Animated motion and effect modifiers in an intelligent title cache system
1. In a title cache system for managing a title having a template and a plurality of variables associated with the template, a method comprising:
- instantiating a title cache object by a title instance manager in response to a title instance event;
managing access to the title cache object, including routing a plurality of internal calls and external calls through a scheduler to control a playback of a plurality of frames;
rendering the plurality of frames into a series of rendered frames by a render engine;
applying an animated modifier to a layer defined in the template, wherein the animated modifier includes a loop modifier or a crawl modifier; and
retrieving the plurality of frames by one or more output devices for streaming the playback, wherein the scheduler is configured to manage a queuing during the playback and configured to forward a plurality of schedule requests to the title cache object to control the playback.
Embodiments of the invention relate to an intelligent title cache system and more particularly to an animated motion and effect modifiers, including loop and crawl modifiers that are applied to a segment of text, graphics, or video in a title template. These animated motion modifiers may define, control and manipulate a sequencing of frames of a title template, adding visual dynamics and enhanced animated behavior to the overall intelligent title cache system.
|IMAGE DISPLAY CONTROL APPARATUS|
Patent #US 20110157474A1
Current AssigneeDENSO Corporation
Sponsoring EntityDENSO Corporation
|RENDERING IN A MULTI-USER VIDEO EDITING SYSTEM|
Patent #US 20100262710A1
Current AssigneeAvid Technology Canada Corporation
Sponsoring EntityAvid Technology Canada Corporation
|METHOD, APPARATUS AND COMPUTER PROGRAM PRODUCT FOR GENERATING ANIMATED IMAGES|
Patent #US 20130300750A1
Current AssigneeNokia Technologies Oy
Sponsoring EntityNokia Technologies Oy
- 1. In a title cache system for managing a title having a template and a plurality of variables associated with the template, a method comprising:
instantiating a title cache object by a title instance manager in response to a title instance event; managing access to the title cache object, including routing a plurality of internal calls and external calls through a scheduler to control a playback of a plurality of frames; rendering the plurality of frames into a series of rendered frames by a render engine; applying an animated modifier to a layer defined in the template, wherein the animated modifier includes a loop modifier or a crawl modifier; and retrieving the plurality of frames by one or more output devices for streaming the playback, wherein the scheduler is configured to manage a queuing during the playback and configured to forward a plurality of schedule requests to the title cache object to control the playback.
- View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9)
- 10. A title cache system for managing a title having a template and a plurality of variables associated with the template, the system comprising:
a title instance manager configured to instantiate a title cache object in response to a title instance event, wherein the title instance event includes a load event and a create event; a scheduler configured to manage access to the title cache object, including routing a plurality of internal calls and external calls through the scheduler to control a playback of a plurality of frames; a render engine configured to render the playback of a plurality of frames into a series of rendered frames; an animated modifier configured to modify a layer defined in the template, wherein the least one animated modifier includes a loop modifier or a crawl modifier; and one or more output devices configured to retrieve the plurality of frames for streaming the playback.
- View Dependent Claims (11, 12, 13, 14, 15, 16, 17, 18)
This application claims the benefit of priority of U.S. application Ser. No. 15/494,502, filed Apr. 23, 2017, which is a continuation-in-part of U.S. application Ser. No. 15/441,160, filed Feb. 23, 2017, which is a continuation of application Ser. No. 15/216,397, filed Jul. 21, 2016, which are herein incorporated by reference to the present application.
The present invention relates generally to an intelligent title cache system and more particularly to animated motion and effect modifiers, including loop and crawl modifiers that are applied to a segment of text, graphics, or video in a title template. These animated motion modifiers may define, control and manipulate a sequencing of frames of a title template, adding visual dynamics and enhanced animated behavior to the overall intelligent title cache system.
In the world of title and graphics animation as it applies to 3-Dimensional (3D) Real Time Rendering, the standard approach to rendering complex visual 3D models for real time playback, including real time data updates (for example sports scoreboards, stock tickers, weather, news, and all manner of live graphics used in broadcast production that are meant to display up to date information as it comes in) is to build hardware graphics systems that are powerful enough to accomplish converting data into graphics in real time. However, this always puts an upper limit of aesthetic complexity and necessitates very expensive hardware. Meanwhile, the dramatic rise in low cost computers with decent graphics ability and explosion of live streamed broadcasting necessitates the need for an affordable way to create responsive, yet aesthetically complex, animated graphics on a wide range of systems.
Similarly, in the world of title and graphics animation in 2-Dimensional (2D) Real Time Rendering applications, the standard approach for fast 2D rendering and drawing alphanumeric values or characters efficiently is to create images of each character and then position them at the appropriate position in the display output such as a monitor or television. For example, to display “00:00”, the general method is to take the rendered “0” and draw it onto the display four times, each offset by the appropriate amount. This approach is very efficient, but can have several drawbacks such as 1) It may be very specific, requiring knowledge of everything necessary to position the letters, including kerning and other font information; 2) If the design intends to show different digits in different ways (for example, increasing the size and changing the shading of the leading digit), this information may become even more specific and complicated; and 3) It limits the render to only letters. However, depending on the design of the title and graphics template, it may be desirable to interpret the data value in an entirely different way.
What is highly desirable is a title and graphics animation production and playback mechanism which does not become involved in the specific details of the actual graphics rendering methodology used to create a particular result, but instead can be applied to many different graphics rendering engines with the intended goal of converting such systems, that rarely can render in real time, into fully responsive real time graphics systems. Such a system should have the power of the full 3D system as well as the speed and efficiency of the 2D display system. Additionally, what is needed is a rendering production mechanism that is completely removed from any creative aspects of the title and graphic design itself. Graphic designers should not be required to focus on the implementation and real time performance details of their designs. Conversely, implementers of the systems should not be required to understand and work around the requirements of the specific design and render technology.
