Canon 5D exposure issue
I'm having a somewhat serious issue in my C1 workflow. Due to time constraints, I am forced to shoot simultaneous Raw and Jpeg files. There are often huge exposure variances between the way C1 renders the Raws and how the in camera jpegs look when viewed in Photoshop. Generally my C1 display is at least 1/3 of a stop darker than the corresponding Jpeg in Photoshop. Sometimes the difference is 2/3 or a full stop. The contrast levels are also inconsistent between the two files.
The exposure problems seem to be more pronounced on images with a hot white background. When shooting on grey or black, the exposures are often closer but contrast is always an issue. I won't trifle with the color differences, those are to be expected, but but exposure and contrast should at least be consistent from setup to set up.
Thanks for any help
Canon 5D
Firmware 1.1.0
Picture Style: Faithful
Shooting: RAW & Jpeg (L)
Colorspace sRGB
C1 Pro Ver 3.7.6
Generic 5D profile
OSX 10.4.8
G5 Dual 2.3
The exposure problems seem to be more pronounced on images with a hot white background. When shooting on grey or black, the exposures are often closer but contrast is always an issue. I won't trifle with the color differences, those are to be expected, but but exposure and contrast should at least be consistent from setup to set up.
Thanks for any help
Canon 5D
Firmware 1.1.0
Picture Style: Faithful
Shooting: RAW & Jpeg (L)
Colorspace sRGB
C1 Pro Ver 3.7.6
Generic 5D profile
OSX 10.4.8
G5 Dual 2.3
0
-
hope I can answer this one well without opening a can of worms...
But basically the JPGs in your camera is processed files, the RAW files is unprocessed.
To compare.... Polaroid vs. Film
Polaroid will show up one way, but the film will show up in what ever way you develop the files.
When it comes to contrast this is true and when it comes to exposure.. try checking it your self with lightmeter (not the built in one...)0 -
I'm actually shooting tethered. I can understand some subtle differences from one medium to another, but when these differences approach a stop there is a problem. I'm also concerned that the variance changes depending on subject matter.
I have the feeling that C1pro is ignoring the tagged exposure information and is applying a curve based on the image's histogram. As such, when a large percentage of the image is white, the exposure is changed to produce a more acceptable histogram. I'm basing this in part on the fact that I can see the images being adjusted as they appear. As well, Photo Mechanic displays the Raw and Jpeg files nearly identically.0 -
Open the can of worms !!!!
This is an interesting thread even for a Nikon shooter, some things I have seen in C1 make the imae appear better than say photoshop, others the image is far too bright and I have to adjust down quite a lot.
Perhaps C1 can comment on your guess at the cures adjustment.
An alternative question is if shooting thethered why are you concerned about what the camera displays surely the image on the computer is the important one assuming the screen is calibrated ,etc,etc
?
Lets keep this one going I think that the detail C1 produces from a RAW file is second to none, their algorithms are unique so I am sure they will not ive away too many processing clues0 -
If you need your processed files to exactly match the camera generated JPEG, then your only options are to either use the manufacturer's RAW processing software or a raw converter which is based on the manufacturer's SDK (which I suspect photomechanic is).
You might be able to re-create the general look of your jpegs by creating your own tone curves and saving them as a style. To do this, take a shot of a gretagmacbeth colorchecker, and compare the resultant rgb values for the neutral patches. Create a tone curve in capture one by adding curve points for each patch, and adjust the values so that they match those of your jpeg.
>I have the feeling that C1pro is ignoring the tagged exposure information ...
As far as I know, no raw processing software uses the exposure info tagged on a raw file. Afterall, why should it? Exposure compensation adjusts how the image is captured (how much light falls on the sensor), not how it is processed. (perhaps I have misunderstood you here?)
>....and is applying a curve based on the image's histogram.
My experience with 1Ds files suggests that there are no automatic adjustments based on the histogram occurring by default - the same tone curve is applied to all images as set in the exposure section of capture one. Changing exposure produces a very predictable change in the previews and histograms of my images.
