the High, low light ISO Myth and EOS DSLRs

Discussion in '35mm Cameras' started by anonomous individual, Dec 18, 2005.

  1. Some time back I made the outrageous claim that a Panasonic FZ20
    'Digicam' took better pictures in low light than a Canon 20D. At the
    time I didn't have any interest in documenting the claim. I just
    accepted that there were unshakable believers in the God of EOS who cry
    foul whenever anyone questions the ability of a Canon product, and went
    on with my life.

    Today, I have a 5D to compare the FZ against. I really like this 5D but
    I gotta tell you, I'm having some issues trying to get the best from it.
    No doubt I will overcome them but along the way, I'm re-discovering what
    I found with the 20D in low light is pretty much the same for the 5D.

    http://www.photosbydouglas.com/ISO-myth.htm

    Please yourself if you choose to believe or not believe my sincerity.
    Choose yourself too, if you see this as a beat up on Canon or a genuine
    attempt at enlightenment. In the event you disagree, see if you can find
    someone who cares, eh?
     
    anonomous individual, Dec 18, 2005
    #1
    1. Advertisements

  2. anonomous individual

    Scott W Guest


    So it looks like the 20D got 1/40 of a second and the Panasonic got 1/4
    second?

    I am not sure what you are trying to show here.

    Scot
     
    Scott W, Dec 18, 2005
    #2
    1. Advertisements

  3. anonomous individual

    Paul Furman Guest

    The noise looks worse in the Panasonic (ugly purple & yellow blobs in
    the shadows) even after sharpening the Canon crop.
     
    Paul Furman, Dec 18, 2005
    #3
  4. That's OK Scot, I'm not sure what your comment is about either.
    For my interest, the use of ISO 1600 when the FZ does basically the same
    thing at ISO 200 is the highlight the whole thing. We all know SLRs
    vibrate when the mirror swings away but did we know how much exposure
    value we lose in making allowances for this?

    Perhaps a point worth considering is with the 2 noise example images.
    The Canon image is from 12 megapixel, cropped at 650 pixels. The
    Panasonic image is from 6 megapixels cropped at 640 pixels. Basically
    the Canon image "looks" better because it originated from a camera with
    twice the resolution of the FZ20.

    Give me an opinion Scot. Would you think a DSLR which shudders it's
    mirror so hard as to need this much increase in EV values to get a sharp
    picture, is good or bad? Forget the brand bashing stuff. look
    objectively at it.

    All this started when I priest told me in direct terms, in the middle of
    a ceremony to "stop all that noise and flashing" or get out of his
    church when I followed the couple's instructions to take pictures of the
    ceremony as it progressed.

    After that I bought one of these FZ20s because I can switch off all
    sound and hand hold it at 1/15th, sometimes 1/10th and still get sharp,
    clear images. I don't have a problem about the mirror slap of any of my
    DSLRs. I'm just curious why, with all the advances in camera design and
    electronics, the bloody mirror slap still stuffs up natural light
    shooting when it should have been overcome by now.
     
    anonomous individual, Dec 18, 2005
    #4
  5. anonomous individual

    Scott W Guest

    Ok I understand now. How much the image stabilization helps is
    normally given as 2 to three 3 stops, so in one way ISO 200 with out IS
    would be somewhat like ISO 800 - 1600 without. But shooting at 1/4
    of a second has its own problems in terms of people moving, I might get
    the background sharp but have blurred people. I run into this all the
    time when I am using a tripod, the background looks great, people not
    so much.

    Then again if needed you can buy IS lenses for the DSLR, being cheap by
    nature I have not purchased any of these yet.

    There is a mode on the 20D that allows for the mirror to be swung out
    of the way with the first shutter push and then the photo taken with
    the second. The viewfinder goes black after the first push but you can
    do the second push just a bit after and so this is not a great problem.
    I get just a bit sharper photos this way then with out but it is not a
    huge difference. There is another mode for tripod mounting where the
    mirror swing out at the shutter push and three seconds later the photo
    is taken, my wife hates it when I leave the camera in this mode. This
    get ride of both the shutter slap and the pushing of the shutter.

