Another RAW Question here (1)

Discussion in 'Digital SLR' started by Steve Franklin, Oct 4, 2005.

  1. I am just experimenting with RAW and Photoshop CS2 and have a couple of
    questions too about it that perplex me.


    I understand the advantages of RAW but as far as I can see the photo you end
    up with is reliant on your memory of what the scene look like. What would
    happen if you shot RAW and gave it to someone else to post process? How on
    earth would they know what the original colours were like? At least when you
    shoot Jpeg the in camera processing more often than not gives a pretty good
    approximation of what the scene should look like.

    I have read through the Adobe primers and this question is never
    mentioned...am I missing something here?
     
    Steve Franklin, Oct 4, 2005
    #1
    1. Advertisements

  2. It depends what your are looking for in your photograph. Are you trying
    to capture exactly what you saw? Are you trying to compose a piece of
    art? Are you telling a story and trying to convey a mood in the
    photograph. Post processing is useful with all of these things. You
    could still do it with film, but it was usually accomplished with the
    aid of filters like an 81 series warming filter.
    Yep ... the goal of your post processing. It is up to you to determine
    what you are trying to create in the digital darkroom, not photoshop.
     
    Thomas T. Veldhouse, Oct 4, 2005
    #2
    1. Advertisements

  3. ok...well lets look at it from a perspective of how to shoot RAW and get the
    most accurate colour rendition of the scene...it could be months before the
    photographer gets around to post processing images...how do you get the most
    accurate rendition?

    Worse how do you give it to a person for post processing that wasn't present
    when the shot was taken?
     
    Steve Franklin, Oct 4, 2005
    #3
  4. Normally, all I have to do is bring up the shadow level. Everything
    else is generally OK. You'd have to assume that the person knew that
    grass was green and the sky was blue. With RAW, you shouldn't have to
    tinker with colors. I suppose it's a matter of training like anything
    else. You train the person to create images the way *you* want them.
     
    Randall Ainsworth, Oct 4, 2005
    #4
  5. "A disturbing new study finds that studies are disturbing"
    I tend to agree here, I found that after working with RAW for a while, I
    wind up doing just about the same thing to every image with only varying
    degrees of difference based on the situation. So in fact the RAW image is
    very close to accurately capturing the scene but I have the advantages of
    correcting WB if needed or to enhance shadows, or to improve contrast where
    needed. In other words, I'm not drastically changing the photo with RAW
    manipulation and conversion, just enhancing what's provided... Adding a
    little punch here and there. Most importantly, after doing it for a while,
    I can get through a large number of shots fairly quickly so its not often
    that I forget what the scene looked like.
     
    Robert R Kircher, Jr., Oct 4, 2005
    #5
  6. How would you do it with film and a darkroom? Same answer.
    No different than giving your film to a pro lab.
     
    Thomas T. Veldhouse, Oct 4, 2005
    #6
  7. Wait until you take a picture with flash, or outside on a cool
    clear winter day, and accidentally have your camera set to color
    correct for incandescent light.

    With a RAW file, that is ignored, and you (must) post process
    for whatever white balance you like. With a JPEG, it will be,
    errr, fun...
    I don't use MS-Windows or Photoshop, so I can't speak directly
    to either of those. GIMP, /ufraw/, and /dcraw/ are the programs
    that I do use, and while I can't remember if ufraw is available
    for MS-Windows or not (and I suspect not), the other two are.

    I'm not sure if GIMP and /dcraw/ provide this without /ufraw/,
    but using /ufraw/ as a GIMP plugin to convert from RAW to
    something GIMP can work with involves an initial step where red,
    green and blue gamma curves are overlaid on a single graph.
    There is a slider to adjust color temperature, which moves red
    and blue in opposite directions. Another slider adjusts only
    the green level.

    Plus /dcraw/ can be set to any of several different preset white
    balance options.

    Generally if something was taken, for example, using
    incandescent light, if that is selected (or if auto is selected)
    with /dcraw/ the gamma curves above will in fact be right on top
    of each other. On the other hand, if no white balance is
    selected, in my experience the result of moving the color
    temperature and green intensity to cause the curves to match is
    almost always as close as I can remember to whatever the scene
    actually was. In fact, without even actually viewing the effect
    on the image, that is the way that I color correct almost all of
    my images. The reason is because with GIMP and the /ufraw/
    plugin the preview screen for the image while adjusting the
    curves is too small to be useful, and I don't bother inspecting
    it for color correction until I am finished with that stage and
    click on "okay" to convert the image for GIMP manipulation. At
    that point I get a full screen image to look at. I use a pair of
    19" LCD monitors, so that gives me an image which is larger than
    the final print will be. It happens that I rarely do anything
    more to for color correction, as the blind-and-by-the-curves
    method works very well.

    You might want to check out what GIMP can do, and if /ufraw/
    isn't available for MS-Windows (or even if it is) you might want
    to consider using a Linux box to process images.
     
    Floyd Davidson, Oct 4, 2005
    #7
  8. The photo is what it is; surely it is that your opinion as to whether it
    is correct that depends on your memory.

