rjschell
Established
Is anyone having trouble getting realistic rendition of red roses? They look like they're pasted on the image. I'm shooting RAW.roses008.jpg
Attachments
Darthfeeble
But you can call me Steve
I've found red very difficult in digital, I've gone to underexposing a little and rely on Shadows/Highlight adjustments. In bright light a shade would help too.
jordanatkins
Established
Yes, it's extremely easy to clip reds in digital... I think the red in sRGB and AdobeRGB are quite small compared to green and blue.
I've red somewhere that canon cameras only meter for green, whereas Nikon meters for all colors. Not sure how other manufacturers process their metering. I'm curious to know what camera you're using...
I've red somewhere that canon cameras only meter for green, whereas Nikon meters for all colors. Not sure how other manufacturers process their metering. I'm curious to know what camera you're using...
rjschell
Established
Jordan- I'm shooting NEX-6. The lens was a single coated G. Zuiko 50mm f/1.4 @ f/4.
Reducing red and/or magenta saturation in post-production doesn't help much. It simply makes the rendition muddy and leaves a white halo edge.
Reducing red and/or magenta saturation in post-production doesn't help much. It simply makes the rendition muddy and leaves a white halo edge.
charjohncarter
Veteran
Eggleston said that red was 'at war with all other colors' and that was using film. My solution when I shoot reds (digital) is to shoot RAW, and underexpose. Then use the PSE or Photoshop RAW converter to increase the exposure slider to the red level you want (watch the histogram on the RAW converter).
I try not to really shoot much digital, but this is how I did this shot:
and this:
I try not to really shoot much digital, but this is how I did this shot:
and this:
x-ray
Veteran
You didn't say whether it's film or digital. Also is this an issue when viewing on your monitor or in print?
If film, every film has a specific color rendition. Ektar is different than kodak max film for example. Scanners each have unique characteristics from model to model and brand to brand not to mention whether it's profiled or not. Every monitor must be profiled and each make / model will render. Color different even if profiled. Printers are equally unique and papers as well.
Digital cameras need to be profiled for best color. Every digital camera I own is profiled under different lighting conditions such as sun light, shade, overcast, tungsten and fluorescent. I carry a mini color checker pastport with me and shoot it under the light I'm working in. If I don't have a profile I make one on my computer when I get back to the office. I also use the pastport to balance color in lightroom.
Ok, I shoot raw only. I work in lightroom and apt the correct camera profile to my image when editing and then color balance. My monitor is properly profiled as well.
By applying your camera profile you're correcting for color deficiencies of the CCD or CMOS. It's not perfect but it's much closer than without.
I used canons for years and never liked the reds. They always were on the yellow side to slightly orange. When I started profiling my reds greatly improved. I'm now using nikon and do the same and get the best reds I've ever had. Don't forget you must have a good and profiled monitor to judge color on.
Look into the cor checker pastport and software. It makes all the difference.
If film, every film has a specific color rendition. Ektar is different than kodak max film for example. Scanners each have unique characteristics from model to model and brand to brand not to mention whether it's profiled or not. Every monitor must be profiled and each make / model will render. Color different even if profiled. Printers are equally unique and papers as well.
Digital cameras need to be profiled for best color. Every digital camera I own is profiled under different lighting conditions such as sun light, shade, overcast, tungsten and fluorescent. I carry a mini color checker pastport with me and shoot it under the light I'm working in. If I don't have a profile I make one on my computer when I get back to the office. I also use the pastport to balance color in lightroom.
Ok, I shoot raw only. I work in lightroom and apt the correct camera profile to my image when editing and then color balance. My monitor is properly profiled as well.
By applying your camera profile you're correcting for color deficiencies of the CCD or CMOS. It's not perfect but it's much closer than without.
I used canons for years and never liked the reds. They always were on the yellow side to slightly orange. When I started profiling my reds greatly improved. I'm now using nikon and do the same and get the best reds I've ever had. Don't forget you must have a good and profiled monitor to judge color on.
Look into the cor checker pastport and software. It makes all the difference.
YYV_146
Well-known
Red is an issue with Sony-sensored cameras...well, red is an issue with ALL digital cameras, but Sony gets it worse than CMOSIS Leica or Canon. The reds are typically magenta-ish, too strong and lacking detail.
If you want to create a general profile for the NEX-6, try slightly desaturating the reds and tweaking the color balance. Alternatively you can try to find a camera profile online - the red channel issue is well-known, and should be addressed by most profiles. But of course a camera-and-lens specific profile will always work best.
If you want to create a general profile for the NEX-6, try slightly desaturating the reds and tweaking the color balance. Alternatively you can try to find a camera profile online - the red channel issue is well-known, and should be addressed by most profiles. But of course a camera-and-lens specific profile will always work best.
YYV_146
Well-known
I desaturated the pinks here, but the image was slightly overexposed in the first place, so you can *just make out the tearing effect of a channel being re-toned too far. This is the A7, though, it's a common Sony sensor issue.
jordanatkins
Established
Yep, as others have said... underexpose a little, especially in daylight. I've used an NEX-7 and I now own an A7, and -2/3 of exposure correction in daylight gives me good results. Worked when I shot Nikon too. Should prevent any highlights from blowing, and if not you should have enough headroom in raw to pull them back in.
