01          In this lesson we will learn how to do Color Keying inside After Effects. We will go through the different stages of keying out the green screen or a green background from a footage. Open up the Color Keying 01 after effects file from the projects file folder. In the scene we have our actor in the center of a green screen along with some elements outside of the green screen that we don't want to show in the scene.

02          Before applying the green screen filter we will try to reduce the amount of work that it has to do, like in this case to get rid of the variations in the colour of the green background and the objects outside of the green screen. So in order to do that we will draw a loose Garbage Mask feathered at 20 pixels around the area which we want to get rid of. We will then animate it over the length of the shot so that we can make sure that we get only the things that we want to keep in the shot.

03          Once we are ready with the Garbage Mask around the areas of the footage that we want to keep in the shot, we will then Pre-Compose the layer with the name 'Keying Filter' and with the setting of 'Move all attributes into the new composition'. Now we are ready to apply the Keylight 1.2 effect onto the layer from the Effects and Pre-sets menu as this is the effect that we are going to use to pull out the green matte from the footage.

04          As the first step make sure that the 'Un-premultiply Result' option is checked which will prohibit the background colour to get mixed or blend in with the edges of the foreground object, as otherwise we might get a green haze around the edge of the object. The other important step is to set the colour that we want to get rid of from the 'Screen Colour' option. Select colour picker tool from the screen colour option and select the green colour close to the edges of our actor.

05          After the last step we got rid of the most part of the green screen at the back of the actor but if we look closely we will notice that there are still some parts in the background with the green screen information. So in order to get rid of those parts we will increase the value of ‘Screen Gain’ which actually controls the amount of screen colour being removed to make the screen matte. But increasing it too much will lead to the background showing through the foreground and edge detail will be destroyed. Therefore a better way of improving the matte is by using the Clip Black and White options.

06          Set the ‘Screen Matte’ option back to ‘100’ and switch to 'Combined Matte' inside the 'View' options in order to see the matte result. In this view black area denotes the transparent area on the footage while white area shows the colour channels. If we look closely we will notice that the background is not 100% black which doesn’t make the background completely transparent. So to improve the quality of our matte we want the whites to be white so that we can't see through it and we want the blacks to be black so that it becomes completely transparent.

07          Dive inside the 'Screen Matte' parameters inside the 'Key Light' options and play with the Clip Black and Clip White option in order to balance them for a better matte. We always need to be careful while working on the balance of Black and White as we may loose the detailing over the soft edges of the object like hair and fur. Increase the 'Clip Black' value to 25 in order to decrease the grey spots on the black area and decrease the 'Clip White' value to 85 in order to decrease them on white areas.

  

08          Switch back to ‘Final Result’ option from the ‘View’ parameter. Pulling out a default colour key from the footage might leave us with the background colour spill around the edges of the object. So in order to reduce the spilled colours on the foreground object we need to pick the normal colour pixels on it from the De-spill Bias Colour picker tool.

09          In case we end up destroying the edges of the foreground object, we can use the ‘Clip Rollback’ option to put back the lost details in the edges by putting up back the pixels comprising the edges. But for this example we will leave this value to 0 and will try to remove the black halo around the edges of the main actor with the help of Screen Shrink / Grow option. This option will enable us to shrink down the alpha in the screen matte to produce a black border less output as we reduce the value of Screen Shrink / Grow parameter down to ‘-1.0’.

10          Despot parameters help in coagulating similar regions on the matte, for example in this case we are getting black spots in the middle of the white matte of the actor which we can easily blend into the surrounding white areas with the help of these values. Increasing the ‘Screen Despot Black’ value to 2 will help us in removing those isolated spots of Black in the White matte of the actor. Similarly increasing Screen Deposit White value will remove the isolated spots of white in the background.

11          If we look closely at the arm of our actress we will notice a thin black line which we would like to soften up. We can always soften up the edge in one go but that might erode the pixels that we need to maintain the crispness on the edges. So in order to focus on a particular area we will isolate it with the help of mask and separate layers. As we are going to focus on the arm of the actress we will draw a garbage mask around it and animate it following its position on different frames.

