joe_c5

<p>I understand that the problem with clipping the red channel is real, however if a red flower clips the red channel, then a white object would also clip the red channel with the same exposure and white balance.</p>

<p>Using raw format might help, since it is possible the in-camera processing might take a red channel that was not clipped at the sensor and clip it in the jpeg.</p>

<p>And yes, in theory you could add a filter to reduce transmission of wavelengths longer than red and then alter the white balance. This would require what sounds like an unusual filter, would not work for white flowers, and risks clipping green or blue instead, so I do not recommend it.</p>

<p>That looks about the same as I get for noise with a 5D at ISO 100 in Lightroom with noise reduction off. Jpegs straight from the camera probably have less noise than that.</p>

<p>And while the noise on the red Subaru may be visible, it looks absolutely nothing like ISO 1600 from my 5D.</p>

<p>For a sharpened 1:1 crop with boosted contrast that looks about like I would expect for ISO 400.</p>

<p>If you want to lower the noise, you should be able to go to at least at ISO 200 without a tripod. Noise reduction in post processing could help too, at the cost of losing some detail.</p>

<p>If you really want to crush the blacks, a tone curve that bends sharply down instead of going all the way to the lower left hand corner should do the trick. That will probably make the image look more similar between the three monitors.</p>

<p>Enough of the white card is clipped that I would not recommend using it for white balance.</p>

<p>Of the three images you posted, only the second image has a gray card that is actually neutral gray, at approximately RGB 164 164 164.</p>

<p>So long as your gray card is actually neutral gray (which is difficult to verify) then only that second image, which was white balanced using the gray card, is correctly white balanced.</p>

<p>If the white card were not overexposed to the point of clipping, then it could probably be used to obtain a similar white balance to the gray card.</p>

<blockquote>

<p>3 - Regarding the 2nd photo, the one set with gray card: How does the red tail lights and flowers look to you? They look a bit over saturated with too much red to me. My shirt looks like it is glowing and while it is a bright orange shirt, it just doesn't look natural to me.</p>

</blockquote>

<p>The shirt looks fake, yes, but most if not all of that is the nonlinear tone curve being used by DPP. I tried using the gray card for white balance with a linear tone curve and it looks much more like a real shirt.</p>

<blockquote>

<p>4 - Yes, the first photo, camera's auto white balance, looks a bit cool to me also.<br /> But the big issue, to me, is that if the camera is doing its job of determining proper white balance, and gray and white cards are valid ways of determining proper white balance, shouldn't they all look the same?</p>

</blockquote>

<p>The white and gray cards are more valid methods than auto white balance. The auto white balance just has to guess without enough information. In particular, auto white balance knows neither the color the light nor the color of the gray and white cards.</p>

<p>Depending on what the things you want are, either GIMP (terrible name—I should recommend that they change it), which is free, or Adobe Lightroom, which is expensive but cheaper than Photoshop, might be possible alternatives.</p>

<p>I am sure there are others too, but these are the two that I use instead of Photoshop, and for my purposes they suffice.</p>

<p>I would prefer 32-bit integer over 32-bit floating point. With linear encoding, individual photons would often be somewhere around the 16th bit, leaving plenty of extra precision without needing potentially slower floating point math.</p>

<p>For many purposes, 16-bit is enough even for editing.</p>

<p>I am not a lawyer, but I would not recommend using the photos in that manner on that site without a signed model release. It is likely to still be considered commercial use, with the fact that it is free being irrelevant.</p>

<p>On your monitor, the image in Photoshop is the correct one. On some other monitors, the image in windows is correct. On many monitors, neither is correct.</p>

<p>If your calibration and monitor profile worked correctly, it will print more like the Photoshop version.</p>

<p>If you embed an sRGB color profile in the image, and then use either Firefox or Safari to view it on Smugmug, it should look the same as in Photoshop.</p>

<blockquote>

<p>Quality of "90-95" is way too high.</p>

</blockquote>

<p>I admit that it varies with the images, but visible jpeg artifacts look unprofessional on any website, let alone one dedicated to photography. I stand by 90 quality as the lowest I would use blindly on a photography website. Lower quality settings should be previewed to check for artifacts.</p>

<p>Also, the images in your photo.net gallery appear to be somewhere around 90 or 95 quality.</p>

<p>Keep in mind that this is for web images viewed at 1:1.</p>

<p>Save the images in sRGB color space, preferably with the sRGB profile embedded. Use jpeg format, probably at a quality of around 90 to 95.</p>

