Help me choose my first dslr

Discussion in 'Digital SLR' started by Padu, Aug 14, 2006.

  1. I feel some of them are quite lovely and near perfect.
    You only get dark shadows in dark places or with high ISO settings?
    I though you get the darkest shadows in very high contrast
    situations, like:
    - shooting the full moon at night
    - shooting under the blazing sun, when some things are in the
    shade (which is why people who know what they do so often use
    fill flash in these situations)
    - shooting with strong light sources in the dark (bonfire in
    the night, street at night with street lanterns, ...) when
    exposing for the highlights.
    In the darker parts, sure, you care about shadow.
    There is no 'bad' light. There's only a technical problem of
    creating the right light (like moving the sun behind a cloud, using
    reflectors, flash units, the right filters, etc ... depending on
    the circumstances. OK, moving the sun or the cloud may be a
    bit hard for a mere mortal.)

    I do deem a single candle enough for some types of face-only
    portrait; and hey, this is rec.photo.digital.slr-systems. See the
    "digital"? You can delete the bad picture, and learn something
    about your and your cameras' limits in the process.
    You're completely wrong!
    The rule is: Don't show your bad pictures, and it doesn't matter
    if you shot them at ISO 100 with perfect light or not.
    So? I haven't seen you ever, so may I conclude you don't exist?
    I see ...

    -Wolfgang
     
    Wolfgang Weisselberg, Aug 24, 2006
    #41
    1. Advertisements

  2. What sort of 'concert musician pictures' can't you stand, and why?

    Did you ever try shooting animal life, especially nocturnal beasts?
    -Wolfgang
     
    Wolfgang Weisselberg, Aug 24, 2006
    #42
    1. Advertisements

  3. Basically, your sensor counts photons in each sensor cell by
    Some of it, maybe, and your description of the process seems quite
    informative, but I'm not convinced that things are exactly as you are
    representing them. Since, as you say, the ISO setting done *before* the
    A/D conversion, it isn't really a "multiplier" in the literal sense. A
    literal "multiplier" would mean we already had a discrete count of
    electrons that is multiplied by a discrete numeric value, which would
    yield a completely predictable result, and the final image would indeed
    have no no more noise than the original (although the original noise
    would be more prominent, of course). On the other hand, since the
    amplification is done through analog means, it sure seems to me that
    this in itself would be likely to introduce some noise of its own. That
    is, two cells that collected the exact same "number of photons" (which
    is to say, I believe, "accumulated an electrical charge that to the best
    of our measuring abilities is the same") might well, after
    amplification, end up with measurably *different* electrical charges.
    Meaning even a perfect A/D converter would calculate different values
    for these two cells. No?

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 24, 2006
    #43
  4. Padu

    Craig M Guest

    Whew, that was a post, did you swallow a camera manual :)
    no there is several, I was thinking of one zoom, the AF-S DX Zoom-Nikkor
    55-200mm zoom, or maybe the AF Zoom Nikkor 28-200mm, will have to see if the
    last one will work on the D-50 since it does not have the DX on it.

     
    Craig M, Aug 24, 2006
    #44
  5. Padu

    cjcampbell Guest

    Almost any kind. Such pictures look too static to me. A picture of an
    orchestra is usually uninteresting, unless you get close enough to
    emphasize a single individual or very small group.

    Anything outside of classical music simply does not interest me and I
    find some of the performers revolting in their appearance.

    I have a few pictures of choral groups taken in a darkened auditorium
    and I never had to go above ISO 400 to get them.
    Actually, yes. But I still have never had to go above ISO 400. The big
    problem I have always had with animal life is getting close enough. I
    am terrible at stalking the critters.