This invention defines a title caching mechanism for replacing potentially lengthy graphics rendering with very fast assembly of 2D pre-rendered images into a complete video frame. This mechanism may be implemented and performed using a software or firmware application. It offers this capability through a flexible but consistent mechanism that can easily be used by a wide range of applications with differing requirements. In particular, an objective in this document is to provide an intelligent title caching system which does not itself engage in the actual rendering process of text, graphics and animation, but when applied to a specific graphics rendering engine, manages the rendering of text and graphics in such a way as to achieve improved optimization, efficiency and speed enhancement of the overall title processing system.
In particular, in the intelligent title caching system a Title is composed of a template and a number of variables in the template and an Input may determine which variables are dynamic. In this system, dynamic variables are each rendered in their own layer. All other variables may be set once and rendered in a background layer. In addition, dynamic variables can play in sync with the rest of the title, or independently. In some instances, playback of the title may be from the start to a pause point, and then to the end. Similarly, dynamic variables may start, pause, and stop payback independently. Playback of a variable considers its relation to the pause point. For example, to change a visible value, the system is configured to play from pause to out, and then inserts a new value, playing from in to a pause point. To momentarily view the value, the system plays it from in to out, without stopping. In another instance, the Input can specify a set of expected values for a variable, so they can be precomputed and cached. When variables are not cached, the system may play back the request by managing the rendering of the missing variable from the rendering engine, waiting for the new value to render, and then playing it.
Benefits and Advantages of system include, for example, 1) data independent rendering; 2) highly efficient playback driven by intelligent understanding of the component data as well as knowledge that full 3D render of a scene is much more compute intensive than overlaying precomputed bitmaps; and 3) By offloading the 3D graphics engine, it is possible to run the system on a wide range of computers, from simple processors with onboard graphics software to powerful GPU systems.
Various other objects, features and attendant advantages of the present disclosure will become fully appreciated as the same becomes better understood when considered in conjunction with the accompanying drawings, in which like reference characters designate the same or similar parts throughout the several views, and wherein:
The computing environment 100 may include a system board 110 (alternatively known as the mainboard, motherboard, baseboard, planar board or logic board) which implemented on a printed circuit board (PCB). It provides communication between many of the electronic components of the system operating components, such as a central processing unit (CPU) and memory, and provides connectors for other peripherals. Hardware elements related to the system board 110 include, for example, memory components defined by a system memory using Read Only Memory ROM 131 and Random Access Memory RAM 132 circuitry, a central processing unit CPU 120 being defined by a microprocessing circuit, a system bus 121 having physical wiring bus elements to transfer binary data between each hardware components, and multiple external interfaces including a video interface 190, an output peripheral interface 195, non-removable and removable memory interfaces (140, 150), a user input interface 160, and a network interface 170. External hardware components of the system board 110 may include a display monitor 191, printer 196, speakers 197, keyboard 162, a pointing device or mouse 161, and a local area network 171 interfacing to remote computers 180. Software, programs data, and firmware may be applied to and installed on the system memories (131, 132) and provide instructions to operate the computing system 100. A BIOS 133 for providing a set of computer instructions in firmware that control input and output operations of the computer system 100 may be installed in the ROM 131. Applications related to software include an operating system OS 134, applications programs 135, other program software modules and drivers 136, and program data 137. Software may also be loaded and operated via storage devices such as hard drive 141, disk drive 151 via disk 152, and compact disk drive 155 via compact disk CD 156. The storage device may include and execute software such as an operating system 144, application programs 145, other program modules 146, and program data 147.
Aspects of the subject matter described herein are operational with numerous other general purpose or special purpose computing system environments or configurations.
Examples of well-known computing systems, environments, or configurations that may be suitable for use with aspects of the subject matter described herein comprise personal computers, server computers, hand-held or laptop devices, multiprocessor systems, microcontroller-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, personal digital assistants PDAs, gaming devices, printers, appliances including set-top, media center, or other appliances, automobile-embedded or attached computing devices, other mobile devices, distributed computing environments that include any of the above systems or devices, and the like.
Aspects of the subject matter described herein may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and so forth, which perform particular tasks or implement particular abstract data types. Aspects of the subject matter described herein may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The computing system environment 100 may also have or may execute a variety of operating systems OS, including an operating system used on a desktop platform or a mobile operating system used mobile devices such as smartphones, cellular/mobile phone, tablets, personal digital assistance PDA, laptop computer, smart watches, and the like. The computing system environment 100 may also include or may execute a variety of possible applications or “Apps”, such as music streamers/players, e-book readers, utility Apps, and electronic gaming apps. The application may provide connectivity and communication with other devices or a server over a network, such as communicating with another computer via a wired or wireless Internet or Intranet network for online interaction such as electronic gaming or online collaboration.
Intelligent Title Cache System—Object Model
1) Data Input Handler and Input Profiles
The input profiles (219, 221) of the data input handler 203 are also known as “Behaviors”, and in some applications matching an input to a title may be considered important. For example, an input that sends the time of day may not be very useful when connected to a title that is designed to show the name of a contestant in a bike race. Thus, matching the input to the template is like matching the data to the display. The Input Profile (219, 221) provides a method for a data source to represent the variables that it manipulates, and in what methods, so it can then be matched to a title correctly and also perform efficiently in the overall rendering process, making sure that the data it is sending can be quickly turned into graphics which are displayed. For example, nothing much happens when a football scoreboard is connected to a weather reporting graphic. You want the data to be matched to the display. And, within that, you want the data which is going to change quickly and frequently to be capable of doing so, and that is where the caching mechanism excels, as will be further described in detail below.
In operation, each input profile (219, 221) is generally required to specify which variables it sends while the title template 235 specifies the variables to which it can respond. Thus, the first objective of the input profile is to match the behavior with a title that supports it. Next, the input profile is configured to provide some intelligence about how it intends to work with those variables, significantly impacting the real-time performance of system 200.