    For me the 20D works well, with the 50mm f/1.8 and running at ISO 1600
    I can take photos inside that I would have a hard time getting
    otherwise. I am normally trying to keep the shutter speed at 1/60 or
    faster to try and avoid motion blur from people moving. I need to get
    at least one more prime for inside work, I am looking at the 28mm
    f/1.8, the 50mm is not always wide angle enough.

    As for the noise of the shutter I can see where this could be a
    problem. In the shooting I have done it has not been an issue to date.

    The one area where the higher ISO does not live up to every thing it
    could be is where you need a given DOF. In this case the smaller
    sensors can work with much lower f/numbers. I can shoot with an f
    number that is 2.5 time bigger on my F828 then I can on the 20D, and
    still get the same DOF. This give me 6.25 times more light for the same
    DOF as what I get with the 20D. So now my ISO 64 is like ISO 400 on
    the 20D. The 20D still comes out ahead but not by as much. This only
    is the case where I need large DOF, otherwise the 20D blows the F828
    out of the water. And if I need a shot with a really large dynamic
    range I need to use the 20D.

    I don't think the small point and shoot cameras are as bad as many
    people try to make them out to be but a DSLR will win out just about
    every time.

    Scott
     
    Scott W, Dec 18, 2005
    #5
  6. anonomous individual

    JimKramer Guest

    And this is what makes you a professional wedding photographer?
    Nothing like asking the pastor about house rules before the ceremony.
    You, yet again, amaze and astound.
     
    JimKramer, Dec 18, 2005
    #6
  7. anonomous individual

    JimKramer Guest

    Well Douglas you've finally inspired me.
    Thank you.

    A New Treatise on the Scientific Method: Modeled on the Illustrious
    Methods of D. J. MacDonald

    I've decided to compare two pies; one is apple, one is cherry. Both
    will be made from scratch and baked by myself. I noted in baking the
    pies, that the apple pie had a decidedly apple smell when cooking that
    the cherry pie lacked. A liberal amount of ice cream was served with
    the pie to mask any imperfections in the cooking process. After
    consuming the pies I noted that each made a satisfactory amount of
    crap.

    Next investigation: Cream Pies and the Clown: To Throw or to Plant?
     
    JimKramer, Dec 18, 2005
    #7
  8. anonomous individual

    no_name Guest

    I don't see where you lose any exposure value.

    Changing ISO changes the camera's processing algorithm. For a given
    amount of light the camera recommends f/4 @ 1/125 at ISO 200. Changing
    the ISO to 400, the camera changes it's processing to f/5.6 @ 1/125. Or
    you can stay at f/4 and go to 1/250.

    Changing the ISO just makes the camera more responsive to what you want
    to do. If you need a higher shutter speed under lower light conditions,
    changing to a higher ISO makes the camera do what you want it to do with
    less quibbling.
     
    no_name, Dec 18, 2005
    #8
  9. anonomous individual

    Scott W Guest

    Changing the ISO changes the analog gain before the signal gets to the
    A/D.

    Scott
     
    Scott W, Dec 18, 2005
    #9
  10. anonomous individual

    Alan Browne Guest

    no_name wrote:

    Algorithm's don't change with values. Gain values (inputs to the
    algorithm) do.

    ISO settings actually change the analog gain prior to the A/D converter.
    Above 800 or 1600 or so, then numeric gain is also applied after the
    A/D conversion. Varies by sensor and camera implementation.

    Cheers,
    Alan
     
    Alan Browne, Dec 18, 2005
    #10
  11. Well I sure as hell know where to find the clown Jim. What are your rates?
     
    anonomous individual, Dec 18, 2005
    #11
  12. Douglas, I really don't see how what you are saying proves anything. I'm
    not a believer in the God of EOS, but your comparison is done in such a
    different manner for each camera that nothing is proven whatsoever. To
    my eye, the Panasonic has big chunky hunks of noise, the canon has
    smaller more speckled noise - so what. Each image would be usable,
    although I'm actually leaning toward the Canon in that example. I
    _think_ what you are trying to say is that the Pana can do a handheld
    low-light shot by virtue of IS, whereas to get a handheld low-light shot
    on the Canon requires upping the ISO. Yeah, that's two different ways to
    get a low-light shot. Another way is to buy an IS lens for your Canon,
    or you could use a tripod, or you could use a faster lens on the Canon.