    I figure that if I ever get around to photographing the bats around
    here I will have to use flash anyway, and probably a wireless shutter
    trip, too.
    Well, I have had plenty of black eyes in my day, but they have never
    made me any smarter or better informed.
     
    cjcampbell, Aug 25, 2006
    #45
  6. Almost any kind. Such pictures look too static to me. A picture of an
    And, assuming that getting close enough means using a telephote lens as
    opposed to physically getting to the front row or being on stage, this
    is precisely the sort of thing that can require higher ISO.
    OK, that's your right, but since apparently it *would* be interesting if
    it were classical music, this suggests even you would agree that such
    pictures *can* indeed be worth taking.

    Here's an ISO 1600 picture you might not find too revolting - a
    classical violinst named Chris Luther:

    http://www.outsideshore.com/marc/photo/music/images/image/MJS_060428_0859.JPG

    This was in a fairly well-lit recital hall, and it still took ISO 1600
    shooting wide open for the lens I was using to get me a shutter speed of
    1 / focal-length, the usual standard for handheld shots (tripods are not
    normally an option in this type of situation). Given the unusually good
    lighting, I could perhaps have gotten away with ISO 800 had I had either
    a faster lens or image stabilization. However, had the lights been more
    typical concert lighting, then I'd have been struggling at 1600 even
    with a faster lens and stabilization, and indeed, even with a tripod. I
    don't care how steady that camera sensor is, a shutter speed of below
    1/60 is likely to produce subjct motion blur when dealing with
    performing musicians. But I'd claim the picture would still have been
    worth taking, even though there would have been little detail in the
    black shirt, and the background would have gone darker. I have many
    examples of such pictures, including ones that are far less "static"
    than this one, but they aren't classical music, so I wouldn't want to
    revolt you by posting links. But for many of us, it is still worth it
    for the detail that *is* present in the lit areas.

    I find it ironic, BTW, that photographers in general seem so concerned
    with shadow detail. One of the first things you learn in painting is to
    *avoid* placing too much detail in the shadows, as that kills the sense
    of light. Photographers go to great lengths with fill lights and
    reflectors and so forth to get a relatively evenly lit portrait, yet
    paintings will typically go to equally great lengths to create and
    exaggerate a more dramatic "chiaroscuro" type of lighting.

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 25, 2006
    #46
  7. Padu

    cjcampbell Guest

    Thanks. :)
    I have to agree with you there. This picture has very little in the way
    of shadow detail. It was taken at ISO 100, I believe.

    http://www.treklens.com/gallery/Asia/Philippines/photo175007.htm
     
    cjcampbell, Aug 25, 2006
    #47
  8. Padu

    Craig M Guest

    I think the lighting you refer to is because painting is a rendering of what
    the artist sees, and a photograph is what the camera sees, and cameras
    reproduce what they see with no bias at all, artist with a brush can bias
    the painting, the camera can show detail in shadow, the brush, well, its in
    the hands of the artist.
     
    Craig M, Aug 25, 2006
    #48
  9. I find it ironic, BTW, that photographers in general seem so
    Great shot. But according to the EXIF data, it was ISO 400. I checked,
    because I thought it looked far too noisy/grainy to have been 100. I'm
    still surprised it was only 400- was that the result of underexposing
    the picture and then pushing it in photoshop? Or did you add noise in
    postprocessing to simulate film?

    Anyhow, if there *was* enough light to get a decent exposure at f/5.0 at
    1/60 with ISO 400 as the EXIF data reports, then the light in this scene
    was far better than most concert or night scene photography, and yet the
    "chiaroscuro" effect is about the same. Meaning, if this shot works,
    then so potentially do any number of concert/night shots.

