Zum Hauptinhalt gehen

⚠️ Please note that this topic or post has been archived. The information contained here may no longer be accurate or up-to-date. ⚠️

Canon 600D "blown" highlights turn pink after exp. reduced

Kommentare

13 Kommentare

  • Paul Steunebrink
    Hi Grant,

    The behavior you describe sounds familiar for Canon shooters, and from the top of my head, I like to give a few suggestions.

    It all boils down to use 'real' ISO settings and avoid the Canon tweaks as in:
    - extended ISO like low and high
    - highlight reduction settings (custom functions).
    Also you might want to avoid reduced raw formats for a start (introduce them later, if applicable).
    0
  • SFA
    [quote="Paul_E" wrote:
    Hi Grant,

    The behavior you describe sounds familiar for Canon shooters, and from the top of my head, I like to give a few suggestions.

    It all boils down to use 'real' ISO settings and avoid the Canon tweaks as in:
    - extended ISO like low and high
    - highlight reduction settings (custom functions).
    Also you might want to avoid reduced raw formats for a start (introduce them later, if applicable).



    Hi Paul and thanks for the response.

    None of the images are, at point of taking, using extended ISO, highlight reduction (it''s off if that refers to Auto Lighting Optimiser) or reduced RAW formats. Or at least if they are it's not obvious nor intended from the settings! However that gives me something to experiment with to see if I can induce the symptoms from a less obviously wide dynamic range image. Thanks!

    The only thing I have set is Auto ISO with a max of 3200.

    Grant
    0
  • Jim MSP
    >> "any whites which are at right around the burn-out point will tend to turn beige going to pink is I reduce the exposure of an image "

    I use a Canon 60D, which I think has the same sensor.
    I have seen a similar effect, except that the color can sometimes go gray. I have not done a controlled experiment, but I think the color is related to the color of the surrounding material.

    Jim
    0
  • SFA
    [quote="Jim MSP" wrote:
    >> "any whites which are at right around the burn-out point will tend to turn beige going to pink is I reduce the exposure of an image "

    I use a Canon 60D, which I think has the same sensor.
    I have seen a similar effect, except that the color can sometimes go gray. I have not done a controlled experiment, but I think the color is related to the color of the surrounding material.

    Jim



    Hi Jim,

    I think you are right about the sensor though I suspect that there may be a little difference in the way the output is processed in camera. However one might hope that there are not too many differences in a RAW file.

    I think you may also be right about the colour issue.

    I first noticed the phenomenon when processing some photos of Cricket - an odd sport to many - where the players traditionally where 'whites' The clothes are actually slightly off-white. Whn blown or at least at the limits reducing the exposure tends to leave a beige colour patch. Typically the nearest surrounding colour will be green/yellow (grass) but the colour does not seem to have to be immeditaly adjacent to the blown patch.

    The very strong pink I noticed over the weekend is next to a very deep blue sky but also some very white looking clouds that are not affected. The blown area, before processing, shows max colour channel values as expected. After processing the balues are all reduced but red and blue are higher then green. Or maybe that should be thought of the other way around ...?

    Interestingly this morning I realised that I was processing some more shots from a camera/lens combination that exhibits the phenomenon but although I could see that the files had whites with 255 values for all 4 channels making the adjustments did not show the anomaly. 255 sections that stay 255 (really blown) or drop by a consistent amount they remain white or drop to grey, as one expects.

    However when start with, say, R, B and Grey on 255 and Green on 254 then the imbalance can appear. So making an extreme adjustment that turns the white cloud pink (to make it completely obvious I used -2.50ev) the final values
    for the colour channels seem to be Red and Blue 175 (a consistent value drop) Grey showing as 167 but green now 162.

    I assume there is a way to adjust the green value and assess the result but so far I am struggling to spot something that seems to work. In any case this seems so specific that there has to be an alternative solution but I'm not yet sure where in the process it resides.

    Looking at the Histogram as one adjusts the exposure value (or uses other tools that will have a subsidiary exposure effect) it's clear that the green curve has different peaks to the Red and Blue although where and by how much (relative of course) varies from image to image.

    I have some 7D files from a couple of years ago. Also the same base sensor iirc. I checked them quickly and did not see the same phenomenon but will now revisit having seen some specific circumstances that sem to be required for the phenomenon to appear.

    Jim, if you get the chance to check some of your files and see something similar in terms of the values when the issue appears maybe we can pool the observations and ask P1 if they have any suggestions about avoiding the problem - preferable not by shooting to the left! 😊


    Grant
    0
  • SFA
    My favourite RAW converter, prior to C1, disappeared as a product about a year ago. However a small group of enthusiasts maintain some level of support - notably for new camera releases - and so I make use of it from time to time As might be expected there are some things that it does that C1 doesn't and a few things that C1 does that the old favourite addresses but cannot match so easily. If only I could combine the two ...

    More pertinently I have updated some of the old product's components today - dcraw and the some new camere colour mapping profiles - and found that the latest, user community based, releases address most of the issues I mentioned previously. Which is interesting. Where I was getting very similar 'pink' areas on exposure reduction, now I don't.

    I suspect that there might be something very 'special' about the way that Canon have introduced colour mapping in some of their recent products.

    Grant
    0
  • Permanently deleted user
    [quote="SFA" wrote:


    I suspect that there might be something very 'special' about the way that Canon have introduced colour mapping in some of their recent products.


    I used to use a Canon 300D and it also had this pink blown out issue in CO, so it's not just recent products
    0
  • SFA
    [quote="Alex1111" wrote:
    [quote="SFA" wrote:


    I suspect that there might be something very 'special' about the way that Canon have introduced colour mapping in some of their recent products.


    I used to use a Canon 300D and it also had this pink blown out issue in CO, so it's not just recent products



    That's interesting.

    I have yet to see it from the other RAW files I have to play with - 400D, S90, G11 and 1D3. But the situation needed to induce the problem does seem very specific.

    Grant
    0
  • Drew Altdo
    Any chance someone would care to create a support case and provide an example of this occurring?
    We do not own a 600D and the files we have do not show this.
    0
  • SFA
    [quote="Drew" wrote:
    Any chance someone would care to create a support case and provide an example of this occurring?
    We do not own a 600D and the files we have do not show this.



    Hi Drew,

    I will do so if I feel fairly confident that there seems to be a genuine anomaly and I can describe it accurately and provide decent sample files. The circumstances do seem to be quite specific.

    As of last night I got to the point where I had updated to the 'other' application with a revised profile homebrewed by someone else to provide 600D support (also the version for 60D support) and using those the previoous similar problem has been rectified - at least in the sense that the anomalous area changes to grey as might be expected and avoids the colour cast.

    I have a few 60D files I can check but they are not my images - maybe Jim can provide something? I'm not sure the files I have would contain the specific anomaly anyway.

    More soon.


    Grant
    0
  • Jim MSP
    [quote="SFA" wrote:
    [quote="Drew" wrote:
    Any chance someone would care to create a support case and provide an example of this occurring?
    We do not own a 600D and the files we have do not show this.

    ...
    I have a few 60D files I can check but they are not my images - maybe Jim can provide something? I'm not sure the files I have would contain the specific anomaly anyway.

    .....

    I may be able to do that in the next day or so. As luck would have it, I'm installing a new laptop at the moment, as I travel on Sunday.
    Interestingly enough, I was looking at a file yesterday that showed both a pink and a blue-green area from a blown out white. These were adjacent and both, I think, started with RGB of 255.
    0
  • SFA
    [quote="Jim MSP" wrote:
    [quote="SFA" wrote:
    [quote="Drew" wrote:
    Any chance someone would care to create a support case and provide an example of this occurring?
    We do not own a 600D and the files we have do not show this.

    ...
    I have a few 60D files I can check but they are not my images - maybe Jim can provide something? I'm not sure the files I have would contain the specific anomaly anyway.

    .....

    I may be able to do that in the next day or so. As luck would have it, I'm installing a new laptop at the moment, as I travel on Sunday.
    Interestingly enough, I was looking at a file yesterday that showed both a pink and a blue-green area from a blown out white. These were adjacent and both, I think, started with RGB of 255.


    I have a few files where the boundary between 'blown' white and a blue sky shows a blue green band between them after extreme adjustment but I sort of expect that.

    I have yet to notice anything with both colours!

    Using a number of colour readouts I have ascertained that whilst a 255,254,255,255 starting point will seemingly always show the pink anomaly some, but not all, starting points with values of all 255 will still show a pink anomaly and what might be thought of as inconsistent reduction in the green channel value (as reported by the Colour readout.)

    Moreover I can find examples of almost identical images taken seconds apart where one will show the anomaly when adjusted and the other won't. Whatever is going on seems to be rather obscure.


    Grant
    0
  • Permanently deleted user
    If one does a search of "blown highlights pink raw' one will find lots of links to this issue that cuts across raw converters and camera types.

    for example: (in not particular order or usefulness)
    http://people.zoy.org/~cyril/ufraw_highlight_recovery/
    http://forum.getdpi.com/forum/image-pro ... e-one.html
    http://www.luminous-landscape.com/forum ... 100.0;wap2
    0
  • SFA
    [quote="Alex1111" wrote:
    If one does a search of "blown highlights pink raw' one will find lots of links to this issue that cuts across raw converters and camera types.

    for example: (in not particular order or usefulness)
    http://people.zoy.org/~cyril/ufraw_highlight_recovery/
    http://forum.getdpi.com/forum/image-pro ... e-one.html
    http://www.luminous-landscape.com/forum ... 100.0;wap2



    Interesting items, especially the second one. That said the sample files posted to illustrate the problems simply don't show anything very evident (mostly) on my notebook screen ...

    My suspicions are related to ISO values especially the processed 'intermediates'. However so far I have not found any consistency to support such a view. And I've only seen the problem on the 600D or at least if it is present on the other cameras it will have been to a much lower and less than worrying degree that I have overlooked.

    With the 600D it is very much to the forefront.


    Grant
    0

Post ist für Kommentare geschlossen.