    You said that the Canon glass required you to use the 50 lens stopped
    down to F2.8 to get the same clarity as the Pana - could it be that what
    you are really observing is the fact that the canon has twice the res?
    Downsampling the canon to the same res as the Pana, would likely remove
    any apparent loss of sharpness. It would also remove a lot of the noise.
    Your Canon shot is showing roughly double the field of view of the
    panasonic, and in the crop it is still showing greater apparent
    sharpness. Take the image with the same field of view, downsample the
    Canon to be a match for a Pana (or upsample the Pana image, either one),
    and see which one has less noise and sharpness - I think the results
    will be very apparent in favour of the Canon. Redo this with any of the
    current crop of DSLR's (even the E300 which as you say has a reputation
    for noise), and I'm sure the SLR will win every time.

    I think I know what you are trying to say, I just don't think the
    testing methodologies are fair. I'd suggest doing it again with an IS
    lens on the Canon, and with the same field of view in both shots - that
    way you are comparing on a closer playing field. Actually while you are
    at it, why not do the test with some of your other cameras. I would love
    to see a fair and objective comparison of noise, lens sharpness etc from
    different models. I'm sure you understand also that IS is not a
    cure-all for low-light. If the subjects are moving, IS is pointless, and
    you need fast lenses and/or high ISO. If you _need_ 1/40 to get
    something close to stopped motion, can the Panasonic do it? What does
    the noise in the image look like then?

    Your point about the mirror slap and noise of the SLR is a fair one.
    This has been a criticism of SLRs since they were first made. I recently
    bought a 40yo 2nd hand rangefinder, and I love the low noise compared to
    my SLRs. The rangefinder has it's own unique benefits over the SLR, but
    at the same time it has it's drawbacks too. It's all a matter of the
    right camera for the job. Heck sometimes the camera in my mobile with
    it's huge 352x288 resolution (50% of which is noise, even in sunlight)
    is the right camera for the job.
     
    Graham Fountain, Dec 18, 2005
    #12
  13. anonomous individual

    JimKramer Guest

    I think most of the readers here know who the clown is. What I can not
    understand is why you insist on repeatedly demonstrating it?
     
    JimKramer, Dec 18, 2005
    #13
  14. Thank you for your input Graham.
    The test missed the mark of what I am trying to convey, for sure. My
    intention was to show that High ISO in itself, is not a requirement for
    unaided, low light photography anymore than the cost of equipment is.

    I tried to show that manufacturers marching to a different beat can
    achieve basically the same thing in different ways. At the same time I
    hoped to get across that the difference in image quality between a $800
    outfit and a $6500 one is not significant enough to warrant that kind of
    spend just for the High ISO performance in itself. Neither test would
    work on a moving subject but the FZ makes a fine indoor camera for some
    surprisingly critical work. I recently shot an entire catalogue of
    jewelery with it and never one used a flash.

    I can see now that there are those of us who see this as a lop sided
    test and those who see it as something they can get a handle on natural,
    low light photography with. Your suggestion of using an IS lens on the
    Canon carries with it the premise that cost is not a consideration in
    obtaining marginally better results.

    The simple fact that not a lot of people can afford a 5D, much less an
    IS lens of similar quality as the Leica on the FZ, should not be
    overlooked. The FZ20 cost me $825 AUD, complete with lens, spare
    battery, flash card and inbuilt flash.

    The 5D cost $5000 AUD with no flash card and one battery. The latest
    flavour of IS "L" series lens is another $1800. Somewhere in this you
    need an extra $1000 AUD for a grip and speedlite. Basically it's a David
    and Goliath test with David holding his own pretty well.

    If I didn't gain value from the Canon's other benefits - namely wide
    (currently the widest) dynamic range and it's frame buffer, to name just
    2 of it's many attributes, I probably might consider buying a new car
    and making do with a FZ20. As it is, I think the 5D is the hight of
    DSLRs development right now. If they get their quality control right,
    they will indeed lead the world in DSLRs.

