Tom Batinich
Senior Member
I think some are missing your point.
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
5d exposure whacked?
- Thread starter wilkman
- Start date
thw
Veteran Member
-------------------------------------------
See the colors of my world in:
thw.smugmug.com
See the colors of my world in:
thw.smugmug.com
Doug Pardee
Veteran Member
I downloaded those sample JPEGs and examined the median luminosity values for a 9% circle in the center. Here are the results:
35mm f/2: 125 (20.5%)
35mm f/5.6: 139 (25.8%)
135mm f/2: 126 (20.9%)
135mm f/5.6: 131 (22.7%)
The big difference is obviously on the 35mm lens. The difference is a factor of 1.26, or just barely under 1/3 stop. 1/3 stop is the minimum metering and exposure increment on the 5D.
My conclusion, then, is that it does seem to be entirely the vignetting. The camera is metering for the center of the image (center-point AF with evaluative metering) and is getting consistent light readings. The subsequent exposure is giving consistent results in the centers of the images. The images appear lighter or darker depending on the amount of vignetting present.
35mm f/2: 125 (20.5%)
35mm f/5.6: 139 (25.8%)
135mm f/2: 126 (20.9%)
135mm f/5.6: 131 (22.7%)
The big difference is obviously on the 35mm lens. The difference is a factor of 1.26, or just barely under 1/3 stop. 1/3 stop is the minimum metering and exposure increment on the 5D.
My conclusion, then, is that it does seem to be entirely the vignetting. The camera is metering for the center of the image (center-point AF with evaluative metering) and is getting consistent light readings. The subsequent exposure is giving consistent results in the centers of the images. The images appear lighter or darker depending on the amount of vignetting present.
Thx for your continuing interest. I have to agree, it does look like vignetting is the culprit.
I was expecting it on wider glass, but I never dreamed it would be an issue at 135mm. Here's less of a test shot, more towards real world... both are exposed the same, f2 & 1/250 vs f5.6 & 1/30. The white tiles in the center area measure pretty close between the two shots, but man do they have a diiferent look when it comes to overall exposure. I have to bump the f2 shot about .5EV in post to get an overall equal look.
f2
f5.6
I've been shooting bball side by side, 5d with the 135, and d2x with the 85. I've been puzzled why the 5d's plain as day iso superiority hasn't been coming through as much as expected (shooting wide open, as I'm always forced to do in the dismal gyms). I think I've been loosing a half a stop with the 5d/135mm due to this vignetting......
It's almost like the wide open and FF don't go together, as far as reaping the full benefits of a larger sensor gathering more light.
What lens is long enough that there should be zero vignetting? I guess this is what I have to try to make sure vignetting is the only issue in play.
best, mark
Lemming51
Forum Pro
It's in the EXIF of each sample.
First pic:
Exposure Time: 1/125 sec
F Number: 2.0
Exposure Program: Aperture Priority
Exposure Bias Value: -1
Doug Pardee
Veteran Member
You're always going to have cos^4 light fall-off at the least. With a digital sensor, this tends to be further exaggerated by the microlenses which are less sensitive to light coming in at an angle.
Mechanical vignetting is a mainly a matter of lens design. Although in some cases the addition of filters can make it a whole lot worse.
Are we looking at the same exif? Neither opanda or DPP show any EV comp that I can see....
Yea, it's kinda frustrating when folks diss you without seeming to understand what you're talking about..... hey, but such is life
best, mark
Thx. Got my 5d for xmas .... only 1200 clicks in ... but I'm pretty good at disecting a cam's performance (if I say so myself
I've seen the recommendations to use centerweighted averaging, but
I'm not really having a problem with metering being off, or knowing how to compensate on any particular shot. The 5d meter, my d2x meter, even my trustworthy Minolta VI meter all agree. The problem in my mind is that vignetting seems to be causing progressively higher underexposure as a lens is opened up, even for lenses that aren't considered wideangle. Not just vignetting in the corners, but a darkening of much the frame away from center.
If my 5d is normal, it seems like I'm going to have to create a list of the EC needed for each lens when wide open. Maybe my 5d is atypical ... I kinda hope so.... I don't think having to manage EC as a function of f-stop is all that cool.
Interestingly, my d2x shows slight overexposure wide open (relative to stopped down, constant exposure). It's much slighter than the 5d underexposure, but it is there. I'm left kinda wondering more, about the interplay between image circle and sensor size.
best, mark
best, mark
Lemming51
Forum Pro
Could be some kind of conflict in the way your data is written to your JPGs and what my IE plug-in (ViewEXIF v.1.9) expects. I haven't noticed it misreading data before, but there's always a first time.
What is your light source?
assuming you are in North America..
it would be 60hz..
so the first shot @ 1/125, you are risking catching the light in shall I say bad lighting..
while the 1/15 you are sure to catch the full power. hence better exposure.
assuming you are in North America..
it would be 60hz..
so the first shot @ 1/125, you are risking catching the light in shall I say bad lighting..
while the 1/15 you are sure to catch the full power. hence better exposure.
Jonas B
Forum Pro
According to SLRGear.com the EF135/2.0L vignettes up to a 1/4 of a stop when tested with a cropped sensor. It doesn't seem unreasonable that this number increases quite some with a FF sensor.
2 cents
--
Jonas
2 cents
--
Jonas
Thx Jonas,
Why i didn't look to dslgear.com to begin with, I dunno ....
Great tests IMO.
Here's the link to their site .... the 135mm's vignetting on FF.
I'll say it again ... never expected vignetting to be an issue at this focal length.
best, mark
Hi Tim, in our kitchen here.... did your wife pick out your wallpaper too?
I could live with something else for sure....
best, mark
you would think especially with evaluative metering, that if vignetting is less light being read by the sensor, that the exposure value would adjust to some extent, no? If it wasn't a lens specific cause, but instead imagine that vignetting was caused by perhaps using too narrow of a lens hood, wouldn't you expect metered exposure value to change?
Thx Jonas,
Why i didn't look to dslgear.com to begin with, I dunno ....
Great tests IMO.
Here's the link to their site .... the 135mm's vignetting on FF.
I'll say it again ... never expected vignetting to be an issue at
this focal length.
best, mark
Doug Pardee
Veteran Member
He was using center-point AF, so the evaluative metering took the center as his subject.
The metering sensor doesn't cover the entire image anyway, so it wouldn't pick up the edges.
Keyboard shortcuts
- f
- Forum