A photograph is expected to capture the exact image seen by the human eye at that moment in time. This is an accurate expectation IF the lighting conditions are acceptable and have balanced exposure, meaning no major shadows, no major bright spots contrasting with dark areas etc. Most of the time, photographs have a good exposure; however, there are a lot of photographs captured where the light is acceptable for the human eye but not for the camera. This results in a “dark image” that does not resonate with the memory of the photographer.
One of our most challenging tasks in image editing and optimization is the “brightening” of dark images. This is a very tricky issue due to the subjectivity of the assessment and the individual’s knowledge of photography. The determination of whether or not an image is “too dark” depends upon:
- Exposure of the image: photos taken indoors with low lighting, outdoors with shadows, or with excessive flash typically suffer from image darkness issues.
- Calibration of the monitor being used for viewing the digital image (for example, your laptop can show a very different image in comparison to a stand alone LCD monitor). In many instances, the image is actually fine but the monitor isn’t calibrated correctly. Run a basic test on monitor calibration.
- An individual’s preference on the degree of brightness/contrast (for example, professionals and amateur photographers have zero tolerance for brightening images whereas average point-and-shoot photographers are only interested in the main object in the image looking nice and bright).
- Lastly, in brightening an image, “grain” and “highlights” will develop that may or may not be acceptable to the viewer. Typically, a professional or an amateur has a very good understanding of highlights and grain.
A picture is worth a thousand words and here I am writing an essay…stop! See the image below. We tested this image on 20 customers and asked them “which one do you like the most?” 2 chose A, 5 chose B, 10 chose C and 3 chose D. Different customers chose different images as the “ideal” one. Professionals leaned towards “A”, amateurs and semi-professionals chose “B or C”, the average point and shoot photographer chose “D”. In our standard offering for scanning and image enhancement services (19 cents for negatives and 24 cents for slides), our graphics artists spend about 3 minutes on each image, which will take you from “A” to “C”.
I chose an image with poor lighting as an example primarily because this type is by far the most prevalent example of a “dark image”. People tend to underestimate the amount of light required indoors and assume that because they can see perfectly, then the clarity will be mirrored in a photo but this isn’t the case. This is why you always see so many lights and reflectors when a professional photographer is shooting.
|Click for a larger view.|
So how do we manage this conundrum: How bright is bright enough?
First of all, all our monitors are calibrated using the latest software and Spyder3 hardware from Datacolor. We also use only diffused ambient lighting in our work spaces. Such lighting is optimal for a graphics work environment.
Secondly, our artists do not use the generic “brightness” tool in an image editor software product such as Adobe Photoshop because doing so brightens ALL the pixels in an image without discretion, which typically ruins the image. The brightness tool will enhance the dull regions in an image, but completely over brightens the brighter regions (see window in image above, this artifact is known commonly as “blown highlights”). This is actually a good reason to use ScanCafe – because it requires the manual touch of a human to make the appropriate level of correction. ScanCafe artists use the “levels” and “curves” tool in Photoshop which allows us to selectively adjust the brightness of pixels based how bright or dark they are to begin with. We prefer to lean on the side of not brightening images too much since the change is irreversible.
Hope this helps in understanding the challenge with dark images. I would like to thank my COO, Dr. Naren Dubey, for contributing to this post. We welcome other technical questions – just let us know what we can do to help.