To illustrate this point in a sporting event application, take for example a scoreboard as a source for the data input 201, and consider four variables defined on the scoreboard: Home Score, Clock, Possession, and Home Team Name. In this scenario,
- The data input 203 expects to set the Score variable infrequently.
- The data input 203 expects to set the Possession every now and then, but it knows that this will only be one of two values: Home or Opponent.
- The data input 203 expects to set the Clock every tenth of a second. It knows that this too can be a predetermined set of values.
- The data input 203 does not set the Home Team Name—that should be done before the game. It is unaware this variable exists.
In some instances, the input profile (219, 221) may handle each of the items above as follows:
- The input profile defines the score variable as one that will change dynamically. Each time the score changes, a new score must be rendered and composited onto the rest of the image.
- For the Possession variable, it provides two image choices, one for each direction.
- For the Clock, it breaks down the string into sets for each digit. So, a 4-digit clock might have 10 states for each digit.
- It completely ignores the Home Team Name. This should be set once and rendered into the background of the title.
Note that all variables, whether dynamic or preset, have timelines, so each of these examples involves a timeline of the text or image first entering and then exiting. These timelines may vary from variable to variable. Moreover, the input profile (219, 221) may denote which variables are required to be in the title template and which are optional.
In summary, the input profile (219, 221) may define a short set of information that the data inputs 203 provides to accomplish the following:
- Describe a particular purpose, or behavior.
- Find and match with appropriate title templates 235.
- For performance optimization, identify which variables it will be changing, and where appropriate, what all the values are required to be.
- Conversely, report back to the host (e.g., a data input service, application, device, etc.) which variables are expected to be manipulated.
In system 200, the scheduler 205 is generally responsible for managing the entire system of playback, and all commands route through it. Some of these commands include, for example, actions to load projects, assign variables, and trigger playback. Internally, the scheduler 205 maintains a set of queues 221, to help sequence commands to the title templates 235.
3) Render Engine
The Render Engine 209 of system 200 is generally responsible for performing all 2D or 3D rendering and animation in the system 200, being the key tool for creating the animated imagery in this titling technology. It may take as input a title template, with its variables assigned, and generates an output series of frames such as, for example, the rendered title. However, it can also be instructed to render only specific layers in other applications. Examples of the Render Engine 209 may include software based render engines or hardware based graphics and animation render engines and APIs, such as Unity, Direct3D and OpenGL.
The compositor 237 takes one or more frame outputs from the Render Engine 209 and blends together various title templates 235 to make a finished frame or video output 239 as shown in
5) Channel Mixer
The video output 239 from the compositor 237 is received by the channel mixer 213 which combines the video output 239 from multiple title templates 235 on the same channel into one and then sends one or more combined video outputs 241 for that channel to the output devices 211. In practice, the system 200 may implement a variety of channel mixers, including software based mixers and hardware based mixers such as, for example, for combining the layers of images from all the titles into a single one. In one implementation, the channel mixer 213 may simply be an alpha blend composite.
6) Variables and Rendering
Referring again to
The rendered title itself may be separated into three types of layers:
- Variables that have predefined values and typically change frequently. These can be rendered once, saved to cache, and used multiple times.
- Variables that always receive new values. These are rendered as required and then discarded, or, if desired, kept and stored for later use.
- Everything else. These are all the variables that don'"'"'t change and all the background graphics, etc.
This determination of the layers is made by matching the input profile (219, 221) with the title template 235.
- The title template templates 235 list all the variables it supports, as well as when they occur and where they are in the layer organization of the title templates 235 (including compositing information so they can be rendered in afterwards.)
- The input profile (219, 221) lists which variables it will manipulate and, when applicable, all preset values the variable can take.
In operation, the scheduler 205 and title 235 use this information to orchestrate the performance intelligently and efficiently.
7) Schedule( ) Command:
In system 200, a Schedule( ) command 243 essential to runtime operation for triggering and controlling the flow animation and graphic events in system 200. The Schedule( ) command 243 may be organized around variables, since variables determine what particular information are required to be updated. For example, by running multiple Schedule( ) commands 243 for different variables on one title template 235, the different variables may independently update, even animating asynchronously within one title template 235. Conversely, the same variable across multiple title templates 235 can be scheduled to play synchronously with one call.
The Schedule( ) command 243 may take several parameters, including, for example:
- a) Which title(s) to send it to. The address can be:
- The name of a specific title. (“title” field)
- The id of a specific title. (“title” field)
- The name of an input behavior or input profile. All titles that were set to this will receive it. (“input” field)
- The name of the variable(s). All titles that use these variables will receive it. (the “variable” field itself)
- b) A set of variables and accompanying values. This can be empty or any number of variables to set all at once.
- c) Markers, indicating optional start and end points within the title for the operation
- These refer to the named markers that are embedded in the templates, e.g., “Start” and “Pause”.
- They can also be discrete time values, measured in seconds, e.g., “0.23”
- d) An Action command, indicating action to take, including modifiers.
- e) A time value. This is used to set the time this will occur. It can be modified by the Action.
- a) Which title(s) to send it to. The address can be:
A named queue (optional): Queues may organize actions to play in an order sequence and can have customized names. The queue may exist as long as there are Actions placed in it, and then eventually discarded.
The Action command defined within the Schedule( ) command 243 may be composed of multiple options and modifiers which are summarized and provided in Table 1 below:
Variables 305 are defined in this document as receptacles for data that is displayed in the layers 303. Variables 305 manage images, text, even colors. Of these, the most frequently used are text variables, which may provide text to place in a layer. In all cases, variables 305 are communicated purely as text strings. The cache system 200 generally does not understand, and is not required to know, anything about the data in a variable 305 other than that it is represented with a text string. For example, a title template 235 may be designed to represent the string “01” with two numeric letters, or it might display the image of a goose head on a spring. The point here is that all the communication occurs as text strings, whether they are floating point numbers, RGB color definitions, file addresses of images, or the words “true” and “you lied”.
Intelligent Title Cache System—Scheduling API
A TitleCache class of the system 200 may optimize playback by determining which components of the title animation can be pre-rendered and which need to be prepared dynamically or “on the fly”. This is all determined by which variables are in use in the template and how they are used.
Variables within a template fall into three categories:
- Fixed. These are variables that are set once by the user in set up and then they are never changed during playback. These include variables that set information that is fixed and unchanged during a broadcast. Examples would be “Home Team” or “Background Color”
- Predefined. These are variables that will change during the broadcast, but they will only pick from a predefined set of states. Examples might include “Inning”, “Strikes”, or “Countdown Clock”.
- Dynamic. These are variables that change during broadcast and can take on an unlimited range of values.
Knowledge of whether a title variable is Fixed, Predefined, or Dynamic is determined by two things:
- The Template itself, which simply defines all of the available variables.
- The Input, which specifies:
- Which variables it uses.
- Which variables are pre-defined.
- For predefined variables, all the different values.
Once the TitleCache class knows the behavior of each variable, it can construct a much more intelligent representation of the title in memory, breaking the representation down into one or more layers, which follow these rules:
- One or more fixed layers, including the background layer, that span the full length of the title and includes:
- All fixed imagery within the title.
- All imagery that is driven by fixed variables.
- Fixed imagery that may appear in front of some of the titles (and so must be a separate fixed layer from the background fixed layer.)
- For each predefined variable, a layer is created that manages all predefined variations of that variable. These renderings span the duration of the variable and typically have transparency so they easily composite on top of other layers, including the background layer.
- For each dynamic variable, a layer is created, but no renderings attached until it is given a value.
- One or more fixed layers, including the background layer, that span the full length of the title and includes:
Prior to playback, the background and predefined variable layers are completely pre-rendered, and their frames are made available as a frame cache, either in memory or in a file. The frame cache may be implemented both in memory as a memory frame cache or disk such as a file frame cache.
In a project work environment, this same cache can be saved to disk and then reloaded when the project is reopened. This cache gets completely regenerated if anything that changes the background layer or the predefined variables occurs. For example, conditions for regenerating the cache may occur when:
- the template is edited or replaced with a different one.
- sizing or positioning of the template is changed.
- any fixed variables are changed.
- the input definition for predefined variables changes (i.e., input changes.)
However, these are infrequent events and likely do not occur during a broadcast. During playback, dynamic variables respond to variable changes from external sources.
CCacheLayer, CCacheValue, CCacheSubFrame
A CCachelayer object represents one layer in the rendering of a title. Typically, this represents a variable that can change dynamically.
The rest of the template, including the design time variables, may all be handled by one or more CCachelayers that have no variables. These fixed layers include, for example, the “background” layer as well as any fixed semi-transparent layers that may appear before some other layers.
Multistate variables and Regular variables are two types of variables having the following attributes:
- Regular variables get generated once and are used typically just once. They include most information that gets displayed on templates and are changed frequently, like Name, Message, Score, etc.
- Multistate variables typically go back and forth between a fixed set of values, and they do it often. Two examples would be a clock and a display of strikes in baseball.
With all this in mind, we have three kinds of CCacheLayers, and rules for their behavior:
- Multistate variable layer. This has two or more Values that represent each variation of the variable and they all get rendered up front and saved to frame cache, so they can be used again.
- Regular variable layer. This has exactly one Value and gets re-rendered when the variable changes. These are rarely saved to the disk cache, because they won'"'"'t last long.
- Fixed/Background layers. These have all the steady state variables, none of which it needs to track. These are re-rendered whenever any of their variables change (which should be infrequently) and always saved to file frame cache.
The CCacheValue keeps track of one value for a variable, managed by CCachelayer. Each CCacheValue carries all the information used to generate it and, of course, the resulting set of partial frames, in the form of an array of CCacheSubFrame objects.
CCachelayer creates a set of one or more CCacheValues to represent each possible variable value.
CCacheSubFrame represents one point in time in a full CCacheValue sequence. It manages the storage of this frame either in memory or to disk. Note that this frame is typically just a subset of the full frame at a specific point in time.
This provides methods to write a frame as well as read it.
It also manages compression of the frame itself. It does so with several mechanisms, all of which must respond in real time such as, for example, a
- “Region of Interest” (ROI): Store only the rectangle within the frame that is in use;
- Data Compression: Fast lossless compression algorithms tend to be efficient with graphics imagery because of the prevalence of repeated pixel sequences; and a
- Frame Duplication: Often, layers don'"'"'t have any changes over a period of frames, and so one value can be stored for all identical frames.
Integration of the TitleCache Class
It also may be memory mapped, to improve performance. If compressed and the compression is fast, then the compressed format might also be the format that is kept in memory, when loaded, and then uncompressed prior to rendering.
In another instance, when the title is new or gets a new template, it may notify the Title Cache to refresh itself. This regenerates the renders and saves to cache, as needed.
Accordingly, when the title is saved, it gets the cache file XML data 701 from the Title Cache.
Next, to properly build a representation of the title, “TitleCache” calls the Title Instance.
- 1. Get the names of all the segments, markers, and their time points (901).
- 2. Get the names of all the variables and the time points for their starts and ends (903).
- 3. Gets the names of all the variables in use by the Input Behavior (905).
- 4. Of those variables, get the list of predefined values for multistate variables (907).
Referring again to
In some instances, the Scheduler 205 maintains a central list of all TitleCaches 703 and then parses incoming commands and sends them to the appropriate TitleCache instance 703. The Scheduler 211 may also manage access to the output of the TitleCaches 703, so that the output device 211 can simply call the Scheduler 205 to retrieve the rendered frames.
During playback, the Scheduler 205 manages queuing and then forwarding Schedule requests to the TitleCache to control playback.
Referring again to
A specialized class handling called CCacheRenderRequest, representing a render request, is a block of frames that are to be rendered for a particular set of variable values of the time segment for the frames.
In some instances, the render engine generates all calls to TitleCache. These calls are thread safe and return quickly. With each frame rendered, the render engine may generate a call back to the TitleCache to pass the frame to it and determine whether to continue or switch to a different render. In another embodiment, the title cache generates calls to the render engine to render the frame sequences.
- 1, Render Engine calls GetNextRenderRequest( ) (1001) which returns CCacheRenderRequest (1003).
- 2. Render Engine prepares the variables (1005).
- 3. Then, for each frame in the request (1007),
- a. Render the frame (1009).
- b. Call TitleCache::SubmitRender (passing it both the Frame and the CCacheRenderRequest) (1011).
- c. Cache searches for the correct CCacheValue and, if it is still there, sends it the frame (1015), which CCacheValue copies in (1017).
- d. Cache returns a success code to continue or a failure code if this is no longer needed. Failure can also mean the render is finished (1019).
- e. If success, Render Engine moves to render the next frame (1021).
- f. If failure, Render Engine calls GetNextRenderRequest( ) (1023).
Note that the Render Engine retains the one instance of CCacheRenderRequest, returning the one instance when done via ReturnRenderRequest( ). In some situations, this typically frees the memory. But, this also provides a way for the title cache to stop a render and then continue at a later time. It also provides a way for the Render Engine to return a render that it has not completed, because the Render Engine needs to work on something else that may have higher priority. However, there can be other equally acceptable ways to organize the queue of render requests between the title cache and the Render Engine.
Finally, the Output Devices 211 component, as shown in
The TitleCache may use a memory cache of two or more CCacheFrameBuffers to ensure that TitleCache is ready to play. A lower priority thread may run in the background, ensuring these are replenished.
The CCacheFrameQueue may generate and store a short queue of fully composited frames for playback. The intention is to make sure that there is always a frame ready to go so the output is uninterrupted.
In some embodiments, there are three external objects that generally drives the framebuffer:
- The output. These include the output devices through which the frames stream. It is constantly asking for a new frame.
- The parent cache. Under the control of the scheduler, this is frequently changing the status, from updating variables to entering and leaving playback.
- The refresh thread. This is a background thread that repeatedly calls all of the frame buffers to refresh their frames.
In other embodiments, there may be three modes of activity: Off, Play, and Paused.
- Off: the title isn'"'"'t even active. Any requests for a frame for output should get back a blank frame. Meanwhile, the buffer should be primed with the first frames for playback.
- Paused: the titles are active, but not moving. Requests for frames always return the same frame, which is at the paused point in the title.
- Play: the title is moving forward. Every time a frame is requested, the next in line is served and the background thread is actively replenishing the queue.
Note that the actual mode is determined by the activity of the title. For example, as soon as something is scheduled to play, the play state becomes Play. The actual mode remains in that state until the last layer is finished sending out frames. Once all layers are stopped, the mode enters with Paused or Off state. The Off state is the special case where all layers have completed playing and entered the off state. The Paused state is the case where one or more layers still have the cursor, but are not moving.
CCacheRenderRequest represents a block of frames that is required to be rendered for a CCacheValue. This object is sent to the render engine, giving it a fully self-contained description of what is required.
Cache Rendering Optimization:
From the above discussion, a summary of attributes for defining and optimizing the cache rendering process in system 200 is provided below:
- A variable is nothing more than a string of characters, i.e., an array of byte characters that specifies a specific graphic (i.e. text or image, etc.) to show over a range of time.
- The data input 203 defines which variables it sends and what values it puts in them.
- The Title Template 235 defines which variables it supports and, for each, how it may translate the incoming string into the appropriate image to render.
- The TitleCache 601 takes the set of variable value strings from the data input 203 and sends them to the Render Engine 209 to render an image, using the title template 235 to define the render.
- Later for playback, the TitleCache 601 matches live data strings from the data input 203 with its cached images and copies it into the frame.
In yet another example of the caching system 200 applied to a sporting event scenario, the data input source 201 may be a sports score controller, sending a numeric value to indicate a game quarter. Depending on its layout and intended usage, the title design might have several choices for how it displays the first quarter:
- 1 (text string of “1”)
- 1st (text string of “1st” where the ST are a much smaller font size)
- First Quarter (translated into a completely different text string)
- [image] (an image that shows the quarter status.)
In all cases, the representation of the original string is defined by the title design and the cache should not be involved. Instead, title design should request a render of “1” and return whatever interpretation the design has for what first quarter should look like. And, as with the previous clock timer example, it may be constraining for the cache to interpret how large or small the rendered image will be or where it should be placed.
Data Independent Rendering
For the game quarter case, the cache may represent all values for all variables as simply text strings, with no comprehension as to their meaning. And so, it is agnostic as to the meaning of any string passed to it by an input source. It simply passes it to the rendering engine, along with the selected template.
The title template may incorporate a translation table, equating specific input strings with specific values to render an image. This allows one configuration to convert a “1” into a simple displayed string, while another configuration to convert it into an image.
With that in mind, the steps for preparing the cached image of the value “1” for variable “quarter” include:
- 1. The input sends the text string “1” for the variable “quarter”
- 2. The cache simply passes it to the title template and requests it to render “1”.
- 3. The render engine consults the translation table in the title template, which converts the “1” into the appropriate rendered text or image.
- 4. The render engine sends back the rendered image.
- 5. The cache stores the rendered image, with optional optimizations of compression and calculating and storing just the ROI.
- 6. On playback, the input sends “1”
- 7. The cache selects the image for “1” and builds the full image by layering this image on top of the remainder of the scoreboard components.
For the game clock scenario, the title cache is also unaware of what it is displaying, similar to the game quarter case. In the game clock case, additional items are considered. For example, if a rendering of every variation of the clock is created, 10,000 frames are potentially created for all the combinations of values in a four-digit clock. In this situation, the cache system 200 may subdivide the problem and address each digit independently. Then, the title cache system 200 may prepare only a smaller portion of different variations, (e.g., 10 variations), one for each numeric value. However, because the title cache does not know how these are drawn or what they really represent, the title cache does not have the ability to reuse the same digit in all four positions.
To address this limitation of the title cache, and still keep the title cache completely independent of the data, a Multipart Variable caching is introduced as another component of the system 200. In one instance, the multipart variable caching may break a string into several sub-segments and create a set of variation renders for each string. Then, on playback, the Multipart Variable caching may select the variations that together match the requested input and merge them into one image. For Multipart Variables,
- a. The Input can specify a variable as composed of individual letters, or related sets of one or more letters, as long as they combine to match all possible input patterns.
- b. The Input specifies all values for each letter (or set), so that each digit can be independently rendered.
- c. The Input may provide a format string that is used to set the spacing of the letters during rendering.
- d. Each letter (or set) is rendered separately with its own time sequence.
- e. When an individual letter (or set) changes, it can play out and then the replacement plays in.
- f. If a letter (or set) does not change, it remains unchanged at the pause point.
- 1. When building the cache:
- a. Break the string down into separate segments. (i.e. the clock digits)
- b. For each segment, provide all possible value variations.
- c. Render each variation for each segment.
- d. Represent each variation as the full string but with empty spaces for the parts outside the segment. (i.e., —“1______”, “2______”, all the way up to “______1”, “______2”)
- e. Store each of these variations as the partial string and its associated image.
- 2. On playback, in response to a text value from the input:
- a. Search through the list of variations.
- b. For each variation that provides characters that match exactly in value and position the input value:
- i. Select the variation.
- ii. Remove the matching characters from the search.
- iii. Continue until all characters are matched.
- c. Create a composite by adding all of the selected variation images together into one image.
- 1. When building the cache:
In the cache system 200, the process may construct a composite of images to correctly represent the string that is passed by layering together all variations, using the matching of substrings to recreate the set of sub-images that build the equivalent image. The solution presented so far only works if the size of each digit is exactly the same. For example, draw the time string “12:34” in a proportional font, then draw just the last digit with the string “4” results in string placement errors. To overcome this placement error, the optional concept of a format string is applied to the system 200. The format string may provide an interpretation of the render that should be used to determine a positioning of each segment during render. Such positioning adjustment is referred to in this document as a Multipart Alignment.
In one instance of the Multipart Alignment, the format string may be simply a string with all characters filled in that shows an appropriately spaced rendering. For example, with the clock format, it could be “00:00”.
To Implement the Multipart Alignment in System 200:
- 1. For each render, include both the substring, i.e., “4”, and the format string, i.e., “00:00”.
- 2. The render operation first organizes the spacing of the proportional characters for “00:00”.
- 3. Then the render replaces the letters of the format with the letters of the substring.
- 4. It then renders with the letters in these positions.
- 5. The ‘4’ is placed exactly in the position of the 4th ‘0’ from the format, as it should be.
Several features, attributes, and benefits of the system 200 are summarized below:
- A Title is composed of a template and a number of variables in the template.
- Input determines which variables are dynamic.
- Dynamic variables are each rendered in their own layer.
- All other variables are set once and rendered in a background layer.
- Dynamic variables can play in sync with the rest of the title, or independently.
- Playback of the title is from the start to a pause point, and then to the end.
- Dynamic variables may start, pause, and stop payback independently.
- Start and stop times for dynamic variables may optionally be the actual start and stop times of the variables themselves within the template.
- Playback of a variable considers its relation to the pause point. For example:
- To change a visible value, play from pause to out (end point of variable in the time line), then insert new value and play from in (start point of the variable in the time line) to the pause point.
- To momentarily view the value, play from in to out, without stopping.
- Input can specify a set of expected values for a variable, so they can be precomputed and cached.
- When variables are not cached, playback request waits for the new value to render, then plays it.
- The playback includes options for waiting for render, waiting to sync several parallel variable renders, playing without waiting, playing part way through, etc.
- When variables are not cached, playback request waits for the new value to render, then plays it.
- Multipart variables
- Input can specify a variable as composed of individual variable subsets such as individual alphanumeric letters or subsets of letters.
- Input specifies all values for each letter/subset, so that each letter/subset can be independently rendered.
- Input may provide a format string that is used to set the spacing of the letters during rendering.
- Each letter/subset is rendered separately with its own time sequence.
- When an individual letter/subset changes, it can play out and then the replacement plays in.
- If a letter/subset does not change, it remains unchanged at the pause point.
- Benefits and Advantages of system 200
- Data independent rendering.
- Highly efficient playback driven by intelligent understanding of the component data and recognition that full render of a complex scene is much more compute intensive than overlaying precomputed bitmaps.
- Isolate different data (variables) into their own layers.
- Things that never change are rendered once and stored.
- Things that have all values known before also rendered once and stored.
- Complicated multiple part data, such as time clocks, broken into parts and pre-rendered.
- Things that are truly dynamic are rendered on the fly
- Typically just a short text sequence—simple and fast.
- Automatic—system inserts the render prior to play.
- Deliberate—requests the render then follows with play.
- Supports complex, multiple asynchronous animations as easily as a still image.
- Render is simply a look up of each variable image at its point in its sequence.
- By offloading the 3D graphics (or other render technology) engine, it is possible to run this on a wide range of computers, from simple processors with onboard graphics software to powerful GPU systems.
- Playback is almost always the same. (Exception is insertion of updated variable value.)
- Preparation/pre-render will take longer on slower machines.
Loop and Crawl Modifiers:
In yet another implementation, the intelligent title cache system 200 may implement animated motion effects to the title to add visual dynamics and appeal to the title template without the loss of CPU performance. Two of these animated motion effects may include loop and crawl modifiers and both may be implemented in the intelligent title cache system 200 by taking advantage of the frame cache and changing the way the playback is sequenced through it, and like other modifier defined in the intelligent title cache system 200, loops and crawls may be defined as variable and layers in a project, having specific and predefined rules for playback.
In the system 200, the loop modifier may be applied to the title for providing an otherwise motionless title some animation for a set period of time. In one example, the loop modifier may be applied to cause the background of a flag to wave in a continuous manner. In another example, the loop modifier may be applied to offer a simple light flare, or “glint”, that periodically shines across a sports scoreboard or lower thirds news headline to draw attention to it.
The crawl modifier provides another mechanism for animating the title by sending a message across a banner so that the observer can read the text. In some implementations, crawl modifiers are typically applied to and advantageous in news reports, stock tickers or any other sort of continually streaming flow of information having text, graphics, and/or video.
Different implementation schemes of loops and crawls may be used in the intelligent title cache system 200. For example, consider a normal playback scenario where the normal playback comprises of playing from a start of a frame sequence to a pause point. It then holds at the play point for the duration of the display of the current value. Once it is time to no longer show this value (typically a text string for a message, numeric digit, etc.) it continues by moving forward from the play point until the end of the sequence, taking the image sequence through to the end, when the text string are no longer visible.
For the crawl modifier, one implementation in intelligent title cache system 200 is to simply play all the way through to the end of the sequence. Unlike the normal playback, there is no stopping at the pause point. In addition, the crawl modifier may be of different durations which may be determined by the length of the sequence of text strings placed in it.
For the loop modifier, operation and application of the loop in the system 200 is completely different than that of the crawl modifier. Like the crawl modifier, the loop modifier ignores the pause point. However, unlike the crawl, the loop modifier plays the same image sequence in a repetitive manner and repeated multiple times. To accomplish this, parameters defining the loop modifier are dependent on certain looping points embedded in the construction of the image sequence defined in the title template accompanied by a delay time to hold at the loop end point before starting again at the loop start.
In both implementations, the loop and crawl modifiers may use the existing intelligent title cache system 200 that routes an animated sequence of frames, by applying new unique playback rules that control the sequencing of these frames, resulting in a visually dynamic and enhanced behavior of the title template.
In one aspect and advantage, loop modifiers may add a critical element of animation to otherwise still moments in the display of a title defined in the intelligent title cache system 200. The effect of this modifier can play a significant role in drawing interest back to the information displayed in the title, among other things. In addition, because certain aspects and features of this modifier build on the intelligent title cache system'"'"'s ability to store design animations as a layer within a multi-layer display, the system 200 enables a very powerful implementation of looped segments that can played asynchronously.
In another aspect and advantage, the additional CPU cost of displaying a looped segment is almost immeasurable, because it may build and extend the actions and behaviors of the existing intelligent title cache system 200, which optimizes playback by rendering once into the frame series (which can then be saved in a cache file). In operation, the loop modifier in system 200 provides the mechanism of choosing which frame from that series to display at any point in time. Overall, the loop modifier does not add any significant processing time to implement in system 200, leaving the graphics processing time in system 200 virtually unchanged.
In another aspect and advantage, crawl modifiers may enable textual information of the title template to be displayed in succession as a series of informational text lines that sequentially display useful information as part of a larger display. Because of the underlying driven creation of the actual crawl itself, this crawl mechanism is unique in its ability to present crawl effects that animate in every conceivable manner and direction.
Like the loop modifier, the crawl modifier is configured to build and extend the actions and behaviors of the existing intelligent title cache system 200. In particular, the implementation of the crawl modifier in system 200 may take the design of a sequence defined in a title template and represent it as a series of frames that can be played in any order at will. Furthermore, the CPU cost of displaying a crawl segment is essentially the same as displaying a fixed visual element that plays in, then plays out, since it is an extension of system 200, which optimizes playback by rendering once into the frame series (which can then be saved in a cache file). The crawl mechanism in system 200 is configured to use different logic for choosing which frame to display at any given point in time.
In Table 2, the following loop parameters are added to a layer within the title template of system 200.
A looped sequence may include several loop time parameters that govern its playback sequence, including, for example, a loop start time, a loop end time, and a delay. A special playback setting menu having a Crawl option and a Loop option may be provided to set and select options related to the loop time parameters. Upon selection of the Loop option, the special playback setting menu provides the user additional input options to manually set the loop start time, the loop end time, and the loop delay, each option having values that are predefined or may be modified by the user. The loop start time indicates the point within the loop sequence to start the looped portion of playback; the loop end time indicates the point (or frame) in the sequence to jump from to go back to the loop start; and the loop delay, which defines a time duration, to wait at the loop out point before starting again at the loop in point.
In practice, the loop time parameters either preset to a certain value within the system 200 or may be modified to different values as determined by the user. Note that unlike the pause point, which is set to be global across all layers, these points are specific to the individual layer. The significance of this feature is to allow multiple layers to loop independently.
Note that the loop start and loop end points can reside outside the actual frame sequence, indicating that the loop is really just a full sequence, like a crawl that plays all the way through, then holds at the loop end time, waits the full delay, and then picks back up at the loop start time.
A looped layer may behave differently from a regular layer in that it may ignore and bypass the pause point. Instead, it plays through to its end point (endTime), waits for the delay time, and then starts playback again from the start time.
Later, when the title is given the play out command, it may:
- If currently playing, continue to play through to the end.
- If currently paused at an end time that is within the layer, play out to the end.
- If currently paused at an end time after the end, then do nothing.
- Optionally, stop immediately.
In sum, the looped playback starts by playing through to the loop endpoint, then, holds and waits the duration of the delay without doing anything. Once that delay has been achieved, the looped playback continues back at the loop start point and play through again.
Implementing of the crawl into system 200 is simple and requires only a minimal set of parameters. In Table 3, for example, the following crawl parameter is added to a layer within the title template. If the m_bCrawlEnabled flag is set, it indicates to system 200 to play all the way through and do not stop at a pause point.
In system 200, the m_bCrawlEnabled flag may be manually set for a predefined animated movement which causes the text to move through as defined by the user. In practice, the application of the crawl effect can be an accomplished with any design approach that sets letters to move on and then off the screen. For example, this can applied to a scripted animation that simply moves all letters from right to left at a constant rate as defined by the user. In another example, it could be applied to a more standard animation that flies letters in, holds a few seconds, and then follows with an animation that flies them back out. Accordingly, both design approaches result in a display of each message that is no longer a standard “crawl” of moving an objects at a constant pace in one direction, yet nevertheless provides its full functionality in that one message is followed by another, sequentially, in the same visual space.
In another embodiment, variable-length crawls may be supported by the system 200. To facilitate variable-length crawls, the following steps are generally taken during the render process. First the duration of the paragraph into which the crawl is rendered is set to be significantly longer than the longest possible crawl. Then, the text of the crawl is rendered into that paragraph. Next, the sequence of frames is pulled into the playback engine cache, which reads the frames until it gets to a completely blank image, indicating that the full sequence of frames has now passed through. Finally, it marks this point as the end of the frame sequence and discards the remaining frames.
A further optimization on this variable-length crawls technique is to have the render process itself stop once it renders a completely blank frame, according to another embodiment. Of course, the process of looking for a blank frame may not start until there have been visible frames rendered, indicating that the crawl has indeed started.
During playback, the display of a crawl element is almost identical in behavior to the display of an updated value except that the usage of the pause point is inverted. For the normal playback, the current value is displayed at the pause point. When it is time to introduce a new value (i.e., replace a message or increment a numeric counter), the arrival of the new value triggers the current value to play out to the end and then the new value plays in, stopping to hold at the pause point.
With the crawl flag being enabled, the pause point is ignored, which means by default nothing is currently holding still at the pause point. When a new value arrives for playback, it simply starts playback and carries all the way through, passing through the pause point without stopping and playing all the way out. Once the sequence of frames has played all the way through, the layer is ready to accept the next crawl sequence.
In yet another embodiment, two crawl sequences may be played at the same time. Simultaneous playback of two crawl sequences, for example, may allow the first crawl sequence to roll off the display on the left, while the following crawl sequence can already start on the right. The cache playback engine of the system 200 supports the simultaneous display of two values because of its support for simultaneous digits (i.e., for a numeric counter) to be playing at the same time.
As with every image sequence in the system 200, each crawl sequence may be represented by a unique text value—the specific text of the words used in the crawl sequence. And so, both the crawl sequence before and the crawl sequence after are available as separate sequences that can be played simultaneously. The objective, then, is to cue the second crawl sequence to start playback while the first sequence is still playing out.
To accomplish this, a time offset is provided to the scheduler, indicating how long to wait after the first message plays out before starting the playback of the second message. In the case of the crawl sequence, this time offset can be set to be a negative value, in seconds, with the result that the second crawl sequence starts while the first crawl sequence is still alive, playing through to its end which the second starts rolling in. In this manner, a series of crawl sequences can play, one after the other, with the spacing from one to the next a consistent distance, because the start time of the new crawl sequence has been timed to play at an offset against the end of the preceding crawl sequence, which was, of course, identified as its first completely blank frame.
These crawl sequences can be triggered by the same “Update” command in system 200 used to refresh a message or counter. Accordingly, after the crawl is enabled, the update results in the next message playing through.
1. On play in of title 1603:
- a. Play forward through frames until loop out point is reached 1605.
- b. Hold at loop out point while decrementing the delay time counter 1607.
- c. Jump back to loop in point and play forward to step 1a 1709.
2. On play out of title 1611:
- a. If currently at loop out point, immediately continue play out from this point forward 1613.
- b. If currently in motion and the time is after the pause point, play through to the end 1615.
- c. If currently in motion and the time is before the pause point 1617, either:
- i. Play through to the end at an accelerated speed to complete playback within the same duration as playing from the pause point to the end 1619.
- ii. Or, force the other layers to wait until the looped progress reaches the pause point, and then all layers commence to play out as designed, in unison 1621.
1. Render of layer 1703:
- a. Set duration of layer to “infinite”, or sufficient duration to ensure the longest message can fit within it 1705.
- b. Render frames until the first non-blank frame, optionally after the pause point 1707 (to ensure that something is shown before this algorithm is triggered.)
- c. Set first blank frame as the last frame in the sequence and return to the cache 1709.
2. Play In 1711:
- a. If layer is already playing another instance 1713
- i. Set the start time to be the end time of the prior instance plus the delay time 1715, which may be a negative number.
- b. Else start playback immediately 1717.
- c. Play through to the end of the frame series 1719.
All patents, patent applications, and other references cited herein are incorporated by reference in their entireties.
What has been described and illustrated herein is a preferred embodiment in this document along with some of its variations. The terms, descriptions and figures used herein are set forth by way of illustration only and are not meant as limitations. Those skilled in the art will recognize that many variations are possible within the spirit and scope of the embodiments in which all terms are meant in their broadest, reasonable sense unless otherwise indicated. Any headings utilized within the description are for convenience only and have no legal or limiting effect.
Other embodiments and modifications of the present document may occur to those of ordinary skill in the art in view of these teachings. Accordingly, the embodiments in this document are to be limited only by the following claims which include all other such embodiments and modifications when viewed in conjunction with the above specifications and accompanying drawings.