    Indeed, many more "vanilla" indoor shots are not this well lit.
    Consider, for example, this:

    http://www.outsideshore.com/marc/photo/people/images/image/MJS_051218_0542.JPG

    Although the EXIF data is incomplete because I was using an old manual
    lens, I know it would have been shot at f/2.8 or wider, or I wouldn't
    have been using such a lens (and actually, based on the appearance of
    the photo and my recollection of the event, I'm thinking it was a
    135/2.8 wide open). Regardless of what anyone else might think about
    the aesthetic quality of this particular photo, the moment captured has
    significant value to the people involved (it was a surprise retirement
    party that ended up also being a family reunion), and they'd certainly
    agree this picture was worth having taken. Yet even shooting at f/2.8
    at ISO 1600, that was still only enough light for a 1/60 second
    exposure. It's not an especially oddly lit scene - indeed, it's pretty
    uniform lighting that still shows reasonable shadow detail, so shots
    taken in this environment have a chance of coming out technically
    acceptable, even by the odd (to me) shadow-detail-loving standards of
    professional photographers. And yet, without ISO 1600, shots like this
    would not have happened at that event.

    Another point worth mentioning is that your shot was taken at 18mm.
    Obviously, you were standing pretty close to the woman depicted. That's
    not always such a great idea in taking candid shots (such as mine
    above). Candids make another broad category of pictures in which, due
    to the desire to avoid subject movement and the relatively long focal
    lengths that are typically involved, ISO 1600 is extremely useful even
    in sitatuons where the light isn't *that* bad.

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 25, 2006
    #49
  10. http://en.wikipedia.org/wiki/Analog_multiplier
    http://en.wikipedia.org/wiki/Analog_computer

    I content that it is a multiplier.
    You mean a *digital* multiplier. It isn't, that's right.
    Nope: the discrete count of electrons is not completely predictable.

    And predictable != correct. (A soccer player that
    consistently misses the goal, a computer that consistenty
    crashes after 20 minutes ... they are very predictable.)

    Also, digital computers struggle with quantization.
    Again, wrong. You'd have (at best) the same signal-to-noise ratio.
    However, both the noise and the signal have been amplified.
    Of course it does!

    But it still gives you less noise that way, assuming correct
    exposure. Try it out: shoot the same scene in RAW at ISO 1600, lock
    aperture and exposure time and switch to ISO 100, then
    multiply the values in your RAW-converter.
    Ah, but they won't collect the same number of photons. There's
    noise even in the *nature* of photons being sent out or reflected.
    While they even out over a large number of photons, we often deal
    with small numbers here.

    Additionally, 2 cells collecting the same number of photons, if
    they could, would not necessarily have the same charge collected:
    the sensor is not noise free either.
    Obviously.
    But will it have any impact on the picture?

    That depends: is it in the bright or dark parts of the image,
    does the A/D-converter produce tons of linear steps (to be reduced
    later into 256 gamma-based steps) or only a few hundred, where
    they may even have to be expanded in the shadows, do other noise
    sources hide the additional noise, ...

    -Wolfgang
     
    Wolfgang Weisselberg, Aug 25, 2006
    #50
  11. Hmmm.
    What would such a picture need to be non-static?
    True. So go for faces or half-body 'portraits'.
    Ah, yes, some musicians prefer to stun visually rather then
    aurally ...
    f/1.2 at 1/30s? :)

    Try catching someone's hand playing an instrument, say a
    concert harp, a percussion instrument or a lute.

    OK; I do have some pics with multi-second exposures of that
    kind of thing, it just isn't easy to get a keeper.
    You only shot whenever they crossed a sunbathed clearing? ;-)
    A flash can help, but, ah, you'll need some idea of what to do
    with the background. I, at least, don't like it all-black.

    But then I will shoot with flash, f/1.4, high ISO and multi-second
    exposures --- free hand. And get good results, if I do things
    right.
    Nobody said it did. But maybe that's the reason for the
    'plenty'.

    -Wolfgang
     
    Wolfgang Weisselberg, Aug 25, 2006
    #51
  12. Ah, but that's a different thing.
    You don't *place* detail in the shadow, you _keep_ detail in the
    shadows. Also, the painter will leave the brush markings even
    in the dark parts, the texture of whatever he paints on will be
    there, etc.