<p>Try not to edit and re-save jpeg files, since in some cases the compression artifacts can be cumulative. Better to start with a fresh copy of the master image.</p>

<p>For image dimensions there may be no correct answer. With heights greater than around 800 pixels or wider than around 1280 pixels, some people will probably see scroll bars or resized images. However, with images that small they will not fill the screen of large monitors.</p>

<p>If you do not want people copying your images, consider watermarking them. If you do not want them making prints, consider low resolutions of no more than around 720x480.</p>

<p>Some people like to sharpen images after they have been resized. I do not recommend doing this to every image by default, but some images can look better with the higher acutance (edge contrast).</p>

<p>Use separate images for thumbnails and full size images, since web browsers may not resize well and the thumbnails would be slow to load.</p>

<p>That ought to get you most of the way there.</p>

<p>By the numbers, in the second version the belts and railings are slightly blue and in the first version they are slightly yellow. I recommend using the eyedropper to set the white balance to match either a belt or railing depending on which one you want to be neutral white; they are slightly different colors.</p>

<p>The monitor's sRGB mode should have a white balance at least somewhere close to the sRGB standard, so in theory calibrating and profiling the monitor should not make a huge difference in the white balance. It is probably still a good idea for other reasons, but may not address this specific issue.</p>

<p>Your eyes will automatically adjust to the white balance of your surroundings. If your monitor were the only light source in a pitch black room, they would adjust to it as well. Since that is probably not the case, the room lighting might affect what you perceive as white.</p>

<p>If it is important that white be white, having equal red, green and blue values is a better bet than what the image looks like.</p>

<p>White balance is far and away the most likely cause. For a single color, white balance will by definition change that color to a neutral gray.</p>

<p>It is possible that something else is wrong, but without an example photo I do not have any guesses.</p>

<p>Any image viewer that looks different from Photoshop will probably continue to look different from Photoshop on a profiled monitor. Calibrating and profiling a monitor should make the colors more accurate in Photoshop, but I suspect that quick view is not color managed and will therefore still look incorrect.</p>

<p>Neither Chrome nor IE8 is color managed. Depending on the monitor, they could look anywhere from correct to ridiculous and that would be expected behavior.</p>

<p>If you consider this to be a problem, the solution would be to use either Firefox or Safari. By default they will only apply color management to tagged images, but this particular image is tagged.</p>

<p>As for Lightroom appearing oversaturated, it is possible that your profile is bad, but I do not have any suggestions for how to tell whether this is the case.</p>

<p>It depends on the lighting. A white balance of 5000K will make incandescent lights look reddish and direct sunlight at noon may look slightly bluish depending on the amount of blue sky.</p>

<p>As said above, if you use raw format the white balance barely matters unless you save time in post processing by setting it in camera.</p>

<p>The most accurate way to set white balance is using a good white or gray card, or a diffuser (which you would point at the light source rather than at the subject), and to set that color to be neutral gray.</p>

<p>The 50D photo of the poster appears to be focused on the front edge of the frame, with such a narrow depth of field at f/2.8 that almost the entire poster is out of focus. Make sure to focus on the center of the poster if it is not parallel with the focal plane, and perhaps try f/5.6 at ISO 400 or f/8 at ISO 800 for greater depth of field.</p>

<p>The landscape photo seems fine at the 1600 x 1067 pixel resolution posted. Depending on what the full resolution version looks like, increasing the sharpening in-camera might help. At f/10 on a 50D, diffraction blur will be starting to set in. Something like f/8 or f/7 should be slightly sharper, though with less depth of field which may not be a net win for a landscape.</p>

<p>From those two photos I cannot draw any conclusion about the quality of the lens beyond that it is at least sane.</p>

<p>It may not be possible to match all colors without editing them by hand because the spectral response of the two cameras is probably different, but I would expect using the same third party RAW converter software to get you most of the way there.</p>

<p>I don’t recommend using that exact profile, no. I looked at it again, and it is not only using a D50 white point, but it also has a color adaptation matrix that seems to convert to an even lower color temperature. I am not sure why, beyond that it thinks your monitor is very red.</p>

<p>The image I posted should look pink to the point of being ridiculous.</p>

<p>D65 is roughly the color of a mix of direct sunlight and skylight. D50 is slightly redder than direct sunlight at noon.</p>

