Well, I started to play with my D300 and wanted to see the difference between 12 bit Lossless compressed vs. 14 bit Uncompressed. Now I have been reading that there is almost no difference or almost imperceptable. So I tried this myself and the answer from my experience is as follows.
First of all, I was almost flattened by the 25 Meg file size of the 14 bit, uncompressed image. When I looked at it compared the the 12 bit lossless compressed, indeed it looked quite similar." Why would anyone want to shoot 14 bit uncompressed?", I thought to myself. I thought to myself, "There's got to be something to this, else why would Nikon design it that way."
So, I tried a couple of things in NX on my 14 bit NEF image and saved the settings, and saved it as a high quality JPEG.
I then applied the very same settings in NX to the 12 Bit NEF image, and save it too as a JPEG.
The image was shot of a couple of white pillows against a wooden headboard.
Excuse me for the large files, but that's where you'lly be able to just.
Amound other things, pay particular attention at the boundary were the pillow ends and the headboard appears. I don't think I have to tell you which is the 14 bit and which is the 12 bit.
Absolutely. The first shot is sharper on my monitor (a 23-inch Apple Cinema). I'm not sure it's that much sharper to justify the resource intensive penalty the 14-bit file imposes.
Great comparison louhand. Thanks for posting these images.
I thought perhaps there was some camera motion on that 2nd one, but if not......than indeed the gradation of a 14bit file is evident. If I were taking a shot I considered to be VERY important I certainly would use it. I have no issues with the "resources'. To me it's all about the final results..........this is not the first evidence I've seen about the contrast and color gradation advantages........it's there.....one must just decide if it's important to them or not............or if the scene is challenging enough to warrant it's use.
Nice comparison, i've tried a few 14-bit shots myself in challenging circumstances. Results where great, i have no idea what the difference with 12-bit is tho, since i did not compare.
gfiksel wrote:
The second image is blurred indeed. Hard to make any comparison.
Let's assume that the second image was blurred. Even so, how do you explain the following:
Look at the black dots in the wooden headboard at the top of each image.
The difference in the first and second image of those same dots, factoring in the "Blurry Factor", is nowhere as dramatic at the boarder between the pillow and the headboard, which seems to be a huge difference.
All I am saying is that "Post Processing" may contribute to an even bigger difference in image quality. That is we should necessarily judge image quality differences, simply out of camera, but see what effect post processing may have down the road.
louhand wrote:
Look at the black dots in the wooden headboard at the top of each image.
The difference in the first and second image of those same dots, factoring in the "Blurry Factor", is nowhere as dramatic at the boarder between the pillow and the headboard, which seems to be a huge difference.
Blur will greatly reduce micro contrast in an area like that and smear small details like the black dots in the wood out of existence.
It looks to me, on the transition between the pillow and head board, that the shake was mostly horizontal, hence the double image of the right edge of the pillow and how bad it looks there.
All I'm saying that tests are notoriously difficult even with very few variables. Among those few you want to keep all but one the same. Even then the interpretation can be difficult. Otherwise, welcome to the wonderful world of assumptions
I think 14bit is better for shadow detail recovery and highly editing the histogram and curves of an image (able to push it further with out destruction)
?