12          In order to avoid the artefacts of using multiple layers always make sure to soften the edges of the masked layer by increasing the value of ‘Mask Feather’ option to 10. Select the ‘Keying Filter’ comp layer and duplicate it while renaming the duplicate as ‘Base Matte’. Dive inside the ‘Mask’ parameters of the composition and hit ‘Invert’ in order to invert the Mask and get the rest of the footage back that doesn’t need modifications.

13          Now if we zoom into the region where we had applied the feather we will notice that we are getting transparency issues. So in order to get rid of this we will create another copy of the composition with the name of ‘Core Matte’. Place it underneath the two compositions while deleting the mask from it in order to fill the areas with transparency issues.

14          But as we are going to deal with our black border problem we will be pulling out the dark edges from the Keying Filter composition, all we are going to see is the Core Matte composition underneath. This means that we need to shrink the edges of the ‘Core Matte’ composition more than the value of ‘Keying Filter’ composition. Therefore select the Core Matte comp and increase the value of ‘Screen Shrink Grow’ option to ‘-3’ and inside the Keying Filter comp increase the same value to ‘-2’ while also increasing the ‘Screen Softness’ value to ‘3’.

In this lesson we will learn how to project a still image onto 3D Geometry in After Effects. We can then introduce some camera move into the scene to make it look like we are actually in the 3D Space. Irrespective of the limitations in the camera movement with this trick, we can create some dynamic camera moves which can bring life to the boring still images.

01          Create a new composition by dragging and dropping the footage of the Street provided in the footage folder onto the ‘New Composition Button’. This will create a new composition depending on the resolution and size of the ‘Street’ footage. Rename the footage layer as ‘Reference’ and make a duplicate of it with the name ‘Projection’.

02          Now we will create a 3D geometry in the scene for which we will bring in a new ‘Solid Layer’ (White) and rename it as ‘Floor’. Apply a 'Grid' effect on to the solid layer in order to make the layers at the back visible. Change this Solid Layer into a 3D Layer and place it flat in the scene according to the road in the footage, by rotating it to the value of 270 degrees in the ‘X’ axis.

03          Create four copies of this solid layer and place them according to the three walls and roof in the scene. Now in order to match the perspective of the Geometry to the Footage we will create a new '24 MM Camera' in the scene which will take us closer to the perspective of the footage. Make sure that we match-up the corner grid lines of the layers to the corners of the street in the footage. We can use the Orthographic Views (Front, Side and Top) for a better result.

04          The new camera will help in projecting the ‘Street’ image onto the 3D geometry along with the Light Source. So now we will require a ‘Point’ light with the ‘Cast Shadows’ option enabled, in order to make the 'Projection' footage to cast shadows onto the 3D Geometry. Select the ‘Projection’ footage and convert it into the 3D Layer. And place the ‘Point Light’ and the ‘Projection’ footage according to the Position of the ‘Camera’ by copying and pasting the Position values.

05          Now we have all the three layers, Camera, Light and Projection Footage on the same place. So now in order to make the Projection Footage work as the footage for the projection we will bring it in front of the Camera view by moving it in Z Axis a little bit. Bring down the scale of the footage in order to match the size of the composition while changing the rotation values to match the angle.

06          Now we will set the 'Projection' footage to project itself onto the geometry using the Light Transmission function inside the Material Options. Increase the 'Light Transmission' value to 100 in order to project 100% of the light from the 'Projection' footage. But in order to project the footage as it is onto the 3D geometry we have to cast shadows from the layer while disabling its visibility. So in order to do that we will change the 'Cast Shadows' parameter of the 'Projection' footage to 'Only'.

07          Right now the 3D Geometry layers of the street scene are being affected by the lights. But in order to make this projection effect work properly we have to make these layers invisible to light. So in order to do that we will turn off the 'Accept Lights' and 'Cast Shadows' options for each 3D Geometry layer. Turn off the Grid Layer effect from each 3D Geometry Layer and the Reference Layer as well.