    With photography, you have the choice between "Black as the inside
    of a bear" or "dark, but textured (unless whatever you shoot _has_
    no texture to resolve)".

    The eye is very adapt even over huge contrast ranges; the painter
    can simply paint with 4 or 5 stops of luminosity, no matter if
    the real thing has only 2 or 12+ stops. He can create local
    contrasts to increase the visual perception of dark and bright.
    The photographer must either play with graduated filters ---
    not always possible with irregular shapes! --- or use reflectors,
    flash, or HDI, etc. to reduce the absolute differences.

    A photographer can later stretch or compress the contrast ---
    within limits --- by pushing or pulling, burning and dodging,
    or using differently graduated paper. But he's limited to the
    few stops the photographic paper can record and display, and to
    how many stops his film/sensor can record with usable accuracy
    and noise behaviour. Actually, doing this right is part of the
    Zone System.

    http://images.google.de/images?q=Ansel+Adams
    should give you an idea that strong contrasts are often wanted,
    but areas without texture (and that means detail, specifically
    shadow detail) are _not_.

    Both have their place, and there are quite a few very high
    contrast low key portraits.

    -Wolfgang
     
    Wolfgang Weisselberg, Aug 25, 2006
    #52
  13. OK, I will grant that the term *is* used in this manner. But then, we
    must realize that even though it can legitimately be called a
    "multiplier", it doesn't necessarily *act* like what one would normally
    otherwise consider a mutliplier (ie, a digital multiplier). In
    particular, one cannot assume it will not introduce noise of its own.
    I never said it was. I was talking about the multiplication of two
    known discrete values - that is, *after* the unpredictable A/D
    conversion.
    True. But integer multplication is so trivially simple to implement, it
    seems not owrth making this distinction. Multiplication of two integer
    quantities *will* yield correct answers.
    And hence, more prominent in the sense I meant. Not necessarily
    "compared to the amplified signal", but "compared to the unamplified
    noise".
    Right, and it sure seemed to me that the point of your previous post was
    to *deny* this. If not, then I think your actual point was lost...
    True enough. Noise comes form many places.
    I don't know, to be honest. It seems *all* of these things are likely
    to contribute to the visible noise.

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 25, 2006
    #53
  14. Last I checked, the canvas started out blank white, and a piece of film
    or a digital sensor starts out with no image represented. Any detail
    recorded in the final image was most definitely placed there.
    He *can*, but the usual training (which I think is perfectly appopriate)
    tells you not do do so, or you will lose the proper sense of light.

    Anyhow, my point isn't so much about the process but the final
    appearance. Painters tend to value images in which there is a strong
    sense of light and shadow that is supported by the presence of detail in
    one area and its relative absence in the other. Photographers tend to
    want detail everywhere.
    True enough. And correspondingly, there are quite a few low contrast
    middle-value paintings. But on the whole, it seems these are *not* the
    norm of what is usually considered desirable in the respective genres.
    I just find that interesting.

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 25, 2006
    #54
  15. Padu

    AaronW Guest

    What is HDI?

    http://digitcamera.tripod.com/#slr
     
    AaronW, Aug 25, 2006
    #55
  16. OK, I will grant that the term *is* used in this manner. But then, we
    must realize that even though it can legitimately be called a
    "multiplier", it doesn't necessarily *act* like what one would normally
    otherwise consider a mutliplier (ie, a digital multiplier). In
    particular, one cannot assume it will not introduce noise of its own.
    I never said it was. I was talking about the multiplication of two
    known discrete values - that is, *after* the unpredictable A/D
    conversion.
    True. But integer multplication is so trivially simple to implement, it
    seems not owrth making this distinction. Multiplication of two integer
    quantities *will* yield correct answers.
    And hence, more prominent in the sense I meant. Not necessarily
    "compared to the amplified signal", but "compared to the unamplified
    noise".
    Right, and it sure seemed to me that the point of your previous post was
    to *deny* this. If not, then I think your actual point was lost.
    True enough. Noise comes form many places.
    I don't know, to be honest. It seems *all* of these things are likely
    to contribute to the visible noise.

