is this IR on EOS10D?

Discussion in 'Digital Photography' started by dylan, Nov 12, 2003.

  1. dylan

    dylan Guest

    Whilst waiting for the total luna eclipse the other night I decide to try
    out my 10D at 1600 ISO. I took an image of a white van under a sodium
    streetlight. When you split the image into RGB, in PSP, you get a red and
    green channel image and no blue detail, as expected from a sodium
    streetlight BUT if you brighten the blue channel and contrast stretch it you
    get an image that appears inverted. Is it IR it is recording ?.

    Original, RGB and Blue enhanced at www.knighttrain.freeserve.co.uk/1600.htm
    dylan, Nov 12, 2003
    #1
    1. Advertising

  2. dylan

    dylan Guest

    Or maybe it's UV seeing it's at the blue end of the spectrum ?

    "dylan" <> wrote in message
    news:bou4f7$f29$...
    > Whilst waiting for the total luna eclipse the other night I decide to try
    > out my 10D at 1600 ISO. I took an image of a white van under a sodium
    > streetlight. When you split the image into RGB, in PSP, you get a red and
    > green channel image and no blue detail, as expected from a sodium
    > streetlight BUT if you brighten the blue channel and contrast stretch it

    you
    > get an image that appears inverted. Is it IR it is recording ?.
    >
    > Original, RGB and Blue enhanced at

    www.knighttrain.freeserve.co.uk/1600.htm
    >
    >
    >
    dylan, Nov 12, 2003
    #2
    1. Advertising

  3. dylan

    Jim Townsend Guest

    dylan wrote:

    > Whilst waiting for the total luna eclipse the other night I decide to try
    > out my 10D at 1600 ISO. I took an image of a white van under a sodium
    > streetlight. When you split the image into RGB, in PSP, you get a red and
    > green channel image and no blue detail, as expected from a sodium
    > streetlight BUT if you brighten the blue channel and contrast stretch it you
    > get an image that appears inverted. Is it IR it is recording ?.
    >
    > Original, RGB and Blue enhanced at www.knighttrain.freeserve.co.uk/1600.htm


    No.. You just got the blue that was always there. There's a lot of noise from
    stretching the contrast so far on an ISO 1600 exposure :)

    You'll find you can do this with all the colors combined or separated.

    Put the camera in manual mode and shoot a scene so that it's extremely
    underexposed (so you can barely see the detail). Then play with levels and
    contrast.. You'll pull out a similar image.

    FWIW.. If you shoot in RAW, and convert the image to 16bit TIFF and use a 16bit
    editor, you can get even more detail from an apparently black image.
    Jim Townsend, Nov 12, 2003
    #3
  4. "dylan" <> wrote in message
    news:bou58k$rnu$...
    > Or maybe it's UV seeing it's at the blue end of the spectrum ?
    >
    > "dylan" <> wrote in message
    > news:bou4f7$f29$...
    > > Whilst waiting for the total luna eclipse the other night I decide to

    try
    > > out my 10D at 1600 ISO. I took an image of a white van under a sodium
    > > streetlight. When you split the image into RGB, in PSP, you get a red

    and
    > > green channel image and no blue detail, as expected from a sodium
    > > streetlight BUT if you brighten the blue channel and contrast stretch it

    > you
    > > get an image that appears inverted. Is it IR it is recording ?.
    > >
    > > Original, RGB and Blue enhanced at

    > www.knighttrain.freeserve.co.uk/1600.htm




    Are you sure the van was white, not yellow?

    - jz
    Jeff Zawrotny, Nov 12, 2003
    #4
  5. dylan

    dylan Guest

    Why is the van black then and not noisy like the rest of the image ?. It was
    definitely a white van so there should be some more blue present if you are
    correct.

    "Jim Townsend" <> wrote in message
    news:...
    > dylan wrote:
    >
    > > Whilst waiting for the total luna eclipse the other night I decide to

    try
    > > out my 10D at 1600 ISO. I took an image of a white van under a sodium
    > > streetlight. When you split the image into RGB, in PSP, you get a red

    and
    > > green channel image and no blue detail, as expected from a sodium
    > > streetlight BUT if you brighten the blue channel and contrast stretch it

    you
    > > get an image that appears inverted. Is it IR it is recording ?.
    > >
    > > Original, RGB and Blue enhanced at

    www.knighttrain.freeserve.co.uk/1600.htm
    >
    > No.. You just got the blue that was always there. There's a lot of noise

    from
    > stretching the contrast so far on an ISO 1600 exposure :)
    >
    > You'll find you can do this with all the colors combined or separated.
    >
    > Put the camera in manual mode and shoot a scene so that it's extremely
    > underexposed (so you can barely see the detail). Then play with levels

    and
    > contrast.. You'll pull out a similar image.
    >
    > FWIW.. If you shoot in RAW, and convert the image to 16bit TIFF and use a

    16bit
    > editor, you can get even more detail from an apparently black image.
    >
    >
    >
    >
    dylan, Nov 13, 2003
    #5
  6. dylan

    dylan Guest

    definitely a white van (ford polar white ?)