    Mine has a back focus problem. It focuses 3" past the point of focus at
    3 feet and it constantly underexposes along with the auto white balance
    not working properly. Compared to the problems I had with 20Ds, these
    are trivial and I'm sure Canon will attend to them in their usual
    secretive but complete manner just as soon as the holidays are over.
     
    anonomous individual, Dec 18, 2005
    #14
  15. OK, I missed that point. Now that you've made it I somewhat agree. The
    canon is clearly better, but does that improvement justify 8x the
    expenditure? For some people it definitely would, for others no way in
    hell. I don't believe the Pana would be quite as versatile as the Canon,
    so for some people the Pana would suffice, for others it wouldn't.
    I don't doubt it is capable of good work, provided that the work you
    give it is within the scope of the camera's abilities (and the person
    behind the camera). But then even the $99 Kodak's that hardly normal are
    selling are also capable of good work - provided what you throw at it is
    within it's limited capabilities. No one camera is going to be ideal in
    all situations, and as a general rule, the more money you throw into the
    camera the more situations that it will be capable of handling. I guess
    for a buyer, it is a tradeoff between what you can afford, and the
    capabilities you need. For most mum & dad shooters, even if they could
    afford a 5D, it would be wasted because a $200 P&S will better handle
    their needs due to it's pocket size.
    I would have thought that comparing a $6500 camera with an $800 camera
    was already lopsided and demonstrated money is no object. Perhaps you
    could also go as far as saying out of the box the 5D is useless compared
    to a Pana, because out of the box all the 5D can do is wierd masses of
    diffused colour. Only when you pay more and buy a lens can it actually
    take a photograph. :p
    I've almost only heard good about the Leica/Panasonic cameras. I had a
    bit of a play with the old FZ10 and they were a beautiful little camera
    (although I found the user interface a bit awkward - perhaps if I read
    the manual). I have also held (never used) a new one - not sure of the
    model. My first impression just from holding it though was plastic heap
    of junk. I'm sure it'd take nice photos, but it seems they skimped on
    the construction. I have heard 1 person give a negative on the
    Leica/Panas, but it came from a devoted EOS worshipper. He claimed (not
    sure of the truthfulness), that the Leica lens wasn't made by Leica at
    all, but by a chinese factory who licensed the design. Either way, proof
    is in the pudding, and they are a significantly better lens than most of
    the stuff that finds itself with the C word plastered on it.
    Agreed. In this case Goliath won the battle, but David put up a damned
    good fight.
    I find it curious that Canon have indeed leapt out in front with their
    EOS digital range. In film it is a much fairer battle, with the Nikon
    and Konica Minolta's easily on par in the high end, and I think Canon
    get a serious hiding by Nikon, KM & even Pentax in the low to mid end.
    One can only assume that the extra income they get from selling
    overpriced ink to their printer customers has funded more R&D to get
    them ahead of the other camera companies.
    I've heard (admittedly from a follower of the Nikon God, so again
    perhaps unreliable) that focus has been a common problem with all their
    EOS cameras - film and digital included. I was given a technical
    explanation (which again I am unsure of the truthfulness of, so perhaps
    someone more in the know can clarify), that it is all to do with the way
    they focus their lenses. Apparently with Nikon (and everyone else except
    Canon), the camera body has full control over the focussing, and adjusts
    the focus until the body measures in-focus. Apparently with Canon, the
    body measures how far out of focus the image is, and tells the lens how
    much it has to adjust itself by, with no continued feedback from the
    camera. As a result the Canon system is much faster than the others
    (which is definitely true), but is more prone to errors, especially if
    the lens has a bit of age on it. As I say I don't know the truthfulness
    of this, as it came from a Nikon-head, but it could explain why focus
    issues do seem to be more prevalent from Canon.
    Do you think it is a case of bad quality control, or rushing a product
    to market before all the bugs have been ironed out?