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 25, 2006
    #56
  17. Padu

    cjcampbell Guest

    Neither. The photo is a small crop badly oversharpened to compensate. I
    will have to be more careful with that. Anyway, so much for relying on
    an aging memory for the data on the film.
    Well, and there you have it. Not every photo is necessarily a work of
    fine art, nor should it be. Such a memory is, in the long run, worth
    many artistic awards.

    However much we may disagree on whether ISO 1600 was really needed, the
    photograph was needed.
    You are right. But the picture is not a candid and was not intended to
    be such a close-up portrait. The frame actually shows the entire person
    and her wares, but in th end I thought this crop was more interesting.
    That is the reason for the low technical quality. If I had it do over
    again I would have used a longer lens (I was using the 18-200mm VR,
    after all) and done the crop in-camera. It would have resulted in a
    technically much better picture. The thing that irritates me is that I
    did not see this picture when I was there.
     
    cjcampbell, Aug 26, 2006
    #57
  18. Ah. The small crop would explain why it also didn't really look like
    the sort of perspective I'd expect with a wide angle lens from that
    apparent distance. Had I wanted to take the same shot - from the same
    vantage point to get the same perspective, and framed the way you have
    it - I'd have used a telephoto lens. And then I may well have needed
    higher ISO to get a fast enough shutter speed to overcome camera shake.
    Maybe so, but it's not a bad look, in my opinion. But then, I'm
    accusomed to shooting at ISO 1600...

    ---------------
    Marc Sabatella


    Music, art, & educational materials
    Featuring "A Jazz Improvisation Primer"
    http://www.outsideshore.com/
     
    Marc Sabatella, Aug 26, 2006
    #58
  19. It does act like one within it's limits and disregarding a
    digital multiplier's limits.
    A digital multiplier will also introduce noise: it has only so
    many bits. Any input number not representable by these must be
    rounded; any output number not representable by these will be wrong
    (and if we are lucky, we do get an error or overflow flag set).

    Once you turn to discrete (and _wrong_) values (they have to be
    rounded ...), you should get predictable outcomes.

    However, predictable does not indicate correct.
    The conversion is predictable, as the noise itself ist
    statistically predictable. You simply will have to allow for
    a known margin of error due to noise.
    Please calculate 4000 x 5 with 12 bits.
    4000 is 1111 1010 0000
    5 is 101
    The result is 1110 0010 0000
    which evaluates to 3616.

    Oops. Wrong answer: Overflow.

    Worse, we are not necessarily multiplying be integers, nor
    necessarily by numbers >= 1.
    And in what sense would that measurement be of any value?
    Driving the A/D converter to use it's maximal range yields
    less noise than not doing so and relying on 'digital'
    multiplication later on.
    But which one is the main factor/are the main factors?

    Go ahead, shoot a scene at ISO 1600 and with the same aperture
    and exposure at ISO 100 and digitally push the image later.

    -Wolfgang
     
    Wolfgang Weisselberg, Aug 28, 2006
    #59
  20. So you have to _place_ detail in the shadow, even put in the
    shadow.
    Yes. So what do you do, go and paint the shadow detail _in
    the scene_ away? Or do that in post-processing?
    Any detail recorded there was in the three dimensional scene
    that was shot. Not always can you control the scene.
    Hmmm, so what is the difference between the lightest and the
    darkest part of the average painter's work?

    You can surely raise the black point (and get everything look
    like greyish haze) or lower the white points (and get grey
    instead of white). I ... doubt ... that that will give you
    the best of images most of the time.
    Not really. Even photographers can place things in Zone I or
    0 or IX or X. But they prefer to be able to _choose_ to do so.

    -Wolfgang
     
    Wolfgang Weisselberg, Aug 28, 2006
    #60
    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.