accurate scene reporduction, with no "look" added to it, by undoing the
    deficiencies in accurate color reproudction in the chemical system, such
    as what was done with photo cd (not picture cd), there are probably some
    patent considerations in this arena which do not expire until 2007

    so, in summary, to scan negatives, good, you need apriori knowledge of

    1) the design assumptions and system characteristics of the film 2) film
    characteristics, such that you can unbuild the film, to a scene or
    accurate reproduction, or different "look" or "intent"

    because negative has an intent, and the intent is NOT to be looked at by
    the eye, or a scanner, and to arrive at that intennt, on a scanner, you
    must know the design characterists of the system (spectral sensitivities,
    dye absorptions, of both paper and film, chemical interlayer/imterimaeg
    effects of both paper and film, printer light source spectral radiation,
    and filtration absorptions, etc., etc.)

    so, it is VERY unlikely that in the near future you will get an "accurate"
    scan from a negative, perhaps a pleasing one, but not an accurate one, in
    terms of the scene, or intent of the film and system

    slide, on the other hand, is very easy, but it is not a good accurate
    reporduction, do to chemical imperfections, these imperfections can be
    corrected upon scanning, IF, you have film data as described above,
    otherwise you do not know what the scene looked like, you only know what
    the slide looks like, and this does seem to suffice for most applications,
    but digital capture color reproduction, is far and above the most
    accurate, since it entails one set of filters on a CCD to get from the
    innate sensitivity of the CCD to CIEXYZ or CIELAB, the response of the eye

    see my friend Average Joe's site
    the aristocracy was the problem in 1776
    the aristocracy is the problem today
    we must close the door by which aristocracy arises

    "Does God want goodness? or the choice of goodness?
    Is the man who chooses bad, somehow better,
    than the man who has the good forced upon him?"
    a quote from the movie, A Clockwork Orange, Kubrick

    Enlightenment is man's release from his self-incurred tutelage.
    Tutelage is man's inability to make use of his understanding
    without direction from another. Self-incurred is this tutelage when
    its cause lies not in lack of reason but in lack of resolution
    and courage to use it without direction from another. Sapere aude!
    "Have courage to use your own reason!" - that is the motto
    of enlightenment.
    Kant -- What Is Enlightenment? 1784

    Ayn Rand just professed laissez faire, another Jewish mysticist
    of Zionist status quo social order, and aspiring to such, she
    never challenged "the establishment" she was a radical crony,
    not a radical individual, the very anti-thesis of individualism
    Comrade, Jul 13, 2003
  4. Yi-Zen Chu; Yiren Qu

    MA Guest

    2) What can I do to make the image look good on the screen? I hope to
    I have no experience with scanning negatives, but it seems to me the images
    are far more noisy than you should expect.

    Using Corel Photopaint, I made a few experiments and managed to, what I
    think, improve the images a lot. What I did was using 1) lowpass filtering
    and 2) "remove noise". Both filters did a great job removing the noise, but
    one has to be aware that also some image details will be lost during such a
    process. Have a try with Photoshop; the filters/effects probably have
    similar names.

    Best regards, Magnus
    MA, Jul 13, 2003
  5. Yi-Zen Chu; Yiren Qu

    JPS Guest

    In message <[email protected]_y_a_l_e_._e_d_u>,
    Are you viewing the pictures pixel-for-pixel ("Actual pixels" or "100%"
    in Photoshop? That makes the grain from the film much larger than it
    was intended to be viewed at; you're looking at your picture as if it
    was being printed at 47 inches wide, and 31 inches tall!

    Even if you use a shrunken view in Photoshop, at certain zoom levels the
    grain doesn't disappear as well as others, and percentages close to 100
    will always look grainy with your sample. If you view it at 25% zoom,
    the grain pretty much disappears.

    Make copies of your scans and shrink the copies to small sizes for
    on-screen viewing; this will smooth out the grain. There are many ways
    to attempt to get rid of the grain, and there are special programs just
    for getting rid of noise.

    Just using Photoshop's "Image size" and changing the size to something
    like 640*426 with "resampling" enabled will probably do the trick,
    though. A downsizing that drastic usually gets rid of most grain and
    noise. After downsizing, you might want to use "Unsharp mask" with a
    radius of 0.5 to 1.5, and a strength of 50 to 150 to make the final
    image sharp.

    Again, don't downsize the original; do it to a copy or duplicate.
    JPS, Jul 15, 2003
  6. Yi-Zen Chu; Yiren Qu

    JPS Guest

    JPS, Jul 15, 2003
  7. Well, it is destined for a web page, you know. Not a 16"x20" print.
    Tony Whitaker, Jul 16, 2003
  8. Yi-Zen Chu; Yiren Qu

    JPS Guest

    In message <[email protected]>,
    Funny, I see a 47" x 31" blowup on my computer monitor!

    JPS, Jul 16, 2003
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.