08          We will notice that we have started getting a blurry view of the footage when we orbit through the scene. This happens because we are casting a shadow of the 'Projection' footage onto the 3D Geometry for the Projection to work, and After Effects uses a Shadow Map of the footage to cast shadow. So in order to overcome this problem we will increase the Size of the Shadow Map to '4000' from the Advanced options inside the 'Composition Settings' window.

09          Now if we orbit through the camera we will get a pretty decent detail which will help us in creating the illusion of 3D world. We are using the light to project the image seen through the camera at the point in time where the picture is taken place. Therefore in order to match the perspective of the geometry to the scene we will now link the Projection footage and Light to the Camera before starting to orbit the camera to match the perspective.

10          This will enable us to match the perspective of the scene by adjusting the camera using the Camera Tools as the things that are being projected will not change. We can also turn on the Grid effect of the solid layer while turning off the visibility of 'Projection' footage. Finally turn on the visibility of the 'Reference' layer in order to use it as a reference while matching up the perspective of the scene.

Stabilizing can be defined as a process where motion tracking data from a shot is registered for removing the camera motion from it. In this process an object's movement in the shot is tracked and then each frame is re-positioned accordingly to hold the picture steady. But as the frames are re-positioned in order to hold the picture steady the edge of the picture move into the frame, leaving a black margin where the picture used to be. We can solve this problem by scaling up the size of the footage in order to push out the black margin from the edges of the footage.

01          Before starting the tracking procedure always make sure to go through the footage and look for good a tracking target that stays in the frame till the end. The best Tracking Targets should have good contrast and a clearly defined shape with distinct edges that stays the same throughout the shot. We also need to make sure that our Tracking Targets are not getting obscured by another object in the scene. For improving the chances of getting a good tracking data we will increase the contrast in the shot by applying the Brightness and Contrast effect and increasing the Contrast value to 16.

03          Now we will move these Tracking Features to a good location on the footage like a good contrasting point which remains visible in the footage with a constant shape and color for a longer duration of time. Also make sure that the two Tracking Features are having a fair distance between each other in order to make the 'Rotation' tracking parameter to track and store the data correctly. We will also increase the size of the Tracking Features in order to make them sustain the sudden shift of the pixels that might come due to the jitters of the hand held camera moment.

04          Select both the Tracking Features in the view-port while making sure that the Time Slider is on the first frame of the composition and press 'Analyze' forward. The program will step through all the frames in the footage, moving the Tracking Features in each frame to keep them locked onto the Tracking Targets. This is the data collection phase, therefore finding good tracking data is obviously a prerequisite for a good motion track.

05          Now in order to apply the Tracking Data onto the Target Layer hit 'Apply' and select 'X and Y' as the 'Apply Dimensions'. What After Effects has done to stabilize the footage is that it has moved everything around at single point which has also resulted in the loss of footage. So in order to overcome this problem we will now smooth out this tracking data and will also limit the intensity of offsetting of the footage.

06          In this step we will connect a 3D Camera to the tracking data in order to offset it in the same amount that it was previously been offset. Create a new Null object along with a new Camera and turn on the 3D Layer Switch for the video footage. Copy all the Anchor Point data, Position data as well as the Rotation data from the Video Footage and paste it onto the Anchor Point parameters, Position parameters and Rotation parameters of the Null object. And now in order to complete this step we will link our camera to the Null object which will take away the stabilization.

07          By giving a smoothing expression to the Tracking Data parameters of the null object we can now easily smooth out the jerk and reintroduce some of that tracking data to it so that the shot doesn’t have the jerky movements. Press alt and click onto the stop watch of the Anchor Point parameter and type in ‘smooth(1,5)’ which means that this expression will smooth every 1 second and 5 samples. Similarly repeat the step with Rotation parameters.

08          Now we are left with the elimination of the borders of the footage, which can be done by Scaling up the footage or applying the Motion Tile effect to the footage. Ram Preview the footage in order to see a very well stabilized shot.