Canon 5D2, Canon 24-70 f/2.8 at 66, f/13 (no idea why), ISO 400, 1/180 sec. Reprocessed in the current LR (with gratitude for the Show Spots feature) and into PS for NR and a bit more tonal contrast. Full frame.
Beautiful composition and processing! I notice a little gradient banding (is there another term?) in the image. I sometimes have problems with this processing artifact, too. Anyone have suggestions for correcting that? I imagine some sort of Gaussian blur filter? In any case, thanks for sharing this one. It’s really compelling!
Thanks, Jeff! I’d love an answer to that, too. The problem with blur is that the banding is being brought out in areas where you want to really dig out tonal detail. It’s improved a little with later sensors but still a problem. I have a total solar eclipse HDR (maybe 6-7 exposures, tracked) with the Canon 1DX 2 where it limited bringing out tonal detail in the outer corona. I’d love to reprocess the subs with any new tools. (Also have the accurate star BG that was shot 6 months before when that sky area was in darkness, and an earthshine overlay. It came out good but it’s worth more work if possible. Shooting it was a tour de force.)
Stunning, Diane! I am so taken with the awesome feeling of this sunset I hardly noticed the banding. Being the novice I am, I am also struck by how much there is to learn about capturing and processing “banding” . . . I had no idea. A bit overwhelming but thanks for sharing for your processes and your photo.
Yes, there does seem to be some banding. I noticed it changes location when I move closer or further away. I wonder how much is digital and how much is my “optical systems?” Either way, this is a stunningly beautiful image. Thank you for all the work you put into each and every one.
Thanks @JohnSnell, @Mark_Seaver, @linda_mellor and @jefflafrenierre! I was in a rush yesterday when commenting on banding, and answered the wrong question. The issue with the eclipse was sensor banding revealed by digging too far into dark tones. But Jeff was talking about “gradient banding” which is a much more common occurrence. And it can be exacerbated on conversion to JPEG, which is limited to 8 bits (per channel) of tonal steps. And there could be some further tonal compression of an image being processed for displayed here – who knows?
This image was overexposed somewhat at capture, in an attempt to capture a little more tonal detail in the sky, and darkened in raw conversion (LR Develop module) where there is a little more tonal depth than after conversion to PS, even at “16 bits”. In PS, when further detail was dragged out of the sky (tonal separation or stretching of the histogram), it’s possible that banding could occur. But I opened the JPEG in PS and the histogram doesn’t show the “comb teeth” of tonal banding.
Something we don’t see mentioned often is that monitors and graphics cards have bit depths too. My NEC Multisync is 30 bits (10 per channel) to maximize color gamut. But most monitors are 8 or even 6 bits/channel, for faster refresh speeds with less hardware load. Brightness and calibration might also make a difference in viewer’s perception, more in line with the “tired eyes” sort of thing.