    I would suggest that a simple way to check that your results are at
    least reasonably close would be to take a Macbeth Colour Checker with
    you and photograph it in the same lighting conditions. You can then
    compare your results with the original....
    ......And, since it is an industry standard, the person doing the
    processing should probably (if a serious worker) have a Macbeth chart of
    his own to do a similar comparison.

    It's not the same as a thorough system calibration, but it should give
    you some reassurance or point up any material differences.

    David
     
    David Littlewood, Oct 4, 2005
    #8
  9. Steve Franklin

    bmoag Guest

    It is difficult for people with no darkroom experience to understand why raw
    is the best format to shoot in. If your concept of photography has been
    conditioned by the film/minilab/take what you get paradigm then you have no
    understanding of how color, tone and contrast can be manipulated globally
    across the image or how techniques can be used to enhance just one part of
    the image. Most people do not realize the potential that is contained in
    their images, digital or film, because they have never experienced a well
    finished and well printed image. The digital darkroom brings high quality
    image making within the grasp of most consumers but the process is not
    automatic and you have to decide whether or not it is worth it to you to put
    the effort into learning the process.
     
    bmoag, Oct 4, 2005
    #9
  10. Steve Franklin

    larrylook Guest

    This is a great question, and with digital photography you often have almost
    too many options as to what the final shot will look like. Some want the
    shot very saturated, some don't. Some want if more contrasty, some less.
    The photographer has so many choices - almost too many!

    Here's what I do with the D70. I feel that the experts at nikon must know
    more about photography than me and have some idea what the shot should look
    like. So I don't want PS or PSE to do any "automatic" or "optimizing" at
    least initially on my raw shot. I open raw with View (free and I'm too
    cheap right now to buy capture). It applies all incamera settings and makes
    the shot look like the HQ jpeg would have looked like had I shot jpeg. I
    think it's a reasonable starting point. Then I fool with possible WB,
    exposure, sharpness, levels settings in View. Then open in PS (or save as
    tiff and open in PSE), and do more editing (contrast, saturation, levels,
    cropping) there if it seems helpful. Then print. What camera do you have?

    Most shots won't look right if you just let PS or PSE do automatic or
    optimization - it's too thoughtless of a process. Most shots out of my
    camera if not processed would would look a bit dull and dark if just printed
    on my i9900 and ilford paper. I would never think of printing a raw shot
    out of my D70 without at least running levels on it in photoshop and
    slideing the R bar over to the L a bit and playing with the middle bar to
    get the midtones. This may sound extreme, but I think it would be a crime
    to print a photo (and potentially waste paper/kill trees) without running
    levels (image -> adjustments -> levels in PS) and playing with the 3 sliders
    there.
     
    larrylook, Oct 4, 2005
    #10
  11. Steve Franklin

    McLeod Guest


    True, but you can always set the white point and the black point in
    the scene and do a better colour correction than the camera would.
    You can also adjust the levels histogram to make sure there is no
    clipping. What else do you need to do?

    Think of raw files as negatives. When you took your film to the lab
    did you stay and watch the printer and tell them how to adjust every
    file? Did you know that in colour printing the operator decided what
    colour your print should be and what density? The machines averaged
    the density for every negative and tried to print a certain colour
    balance. Anything unusual like a night shot, or snow shot, could fool
    the machine just as easily as your camera can be fooled when shooting
    jpegs or raws. But at least with raw files you can get more post
    processing leeway.
     
    McLeod, Oct 4, 2005
    #11
  12. Steve Franklin

    McLeod Guest


    I suggest "Photoshop CS for Photographers". There is excellent
    tutorials on how to balance the colour in your images.
     
    McLeod, Oct 4, 2005
    #12
  13. Steve Franklin

    nv Guest

    In
    Yes, it is. I have just obtained a copy of UFRaw for my Windows XP based
    setup to try out dcraw's new incorporation of the AHD de-noising algorithm.
    After a very quick tinker about with it using my Kodak SLR/n RAW files, I am
    seriously wondering what all the fuss has been about. It's supposed to
    deliver unparalled sharpness combined with much superior de-noising
    characteristics, but no matter whether I use the AHD, VNG or Bi-lateral
    settings, I can't see any difference between the images even at 300%
    enlargement. So far SilkyPix knocks spots off my proprietary software (Kodak
    Photdesk) and Adobe ACR for eliminating colour aberrations, noise and
    producing the sharpest image overall. I even did a quick comparison test for
    a thread in DPReview.com. You can check the images out here:
    http://forums.dpreview.com/forums/read.asp?forum=1026&message=14913976

    Wow. Talk about a sledgehammer to crack a nut approach. ;-)
    You Linux types will take any opp to ram the message home. :p

    Regards,
    Nigel
     
    nv, Oct 4, 2005
    #13
  14. Steve Franklin

    no_name Guest

    Essentially RAW has all the data from the sensor. The JPEG does not.

    Most Photoshop Pros will used calibrated monitors & printer profiles for
    each kind of paper they print on (turning off any built in color
    management from the printer).

    And as for "how does someone else know what the original colors were
    like?"

    You tell them ...

    .... just like you used to tell the guy at the mini-lab when you looked
    through your prints & found one where the color needed correction.

