Introduction
In my first post on this subject I resolved the image problems I had related to scanning and getting the color space wrong. In the end I made C-prints to evaluate the differences and figured out which colorspace got me the best results in a print as well as online.
Looking at even the best print I thought it was not as 'bright' as I expected. I decided to make another round of experiments. This time I would attempt to 'brighten' the files in various ways in order to see what happened when they were turned into C-prints. Was the original image the best that could be done with the limited color gamut of the C-print process or could I eke out more? I also wanted to know what would happen to a less than optimal image if I let the color lab make their 'lab optimizations'.
Methodology
 |
Original Image
|
I used the same base digital file which was my favorite: Epson color space to sRGB conversion. This master image was a 3200 dpi scan of an Ektar 5x4 negative. Once I applied the 'enhancements' I then decimated the image to a 600 dpi 5x5 image with 1/8 inch white borders. I would add the text of each file name to the respective border to keep track of each image.
The enhancements used three different manipulations and I made 2 'strength' versions of each for a total of 6 files. I then added a 'lab optimized version based on my earlier AdobeRGB image which was not as bright as the best image which was the Epson to sRGB converted file. My reference would be the original Epson to sRGB image from my last round of tests. To summarize...
- +10% Shadow Brightness (Photoshop Elements)
- +20% Shadow Brightness (Photoshop Elements)
- +20 Brightness (Photoshop Elements)
- +40 Brightness (Photoshop Elements)
- Black Level -1.0 (ColorPerfect)
- Black Level -0.5 (ColorPerfect)
- Lab Optimized AdobeRGB Image
Results
When the prints arrived I was initially hard pressed to see much difference and my first impression was I was up against the color gamut of the print process. However I decided to scan them and see if the scanner could measure a difference. I used the same scan settings for all the prints. After scanning I selected the majority of the image excluding the white borders and screen captured the histograms in each case. I used the average pixel value to evaluate any changes in overall brightness in the luminosity histograms. There was indeed some differences in the results.
I will summarize below. I have grouped pairs of scanned images according to the type of enhancement used. Below each image is the histogram for the selection of that image.
 |
+20% and +10% Shadow Brightness
|
 |
+40 and +20 Brightness
|
 |
| Black Level -1.0 and -0.5 |
Finally the original and lab optimized versions.
 |
Lab Optimized and Original
|
I summarized the Histogram results in the table below.
Analysis
A few observations can be made. - Every enhancement showed measurable differences in the scanned image. So adding brightness brightened the images.
- With respect to the original image every enhancement brightened the print.
- The mean to median difference changed really only in the brightened shadows enhancement. This is to be expected as brightening the shadows only will compress the lower values towards the median reducing this difference.
- The lab optimized version increased the brightness but modestly. This might be conservatism on the part of the lab as different images may have negative impact in terms of blowing out the highlights for instance.
Evaluating Negative Impacts of the Enhancements
To see the impact in the highlights I sampled and area at the top center of the image where the sunlight is quite intense. These I cropped and put together. I took the brightest image 'black -1.0' and '20% shadow brightness' as illustrative of the impact. Obviously boosting the shadows has a smaller impact on the highlights. While these crops show the impacts of course the overall effect on the image must also be assessed.
 |
Highlight Assessment
|
Comments