    "Jeff Zawrotny" <> wrote in message
    news:...
    >
    > "dylan" <> wrote in message
    > news:bou58k$rnu$...
    > > Or maybe it's UV seeing it's at the blue end of the spectrum ?
    > >
    > > "dylan" <> wrote in message
    > > news:bou4f7$f29$...
    > > > Whilst waiting for the total luna eclipse the other night I decide to

    > try
    > > > out my 10D at 1600 ISO. I took an image of a white van under a sodium
    > > > streetlight. When you split the image into RGB, in PSP, you get a red

    > and
    > > > green channel image and no blue detail, as expected from a sodium
    > > > streetlight BUT if you brighten the blue channel and contrast stretch

    it
    > > you
    > > > get an image that appears inverted. Is it IR it is recording ?.
    > > >
    > > > Original, RGB and Blue enhanced at

    > > www.knighttrain.freeserve.co.uk/1600.htm

    >
    >
    >
    > Are you sure the van was white, not yellow?
    >
    > - jz
    >
    >
    dylan, Nov 13, 2003
    #6
  7. dylan

    Guest

    In message <bp04se$c0s$>,
    "dylan" <> wrote:

    >Why is the van black then and not noisy like the rest of the image ?. It was
    >definitely a white van so there should be some more blue present if you are
    >correct.


    I suspect that either your software is broken, or it doesn't do what you
    think it does.
    --

    <>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
    John P Sheehy <>
    ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
    , Nov 14, 2003
    #7
  8. dylan

    abc Guest

    Not sure that's the answer but it's possible. Looking at it again the noise
    in the blue image (normal or enhanced) occurs where the light levels are low
    NOT just the blue level. It appears as if the noise in the van area in the
    blue image (black) is somehow supressed by the fact that there is a high
    level Red and/or Green signal ie it's related to the combined light level.

    Maybe knowing the answer won't prove anything but I was interested the
    mechanism of how it happens.


    <> wrote in message
    news:...
    > In message <bp04se$c0s$>,
    > "dylan" <> wrote:
    >
    > >Why is the van black then and not noisy like the rest of the image ?. It

    was
    > >definitely a white van so there should be some more blue present if you

    are
    > >correct.

    >
    > I suspect that either your software is broken, or it doesn't do what you
    > think it does.
    > --
    >
    > <>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
    > John P Sheehy <>
    > ><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
    abc, Nov 14, 2003
    #8
  9. dylan

    jam Guest

    Dylan,

    Short answer: No, it's not definitely not any kind of IR. Nor is it UV
    (no source here). I'm betting on the camera's noise reduction and
    white balance algorithms.

    Long answer: We've established that

    o the van is white in visible light.

    o the main light source was a sodium vapor street lamp (probably of
    the usual low-pressure type)

    Low-pressure sodium vapor street lamps emit a very narrow band of
    visible light centered in the yellow at ~590 nm with no red or blue
    output. They also emit a small longwave near IR (NIR) side lobe
    centered at ~810 nm. (For a look at the spectrum, scroll down to Fig.
    7 at
    http://www.chemeducator.org/sbibs/samples/spapers/34samplejg.htm.) We
    don't know the precise reflection spectrum of the van's paint, but
    it's unlikely to absorb in the NIR for quantum mechanical reasons.

    The appearance of the van itself seems easy to explain. Since there
    was no blue in the light source, the van reflected none and showed up
    black in the blue channel, even in the pumped-up version. The sodium's
    main yellow emission line probably falls closer to the transmission
    spectrum peak of the red Bayer filter's than to the green's; hence,
    the van's brighter in the red than in the green channel and shows up a
    ruddy orange in the RGB image.

    I think the camera's noise reduction and white balance algorithms
    created the very faint inverted image in the blue channel. The
    software saw enough light coming from the van, the curb and some of
    the leaves to convince itself that there really wasn't any blue light
    there. Elsewhere, in the shadows, it wasn't so sure, so it made some
    statistical guesses about how much blue there should be in the shadows
    and threw it in, in part to even out noise in the shadows. The image
    may well have been constructed using red and green channel data to
    some extent.

    Note that no real reflection or emission of light could produce that
    last image. Since the shadows are the most "illuminated", the "light"
    can't be coming from the street lamp. Thermal IR can't be the answer
    because objects under several hundred degrees C don't radiate IR at
    wavelengths silicon-based sensors can detect. Faint visible light from
    a nearby building would have reflected from the van at least as much
    as from the sidewalk. Longwave NIR reflected off the sidewalk from the
    street lamp is also out. For one thing, the van's white paint is
    unlikely to selectively absorb such NIR as stated above. (My digital
    IR experience bears this out: Most visible white objects tend to be
    fairly NIR-bright as well.
    --
    Jeremy McCreary
    Denver, CO
    www.cliffshade.com/dpfwiw/
    -------------------------------------------