Ranchu
Veteran
May be some help. A chapter from one of Dan Margulis's books.
http://chat.ledet.com/margulisassets/pdfs/PLC_Ch09.pdf
http://en.wikipedia.org/wiki/Dan_Margulis
http://chat.ledet.com/margulisassets/pdfs/PLC_Ch09.pdf
http://en.wikipedia.org/wiki/Dan_Margulis
willie_901
Veteran
I also think it is easy to clip the red channel even though the total histogram looks ok. Even a small percentage of clipped red pixels can have a large effect.
Different color filter arrays (not semiconductor technologies) can have a significant impact even when exposure for all three channels is perfect.
Manufacturers are tempted to sacrifice microlense frequency selectivity for transmission efficiency (less filtering means more light.. which in turn means better dynamic range and high ISO performance). However the Bayer rendering interpolation algorithms depend on frequency selectivity (red lenses only pass red, etc). This means some brands will render reds better than others.
Different color filter arrays (not semiconductor technologies) can have a significant impact even when exposure for all three channels is perfect.
Manufacturers are tempted to sacrifice microlense frequency selectivity for transmission efficiency (less filtering means more light.. which in turn means better dynamic range and high ISO performance). However the Bayer rendering interpolation algorithms depend on frequency selectivity (red lenses only pass red, etc). This means some brands will render reds better than others.
Jack Conrad
Well-known
Which brands render reds better?
YYV_146
Well-known
For modern digital cameras the concept of Jpeg colors (mostly a result of company firmware) and RAW colors (mostly a function of the sensor) are different. All RAW colors are greatly malleable, if you always use profiles, you might as well treat every one of your cameras as the same - this is what I do.
For RAW colors, I think the modern CMOSIS Leica (M type 240) and the Fuji X-trans cameras have pretty good, if not perfect, reds. Leica tends to orange and Fuji to purplish, but very slight. Canon's sensors are also less prone to red channel blowouts compared to all Sony sensors; you can see this very clearly between the D600 and 6D. But again, I am nitpicking. Sony's red tones are actually not bad, it's just that they have a tendency to be too strong compared to the other colors.
f16sunshine
Moderator
I recall earlier Pentax DSLR models rendered very well. The KD10 and KD20 models a friend was running in particular.
We compared images to my 5Dii when I picked it up years ago and I was quite unsatisfied with reds after viewing the Pentax files.
I'm not too terribly enthralled with the Fuji Reds/oranges from either x100 or xpro1 RAW files.
They look better printed than on screen I will give em that. Hard to know what you are going to get though since editing is done on screen.
semilog
curmudgeonly optimist
I'm not sure what you mean here. The red channel is a single channel. If you mean that the rendered images have the features you describe, that's going to be due to the camera's JPEG engine or the RAW developer that you are using — and both of these come down to camera profiles. Software, not hardware.
YYV_146
Well-known
I mean the raw images. But areas we actually see as red consists mostly of red information. So I suspect there is something with the sensor as well. Everything loosely red-ish on a Sony sensor is more saturated than it should be, which is why I say sensor design may be in play.
daveleo
what?
I really shouldn't be here, as the discussion is several technical levels above my education, but ....
Is this something that could be helped by a filter on the lens? Didn't Leica correct a magenta problem with a filter on the M8?
Is this something that could be helped by a filter on the lens? Didn't Leica correct a magenta problem with a filter on the M8?
Samouraï
Well-known
Would the Sigma Merrill cameras not produce the best reds? assuming you like the Merrill-Foveon look. Also, do blown out reds occur more often when green or blue are the dominant colours in the frame, like OP's?
f16sunshine
Moderator
Actually you bring up an interesting point Dave.
Although reasonably controlled by a filter directly over the sensor of most cameras, IR "pollution" is still present especially in the new breed of short lens register Mirrorless bodies.
I tried a 55mm IR/UV cut filter on a 50 Planar mounted on my 5Dii. There was an increase in contrast as there should be.
I'm not sure I could say the colors were "cleaner" but, It was easy to see the image did improve in apparent sharpness.
This may explain some part of the issue although, I bet most of it is a result of color filter arrays and sw algos that don't due an adequate job.
Reds are difficult nuff said.
I can't speak more technically to the issue of IR pollution and problems it may incur regarding color correctness and rendering. Hopefully others will chime in.
Although reasonably controlled by a filter directly over the sensor of most cameras, IR "pollution" is still present especially in the new breed of short lens register Mirrorless bodies.
I tried a 55mm IR/UV cut filter on a 50 Planar mounted on my 5Dii. There was an increase in contrast as there should be.
I'm not sure I could say the colors were "cleaner" but, It was easy to see the image did improve in apparent sharpness.
This may explain some part of the issue although, I bet most of it is a result of color filter arrays and sw algos that don't due an adequate job.
Reds are difficult nuff said.
I can't speak more technically to the issue of IR pollution and problems it may incur regarding color correctness and rendering. Hopefully others will chime in.
semilog
curmudgeonly optimist
What is the evidence for this assertion?
Lens-for-lens, the characteristics of the sensor coverglass and the sensor intrinsic frequency response are the the sole determinants of a body's susceptibility to IR "pollution."
Lens register should make little or no difference. Coverglass thickness, and transmittance/absorbance (a function of thin film interference coatings plus coverglass thickness) should be the critical parameters. There is no technical reason I can think of that mirrorless bodies should have more trouble with respect to through-the-lens IR versus DSLRs.
Am I missing something here?