Low Light Experiment and Data

barjohn

Established
Local time
4:54 PM
Joined
Dec 22, 2006
Messages
198
As promissed I set about conducting a low light experiment and test. The results may surprise some of you. First, the set up. I set the camera on a tripod using a cable release. I focused a 28MM f2 lens set to f2 on the distant shutters. I took 4 shots in each series. The first four shots are taken with the shutter speed set to AE and starting with ISO 200, then 400, 800 and finally 1600. The next four shots are with the shutter speed set to +1 and again at 200, 400, 800 and 1600. The last four were shot with the shutter set to +2 with the same sequence of ISO settings. All other setting were left the same. All shots were in RAW + JPG so there are two images per shot. The JPG-Film setting was Normal.

The interesting thing you will find is that the ERF files all look nearly alike in the thumb nails. Only the JPGs show the result of the change in shutter speed. How many of you knew that? I would challenge you to identify by quadrant which shots are JPGs and wich are ERF. Use letter across the top A-P and number down the side.
 

Attachments

  • Snap1.jpg
    Snap1.jpg
    127.3 KB · Views: 0
In the attachment below you will see the results of some errors and the interesting difference between the JPG and the ERF files. In the first photo of the newspaper I inadvertantly forgot that I had the ISO set to 1600 instead of 200. The first picture is a JPG and it is over exposed. The second picture of the newspaper is the ERF of the same shot. Notice, it is not overexposed. What gives here? The next pair, I did not use the flash but forgot that my shutter speed was set to 1/125 for flash. The lens was set to f2. The first picture of the calendar is way underexposed but look at the ERF of the same shot. The last two were taken with flash at ISO 200 and f11. The flash overexposes the highlight areas in the JPG but not in the ERF files.

I am going to upload both set of images to my web site. The first will be available at http://www.barjohn.com/phototest1 and the second set at http://www.barjohn.com/phototest2 They are all original files with no editing at all.
 

Attachments

  • Snap2.jpg
    Snap2.jpg
    234.2 KB · Views: 0
One thing you've got to observe when comparing JPEG to raw files is that most raw-file converters make some attempt at automatic image adjustment.

Since the raw-file converter has more bit depth to work with, and a whole desktop computer's worth of CPU power to use, it can often do a better job of "saving" slightly underexposed or overexposed shots than the camera's internal JPEG conversion engine. And if you don't like raw converter's guess, you can disable automatic correction and tweak it manually.

If you're saving JPEG files directly within the camera, the camera's internal processor does all this without giving you any opportunity to intervene, so your initial exposures have to be a lot more accurate.
 
JLW,

You are correct. That is what appears to be happening in Adobe Elements. It is tweaking and correcting the ERF files. When I went to Epson's Photoraw I saw what the unaltered ERF files looked like. And things like my white balance setting being on incandecent during flash photos became evident. Still, the interesting thing is the effect on noise reduction and the fact that the way overexposed highlights in the JPGs are quite controlled in the ERF files and exhibit far lower noise.
 
good info jlw

good info jlw

but please note that if the camera is doing white balance in the analog domain, this will be preferred to WB post processed in raw. See my note to didier in a related thread here in the RD1 section.

jlw said:
One thing you've got to observe when comparing JPEG to raw files is that most raw-file converters make some attempt at automatic image adjustment.

Since the raw-file converter has more bit depth to work with, and a whole desktop computer's worth of CPU power to use, it can often do a better job of "saving" slightly underexposed or overexposed shots than the camera's internal JPEG conversion engine. And if you don't like raw converter's guess, you can disable automatic correction and tweak it manually.

If you're saving JPEG files directly within the camera, the camera's internal processor does all this without giving you any opportunity to intervene, so your initial exposures have to be a lot more accurate.
 
Back
Top Bottom