    "dylan" <> wrote in message
    news:bou4f7$f29$...
    | Whilst waiting for the total luna eclipse the other night I decide
    to try
    | out my 10D at 1600 ISO. I took an image of a white van under a
    sodium
    | streetlight. When you split the image into RGB, in PSP, you get a
    red and
    | green channel image and no blue detail, as expected from a sodium
    | streetlight BUT if you brighten the blue channel and contrast
    stretch it you
    | get an image that appears inverted. Is it IR it is recording ?.
    |
    | Original, RGB and Blue enhanced at
    www.knighttrain.freeserve.co.uk/1600.htm
    |
    |
    |
    jam, Nov 16, 2003
    #9
  10. dylan

    dylan Guest

    Thanks for your answer, I'll put it down to the camera algorithms generated
    the image but not too convinced by "statistical guesses" but you could be
    right.

    "jam" <> wrote in message
    news:keEtb.12063$Dw6.71581@attbi_s02...
    > Dylan,
    >
    > Short answer: No, it's not definitely not any kind of IR. Nor is it UV
    > (no source here). I'm betting on the camera's noise reduction and
    > white balance algorithms.
    >
    > Long answer: We've established that
    >
    > o the van is white in visible light.
    >
    > o the main light source was a sodium vapor street lamp (probably of
    > the usual low-pressure type)
    >
    > Low-pressure sodium vapor street lamps emit a very narrow band of
    > visible light centered in the yellow at ~590 nm with no red or blue
    > output. They also emit a small longwave near IR (NIR) side lobe
    > centered at ~810 nm. (For a look at the spectrum, scroll down to Fig.
    > 7 at
    > http://www.chemeducator.org/sbibs/samples/spapers/34samplejg.htm.) We
    > don't know the precise reflection spectrum of the van's paint, but
    > it's unlikely to absorb in the NIR for quantum mechanical reasons.
    >
    > The appearance of the van itself seems easy to explain. Since there
    > was no blue in the light source, the van reflected none and showed up
    > black in the blue channel, even in the pumped-up version. The sodium's
    > main yellow emission line probably falls closer to the transmission
    > spectrum peak of the red Bayer filter's than to the green's; hence,
    > the van's brighter in the red than in the green channel and shows up a
    > ruddy orange in the RGB image.
    >
    > I think the camera's noise reduction and white balance algorithms
    > created the very faint inverted image in the blue channel. The
    > software saw enough light coming from the van, the curb and some of
    > the leaves to convince itself that there really wasn't any blue light
    > there. Elsewhere, in the shadows, it wasn't so sure, so it made some
    > statistical guesses about how much blue there should be in the shadows
    > and threw it in, in part to even out noise in the shadows. The image
    > may well have been constructed using red and green channel data to
    > some extent.
    >
    > Note that no real reflection or emission of light could produce that
    > last image. Since the shadows are the most "illuminated", the "light"
    > can't be coming from the street lamp. Thermal IR can't be the answer
    > because objects under several hundred degrees C don't radiate IR at
    > wavelengths silicon-based sensors can detect. Faint visible light from
    > a nearby building would have reflected from the van at least as much
    > as from the sidewalk. Longwave NIR reflected off the sidewalk from the
    > street lamp is also out. For one thing, the van's white paint is
    > unlikely to selectively absorb such NIR as stated above. (My digital
    > IR experience bears this out: Most visible white objects tend to be
    > fairly NIR-bright as well.
    > --
    > Jeremy McCreary
    > Denver, CO
    > www.cliffshade.com/dpfwiw/
    > -------------------------------------------
    >
    > "dylan" <> wrote in message
    > news:bou4f7$f29$...
    > | Whilst waiting for the total luna eclipse the other night I decide
    > to try
    > | out my 10D at 1600 ISO. I took an image of a white van under a
    > sodium
    > | streetlight. When you split the image into RGB, in PSP, you get a
    > red and
    > | green channel image and no blue detail, as expected from a sodium
    > | streetlight BUT if you brighten the blue channel and contrast
    > stretch it you
    > | get an image that appears inverted. Is it IR it is recording ?.
    > |
    > | Original, RGB and Blue enhanced at
    > www.knighttrain.freeserve.co.uk/1600.htm
    > |
    > |
    > |
    >
    >
    dylan, Nov 16, 2003
    #10
    1. Advertising

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

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Frank v.d.Venne

    Sigma 12-24 mm/eos10d

    Frank v.d.Venne, Dec 21, 2003, in forum: Digital Photography
    Replies:
    16
    Views:
    632
  2. CAT

    Canon EOS10D samples from first week

    CAT, Feb 14, 2004, in forum: Digital Photography
    Replies:
    15
    Views:
    480
    H Aaen
    Feb 16, 2004
  3. Gremlin

    EOS10D focus problems

    Gremlin, Aug 27, 2007, in forum: Digital Photography
    Replies:
    1
    Views:
    335
    Malcolm Stewart
    Aug 31, 2007
Loading...

Share This Page