When Is A Mask "Good Enough" in Photoshop CS5

Today's image, obviously, is a series a images taken using continuous high speed shutter.  Before we get into it I'll talk a little bit about Nikon's (or any manufacturer's) continuous mode of shooting.  It kind of cracks me up when some friends say with certainty that their camera can shoot at six, seven or nine frames per second.  In the case of today's image(s) that's most likely the case.  But, if you're in low light and your shutter speed is one second, you can shoot at a maximum of --- one frame per second.  If you're shooting a seven shot bracket, and your "normal" shutter speed is 1/100 of a second, you'd have 1/12 of a second as your longest shutter speed.  Add up 1/12th +1/25th +1/50th +1/100th + 1/200, +1/400, and 1/800th and you'll find you've used up just about that whole second.  Start at 1/50th and that one second is long gone before you get that six frames per second over.  A friend of mine who has a D3 (not a D3X) found that out the hard way when he heard my D300 start to labor during a seven shot bracket.  He wanted to show me how fast his D3 was.  I told him to use the same Aperture I had, in Aperture Priority Mode.  His camera went click, click. click, click,  click,    click,        click, same as mine had.  There are some rules of physics that you can't break no matter how good your camera is.  But, back at today's image, it's a burst of exposures in high speed continuous mode with plenty of sunlight.  It was clicking away as fast as the mechanism would go.  Seeing as the Nikon D300 can run at six frames per second, today's image should be about a one second interval.  Putting the sequence together in Adobe Photoshop CS5 is easier than you might suspect.  To learn how it was done, hit the "Read More".

The first thing to do was to bring the six frames into CS5 from Adobe Photoshop Lightroom 3 as stacked Layers (Photo/Edit In/Open As Layers In Photoshop).  This will save the hassle of opening each image and doing any sort of Copy and Paste to get all the images into one document.  They will go into CS5 in the order they were taken.  From back to front (from lowest number in the sequence to the highest number).  That puts the rear image on the top of the Layer Stack.  Just the opposite way we'd need them.  So, the first thing (in CS5) that needs to be done is to rearrange the order so that the closet image is at the top of the stack and the stack descends back to the furthest image being on the bottom. 

Because I was panning as a followed the rider the edges of the shots don't lineup.  They look a little like a stair step sequence going from low at the left to high at the right.  We'll fix that later.  To figure out where the alignment falls we need to pairs, starting at the two bottom Layers and turning off all the others.  To get into the general area I lowered the Opacity of the upper Layer to about 50%.  The ruts above the rider remained in one place.  I used that as a key and moved the upper image up to match ruts.  One I got it close I brought the Opacity back up to 100% and changed the Blend Mode to Difference.  With the Move Tool (V) selected I used the left, right, up and down cursor control keys to nudge the Layer until the area over the ruts went back.  To check I went one click further in each direction to confirm the placement. 

The next step was to make a Layer Mask.  Using the Quick Selection Tool (W) the target cursor was pulled down through the rider on the lowest Layer.  Once the selection was cleaned up I right clicked inside the selection.  This brought up a dialog box and Save Selection was clicked.  This produces an Alpha Channel.  If you are using Adobe Photoshop Elements 9 you can go to Select/Save Selection and you too will have an Alpha Channel.  It's the way selections are saved, if Adobe lets you see the Channels or not.  This process was repeated up through the entire set of Layers.

From there, each Layer was given a Layer Mask based on the Selection that had been made for that particular Layer.  This showed each iteration of the rider.  Each Mask was then painted with Black using the Brush Tool (B) beneath the rider, revealing the rider's shadow.  This is actually the trickier part of the masking. 

One thing that happened was that the top of the lower left was lower than the head of the rider in the third frame.  Therefore the canvas had to be expanded.  Rather than going to the dialog box (Image/Canvas Size), an easier method (when the dimensions are unknown) is to use the Crop Tool (C).  With the entire visible image selected, grab the upper middle "handle" and pull up out of the image rather than down into the image.  As soon as the mouse click is lifted additional canvas is added and the rest of the image revealed.  I had to do this twice as I was building today's image.  The second time I had to go up and to the right.  I just grabbed the handle in the upper right corner and pulled.

The "stair step" was handled using Content Aware Fill.  The first try resulted in riders heads in the lower right corner.  A Layer Mask was put on the image masking out the riders.  That gave a better result, but not great.  The next attempt was taking small chunks and doing CAF to about a half dozen pieces to make it look right.

The image was brought back into LR3 and tones, brightness, colors, sharpness, clarity, etc brought up there.  The upper left and lower right were darkened to put the emphasis on the riders to complete the scene.