<p>Gamma 2.2 is just the standard tone curve for sRGB, or really sRGB uses a curve very close to that made with gamma 1.0 near black and gamma 2.4 above that. Tone curves specify how bright each shade of gray is compared to black and white.</p>

<p>D65 and gamma 2.2 are what you want, yes. I don’t know if the Huey software lets you profile the monitor without calibrating it, using the monitor’s native white point and tone curve, but that is how I profile mine.</p>

<p>You might try the Pantone help center at http://pantone.custhelp.com/app/home. The fourth popular help topic is “Color cast on monitor after profiling with PANTONE huey/hueyPRO”, and might apply to you (they send a replacement unit if they confirm the problem). Certainly if I try that profile on my monitor, the calibration makes everything greenish blue.</p>

<p>The primary colors on that profile seem sane, but the white point is D50 rather than D65. Basically the profile thinks that white is redder than normal.</p>

<p>The attached image shows what the profile thinks the photo looks like in the sRGB color space, with less difference between white and red. Since color management usually converts to the output device’s white point, the red is faded instead to be closer to the white.</p>

<p>Even though the profile description claims to be D65, you might want to double check the color temperature when you create the profile.</p><div></div>

<p>I had meant that the profile should be embedded in the image you post, rather than a screenshot of it. The image can be anything—one of these screenshots, blank, anything.</p>

<p>The camera is not necessarily accurate either, so a comparison between the monitor and the actual subject of the photo still does not show whether the monitor is the right color or not.</p>

<p>What is the model of the Dell monitor? Basically if the monitor has particularly saturated colors, then the profile is probably doing exactly what it should, and if the monitor doesn’t, then there is probably something wrong with the profile.</p>

<p>You do not ever want to assign your monitor’s display profile. On your monitor that will make a photo look as if color management was turned off, but on nearly everybody else’s monitor it will look completely wrong.</p>

<p>GIMP is color managed, but not automatically, you have to set it in Edit > Preferences > Color Management.</p>

<p>Either there is a problem with your monitor profile, or you misunderstand what the images are supposed to look like, or both. I understand that isn’t very helpful, but it is difficult to determine which without a known good color reference.</p>

<p>Could you post an image with the Huey profile assigned? Yes, I know I said you do not ever want to do that, but in this case I want to see the profile rather than the image, and on photo.net I think you can only attach images.</p>

<p>Also, do you have a wide gamut monitor? If so, then it might make sense for the change to be that dramatic.</p>

<blockquote>

<p>So, I tried the onboard flash, not enough, so then went to the external flash on high sync.</p>

</blockquote>

<p>If you can get enough power to the flash to drown out the ambient light, then you probably will not want high speed sync. In normal sync mode, the flash itself can be extremely fast, perhaps 30 to 100 microseconds. In high speed sync mode, the flash will fire repeatedly at lower power in order to keep the flash lit up from before the first curtain opens until after the second curtain closes.</p>

<p>So if your flash is bright enough, which it probably is at close enough range, try a smaller aperture in <em>normal </em>sync mode, not high speed. Keep the shutter speed at the 50D’s max sync speed of 1/250s.</p>

 

<blockquote>

<p>I was already at f/3.5 to get just enough DOF for the Hummer, and shooting at the 50D's lowest ISO of 100.</p>

</blockquote>

<p>These two settings are at cross purposes. The low ISO decreases exposure, but the large aperture increases it. This makes sense for high quality images, but to shoot in low ambient light you would want to use a large aperture <em>and </em>a higher ISO. Or in this case where the ambient light is only causing blur, you would want to use a small aperture <em>and </em>100 ISO.</p>

<p>The image you posted here is in the ProPhoto RGB color space, with an embedded profile, so it looks correct in some browsers and wrong in others.</p>

<p>The same image in your portfolio has no embedded profile, even though the data looks like it is also ProPhoto RGB, so it looks wrong in all browsers.</p>

<p>I recommend using Photoshop to save a copy of the image converted to sRGB color space for posting on the web.</p>

<p>6000K versus 6500K should have almost nothing to do with color or brightness. Could you be more specific about what is wrong with the prints?</p>

<p>If the prints look too dark, the monitor is too bright. If the prints look too bright, the monitor is too dim. If whites in the prints look too blue, the monitor color temperature is too low. If whites in the prints look too red, the monitor color temperature is too high.</p>

<p>If anything else is wrong, it might be the printer profile, the monitor profile, or both.</p>