This post will be way shorter as I’ll just give a quick walk through of one of my favorite shots being put together. Sadly I think I deleted my old .psd file so I’ve lost the original editing, but I’m pretty basic with my editing anyway so I did a quick redo. The original JPG in all its high res glory, though can be found: HERE.
First, I start with shooting the 30 shots I end up using with my 5D2 + 135L. You’ll want to focus on the subject, then switch to manual focus or if you’re using back button focus you’re fine. Basically you’re doing a panorama, but with a shallow dof. So, you’ll want to use a camera + lens combo with a shallow dof and get in just close enough that the model fills your frame, but still leave some space around the model to make things easier for you. You can get REALLY insane with this if you want to do close up portraits. But it also gets tricky because of movement and perspective.
Note: use manual exposure so your exposure stays exactly the same throughout the pano.
This ends up giving you kind of a pseudo bigger sensor. Your dof stays the same, but you end up with a wider field of view. If that makes sense. You definitely want to shoot farther out than you want the final shot to be because of tricks of perspective and stuff, otherwise you’ll end up with a screwed up pano stitch no matter how good you get at this. Don’t ride the edge. Give yourself lots of buffer.
The collection of shots:
So the way that I do the initial editing is to do my WB and tone sliders and batch copy + paste them so all of the pictures have the exact same editing done to them. Then I export them as 2048 or something jpgs and work with those. I don’t really need to do a full res photo merge because even with smaller JPGs, the combine res of the pano of 30 images is pretty damned big.
Next, go into photoshop and do a photomerge with the JPGs
You’ll end up with something like this:
See what I said about giving yourself a lot of buffer? Perspective is a helluv a thing. The correct way to do these stitches is to use a rig that keeps the sensor plane perpendicular to the plane you’re capturing… but that’s too much work. Just give yourself room.
Usually from there if there’s nothing to fix, I’ll do a stamp visible layer (alt + cntrl + shift + e) to make it easier to work with. For this specific instance I had to fix some of the lack of flare on the tree. Since I’m covering such a wide range with such a tight lens, some parts of it just didn’t have the sun hit the lens like the rest, which resulted in darker, non-washed out area of the tree. To fix this I just washed the blacks out a little and painted it in. (Honestly with the original I linked above, I spent way more time on this getting it to match the rest of the surrounding tree. But for the purposes of this walk through, this will do). See the next photo:
Then I add my selective color and orange/yellow radial gradient and a quick contrast bump w/ an S curve:
Quick shot of my layers:
Then just crop out the extra stuff. Or fill it in using Photoshop… but if you know me, you know I’m too damn lazy for that. So, crop it is: