Mocha Pro 2019 User Guide

Tracking and rotoscoping are part of almost any visual effects project. For 2D tracking, point trackers are most commonly used, but to get good point tracks requires a mix of experience and luck. You often have to “prime” a clip for optimum tracking using color correctors and other image manipulations. If the point being tracked exits frame, you get into offset tracking, which presents its own set of challenges. If it all fails, you are into hand tracking, which is time consuming and very hard to get accurate.

Mocha is a 2D tracker that requires less experience and luck to be successful with, does not require the image to be primed and is less likely to require a lot of tricks or hand tracking on difficult shots.

In Mocha splines are used for both tracking and rotoscoping. This is a different method from standard 1-point or multi-point tracking tools.

Traditional tracking tools require that you locate “points” that remain consistent throughout the entire shot in order to track movement. This is itself a difficult task, especially when tracking a shot that was not originally designed to be tracked. If you wish to also track rotation, perspective and shear you need even more clear and consistent points to track.

Traditional roto methodology would have you outline a shape with the minimum number of points necessary then either manually move the control points or track the shape with a point tracker to “get it close”. Even when using multi-point trackers to impart rotation and scale to the roto spline, the results are often unusable if there is any perspective change during the shot.

Instead, Imagineer’s Planar Tracker tracks an object’s translation, rotation and scaling data based on the movement of a user-defined plane.
A plane is any flat surface having only two dimensions, such as a table top, a wall, or a television screen. Planes provide much more detail to the computer about an object’s translation, rotation and scaling than is possible with point-based tracking tools. Even as an object leaves and enters a frame, there is usually enough information for the Planar Tracker to maintain a solid track of the object.
When you work with the Mocha tools, you will need to look for planes in the clip. More specifically, you will need to look for planes that coincide with movements you want to track. If someone is waving goodbye, you can break their arm into two planes - the upper and lower limbs. Although not all of the points on the arm sections actually lie on the same two-dimensional surface, the apparent parallax will be minimal.

Mocha layouts are modifiable, allowing you to hide or reveal many parts of the interface.

These layouts are clustered into 4 default types you can build from.

You can add or remove many parts of the Mocha interface, such as:

These can either be access by right clicking the area of the interface and choosing a GUI element to show or hide, or selecting from the View menu.

Any changes you make to a layout will not be saved unless you choose View | Layout | Save Current Layout.

For example if you like the Essentials layout, but would like the Advanced toolbar from the Classic layout:

This will now save the Essentials layout with the new toolbar.

If you have made changes to a saved layout want to revert back to the saved version, just choose View | Layout | Revert to saved.

If you want to revert back to the original default layout, just choose View | Layout | Revert to default.

At the very top of the interface you have the tools that form the brunt of your time inside Mocha.

In Essentials Mode, only a basic set of these tools is shown, to simplify the interface.

See descriptions in Advanced Toolbar above for the rest of the tools.

These controls cover what can been seen or hidden while working in the Mocha viewer.

The timeline controls cover frame range, playback, tracking controls and key-framing.

The top left hand panel contains the tools to manage layers.

The section under the Layer Controls panel contains the properties for each layer.

In Mocha v5 we introduced manual cache clearing to allow you to clear the Mocha cache at the project, render or global level.

You can access the Clear Cache option from the file menu under File → Clear Cache…​

You can check the following options:

Only clear the Global Cache if you are certain you don’t want any of your existing project caches to remain.

In stereo mode you will see 3 buttons in the View Controls next to the clip view drop down on the left:

You can preview stereo work at any time by turning on the 3D button in the view controls.

Clicking and holding on the 3D button will give you a range of stereo view options.

Two new icons appear in the timeline to assist with tracking and roto in stereo:

Both these buttons and their uses are covered in the Stereo Tracking and Stereo Roto sections of the User Guide.

The Mocha Pro plugin supports different types of 360 and Stereo footage via the "Views" drop down:

The first 4 Views options are for non-360 footage:

Choosing one of the 360 options automatically sets your Mocha project to be Equirectangular 360. This will enable VR features:

You can also choose to Swap the Left and Right eye input by checking the Swap Views checkbox.

Some hosts will require different handling for Stereo sources:

As you go through the user guide, you will see sections on how to apply Mocha techniques to your stereo footage where relevant.

The Mocha Pro Plugin for Adobe appears in the Effects menu like every other effect.
Simply apply the effect to the layer you want to work with.

The general workflow for the Mocha Adobe Plugin is as follows:

The Mocha Pro Plugin for Adobe appears in the Effects menu like every other effect.
Simply apply the effect to the layer you want to work with.

The general workflow for the Mocha Adobe Plugin is as follows:

The Mocha Pro Plugin for Adobe appears in the Effects menu like every other effect.
Simply apply the effect to the layer you want to work with.

The general workflow for the Mocha Adobe Plugin is as follows:

If you have a license for the OFX version it will work in any of the OFX hosts listed below.
Keep in mind that while the Mocha OFX plugin is designed to be used in multiple applications, it does not support all OFX hosts.

In many cases some functionality may be possible for unsupported hosts, but there is no guarantee of functionality or stability, so please take care when experimenting!

Inside Nuke, the Mocha Pro Plugin for OFX appears in the toolbar menu like every other effect.

You can also call the Mocha Pro effect from the Tab key by searching for 'Mocha Pro' or right-click and choose 'Boris FX Mocha' > 'Mocha Pro'.

mocha Pro node in Nuke node graph (Footage courtesy of Chris Heuer):

Once loaded into the node graph, simply plug the image node you want to work with into the 'Source' input of the Mocha Pro effect node.

Inside Fusion Studio, the Mocha Pro Plugin for OFX appears in the Tool menu like every other effect.
Just choose 'Boris FX Mocha' > 'Mocha Pro'.

mocha Pro node in Fusion (Footage courtesy of Chris Heuer):

Once loaded into the flow graph, simply plug the image node you want to work with into the 'Source' input of the Mocha Pro effect node.

Silhouette v6 introduced OFX support, so the Mocha Pro Plugin for OFX appears in the nodes menu like every other effect.
Just choose 'Boris FX Mocha' > 'Mocha Pro'.

mocha Pro node in Silhouette:

Once loaded into the tree window, simply plug the image node you want to work with into the 'Source' input of the Mocha Pro effect node.

Inside Vegas Pro, the Mocha OFX Plugin appears under 'Boris FX Mocha' in the Plug-in Chooser dialog for the following effect chains:

mocha Pro node in Vegas Plug-in Chooser:

Once loaded, you can begin with the 'Launch Mocha UI' button at the top of the effect panel.

One your source clip is hooked up to you Mocha Pro Effect, the general workflow for the Mocha OFX Plugin is as follows:

Once you have applied the Mocha Pro effect, you can click on the 'Launch Mocha UI' button to launch the main interface.

This then becomes exactly like working in the standalone version of Mocha, with a few exceptions.
Firstly, you will notice you don’t need to set up a project like in the standalone version. The source layer is automatically loaded and ready to track in the view.
Secondly you don’t need to save out a project file (unless you want to export it). You just close and save the Mocha view when done and the project is saved inside the effect.
For further details on how to use anything inside the Mocha GUI, see the rest of the User Guide!

Once you have tracked layers in Mocha, you can then control the mattes for these layers back in the plugin interface.

Once you have set up layers in Mocha, you can then control the renders for each module back in the plugin interface.
Note that you do need to have set up and tracked the correct layers in order for a render to work back in the host.

You have the following options to render a module back in the plugin:

There are also parameters for controlling the view in Lens:Distortion rendering for VR 360 footage.

See Using Mocha for 360 workflow for more on how to use the VR Lens controls.

You can use secondary clips in the host application to render tracked inserts into your shots.
See the User Guide Chapter on the Insert Module for more details on manipulating and warping inserts.

To use the Insert input from your host application in Insert renders:

Your Insert should then appear inside the layer where you have placed your surface.

Alternatively you can import an image or sequence directly to the plugin:

The imported Insert should then appear inside the layer where you have placed your surface.

In cases where your input source has an alpha channel, you may wish to change the Alpha view inside the Mocha GUI.

You can either turn Alpha off entirely by toggling off the button, or choose from one of the following options:

Auto alpha may be necessary when working with some source inputs in Nuke.

When rendering back out to the host, there are cases where you may also need to premultiply the alpha using the premultiply options in the plugin interface.

In these cases you can choose an option from the 'Premultiply' dropdown:

You can also choose to premultiply using standard premultiply nodes.

We now include a light version of Mocha, called Mocha Hitfilm, as a plugin in HitFilm Pro 2017 onwards. This includes:

However, you can also use the Mocha Pro OFX plugins in the HitFilm interface.

When you load a project that contains the old Mocha VR plugin, you will notice that Mocha VR is labelled as "Legacy".
This is because all Mocha VR features have been rolled into Mocha Pro and a Mocha VR plugin stub is kept to avoid breaking compatibility with your old projects.

When you want to start a new VR project, we highly recommend using the Mocha Pro plugin rather than the legacy Mocha VR plugin, as this compatibility feature may be removed in future versions.

See (360 VR and Stereo Views Workflow ) above for how to set the 360 VR modes in new Mocha Pro projects.

Mocha workflow is designed around a project structure. It is good practice to only work on one shot per project file to minimize layer management and to keep the work streamlined.

When you start the application you are presented with an empty workspace. No footage is loaded and most of the controls are consequently disabled.
To begin working, you must open an existing project or start a new project.

Select the clip to import by clicking on the Choose…​ button to the right of the top line. This will bring up a file browser, where you can select almost any industry standard file formats. Image sequences will show up as individual frames. You can select any one of the frames and the application will automatically sequence the frames as a clip when importing.

Name

A project name will automatically be generated based on the filename of the imported footage, but you can change it by editing the Name field.

Location

Your project file and cache files will output to a directory called “Results” by default. This is created in the same folder your clip is imported from. You can change this using the Change…​ button or using the dropdown box to set a different relative or absolute path.

Frame Range

The range of frames to import. We recommend to only work with the frames you need, rather than importing very large clips or multiple shots edited together.

Frame offset

This is set to the starting frame number or timecode by default. You can also define a fixed frame (You can set a default for the fixed frame in Preferences).
You also have the option to view as Timecode or Frame numbers. If your clip has an embedded timecode offset and you switch to Timecode, the offset will be used in your project.
If you need to adjust this value later, you can open Project Settings from the file menu. See "Project Frame Offsets and Clip Frame Offsets" below for more details.

Frame Rate

Normally this is automatically detected, but you have options to adjust if necessary. Make sure you check the frame rate before you close the New Project dialog.
If you need to adjust this value later, you can open Project Settings from the file menu.

Separate Fields

If you are using interlaced footage, set your field separation here to Upper or Lower. Make sure you check your fields match your footage before you close the New Project dialog. If you don’t set them correctly, you cannot modify them and will have to restart the project.

Remove Pulldown

If your footage has pulldown, set it here.

Mocha Pro supports Equirectangular 360 Footage. To set the project to be in 360 mode, check the '360 VR Footage' checkbox after you import your clip.

When you start a New Project you are also presented with the option of creating a multiview project in the Views tab.

If you check Multiview project you are then presented with the view names and their abbreviated names.
The abbreviated name is used in the interface for the view buttons, but is also used as the suffix for renders.
You can also choose the hero view. By default this is the left. Defining a hero eye determines the tracking and roto order for working in the views.

If you are using Top/Bottom or Left/Right footage combined in a single frame, select an option from the Split Views drop down:

If you want to define separate streams of footage for the stereo views, you can add additional footage streams view the Add button below the initial clip chooser.

If you forget to set up Multiview when you start a new project, you can set it in the new Project Settings Dialog from the File menu.

Once you are in Multiview mode, you will see a colored border around the viewer based on the current view you are in.
This is to help artists to identify which view they are currently in without having to refer to the buttons

You can switch between Views by pressing the corresponding L|R buttons in the view controls, or using the default 1 and 2 keys on the keyboard.

You can swap views or change the Split View mapping from the View Mapping subtab under the Clip module:

The Mocha Pro plugin has a slightly different project workflow to the stand alone Mocha applications.

This action loads the footage from the host clip you applied the effect to. It automatically applies the correct frame rate and other clip settings, so there is no need for the standard new project dialog.

After you have done the usual work inside the Mocha Pro interface, you simply close and save the Mocha Pro GUI and then you can control the output from the effect editor interface.

For general guide to workflow with the Mocha plugin, see Using the Mocha Pro Plugin.

For setting up a new stereo project with the plugin, see Plugin Stereo Workflow.

Boris FX introduced the Mocha PixelChooser in BCC 10. The plugin has a slightly different project workflow to the stand alone Mocha applications.

This action loads the footage from the host clip you applied the effect to. It automatically applies the correct frame rate and other clip settings, so there is no need for the standard new project dialog.

After you have done your tracking and/or roto work, you simply close and save the Mocha PixelChooser and it applies any visible layers as mattes back to the effect.

If you will only be working on a section of the shot you can use the In and Out points to set the range on the timeline. Note that the In and Out points affect the range of the Überkey button. You can zoom the timeline to only show you the part between you In and Out points by clicking the Zoom Timeline button.

Frame offsets are important to get right in Mocha so that they export correctly to your target program.

There are two kinds of frame offsets:

For the vast majority of cases the Project Frame Offset is the value you want to adjust for working with data.
The frame offset is usually already set correctly at the New Project dialog stage, but there may be cases where offsets change, such as adding new clip frames.

Only import as much as you need

Working with very long files can be time consuming for the artist and can slow down the tracking as it searches for more frames. Try to only use what you need, and work on individual shots, rather than multiple shots in one piece of footage.

Frame rate, dimensions and pixel aspect ratio are important

Make sure these values match the settings in your compositor or editor, otherwise tracking and shape data will not match when you export it.

If you are unsure which field your interlaced footage is in, import it and check

If you quickly start your project with a guessed field order, you can check to make sure it is correct by using the right arrow key to step through the footage. If you footage stutters or steps back a frame while you’re stepping through, it is probably in the wrong field order, or you may have to set pulldown.

Try to avoid interlaced footage where possible

Interlaced footage is painful to work with. For your own sanity, try not to use it unless you have to!

If you are working on a large roto project you will sometimes need to have more than one person working on the same shot.

When it comes time to export out mattes or do final tweaks you can use the Merge Project option to combine any files that have been used on the same piece of footage.

Simply select the Merge Project option from the File menu, and select a project you wish to merge. You can only merge projects that are the same dimensions, aspect ratio and frame length as the shot you are merging into.

NOTE: You cannot merge projects from versions of Mocha earlier than 3.0.0.

In Mocha V5 we introduced support to import Silhouette .sfx project files. To import a project:

The Silhouette project will then convert to native Mocha splines and appear in the project.

Imagineer’s Planar Tracker technology provides 2D tracking data by tracking planes rather than points.

The key to getting the most out of the Planar Tracker is to learn to find planes of movement in your shot which coincide with the object that you want to track or roto. Sometimes it will be obvious - other times you may have to break your object into different planes of movement. For instance if you were tracking a tabletop, you would want to draw the spline to avoid the flower arrangement in the center of the table — it is not on the same plane and will make your track less accurate.

To select a plane you simply draw a spline around it. You can be fairly loose with your spline — the Planar Tracker is intelligent enough to discard the pixel movement that doesn’t conform to the movement of the majority of the pixels within the shape.

Mocha features two spline types, X-Splines and Bézier splines:

One of the most important concepts to understand with the Mocha planar tracking system is that the spline movement is not the tracking data.

It’s best to think of the splines you draw around objects as search areas. Here’s a breakdown of how the tracking works:

To see this relationship, turn on your surface and/or grid in the viewer after you have tracked something. Scrub the timeline and you will see that the grid and surface move with the spline.

Now select all the points of your spline and move it around the viewer. You will notice that the surface/grid will stay in the same place.

This is because the spline is linked to the track, but the track is not linked to the spline. The spline is merely a search area to tell the track where to go next. It is a common misconception that moving the spline while tracking is affecting the movement of the tracking data. It is not. Moving the spline is only telling the tracker to look in a different place and will not directly affect the motion of the tracking.

This makes the tracker very powerful, as you can move and manipulate your spline area around while tracking to avoid problem areas or add more detail for the search.

You can even unlink the spline from the track entirely so that any planar surface passing under the stationary spline area is tracked and you don’t have to move the spline if tracking starts to go off screen.

With the Planar Tracker you simply draw a spline around something, as shown with the screen below.

In some cases there are parts of an image that can interfere with the effectiveness of the Planar Tracker. To handle this, you can create an exclusion zone in the area you are tracking.

For instance, in the phone example we are using, there are frames where there are strong reflections on the screen. These reflections can make the track jump. So we need to isolate that area so the tracker ignores it. Here’s how this is done:

You can also add as many entirely new layers on top of your tracking layer to mask out the layers below.

This is quite common when moving people, limbs, cars, badgers etc. get in front of the object you are trying to track.

In the Essentials layout, tracking Motion parameters are listed in the Essentials Panel:

In the Classic layout, detailed tracking parameters can be accessed by selecting the Track tab. On the left hand side of the Track tab, you will see two sections: Motion and Search Area.

Understanding the parameters section of the Track parameters is vitally important for obtaining good tracks. Here we provide a breakdown of each parameter and how to use it effectively.

Input Clip

This is the clip you are going to track. By default it is the one currently in the viewer.

Input Channel

When tracking, Mocha looks at contrast for detail. The input channel determines where to look for that contrast.
. Luminance looks for contrast in the light and dark of the image
. Auto Channel looks for contrast in one of the color channels.

By default, Luminance does a good job. If you have low-luminance footage or you are not getting a good track, try Auto Channel.

Min % Pixels Used

One of the most important parameters to look at for tracking.

By default, the minimum percentage of pixels used is dynamic. When you draw a shape, Mocha tries to determine the optimal amount of pixels to look for in order to speed up tracking. If you draw a very large shape, the percentage will be low. If you draw a small shape, the percentage will be high.

In many cases, the cause of a drifting or slipping track is a low percentage of pixels. If you want a more solid and accurate track, try setting the Min % Pixels Used value to a higher amount. Keep in mind however that a larger percentage of pixels can mean a slower track.

Smoothing Level

This value blurs the input clip before it is tracked. This can be useful when there is a lot of severe noise in the clip. It is left at zero by default.

Motion

These parameters control what motion you are looking for when you track:

The main difference between shear and perspective is the relative motion. Shear is defined as the object warping in only two corners, whereas perspective is most often needed where the object is rotating away from the viewer significantly in space.

As an example, if someone is walking towards you, their torso would be showing shear as it rotates slightly back and forth from your point of view.

The front of a truck turning a corner in front of you would be showing significant perspective change.

Search Area

This defines ranges for the tracker to search within

Before performing the actual track, adjust the settings depending on the movement in the clip.

Track the plane selected by pressing the Track Forwards button on the right- hand side of the transport controls section.

Stop the track and adjust the shape if it doesn’t seem to be tracking properly. You may keyframe the spline shape so that it tracks only the planar region of a shape by adjusting the shape and hitting Add Key in the keyframe controls menu. Keep in mind that no initial keyframe is set until you first hit Add Key or move a point with Auto-Key turned on.

The spline should be tracked in addition to the clip being cached to RAM. You can play it back and get an idea as to how the track went. F
eel free to change the playback mode in the transport controls to loop or ping-pong your track.

Another trick you can do to check your track is hit the Stabilize button in the View Controls.

Turning on Stabilize will lock the tracked item in place, moving the image to compensate. In the track module, stabilize view is a preview mode to check your track. Actual stabilization output is handled by the Stabilize Module, explained in the Stabilize Overview chapter.

There may be instances where you have already created mattes for one or more objects in the shot, for example using a keyer or another roto tool that would help you isolate areas to track. You can import such mattes by creating a new layer and then using the Matte Clip setting under Layer Properties to assign it to the layer.

Scrub your timeline

When starting a new project, go through your footage a few times to see what your best options are for tracking. You will save yourself a lot of time by making note of obstructions and possible problem areas in advance.

Use edges

When tracking surfaces you will usually get a much better track if you include the edges and not just the interior of an object. This is because Mocha can define the difference between the background and the foreground and lock on better.

For example, if you are tracking a greenscreen, it is better to draw your shape around the entire screen rather than just the internal tracking markers. In some cases this means you can avoid tracking markers altogether and save time on cleanup later.

When in doubt, ramp up your pixels

You can quite often get a great result with default settings, but if you’re getting a lot of drift, try setting the Min % Pixels Used value higher. The processing can be slower, but you will usually get a much more solid track.

Draw more shapes

Remember you are not limited to one shape in a layer. Use a combination of shapes to add further areas or cut holes in existing areas to maximize your search. If necessary, make an additional layer to track and mask out foreground obstructions before tracking the object you need.

Use the grid while tracking

It’s common to use the surface and the grid to line up your corners after you track, but it can be much more advantageous to set up your surface before you track and leave the grid on to watch for any subtle drift while you are tracking. This way you can stop your track early to fix any issues and spend less time trying to find them later.

Track from the largest, clearest point

In order for Mocha to keep the best possible track, it is usually best to scrub through the timeline and find the largest and clearest area to begin tracking from, draw your shape there, then use backwards and forward tracking from that point.

For example, if you have a shot of sign coming toward you down a freeway, it is usually better to start at the end of the clip where the sign is largest, draw your shape and track backwards, rather than start from the beginning of the clip.

A planar surface does not necessarily have to be flat

We have a Planar Tracker which specifically tracks planes of motion, but this is not limited to tables, walls and other flat objects.

Distant background is considered flat by the camera where there is no parallax. Faces can be tracked very successfully around the eyes and bridge of the nose. Rocky ground, rumpled cushions, clumps of bushes, human torsos and curved car bodies are all good candidates. The key is low parallax or no obvious moving depth.

When in doubt, try quickly tracking an area to see if it will work, as you can quite often trick the planar tracker into thinking something is planar.

In the end, there is no magic bullet

Mocha is a very flexible tracker and will save a lot of time, but you will eventually run into a piece of footage that just will not track. Large or continuous obstructions, extreme blur, low contrast details and sudden flashes can all cause drift or untrackable situations.

If something just isn’t tracking no matter what you try, consider using Mocha to track as much as possible then move to manual work. You can often get a lot more done fixing shots by hand or using AdjustTrack in Mocha rather than trying to tweak your shapes and parameters over and over again to get everything done automatically.

There will be times when tracks can drift due to lack of detail or introduction of small obstructions.  When this occurs, manual refinements can be made by using the AdjustTrack tool.

AdjustTrack is primarily used for eradicating drift by utilizing the four-corner surface area to generate keyframable data to compensate.  It is generally not practical to use it to remove jitter.

To achieve an adjusted track you would ideally line up the surface area where you want to place your insert or lock down your roto. In situations where you don’t require an insertion you should place the corners of the surface area in distinctive locations, as you will need to refer to these locations when you add keyframes to correct the drift.

When you have the Surface where you want it to stay locked and are ready to refine the track, flip over into the AdjustTrack module by hitting the AdjustTrack tab.

Once you select the AdjustTrack tab, a key frame with four reference points is created.

As you play though the sequence you will be able to manually adjust the position of each point as drift occurs.

If your track is spot on, these reference points should line up properly throughout the shot. If you see a Reference Point drifting, that will indicate the track is drifting. Find the frame where the drift is worst and move the Reference Point back to the position it had in the Master Frame and the track will automatically be adjusted based on your correction.

When you perform an adjust track and you begin to move a newly created reference point, you will notice the dashed lines which connect all of the reference points. These lines change in color to represent the quality of positioning of any given reference point. For best results keep reference points away from one another.

When adjusting the track try to always get at least yellow but shoot for green for a more solid adjust track.

The red lines indicate that this reference point position is a poor choice.

The green lines indicate that this reference point position is a good choice.

Often there are times where your reference points are either obscured or exit frame. In AdjustTrack you have the ability to create multiple reference points per surface corner that can be positioned in alternate locations to handle these situations. Simply click the New Ref button to create a new reference point for the selected corner.

You cannot keyframe the Surface — only the Reference Points. The original track and any refinements you make in AdjustTrack cause the Surface to move however.

Every so often a shot will come along that is easier to track backwards than forwards. This is fairly simple when running the tracker backwards, but introduces some rather obtuse concepts when keyframing is involved. This is why there are two “New Ref” buttons provided. If you are working backwards and wish to set a new reference point, you will probably want to use the “<- New Ref” button instead of the forward-thinking “New Ref ->” button.

Because keyframing “thinks” forward, hitting “<-New Ref” will not create new Master Reference points on the current frame, but will go backwards in time, looking for any existing keyframes and set new Master Reference points on the frame directly after. For example, if you decide to create a new backwards reference point at frame 20, a new master reference will be created at frame 21.

Some people may be more comfortable doing this manually by moving the playhead themselves and using the traditional “New Ref ->” button. Others who do a lot of tracking and find themselves working backwards often may find the backwards-thinking New Ref button helpful.

Every Reference Point has one frame in which its initial placement is determined without causing any adjustment to the track. This is called the Master Reference Point; if you step forward or backward in time you will notice the red X change to a red dot.

The red X indicates that this particular frame is the starting point for calculating adjustments. Moving a Master Reference Point will NOT change the tracking data.

Go ahead and experiment - move the Reference Point when it is a red X (a Master Reference Point). You will notice the Surface isn’t adjusted at all. Step forward a frame and move the same point - this time the surface will move because you are now adjusting the track.

By default, the frame in which you create a Reference Point is its Master Reference frame. This Master Reference can occur on a different frame for each reference point. You can change the Master Reference frame by selecting a Reference Point, going to the appropriate frame and hitting the “Set Master” button.

You may set a new Master Reference Frame for all active points by hitting the Set Master All button.

One method for selecting different reference points is to hit the “Next” button.

The next button simply cycles through the active reference points for that frame. More fine-grained control of reference points can be obtained through the Nudge control panel, described below.

Deleting Reference Points is done by selecting the point you wish to remove and hitting the delete key. If there are multiple Reference points on a particular corner, the preceding Reference Point will be extended through your time line until a new Reference point is encountered.

The Nudge section allows you to move Reference points in 0.1 pixel increments, much more fine grained than would be possible by dragging the points manually. You can easily select any active Reference Point by selecting one of the corner buttons in the Nudge section.

If you hit the Auto button, a tracker will attempt to line up the selected Reference Point based on its position in the Master Reference frame. The Search Region Size and Maximum Motion parameters can be set in pixels in the Auto Nudge section.

You can quickly select any corner by using the Corner selector buttons in the Nudge control panel. In the image below, the user is selecting the upper right corner in preparation for nudging operations.

The AdjustTrack tab has a View section for cleaning up your AdjustTrack workspace. Deselecting the Inactive Traces button will cause the display to hide the traces of the inactive Reference Points. This is helpful if you have a corner with numerous Reference Points offsetting it.

Deselecting the Unselected Traces button will hide any Reference Point that is not selected.

Finally, deselecting the Search Area button will hide the Search Region Size (in Yellow) and the Maximum Motion search area (in Pink).

Adjust frames at the peak of drift

When you see a drift, carefully cycle through the timeline and look for where the motion starts to change direction. A frame before this, adjust your drift, then go halfway between your master frame and the adjusted frame to check for any further drift.  If you keep working by checking halfway between each keyframe you set, you will reduce the amount of keyframes required.

A huge amount of keyframes is not a good sign

If you end up with adjustment keyframes on a large amount of frames it may be better retry the track.  AdjustTrack is aimed to help reduce small anomalies and fix drift when a tracked corner has become obscured.  If you are fixing every second keyframe it means you have more than a simple drift.

Most often a good matte requires a combination of both keying and rotoscoping techniques.

Good rotoscoping artists often think like animators, reverse engineering the movements, the easing in and outs, the holds and overshoots of objects, and set their keyframes accordingly.

In general, the fewer the keyframes, the better your mattes will look. Too many keyframes will cause the edges to 'chatter' and move unnaturally. Too few keyframes will cause the shapes to drift and lose definition. Finding the right number and placement of keyframes often comes with experience but there are a few things to keep in mind when rotoscoping.

While you can refine a shape you have tracked to do your rotoscoping, the recommended way is to do a rough shape to track something and then link your roto to that track. This reduces the amount of work required when you are dealing with complex shapes, as you will not have to track and refine each shape as you go. It also means that if you have to retrack something you won’t have to redo a lot of layers. Another reason is reducing the amount of data needed in your project file.

It also helps to remember that your spline shape is linked to your tracking data and not the other way around.

For the following examples you can try out the tools using the BMW tutorial files
available from http://cdn.borisfx.com/borisfx/store/clips/rotoscoping.zip

The Überkey is a powerful tool that allows you to offset the positions of control points without destroying their keyframe data.

Use this tool with care, as it is not setting any keyframes per se, it is offsetting any and all keyframe data on the points you move while it is on. Überkey is very useful, but remember to turn it off again when you don’t need it. Use with care.

Überkey affects only those frames between the timeline’s In and Out point. If you wish to make adjustments to a particular range, set the In and Out points to that range.

You can translate, rotate and scale selected points as a group by using the corresponding tools listed in the toolbar.

Or alternatively, use the transform tool to perform all of the above functions in the same tool:

You can turn on and off individual points in a spline. When they are off, you can still see the points, they can still be animated, but they are not contributing mathematically to the spline. This allows you to have a complex spline only when you need it, rather than having to deal with superfluous points in parts of the shot when they are not needed.

To turn off points, select the points on the spline and hit Shift+Delete. You will see the curve change shapes, but the points will remain.

To turn a point back on, right-click on it and select Point | Activate.

If the Autokey button is enabled, a keyframe will be created when you change a point’s active status.

You can use the movement of the individual spline points to determine motion blur. Any movement in the spline, whether through simple X/Y translation or by shape deformation will cause motion blur.

You can control the amount of blur by changing the motion blur value in the Edge Properties panel.

Although not necessary in this example, note that you can change how mattes are blended in the Layer Properties panel. You may make each layer’s matte Add or Subtract and you can also invert the matte.

Note that this can’t be keyframed and that these settings apply to the entire layer, not to individual splines of the same layer.

In the View Controls, several options are offered for viewing your mattes. The Matte drop down is has options to view all mattes, just the mattes you have selected or no mattes.

Select the Matte button and you will see your rotoscoped object against a flat background.

You may wish to rotoscope against a particular color. Select View → Canvas Color…​ and a color picker will allow you to choose a particular background color

When you have your Mattes turned on, you may choose for the matte to be filled with a color instead of cutting out the object, using Colorize.

You can adjust the opacity of the color fill by changing the blend value to the right of the Colorize button.

The color used by Colorize is derived from the Selected and Unselected properties of the Overlay Colors panel, which can be changed per layer.

This is only a preview and will not affect how your mattes are rendered when exporting.

If you want to get a better view of your roto, you can get a better view by turning off some overlays.

For layer spines:

For points:

Hiding points will still allow you to adjust the layer with the transform tool.

In the View Controls panel, you will find a drop-down menu for selecting the clip to view.

This allows you to view the actual rendered mattes, which can be especially useful when tweaking motion blur. The motion blur you normally see in your canvas is an OpenGL preview and can differ slightly from the actual render.

If you’d like to see what the actual motion blur render looks like, switch to viewing the layer whose matte you wish to see.

Because you can choose specific layers for export when you render, a render pass is created for each layer.

Switch the View Clip drop-down back to your source clip to continue working with that clip.

If you want to draw open splines, you can simply hold shift when you right-click to finish the spline. This will open the shape up.

You can also right-click a spline and choose: Spline | Open/Close Spline

Name your layers

Naming layers is very important to save yourself time later, especially if you are doing a heavy rotoscoping job. Get into the habit of labeling each layer with specific names.

Turn off the splines and just work with points and the matte

If you are working on a tight roto it can sometimes be easier to turn the spline off and just see the matte with the control points. To do this:

If your other view options are at default settings you should now see the matte in the viewer with only the tangents and control points visible.

The Magnetic Tool draws a pixel line that snaps to the nearest edge to where the cursor lies, tracing the shortest path from any previous click.

To use the Magnetic Tool:

(Image courtesy of LateNite Films)

Similarly, if you go off the edge of the frame, the magnetic tool will also switch to Freehand mode,
so you can freely continue the shape.

The Freehand pen tool is exactly like the Magnetic tool, minus the magnetic properties.

This gives you the freedom to quickly draw any line.

You can access the Freehand tool in the same tool drop down as the Magnetic tool, but also any time you
press and hold the mouse/pen down in the Magnetic tool, or go off the image frame.

One you have completed a drawn line, you can either click back on the original point, or right-click.

This will generate an X-Spline that follows the originally drawn line.

(Image courtesy of LateNite Films)

X-Splines generated by the Magnetic and Freehand tools have a fairly high point count to match the subtle changes in the line.

If you want to increase or reduce this detail:

If you need to offset your tracking or roto manually (see below), you can use the Difference 3D view to help align the layer.

To export tracking data to After Effects, press the Export Tracking Data button.

Then choose either:

After Effects Corner Pin Data:

The corner pin data records and exports the 4 point x, y information from either the adjusted track or the raw track.
There are three different types of corner pin exports – two for recent After Effects versions and one for CS3 and older versions of After Effects.

After Effects Transform Data:

The transform data exports x and x positions as well as the scale and rotation for the whole surface.

If you click Save, this will display a file browser for saving the tracking data for use later. By default, the files will take the name of the layer, so for a layer name Track_Layer the export will create a file named "Track_Layer.txt".

If you don’t need to save the export, you can press Copy to Clipboard, and then go straight to After Effects and paste the data. If you are saving to file you will need to open the text file you saved with the data, select the entire body of text and copy it.

In After Effects, load the footage you tracked and the footage/image/composition you wish to apply the transform or corner pin data to.

Select the item on the timeline that is the insert object. Paste the data to the selected layer. You can do this by selecting the ‘paste' option in the edit menu or by typing 'command- v' (Mac) or 'ctrl-v' (Windows).

With the layer’s information expanded you can see either the 4-point tracking data for the corner pin, or the position, rotation and scale information from the tracking is now applied to the insert layer.

If you are pasting transform data rather than corner pin data then you will need to delete the anchor point keyframes to see a result. We export both position and anchor point keyframes so that stabilization or tracking can be achieved. See below.

This example illustrates how to export Basic Motion data to Final Cut Pro or Final Cut Express, and use it to matchmove one clip to another, with translation, rotation and scale.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately before export.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Next, choose either the Final Cut Basic Motion or the Final Cut Distort options.

Final Cut Distort:

The distort option exports the 4 point x, y information from either the adjusted track or the raw track. The points exported are the four corners of the surface.

Final Cut Basic Motion:

The basic motion option exports x and y positions as well as the scale and rotation for the whole surface. The position exported is the center of the surface.

Now click Save. This will display a file browser; you need to select a filename and directory for the files to be saved. By default, the files will take the name of the layer, so for a layer name Track_Layer this export produced a file named:

Track_Layer.xml

Now open the Final Cut project where you want to use the tracking data. To import the XML file in Final Cut Pro, click File | Import | XML…​. In Final Cut Express, click File | Import | FCP XML from iMovie…​

You will now see a new clip in your bin, named 'Mocha distort – layer_name' or 'Mocha basic motion – layer_name'. If your original footage was a QuickTime file, then the new clip in Final Cut will point to this footage. If it was an image sequence, the clip will be connected to the first frame of the sequence, stretched to the duration of the whole clip.

In most situations, you want to apply the tracking data to a different clip in your timeline. To do this, you can copy and paste the data from the imported clip onto any other.

Drop your imported clip into the timeline, then select it and press Cmd-C to copy it to the clipboard. Now select the clip you want to apply the tracking data to and press Opt-V (Paste Attributes). You will see a dialog asking you to choose which data to paste.

Depending on the kind of data you exported, tick either 'Basic Motion' or 'Distort' and untick all the other boxes.

You should now see the clip following you tracked.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Next, choose 'Motion basic transform(.motn)' or 'Motion corner pin (.motn)' and click 'Save' to save the file.

Inside Motion, you can either choose to open the exported file as a new project or drag it into an existing project. You will see a Group containing the footage you tracked as well as a blue solid called "Surface".

Then take the following steps to insert your footage:

This section explains how to export tracking data in a format readable by Apple Shake, how to import the data into Shake and how to use it for match move, corner pinning and stabilization tasks.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Next, choose 'Shake Script (*.shk) and click 'Save' to save the script to disk or 'Copy to Clipboard' to simply copy-and-paste the data into Shake.

Now let’s use that data in Shake. To load the tracking data in a file into an existing Shake Script follow these steps:

In the File menu, select ‘Add script'.

Navigate to the Shake script file you exported and select it.

If you have the data on your clipboard, simply press Ctrl/Cmd+V or right-click and choose 'Paste'.

Three nodes will now appear in your script: Stabilizer, MatchMoveLayer, and CornerPinLayer.

At this point you have successfully imported your tracking data from Mocha into Shake. But what exactly did you import? Let’s start with the MatchMoveLayer layer.

The MatchMoveLayer is used to apply the tracking data to a foreground element that you wish to matchmove to a background. It consists of two inputs, the background and the foreground. The foreground element will be the element you wish to apply the match moving data to. The foreground is connected to the left side input of the node while the original tracked background element goes into the right hand side.

The next node is the CornerPinLayer node. This node has a single input and works just like the left input of the MatchMoveNode.

It takes your foreground element and applies the scaling, rotation, and translation data to it and prepares it for compositing into your next layer.

The last exported node is the Stabilizer Node. Just as the name implies, it takes all of the exported tracking data and uses that data to stabilize your input clip.

With this node, for example, you may apply this data to the background element you tracked to make for an easier composite.

Now that you have successfully exported and imported your data into Shake you are ready to continue working on your project.

This section explains how to export tracking data in a format readable by Nuke, how to import the data into Nuke and how to use it for match move, corner pinning and stabilization tasks.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Next, choose Nuke Ascii (*.txt)' and click 'Save' to save the script to disk.

In Nuke, append a tracker to the background clip, e.g. by selecting the node and pressing tab, then typing 'tracker' and finally selecting the 'Tracker' node in the list of nodes.

In the tracker properties window, select the animation submenu button for Tracker 1, then choose File | Import Ascii…​ in the menu.

On the import Ascii dialog, press the folder button next to the file name, browse to the file you exported, select it, click 'Open' and then click 'OK'.

Repeat these steps for tracker 2, 3 and 4.

This section explains how to export tracking data in a format readable by Fusion, how to import the data into Fusion and how to use it for match move, corner pinning and stabilization tasks.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Next, choose 'Blackmagic Fusion COMP Data (*.comp)' and click 'Save' to save the script to disk.

In Fusion, open the .comp file, then drag the tracker node into the right view.

Now import the clip that you want composited onto the background and tie the output of the clip to the input of the tracker node.

In the settings of the tracker node, select the 'Operation' tab and select either 'Corner Positioning' or 'Match Move' to composite the insert clip on top of the background. Notice that you can switch 'Position', 'Rotation' and 'Scaling' on and off for different effects.

This section explains how to export tracking data in a format readable by Autodesk Inferno, Flame, Flint, Smoke and Combustion.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Select 'Autodesk IFFFSE Point Tracker Data (.ascii)' as the format, and save this to a file. Next select the 'Autodesk IFFFSE Stabilizer Data (.ascii)' export and save this to another file.

Importing into Combustion

Let’s look at how we use the data. In Combustion, create a layer with the foreground graphic and set the layer shape to 'Four-Corner'.

Now select all four corners of your layer, enable the 'Tracker' tab and select 'Import Tracking Data.'

Import the single Ascii file with the stabilizer tracking data.

Activate the insert layer visibility and you will see that the insert layer is now tracked to the background element, even though the image is not sized correctly yet. Select all trackers and set the mode to 'Absolute' to resolve this.

If you prefer to import your data one point at a time you can instead select one track point and import the Ascii file with the corresponding tracking data. Remember to switch to 'Absolute' mode once all data has been imported.

Importing into Inferno, Flame, Flint or Smoke for Corner Pinning

Firstly, we will set-up a corner-pin composite in the Action module, to reproduce the basic compositing operation.

Enter Action and delete the default Axis and Image nodes.

Create a new Bilinear Surface. Tap on bilinear1 then press ~ to view its settings. Go to Surface then click the S button next to Track.

You should now be in the familiar stabilizer module. For this example we are doing a corner-pin so we will need to use all four trackers.

Tap Tracker 1 then press 'Imp' under 'Track Y'.

Now browse to the corresponding file, e.g. xxxx_top_left.ascii. The files correspond to trackers as shown below.

You should see the marker for Tracker 1 move into the correct position.

Repeat the process for the other three trackers, making sure that you use the correct files as shown in above. You should also mark each tracker as Active if it is not already.

Tap Return to return to Action, and you should now see your finished corner pin.

Importing into Inferno, Flame, Flint or Smoke for Stabilization

An alternative use for the tracking data is for stabilization. You can use the Stabilizer module to do 1-, 2-, 3- or 4-point stabilization. In this example, we will do a 1-point stabilization to stabilize for position only, using the center point.

Enter the Stabilizer module. Click the Imp button under Track Y, as for the Corner Pin.

Select the _center file, in my example this is PDA_center.ascii. You should now see the key-frames loaded and be able to process

If you want to use more points to stabilize zoom, rotation, affine and perspective moves, just load the corner tracks as described in the Corner Pin section.

To import data into SCRATCH requires SCRATCH v7 or later.

To import the data into Assimilate, do the following:

You can export either Corner Pin or Center Point data out to any Boris FX plugin that supports motion tracking data. These include effects such as the BCC Corner Pin effect and Witness Protection.

The files are import only, so you can’t copy them to the clipboard. You can save the data as a text file and import it.

Importing the track is as simple as locating the Motion Tracker section of your BCC plugin and clicking either the "L" button or selecting "Load…​" from the Import-Export dropdown, depending on the plugin you are using.

This section explains how to export tracking data in a format readable by Avid DS.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Select 'Avid DS Tracking Data (*.fraw)' as the format, and save this to a file. This stores the x/y coordinates of the four surface points defining the track in eight .fraw text files using the following naming convention:

…​._R1x.fraw Lower left X coordinate

…​._R1y.fraw Lower left Y coordinate

…​._R2x.fraw Lower right X coordinate

…​._R2y.fraw Lower right Y coordinate

…​._R3x.fraw Upper left X coordinate

…​._R3y.fraw Upper left Y coordinate

…​._R4x.fraw Upper right X coordinate

…​._R4y.fraw Upper right Y coordinate

Where "…​." is replaced by the name of the exported layer.

On your DS v10.x system, create a Tracker node and open the Animation Editor for it. In the left window check the little Blue Animation box to the left of R1x, R1y, R2x, R2y, R3x, R3y, R4x, and R4y. Now right-click "R1 Tracker Y" and select "Import tracking coordinates". Navigate to the folder containing the eight FRAW files that Mocha created and DS will load them. If you have an older version of DS then you will have to load each of the eight FRAW files individually.

From HitFilm 4 onwards we have introduced Mocha transform and corner pin support, so you can now export Mocha tracking data directly to a HitFilm Composite Shot file.

Exporting tracks to HitFilm 4 is very similar to exporting HitFilm Camera solves or Shapes:

Importing is done via the Compositing section in HitFilm 4:

To import data into the Quantel requires Boris FX’s TrackImport plug-in for Quantel.

You can import tracking data into Quantel systems by using Boris FX’s TrackImport plug-in.

Track an object in the usual way, and use AdjustTrack to correct for any drift if necessary. The exported tracking points will be the four corner points of the Surface, so you should position these appropriately.

Press the Export Tracking Data button on either the Track or AdjustTrack tabs. Select 'Quantel Corner Pin Data (*.xml)' as the format, and save this to a file.

In the Quantel system, select the TrackImport plug-in, and use the plug-ins file browser interface to select the tracking data file to import. Then click 'Settings' and choose 'Tracker' in the settings window and click 'Save'.

The tracking data can now be used in a manner identical to tracking data generated within the Quantel system.

To export tracking data to the Good Spirit Graphics MochaBlend plugin, click the Export Tracking Data…​ button on the Track module, AdjustTrack module or from the File menu.

You can export the tracking data by either saving it to file, or copying to the clipboard:

To import the tracking data into the plugin, you start by selecting an available Data slot and then either paste from the clipboard or open the exported file:

Once imported, you can then adjust your scene to fit the data. It is important to make sure you check the warnings and correct any relevant items before setting up your rig:

You can then go ahead and tweak individual settings in the MochaBlend plugin. See MochaBlend documentation for further information on working with tracking data in the plugin.

Exporting stereo Track data from Mocha is the same as exporting in mono mode, however when you are in multiview mode you can choose the view you want to export.

When your roto work is complete and you would like to export your mattes, choose the menu item File → Export Rendered Shapes…​ to be guided through the rendering process.

The render dialog will give you options to choose either the currently selected layer, all visible layers or all the layers in the project.

Options include:

Colorize output

Choose whether you want to render grayscale mattes or by their layer fill color.

Export to

The export options for Image Sequences and QuickTime movies.

For Image Sequence you have the following options:

Choosing QuickTime movie will bring up the standard QuickTime export options when you click OK.

The render dialog also allows you to render to single footage streams or a combined stream via EXR.

If you have chosen to render the current view or separate footage streams, the view abbreviation will be suffixed to the rendered export automatically, so there is no need to define this in the file name.

Choose a shape (not a layer) and select Export Shape Data…​ from the File menu. A dialog will show with a drop-down containing 3 different saving options.

Choose the target application and hit Save. The data going into the file is not binary, and is shown in the dialog so that you may copy and paste it directly into a text editor if you prefer to work that way.

Because of differences in the way Splines are handled in the applications (Combustion and Flame handle splines differently), maintaining accurate keyframe interpolation between our software and the other applications requires that the exported shapes have a keyframe on every frame. This is not a bug but required to ensure your mattes look right once they’ve been imported into your compositing application.

The Mocha shape plug-in will import the following data into After Effects:

If you wish to export a single layer, select it before proceeding with the export.

When you are ready to export, select Export Shape Data button. On the dialog that opens, choose if to export the selected layer, all visible layers or all layers.

When you have made the right selection, click Copy to Clipboard, then switch to After Effects.

In After Effects, import the same footage and ensure that the frame rate and pixel aspect ratio are the same as those used when generating the shape. You can change these values in After Effects by selecting File | Interpret Footage | Main…​.

Bring the footage into a composition, then select Edit | Paste to add the shape effects to the composition. Each shape exported will come across as its own plug-in effect.

In the plug-in controls you can change the following parameters:

Blend mode

You can choose to change the blend mode from the one assigned to the shape by selecting Multiply, Add or Subtract from the dropdown menu.

Invert

When checked, inverts the matte.

Render soft edge

If you have created the shape with feathered edges but wish to switch these off, untick this parameter. Note that if you have not created feathered edges this will have no effect.

Render type

This setting allows you to choose between various render effects. The default is the 'Shape cutout' which uses the matte to cut out the corresponding area in the background footage. 'Color composite' will apply a single color to the area within the matte, useful when wanting to preview the positions of multiple layers. Note that the Opacity setting affects this color fill, allowing you to blend it with the background footage. 'Color shape cutout' combines the effects of the previous two.

Shape color

Defines the color of the color fill applied when selecting the 'Color composite' render type.

Opacity

Defines the opacity of the color fill applied when selecting the 'Color composite' render type.

Exporting shapes to Adobe Premiere Pro CC is very similar to exporting shapes to After Effects

Importing is just as straightforward:

The Mocha shape plug-in will import the following data into Final Cut:

If you wish to export a single layer, select it before proceeding with the export.

When you are ready to export, select Export Shape Data button. On the dialog that opens, choose if to export the selected layer, all visible layers or all layers.

When you have made the right selection, click Save and select a destination to save the XML file to, then switch to Final Cut.

In Final Cut, import the XML file by CTRL-clicking in the project asset window.

Select 'Import', then 'XML…​' and finally select the XML file that you exported.

Accept all the defaults in the 'Import XML…​' dialog.

Your Mocha shape sequence will now appear in your list of project assets.

To use the shape to composite the rotoscoped object over a new background, simply place the Mocha shape sequence in a video track above the background sequence.

Additional controls

If you want access to the actual matte, individual layers of the matte or control that affect the appearance of the matte, double click on the Mocha shape sequence to reveal the two main sequences it consists of - the original sequence and the ContourSequence.

Double clicking the ContourSequence will reveal the individual layers that the ContourSequence consists of. Dragging a layer into the viewer and selecting 'Controls' will give you access to controls of that layer, as shown below.

Invert

When checked, inverts the matte.

Soft edge

If you have created the shape with feathered edges but wish to switch these off, untick this parameter. Note that if you have not created feathered edges this will have no effect.

Render mode

'Luma' will cause the layer to render itself in the RGB channels, whereas 'Alpha' will cause the layer to render itself in the alpha channel of the generated sequence.

Matte color

Defines the color of the matte being rendered.

HitFilm 3 Pro has introduced mask support, so you can now export Mocha shape data directly to a HitFilm Composite Shot file.

Exporting shapes to HitFilm 3 Pro is very similar to exporting HitFilm Camera solves:

Importing is done via the Compositing section:

You can export a Roto, RotoPaint or SplineWarp node straight to the clipboard or to a .nk script by choosing Nuke Roto Paint (*.nk) from the Export Shape Data dialog.

To export shapes to Nuke:

You have the option of exporting Basic Roto data which bakes the keyframes, or Transform and Shape data which separates the tracking data from the manual keyframes.

The latter makes the data less heavy and is only supported in Nuke 6.2 and above.

The SplineWarp node exports each layer as a joined set of splines with the spline keyframes separate from the tracked data (which is set in each curves transform).

For example if you only have 1 tracked layer to export, Mocha will export that layer to SplineWarp as two joined splines in A.

You can then modify the second spline for the warp, or delete it and choose another.

To export shape data to the Fusion shape format, click the Export Shape Data…​ button in the Track module or from the File menu.

You can export the shape data by saving it to file or by copying it to the clipboard:

To import the shape data into Fusion, either paste directly into the Fusion Flow View or open the comp file from the file menu.
The Mocha layers will come in as separate nodes into the Flow View.

To export shape data to the Good Spirit Graphics MochaBlend plugin, click the Export Shape Data…​ button in the Track module or from the File menu.

You can export the shape data by either saving it to file, or copying to the clipboard:

To import the shape data into the plugin, you start by selecting an available slot in MochaBlend and then either paste or open the data file:

Once imported, you can then adjust your scene to fit the data. It is important to make sure you check the warnings and correct any relevant items before setting up your rig:

You can then go ahead and create the splines under the Objects settings. See MochaBlend documentation for further information on working with shape data in the plugin.

To export shape data to the Silhouette FXS shape format, click the Export Shape Data…​ button in the Track module or from the File menu.

You can export the shape data by saving it to file:

To import the shape data into Silhouette, either start a new session or import into the existing session from the File → Import…​ → Silhouette Shapes menu option:

Exporting stereo Shape data from Mocha is the same as exporting in mono mode, however now you can choose the view you want to export.

Nuke Roto exported as "Export All Views" will paste to Nuke as a combined roto node. If you would prefer the nodes to be separate, export using the different views instead of checking the all views option.

You can export a clip to an image sequence or QuickTime format by choosing "Export Rendered Clip…​" from the file menu.

Options include:

Clip

Choose the clip you want to export.

Frame range

The range of frames you wish to export. If you choose to export the full range but have not rendered all your frames, the next drop down, "Revert to clip" will be used.

Revert to clip

Choose how to export frames that have not been rendered. If you choose None or the current clip to export, black frames will be exported for non-rendered frames.

Save channels

By default, this will just export the flattened render (Color), but if your render has alpha you can choose this also.

Export to

The export options for Image Sequences and QuickTime movies.

For Image Sequence you have the following options:

Choosing QuickTime movie will bring up the standard QuickTime export options when you click OK.

The render dialog also allows you to render to single streams or a combined stream via EXR.

If you have chosen to render the current view or separate footage streams, the view abbreviation will be suffixed to the rendered export automatically, so there is no need to define this in the file name.
'''

The Camera Solve module is used to take planar tracking information and convert it to 3d space.

This is particularly useful if you cannot get a good track on the plane you need, as you can track other planes in the shot and use them to give you track in 3D space instead

As we are dealing with 3D calculations rather than 2D planar projections, the workflow is slightly different to a usual planar track.

In order to get a good camera solve you must first identify what type of track it is. Mocha recognizes three types of camera situations:

Once you have identified the type of camera motion you can start tracking your shot.

Camera solves now also work with Multiview. Like with tracking, a stereo camera solve is designed to be as similar to the Mono process as possible.

New additions to the camera solve for Stereo are:

You can then export out to left and right views, or for Standard FBX, you can export full Stereo cameras. The full stereo camera solve FBX presently works in Maya.

All other exports can only be exported as a single camera view and the solved nulls.

You have two steps for camera solves:

The resulting track drifts or jumps

This can be from the solver not having enough reliable information from the tracks.

The Insert module is where you choose the image you would like to insert into your tracked layer. You can import a still frame or a moving sequence. Once imported, the Insert module provides a comprehensive range of tools for matching this new image to the original background layer. The skill here, naturally, is to make the newly imported image look like it was in the original shot all along.

This is where you choose the Input Clip or background layer and the Insert Clip or foreground layer, and optionally a separate input clip with an alpha channel for compositing.

Background Clip

By default Mocha Pro selects the last clip that was tracked as the input. If you want to change the input, just select a different clip from the pulldown menu.

Insert Clip

This clip mirrors the Insert Clip inside Layer properties, i.e. if you change one, the other also changes.

By default mocha Pro selects None as the foreground input, expecting you to make a choice of your own. To choose an insert select the Import button and use the file browser to locate a still frame or a file sequence that you would like to appear over the tracked background clip.
In the Mocha Pro Plugin, you can choose an 'Insert Layer' placeholder that reads from layers back in the host timeline to render in Mocha.

Here you can select the name of the clip to be output. Once a clip has been rendered it is automatically named Composite_<clipname>.

To choose a different name for the output clip select the New…​ button and choose a new name.

The remainder of the controls in the Insert module dictate both how much of the insert is displayed and how it is displayed during the course of the shot.

The Source region of interest (ROI) can move anywhere on the insert image, including outside it. It is the area of the insert that you would like to be displayed in the tracked layer. As you adjust the ROI, the insert is scaled to fit the surface. There is also a Reset button to reset the ROI to the corners of the insert image. By default, the aspect ratio is locked. Uncheck Lock Aspect Ratio to allow arbitrary ROI changes.

To adjust the ROI aspect ratio to match that of the background image given its current position and shape in the background image, click on Fit ROI to Surface. Conversely, the Fit Surface to ROI button moves the surface position to match the aspect ratio of the insert image, given its ROI settings. These controls will work best if you are in a frame where the insert is as front-on as possible. For them to work correctly, the pixel aspect ratio of both the input clip and insert clip must be correct. A common issue is that if the resolution of either clip was not recognized when the clip was imported, it is assigned the PAL camera type. This often does not give the correct pixel aspect ratio. To check this, switch to the Camera tool and select each clip in turn, checking the appearance on the screen to make sure that each clip appears correct on the screen. If not, change the Film Type. If you know that the pixel aspect ratio of the clip is one (square pixels) select Custom as the Film Type and set the pixel aspect ratio to one.

The number fields are positioned in the menu to relate to the edge of the ROI that they adjust. So, to reduce the height of the insert ROI at the top of the frame, decrease the value in the top ROI number field by dragging or highlighting the current value and typing in a new value. Similarly, to reduce the height of the insert ROI at the bottom of the frame, increase the value in the bottom number field by dragging or highlighting the current value and typing in a new value. The same applies for the left and right edges of the frame with the left and right number fields.

Warp uses a warp mesh to distort your insert. If you turn on "Show Mesh" in the warp tools you can use the dropdown to set the level of detail in the mesh and then warp the grid points in the viewer to adjust your insert, as well as the outer yellow lines to bulge or pinch the edges.

This is especially useful if you have to insert something that does not look completely planar, like a t-shirt logo.

If you want to add a soft edge around the edges of the insert, use the Feather controls. There are separate controls for the left, right, top and bottom edge widths, which are between 0 and 1, where the value 0 indicates no edge and 1 means that the edge covers the whole of the insert.

If you switch on Lock, all four edges are locked to the same edge width. H Lock and V Lock apply lock separately to the left/right and the top/bottom edges respectively.

This section controls a level of corner pin control on top of the adjusted track. This feature is particularly useful for curved surfaces. It allows the corners of the surface to be used in the same way as the warp control points (described below) – by aligning the newly imported insert with a region of the tracked image independently of the blue surface contour, which follows the adjusted track.

The offset region is drawn in yellow underneath the existing blue surface. You can use the new points either by dragging the sliders to increase or decrease the value of the coordinates, or by highlighting the numeric field and typing in a new value. In addition to this you can hold down the Alt and Control keys on the keyboard (Alt+Cmd on a Mac) whilst dragging a surface point or line to achieve the same result. The new points created are offset in a controlled way from the adjusted track.

Hold down Alt, Control and Shift (Alt+Shift+Cmd on a Mac) to gear the changes down 10 times. There is finally a Reset button to return the offset parameters to their defaults.

When using the rotate tool to rotate the offset surface, the pixel aspect ratio of the insert clip will be used to create the correct effect. If it is not correct the offset surface will appear to squeeze or stretch as it is rotated. See the ROI section to see how to fix this problem.

All inserts are warped in stereo if you have tracked both views.
You can render the insert for both views by selecting Operate on All Views button next to the Render buttons on the timeline.

The remove module is designed to remove foreground objects from your footage to create clean plates. A removed object can be anything: wires, poles, signs, people and so on. The only requirement for a good removal is enough background to use.

To understand how Mocha Pro removes a foreground object consider how you would do it yourself. A common method is to select a source image where the foreground object does not obscure the background region you are trying to paint in the target frame. You would then clone or otherwise copy the pixels from the source frame to the target frame. If the background is not in the same position in the two frames you would have to track the patch of pixels into the frame, and if the lighting is different between the two frames you would have to adjust the brightness of the patch.

The remove module attempts to do this for you automatically using a method called motion keying. As it has tracked the background areas you have defined, it can move pixels of these background areas from one frame to another. Hence, if a part of a background area is obscured in one frame, Mocha Pro can search the rest of the clip for a frame where the obscured background area is visible, and move the pixels correctly into the target frame. During this process Mocha Pro will evaluate and compensate for the lighting differences between the source and the target frame.

Object removal requires a few simple steps to work correctly. The key to a good remove is adequate, tracked background and a correctly masked foreground.

1. Identify your foreground and background planes

First of all, scrub through your timeline and identify the planes of depth within your shot. Background often consists of multiple planes of motion, and there can also be obstructing foreground material. For example, in the shot below, the sign is in front of the ground, and the car door is in front of (but not obscuring) the sign.

2. Track and roto your foreground elements

We handle the foreground first as we need to mask them off from the background plane and from each other. You don’t actually need to worry too much about tracking the foreground: If the object is untrackable for whatever reason, just hand-animate a spline around it. You absolutely must track an object if it is going to be used as a background pass to replace foreground. For more information on effective tracking, please refer to the tracking section of the documentation.

In the example below, we have tracked and rotoscoped the sign and the door frame, which would otherwise obscure the ground plane.

3. Make sure your layer order is correct

Layer order matters in Mocha. The closer to the camera an object is, the higher in the layer stack it should be. Far background should be on the bottom layer, foreground objects should be on top, in order of closeness.

In the example here, we show the door layer is above the sign layer, and the sign is above the ground layer in the stack.

4. Track your background plane

Once you have your foreground objects masked out across the timeline, draw as big a shape as possible to cover as much of the possible area you want to remove. Make sure the background layer is under the foreground layers, and track. For multiple planes of background, such as a floor and a wall, you will have to track two separate planes of motion. You must specifically track the background layer in order for the removal to work, even if you are working with a locked-off shot.

Here we can see the background drawn on a layer at the bottom of the layer stack, covering the large area of grass.

5. Select the layer you want to remove and adjust your removal parameters

Switch to the remove tab and adjust your parameters (see Remove Parameters below). Key things to focus on are your search range, whether or not you need to create or adjust a cleanplate, and if you need to use an illumination model.

We recommend trying a remove with Illumination modeling set to None first. This is the fastest setting for removals, and can quite often be all you need if there is not much light difference in the background over time.

6. Render

Once your layers are ready and the parameters are set, click the render button or render forward.

One or more background layers must be defined that include all areas that the foreground object moves across during the shot.

To get the result you expect you should observe the following rules:

Here we provide a breakdown of each remove tool parameter and how to use it effectively.

When changing parameters in Mocha Pro the change only affects the currently selected layer. To change the parameters of a layer other than the currently selected layer that layer must first be selected.

Input Tab

Here you specify the input clip for the remove process and any cleanplates you wish to import.

Input Clip

You can choose from any of the Mocha Pro result clips to be used as the source clip to fill the requested foreground layers, instead of the default, which is to use the originally imported clip as the input clip. This can be useful if you have to do multiple passes to get an effective remove.

Cleanplates

Here you can import cleanplates to replace frames in your footage. If you don’t have enough background to use in your shot, importing your own cleaned up version of a frame can assist the remove tool greatly.

To import one or more cleanplates:

If you import two or more cleanplates, Mocha Pro will try to guess the frame numbers from any numbering in the file name. When using the cleanplates between those frame numbers, Mocha Pro will blend the nearest two cleanplates to produce a smooth transition through the clip.

If you want to change the cleanplate settings after exiting the cleanplate window, click on Edit…​ You would need to do this if you are using the frames on a new machine where the cleanplates are stored in a different location, or just to add new cleanplates. If you re-import files with the same name but different directory to existing cleanplate files, Mocha Pro will update the file to the new directory.

Create (Cleanplate)

One useful option within Mocha Pro is to create a cleanplate from the currently viewed frame.

To do this, make sure you are viewing the frame you want to use and click on the Create button. This will create a clip containing the frame you are viewing, and set the Cleanplates clip to the new clip. You can then touch up this cleanplate from your Results folder.

When you save your edits, it will automatically be updated in Mocha Pro to be used in the cleanplate list.

Use Cleanplates Exclusively

If this option is checked, only the cleanplates will be used by Remove to remove the pixels in the selected layer. If it is unchecked, the normal Remove process will be used, pulling in pixels from other frames in the input clip. The cleanplates will then only be used to remove the remaining pixels.

Output Tab

This assigns an output clip for the removal render. You can create new output clips if needed here.

Stereo Remove works in exactly the same way as Mono Remove above, with the additional bonus of being able to render both views at the same time and also choosing whether or not each view assists the other view during the remove process.

By default "Prefer Same View" is checked on in the search range section of the Remove tab. This will attempt to use the current view rather than both views to perform the remove.
If you have useful information in the other view that may assist the remove, you can uncheck this option.

Some of the object is still visible after removal.

The replaced background region is brighter/darker than the surrounding image.

The background patches don’t line up with the surrounding image.

Remove is slow.

If you have a long clip, especially working with HD or film, Remove can be slow because it has to search over a large number of images with a large memory footprint.

Before you change any settings in the Remove parameters, it is worthwhile checking to see if GPU Processing is turned on in Preferences under the GPU page.

GPU Processing can have a significant impact on your render times.

Remove is the most memory intensive module in Mocha Pro, and it will always benefit the performance to add more memory.
If Mocha Pro can fit all the images it needs in memory, performance will be dramatically accelerated when rendering Remove in multiple frames, because it will minimize the amount of disk accesses.

Your aim should be where possible to change the settings to achieve this:

The stabilize feature in uses the data you have tracked to lock down a moving shot. You can think of it as the inverse of what tracking does: It moves the footage around an area rather than an area around the footage. Stabilization is useful for:

Stabilization is a fairly straightforward operation. Once you have selected the tracked layer you wish to work with you have options to restrict the stabilization to key details.

Input Tab

Select the clip that you wish to use to stabilize here.

Range Tab

Select the frame range for the stabilization.

Fixed Frames

Here you can enter the frames you want to be unaffected by stabilization. Mocha will adjust the stabilization between these frames. You can either enter them into the field, or go to a frame and press the + key.

Fixed frames can be useful when you want stabilization across a shot but would like to keep the general motion of the original shot intact.

Smooth

This is the main section for controlling how you want the shot to stabilize.

All Motion

This checkbox sets all the options below it.

X Translation

Stabilize translation in X

Y Translation

Stabilize translation in Y

Rotation

Stabilize rotation

Zoom

Stabilize the scale/zoom

X Shear

Distort the footage according to the tracked surface’s shear data in X

Y Shear

Distort the footage according to the tracked surface’s shear data in Y

X Perspective

Distort the footage according to the tracked surface’s perspective data in X

Y Perspective

Distort the footage according to the tracked surface’s perspective data in Y

Shear and Perspective stabilization can be useful when you want to straighten out a plane in your footage to work on it flat before restoring it to its original perspective and motion.

Maximum Smoothing

Stabilize across the entire track. Setting this value will override the #Frames value beneath it.

# Frames

Stabilize variation across a certain amount of frames. Setting this to a low value will focus the stabilization to only pick up motion that occurs in short bursts (such as a bump in the road). A higher value will try to adjust longer movements.

One of the most common reasons to stabilize is to remove jitter from a shaky camera shot. With shaky camera footage you are primarily concerned with removing position and rotation data. This means you do not have to use the shear or perspective options when tracking. Here is the common method:

Sometimes you want to be able to completely lock down a section of the footage so that it stays in one place and everything else moves around it. For this you can use more aggressive stabilization:

Exporting Stabilization is similar to exporting tracking. When you hit the Export Stabilized Tracking Data button you will be presented with a dropdown box with options for various applications.

All stabilization occurs in stereo if you have tracked both views.
You can render the stabilization for both views by selecting Operate on All Views button next to the Render buttons on the timeline.

The Mocha lens module helps locate and undistort lens distortion.

To compute lens distortion you need an image with one or more distorted straight lines or a distortion map (Mocha Pro only).

For line calibration you can either use your source material if it has suitable lines or take an image of a calibration grid. Using either 1-Parameter or 2-Parameter radial distortion models will allow you to straighten up the image. There is also an Anamorphic model that will accurately model the distortion in anamorphic and extreme wide-angle lenses.

If you have a Distortion Map, you can also remove or work with distortion by importing your maps instead of calibrating.

Distortion of the images can also influence the performance of other modules in Mocha Pro, because the tracker tools assume that the camera obeys a simple pinhole camera model. Severe distortion effects are most likely when a fish-eye or other short focal length lens is used.

Typical lens distortion makes a square object appear either barrel shaped or pin cushion shaped in the image.

The most obvious way to tell whether your images are distorted is to look at the straight edges of objects in the scene. If lines that should be straight are actually curved in the image, this is indicative of distortion. If there are no long straight edges in the scene (for instance a natural scene) then it is much more difficult to discern distortion, even when it is significant.

With the Mocha lens module you can compute, apply and remove distortion automatically. There are two parts to this process – firstly compute the distortion and secondly correct the distortion.

To compute distortion automatically you will need to have some straight lines in your image(s).

It is useful to toggle the grid on and off during the different stages of calibrating the lens to see a visual representation of what is actually happening. Just select the Grid button in the View Controls bar to switch the grid on and off.

Lens calibration is dependent on working with the parameters, which are described in detail below. The basic workflow for lens calibration is as follows:

You can now choose to render out the distortion, or use the current calibration to assist your tracking and insertion workflow.

For further detail on the calibration tools, see Lens Parameters below.

The first step for working in 360 VR mode is define you are working in Equirectangular space.

Mocha Pro can set this in the New Project dialog on standalone:

Or in the Views drop down in the plugin:

You can also change to Equirectangular in the Lens module:

And that’s it! You’re now ready to work in VR lens space and don’t need any further calibration. If at any time you want to adjust the 360 view numerically, the fields Longitude, Latitude and Field of View are available to adjust in the Lens page.

Note that if you are rendering a 360 lens distortion in the plugin version, you can control the Lens distortion from the plugin interface without having to open the Mocha GUI.

See Equirectangular Lens Workflow with 360 VR for more details on using the Equirectangular lens calibration effectively.

Input tab

Here you have two options:

Output tab

This is a way to re-name a distortion setting if you want to have more than one distortion applied. You can choose to leave the input clip intact and create a new clip containing the output from the Lens module, and then rename the rendered distorted or undistorted file and carry on working. Just select the clip from the pulldown menu and select the New…​ button to name the new clip. The new clip will be the same as the current output if you choose to select that it inherits the attributes of that output clip.

Distortion Map tab

Here you have two options:

Calibration Lines

Line calibration is the most important part of the process for 1-parameter, 2 parameter and Anamorphic calibration: These buttons work to locate and then define the straight lines in your shot. See below.

Locate Lines

Select the Locate Lines button once for the image you want to use for calibration.

The output of the Locate Lines process is an overlay showing all the line segments that were detected in the image. Typically longer lines will be divided into two or more line segments. See “Min Line Length” below.

New Line

Using the New Line button you can start selecting line segments to define the straight lines in your scene.

It is important to remember to select New Line each time you want to select the segments of a new line in the scene.

Min Line Length

Increase the Min Line Length parameter to show only longer lines and simplify the display, or if you only short line details, try a smaller line value.

Calibrate Parameters

Having chosen your lines (see above), you need to select a camera model:

The Grid display option will show a reference grid with the computed distortion added.

Here is an example grid of a 1-Parameter distortion model:

Here is an example grid of a 2-Parameter distortion model:

Here is an example grid of an anamorphic distortion model:

Calibrate button

Once you have selected your lines and set either the 1-Parameter, 2-Parameter or Anamorphic models, you can click this button to start computing the calibration. This will invoke the Mocha Pro camera calibration algorithm which will apply the new calibration parameters to any image you are working on in the current project.

The No Distortion parameter does not compute any distortion and the 1- Parameter Inv model is only for use with RealViz Rz3 files.

If you have a grid image, select the Equidistant Lines box and then select lines appropriately for a grid as explained above.

Note: You only need to render the image and remove the computed distortion if you want to save the output – Mocha Pro will use the calibration data generated without having to use an undistorted clip. To render the clip use the Distort/Undistort controls.

Lens

The image center is naturally set at the center of the image by default. Again the coordinates can be entered manually if you want to eye match the center position of the lens by dragging the Center %X or Center %Y sliders or highlighting the current value and typing in a new value. Alternatively, switch on the Calibrate option so that it will be adjusted automatically when you invoke the camera calibration algorithm.

Distortion

The distortion values for the current camera model can be entered manually if you want to match the lens distortion by eye, or they can be computed automatically using the Calibrate button – which is considerably quicker and easier. To adjust manually simply drag the sliders to increase or decrease the values in the K1/Cxx, K2/Cyx, Cxy & Cyy fields or highlight the current value and type in a new value.

Function

Select whether when you render an image you want to remove or add distortion here by selecting the Undistort radio button to remove distortion or the Distort radio button to add distortion.

Distort/Undistort

These buttons are for rendering the clip with the distortion or undistortion values selected.

Selecting the right arrow renders the clip forward from the current point in time. Selecting the left arrow renders the clip backwards from the current point in time.

Selecting the central button with a square stops Mocha Pro from rendering, which can also be done by selecting the escape or space keys on the keyboard instead.

Selecting the down arrow just renders the single frame at the current point in time. This is useful for testing a single frame before deciding to render an entire sequence.

Selecting the lock button enables re-rendering for this module. If the button is locked Mocha Pro will render the frame each time a parameter is changed.

Export Lens Data

You can export the lens parameters to a variety of formats. See "Exporting Lens Data" below for more information:

A common way to generate accurate distortion models is to take an image of a calibration grid, which not only provides long straight lines that exhibit the distortion clearly, but also restricts the lines to a rectilinear grid. The latter constraint can be used to improve on the simple “straightening lines” technique.

If you are using a grid image, select the Equidistant Lines
button. This will enforce a regular grid structure on the selected lines, by forcing the distance between adjacent horizontal and vertical lines to be a constant fixed value.

To use this feature, select lines a fixed distance apart on the grid. You can use a different separation horizontally and vertically. You don’t have to select the lines in the right order - Mocha Pro will re-order the lines according to their horizontal and vertical position on the image.

You also don’t have to select all the grid lines. All that matters is that the distances on the grid between adjacent sampled horizontal lines are the same, with the same rule applying vertically as well.

Grid image with sampled equidistant horizontal and vertical lines. Image courtesy of Jean-Yves Guillernaut, University of Surrey

Note that when you fit a grid, the Grid display option attaches the rendered grid lines to the image grid, so that you can easily see the result.

In cases where even the radial distortion models are not sufficient, or you have an anamorphic shot for which a radial distortion model is not suitable, select the Anamorphic camera model. This allows for different horizontal and vertical distortion. This is the model used in 3D Equalizer V3, although without the inversion of the model used in that product, and we use the “raw” curvature parameters cxx , cxy , cyx , cyy .

You will probably need a grid image to compute the parameters of this model accurately.

The image center is by default set to the center of the frame, i.e. 50% of the frame dimensions in both directions. If the center of distortion is or may be offset from the center of the frame, these values need to be modified.

You can either select the correct image center manually by eye or switch on calibration for the image center , which will then calibrate for the image center along with the distortion parameters. A yellow cross indicates the image center position.

If you don’t get good results from the Mocha Pro calibration procedure, or you have known distortion parameters that you wish to use in Mocha Pro, you can select the distortion parameters manually.

By manipulating the distortion parameters and observing the effect on the image, choose the parameters that straighten up the curved lines as accurately as possible. This should at least deal with the worst effects of the distortion.

Mocha Pro also has the capability to import camera parameters computed by Realviz’s MatchMove 3D camera tracking software.

If you are working with Distortion Maps (sometimes called UV maps or ST Maps), a lot less calibration is required. You can simply bring in your map and it will automatically set the lens.

An example of a Distortion Map

You can now choose to render out the distortion, or use the current calibration to assist your tracking and insertion workflow, as you would a line calibration.

Keep in mind the hard boundaries of the Distortion map being imported may affect the tracking of your shot.

The Lens Module provides different exports for getting your lens data out to other applications.

When working on large files, drop the proxy scale

If you set the proxy scale in View Controls from Full Res to Half Res it should still give enough information to locate good lines, and will be a lot faster.

Make sure your aspect ratio is correct

Note that for the radial distortion models you need to have selected the correct pixel aspect ratio for the images when you started the project. If you have chosen the correct aspect ratio the image will appear in the correct proportions on the screen.

The film back width and height selected when you started the project determine the pixel aspect ratio. Mocha Pro will normally select the correct film back from the image dimensions, but sometimes it may be necessary to make manual adjustments if the correct match was not found.

You can use the Lens: Undistort and Lens: Distort options in Mocha Pro Plugin to render out 360 patches for easier paint and effect work back in the host.

The patch workflow is a one-way control system in the plugin interface independent of the Mocha GUI, which means you don’t need to open Mocha to control the lens view.

By choosing Lens: Undistort from the render options in the plugin, you can then control the view by adjusting three controls:

Alternatively you can pick a view using the Views dropdown option.
This replicates the 360 views in the Mocha GUI, allowing you to choose between specific angles such as Zenith, Nadir, Front, etc.

The lens distortion workflow for the plugin is:

You can then perform any paint or effect work on the lens-distorted version of the footage and it will be re-distorted correctly on top of the original footage.

The Mocha Pro Reorient Module provides two useful tools for adjusting the VR camera:

The Dope Sheet is used to move, copy and paste keyframes in your layers.