    You DID look through your prints when you picked them up & made sure the
    colors were correct didn't you?
     
    no_name, Oct 5, 2005
    #14
  15. Steve Franklin

    no_name Guest

    Custom White Balance.

    Profiles.
     
    no_name, Oct 5, 2005
    #15
  16. Steve Franklin

    Eugene Guest

    I also was a bit surprised when I first started processing my RAW 20D
    images in CS2. The default settings made them way over-saturated. I
    always found with my G3 images that the RAW files by default would come
    out looking much the same as the jpegs.

    Does your camera support simultaneous JPEG+RAW capture? That's what I
    mostly do now. I set it to capture a low-res low quality (still pretty
    high really) jpeg as well as RAW. This lets me do a really quick and
    easy preview when I get back to the computer before going through the
    process of converting the RAW file.
     
    Eugene, Oct 5, 2005
    #16
  17. That's good! Thanks for pointing that out.
    I am not aware of that, and will go looking for it to see what it does!

    Again... thanks! Good info...
    Really a good demonstration.
    Some *skulls* are so thick that it takes a sledgehammer.

    But, that was not a sledgehammer approach at all. It is sound
    reasoning that makes sense for a rather large number of reasons.
    It is economical and practical, however, primarily because of
    many things unrelated to this particular thread, and therefore I
    did not detail them.
    I pointed out a simple fact, and did not feel it necessary to
    provide *any* detail. Clearly that is *not* trying to "ram the
    message home" in any way shape or form.

    I am much more interested in technical detail than I am in any
    form of advocacy. Anyone who does any serious research is going
    to find plenty of both, so there was no need for it in that
    message. By the same token, if you want to skip the
    advocacy/marketing hype part of it, and discuss technical
    details about why it would make sense, then I will be very happy
    to provide as much detail as I can about any specific questions.
     
    Floyd Davidson, Oct 5, 2005
    #17
  18. Steve Franklin

    nv Guest

    In
    I should make a temporary retraction of my criticism of UFRaw now, as I
    rather ashamedly realised too late before posting, that I had attempted to
    use a down-sized RAW image for evaluation, instead of the full-sized
    version. If you are puzzled by what I mean as to 'down-sized', then the
    Kodak DSLR can produce a smaller RAW image file In-Camera, rather in the
    same way as most digitals can produce JPEG's of varying compression and
    resolution. However, the downside is that most 3rd-party RAW software which
    can read the full-sized Kodak RAW files, cannot do so with the down-sized
    ones. I am actaully amazed that UFRaw was able to read it AT ALL! I shall
    now try and find some full-sized RAW's to evaluate UFRaw with again, and
    post back. But my 2 month-old camera is away being fixed atm, and won't be
    back for another 2 weeks, I have heard this morning! That will make it 6
    weeks away in total. :-((
    But that is another story........

    NP, but Udi Fuch's website (http://ufraw.sourceforge.net/) is a bit
    confusing (at least, to me ;-) ). He states that the current release of
    UFRaw is 0.5, and includes the bullet point in his News item dated 25/09/05
    that "Notice that the new Adaptive Homogeneity-Directed interpolation is
    still not enabled". However, if you download the execuatble, it is version
    0.6 dated 29/09.05, and *does* include the AHD option in the software.
    However, I am not sure if it includes the latest version of dcraw.c v7.70,
    which was optimised to remove 'edge-sensing'. See the repro'd announcement
    here:
    https://sourceforge.net/forum/forum.php?thread_id=1357438&forum_id=434060
    Again, NP.

    You are also excitable! ;-)

    Kind regards,
    Nigel
     
    nv, Oct 5, 2005
    #18
  19. Agreed on the site being confusing (in several ways).

    Based on your comments about newer versions, I downloaded the
    current source... or what I thought was current until I see
    your comments here! It is version 0.5, and I found the
    interface has been significantly changed from the version 0.4
    that I was describing previously. I'm not yet sure that I like
    the difference, though my first impression is that I may want to
    keep both versions available.

    The adjustments of just about everything are different. The
    color temperature adjustment is doing something significantly
    different in 0.5 than it did in 0.4, and the range of adjustment
    for the green level is distinctly reduced. Obviously this is
    a work in progress!

    Interestingly enough one of the things I found myself doing
    previously was making to a lot of effort learn the slider
    adjustments for gamma and black level the way it was done in
    version 0.4, but I often just switched back and forth with XV,
    which uses a live graph. I found the two distinctly different
    interfaces useful in different ways, but now that I'm used to
    having two different ways to do that... the new version 0.5 has
    a live graph interface somewhat similar to XV.

    But, if there is a version 0.6, I need to track that down and take
    a look. Thanks.
    *You* got excited. But I'm going to stay with discussion of
    technical detail...
     
    Floyd Davidson, Oct 5, 2005
    #19

  20. If they have the ICC profile of your camera it is not a problem.

    When the camera changes it to another format it adds the color profile
    you had chosen, if you have one at all.




    *
    Photographs by Christian Bonanno
    http://christianbonanno.com/
     
    Christian Bonanno, Oct 5, 2005
    #20
    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.