    Back to the point of the right camera for the job etc - I'm not
    convinced that the best way to lift ISO is by jacking up the gain. One
    of the big differences between film and digital, is that with film you
    can change the film depending on the conditions you are shooting in and
    the look you are after. Digital is like buying an expensive camera but
    only ever using it with Kodak Gold, and push-processing it when you want
    low-light performance. I think the day will eventually come when we will
    have interchangeable sensors or backs that will allow different
    performance. For example what if you could replace the 8MP sensor of
    your 20D with a 2MP sensor with native ISO6400 performance. Or replace
    it with an 8MP true B&W sensor, that would deliver native ISO 400
    performance because it doesn't lose 2 stops at every sensel from the
    bayer filter. Perhaps you could get sensors that delivered greater
    dynamic range at the expense of resolution or vice-versa. Or maybe an
    Infrared sensor or a UV sensor. At the moment, because all digi's cater
    for the most common denominator (daylight colour photography), they all
    lack in other areas. I'm sure there would be other photographers
    interested in a camera that could deliver smooth 8MP B&W images at
    ISO400 with the same low noise and good dynamic range that the current
    crop of ISO100 cameras deliver. I'm sure there would be photographers
    interested in a camera that could deliver low noise ISO6400 images even
    if they are at a lower resolution. Different sensors would also have a
    different look, just as Fuji and Kodak films have a different look.
    These special needs wouldn't justify the cost of a whole dedicated body,
    but if a way could be made of changing the sensor, just like we
    currently change film, the cameras could truly be called versatile.
     
    Graham Fountain, Dec 18, 2005
    #15
  16. anonomous individual

    Leo Guest

    Tell ya what... test both cameras at ISO 200, or both at 1/40 second... you
    can't shoot people at 1/4 second.

    Your test proves nothing.

    Except maybe you're a Canon basher... I am too, I prefer Nikon, but I don't put
    up bullshit web sites!
     
    Leo, Dec 19, 2005
    #16
  17. What sort of web sites do you put up Leo?
    Any?
    None?
    Only ones with no information?
    which is it mate?
    Maybe you and Jim Kramer ought to get together and make a pictorial
    gallery of all your best efforts? - both of them!
     
    anonomous individual, Dec 19, 2005
    #17
  18. anonomous individual

    Leo Guest

    I have a very large technical site, that gives very specialized information
    dealing with electronics technology, which has been up for several years now,
    and is quite popular as a knowledge base for certain technicians. I don't
    compare apples with oranges, and any comparisons I make dealing with technology
    are equal and fair. I get several hits each day, and lots of Emails asking for
    further information and advice. I make use of cameras, both still and video, to
    supply data for my site. I own or have owned almost every type of camera you can
    name, and shooting in low light is quite important in my field. Comparing
    cameras should be done equally, not with different settings!

    BTW the example pics you showed were kind of poor! They showed nothing that you
    would want to compare! How about showing something with detail, not a side of a
    plastic computer!

    And mucking about with de-noise and other shit is not exactly a fair comparison
    - say I don't have such software? How about I send you two audio samples, one
    from my equipment and one from a competitor, after de-noising and applying
    dynamic range processing... fair? Actually it would be an un-fair comparison
    towards my superior equipment, so I wouldn't do it!

    BTW I apologize for coming down a bit too hard there, I didn't want to appear so
    rough!
     
    Leo, Dec 19, 2005
    #18
  19. Leo, I think said what you did because you assumed too much. I only
    "treated" the top two 'full' images. The crops and noise examples under
    have absolutely no modification whatsoever. They are straight from the
    camera, cropped instead of cut and pasted so the EXIF stayed in place
    and told a true story.

    I did compare both cameras equally. Try and hand hold a 5D at 1/40th and
    see how hard it is to get a sharp picture. It is just as hard to get a
    sharp picture at 1/4th with the Panasonic.

    The FZ20 doesn't work the same way as a 5D and makes it's low light
    pictures differently to a SLR camera. What could be fairer than to use
    the cameras in their best possible setup for a given scene? Each to
    their own, perhaps?

    Thanks Leo.
     
    anonomous individual, Dec 19, 2005
    #19
  20. anonomous individual

    Annika1980 Guest

    The test missed the mark of what I am trying to convey, for sure. My
    Bull. I had to take some people pics yesterday in a poorly lit room
    with fluorescent lighting (yuck!). I was getting about 1/80 @ f/2.8 at
    1600 ISO with my 20D without flash. So explain to me how setting your
    Panasonic to ISO 200 is gonna give me decent pics when I have to shoot
    at 1/10 of a second instead? Yeah, it might work if you're shooting a
    computer sitting there on your floor, but you're kinda screwed in the
    real world. Show us some of your famous wedding pics taken at ISO 200
    in a dark church. This oughta be good.
     
    Annika1980, Dec 19, 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.