Tim Gray
Offline
• • •
Upload & Sell: Off
|
 p.1 #1 · filter test and info - debate? | |
Further to this thread:
https://www.fredmiranda.com/forum/topic/417503
I though I'd respond with a separate thread since I'm taking the opposite view - which is that all things considered, post processing is at least as effective, if not more so than the use of split or gnd filters. (With the exception of PS's HDR tool, of which I'm not a fan - but maybe in the next version?).
My reference shot is the following which was built from 6 separate shots, each 1 stop difference and blended using the ctrl, alt, ~ technique. This shot could not be captured with any kind of filter.
http://www.timgrayphotography.com/galleries/200604wcoast/slides/J3M8391-8394-cat16.jpg
The shot in the referenced post by Jose which is absolutely a fine image is, in fact, a perfect candidate for a bracket.
From a field workflow perspective, regardless of whether a filter is used or the intent is to bracket the first step is to determine if the shot requires such treatment. Often this ends up being the result of the first frame and seeing extensive clipping in the histogram. Alternatively (likely as with Jose's shot) you just look at the scene and know you're going to have to do something. My camera is a 1d2, so I can bracket up to 3 stops on either side, but even if you are limited to 2, or don't want to bother turning on the auto bracket option, it's a simple matter to switch to manual and just cycle through the shutter speeds until you have one shot that doesn't clip the shadows and one that doesn't clip the highlights (exceptions for specular stuff) and all the shots in between at whatever interval you want. My contention is that you can fire off a 2 or 3, or 4, or 5, or 6 shot bracket much faster than opening your bag, installing the filter, adjusting the transition and taking the shot (and then putting the filter away).
Using a filter gives you one choice for the shot based on the one filter (or combination) you decide to install, 2, 3, 4 stops? hard or grad? How many do you intend to carry? - if more than 1 then your potential set up times are further impacted since there will be occasions when you set up the wrong one (or combination) - or you could cycle through the same number of exposures I would in a bracket to nail the range. I suppose you could use just the one filter and then bracket. 
Post processing a bracket of Jose's shot is absolute simplicity. It's a matter of a couple of seconds to layer the shots and apply a gradient mask. The gradient mask gives you an infinite (ok not infinite, but very large) number of options as to how abrupt you want to make the transition, and where you want it to be. This 2 second technique perfectly mimicks a split filter, with the added flexibility of being able to contour the mask to perfectly match the zone you need to tone down.
My raw workflow is based on Pixmantec RSP (the Adobe acquisition thing is another issue entirely) and RSP is sufficiently efficient that the incremental processing of the brackets occurs in the background and overall, the conversion process is not significantly lengthened, particularly if you're selective in which of the day's 400 shots you want to actually process. This is due in large part to the fact that I'm relying on the bracket rather than manual curves, levels and exposure to manage the dynamic range, so raw adjustments are at a minimum.
As Peter (pappy) notes, a filter isn't particularly effective unless the zone of transition is straight - it doesn't have to be parallel to the frame, but hard or soft, it doesn't do a great job for "U" shaped compositions with tall foreground objects framing a lower centre, or vice versa.
The one instance where a filter will be more effective than a bracket is when there is motion in the frame. But this is relevant only if the motion crosses the transition boundary. In Jose's shot if there was a bird crossing the frame in either the top 1/3 or bottom 2/3 that would not be a problem for a bracket (depending on how wide you wanted the gradient to be). Note that in the scenario where you have tree leaves/branches blowing in the wind a filter (if selected for the bright background) is going to make the leaves too dark - so there's really no net improvement over a bracket - neither technique is going to solve the problem (this is where the 1 raw shot processed twice is a life saver - but you still need to know how to process brackets to make it work).
The only other scenario open for debate is shooting hand held. I assume most shots using grads are shot from a tripod. But even for hand held brackets, there are relatively simple ways to deal with registration issues.
Finally, I'm totally open to the experiential justification for shooting with splits - I use the same experience justification for using a 1d2 rather than a 30d etc - I just love the feel - I certainly don't need 8 fps. (the "I hate post processing" is a similar experiential - subjective - argument that I can't refute). But I do debate the suggestion that splits are faster or more effective than brackets.
|