There are differences in the tone curves used by capture one compared to those used by canon. Again, you will have to work out your own compensation factors here. The differences in tone curves are why you see:
>I'm basing this in part on the fact that I can see the images being adjusted as they appear.
What is happening is that the first time an image is imported, the thumbnail shown by capture one is based on the camera-generated thumbnail embedded in the raw file using Canon's tone curve. Once capture one has generated its own preview using its own tone curve, the camera-generated thumbnail is replaced with one that reflects the Capture One settings.
Jason0 -
...and of course shooting RAW only and using the QuickProof algorithm to process JPEGs that are color managed the exact same way the Quality Processed RAWs are will produce exact same results.
These in-camera processed JPEGs are using settings only the Canon program can understand, i.e. Picture Style: Faithful
JPEGs can only be browsed in Capture One and they are not color managed at all.
Cheers,
K C0 -
I was characterizing my 5D today and run into the same issue. When I shot with TTL, it only try to expose so the JPEG and the histogram will turn out exposed correctly while leaving the RAW ~1 stop underexposed (turning the Exposure, Brightness, and Contract all to 0 on CS2 ACR). (The ACR generally need to do some random and sometimes really extreme tuning with the RAW to bring the picture close to correct exposure.) The reading from the camera TTL is also 1 stop under what a Sekonic meter says. In other words, if I want to shot to JPEG, I can trust the camera TTl. If I plan to use the RAW, I have to increase the exposure by 1 stop. Why is that?
The experiment is done with the ColorChecker and I have all the data in a spread sheet in case anybody wants to exchange note.
Thanks for answering.
Jeffrey0 -
Canon (and maybe other) dSLRs base their histogram on the in camera JPEG even if you're shooting in RAW mode. 0 -
Jeffrey - how'd you get the data into the spreadsheet ? D'you have software that reads patches ?
Edmund
-----
I made a special C1 profile for the 5D that lifts the exposure, and confirmed this profile with a second camera; however many people tell me they have no underex problem with C1 and the 5D. I'm beginning to think there may be two different batches of 5D out there.
People who want the lifted profiles are welcome to email me - these are mainly suitable for portraits.
Edmund
edmundronald@gmail.com0 -
Edmund - have you tested your profile under different lighting conditions? That has been the most nerve wracking part of this, the exposure compensation changes dramatically depending on the situation.
In response to the other posts:
I would prefer to use C1pro rendered jpegs always, but often I'm under time limits that don't permit me to render several hundred files. Personally I find the contrast, color saturation, and skin tones as rendered by the camera to be awful.
I would like the display that my art directors are looking at to be a reasonable representation of the files they will receive. Subtle differences in contrast and color sat are acceptable, exposure differences of close to a stop are not.
By the way, I'm using the "Natural" style with the contrast up one click. I have to say, there's nothing natural about the way it handles reds.
I've tried the quickproof jpegs, and they're not acceptable for our use.
My personal solution has been to approach this on a job by job basis. If I feel that there is time, I will process files. If I know that the deadline is extremely short, I will use camera rendered jpegs. Where I run into the most trouble is when I bet on processing files, and end up shooting over 1000. That's about 4 hours of processing, if I'm not shooting another job in the mean time.
Thanks all0 -
Could it have to do with shooting in sRGB instead of Adobe RGB?
I'm assuming you are converting the RAWs to Adobe RGB, and the jpegs from the Camera are in Adobe RGB?
Norbert0 -
[quote="norbertsf" wrote:
Could it have to do with shooting in sRGB instead of Adobe RGB?
I'm assuming you are converting the RAWs to Adobe RGB, and the jpegs from the Camera are in Adobe RGB?
Norbert
If I saw a need, I could try and make a special profile that matches the in-camera Jpegs. Maybe I'll try it at some point. I could also release profiles tailored to certain light conditions.
However for people who really want a perfect match for in-camera Jpegs there is only one bullet-proof solution and that is the manufacturer's software - I think the old EOS File Viewer is the one to use for Canon, at least it used to be for my 1Ds and 1Ds2. DPP had different algorithms at the time/
Edmund0
投稿コメントは受け付けていません。
コメント
11件のコメント