| Author | Thread |
|
|
10/30/2003 05:17:14 PM · #1 |
Took a heap of photos last night, playing around with some cheap flood lamps I got. Still life sort of set up (though not for that challenge).
Kepot changing settings, trying to teach myself more about exposure in particular (and just spotted jmsetzler's tutorial, which I am reading - great stuff there!)
I took a heap of photos, only adjusting aperture, and this over or under exposing the shots. So, I end up with, lets say, 8 photos, identical except for exposure (as far as I am concerned here, not worried about DOF issues).
Obviously they all look different.
Now, usually I might adjust the RAW file using Capture I and can over or under expose it. Alternately I can adjust the levels once it is in photoshop.
As far as I can see, if I do either:
1) Adjust aperture on the camera
2) Adjust exposure of RAW file prior to making a TIFF
3) Adjust the levels on the TIFF in Photoshop
As far as I can see I end up with exactly the same photo via any of those three routes. I can take any of the 8 photos and end up with the the same as if they were correctly exposed in the camera originally.
Can anyone please explain the differene here to me? I am particularly interested in if there is an advantage to changing the exposure on the RAW file with software, or doing it with levels.
Confusedly yours
Natator
|
|
|
|
10/30/2003 05:37:00 PM · #2 |
The only difference there can be is where your exposure puts the grading of information: thus, if you under-expose, there'll be stuff that's disappeared into the black that you'll never recover - it's not there. The same goes for over-exposing and blown-out areas.
Between these, there's the issue of getting the biggest possible range of information between the 0,0,0 black point, and the 255,255,255 white point.
How you get that out of your camera is up to you - but if you're having to compensate at any point for over/under-exposure, that infirmation is still restricted, but just spread over the full range of black/white
ed
Ed
|
|
|
|
10/30/2003 07:50:58 PM · #3 |
There will certainly be a point with either over- or under-exposure where you will begin to lose information, as natator has stated. Within the range where you are neither clipping highlights or losing the shadows, you have 12 bits of resolution, of which you're going to "throw away" 4 bits when converting to an 8-bit-per-channel file. You should see that if your exposure is toward the low end, and you compensate by bringing exposure up, there will be more noise. With the 10D, whoever, the noise levels are very low especially at ISO 100, so the effect won't be drastic for moderate exposure compensations.
|
|
|
|
10/30/2003 10:17:33 PM · #4 |
just on this subject...
when you take a photo, and have a real tight curve in the histogram (ie, using only 1/2 of the available area.)
and when put it on the computer and all the co0lours are unused down the bottom and top end, is there any settings in camera/lens to expand that area?
it seems a waste having 255 possible gradients and only getting use of 150 or so of them.
normally its full, but just those off times it doesn't fill the graph. |
|
|
|
10/30/2003 10:24:28 PM · #5 |
Originally posted by cvt_:
just on this subject...
when you take a photo, and have a real tight curve in the histogram (ie, using only 1/2 of the available area.)
and when put it on the computer and all the co0lours are unused down the bottom and top end, is there any settings in camera/lens to expand that area?
it seems a waste having 255 possible gradients and only getting use of 150 or so of them.
normally its full, but just those off times it doesn't fill the graph. |
If you take a photo that has very limitied dynamic range, that's what the histogram will look like. Try taking a picture of an evenly illuminated gray card and you'll see a very thin distribution.
Interpretation of the histogram does take some getting used to, it definitely needs to be taken in context with the subject matter.
|
|
|
|
10/31/2003 12:43:34 AM · #6 |
Anyone know a good tutorial page, explanation, or whatever on the histogram?
I'm also looking for one on curves, particularly the curve tool in Photoshop, and how to use it effectively (my current policy is toi gran the curve, drag it around for a while, stuff it up completely, hit cancel, and then ignore it).
|
|
|
|
10/31/2003 02:10:08 AM · #7 |
|
|
|
10/31/2003 02:29:34 AM · #8 |
I read an article recently which concluded that you should always expose as light as possible, without overexposing. So, get the histogram as much as possible to the right.
The article is here
It has to do with the signal to noise ratio of image sensors and the higher sensitivity for the highlights. So you will get less noise if you follow this guideline.
As for your experiement : you will only get 8 identical pictures if all the values still fit within the dynamic range/historgams. If you have a wide range from real black to real white it will not work. You will loose info on both the underexposed and overexposed pictures.
|
|
|
|
11/01/2003 08:19:13 AM · #9 |
Thanks guys, great articles!
I now have an infinitely better understanding of curves and histograms, and will be using them to far more effect in future.
Thanks :)
|
|
|
|
11/01/2003 08:30:08 AM · #10 |
thanks for the links!
very helpful!
|
|
Home -
Challenges -
Community -
League -
Photos -
Cameras -
Lenses -
Learn -
Help -
Terms of Use -
Privacy -
Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 08/29/2025 03:43:09 PM EDT.
