Google is committed to advancing racial equity for Black communities. See how.

SurfaceTexture

SurfaceTexture is a combination of a surface and an OpenGL ES (GLES) texture. SurfaceTexture instances are used to provide surfaces that output to GLES textures.

SurfaceTexture contains an instance of BufferQueue for which apps are the consumer. The onFrameAvailable() callback notifies apps when the producer queues a new buffer. Then, apps call updateTexImage(), which releases the previously held buffer, acquires the new buffer from the queue, and makes EGL calls to make the buffer available to GLES as an external texture.

External GLES textures

External GLES textures (GL_TEXTURE_EXTERNAL_OES) differ from traditional GLES textures (GL_TEXTURE_2D) in the following ways:

  • External textures render textured polygons directly from data received from BufferQueue.
  • External texture renderers are configured differently than traditional GLES texture renderers.
  • External textures can't perform all traditional GLES texture activities.

The main benefit of external textures is their ability to render directly from BufferQueue data. SurfaceTexture instances set the consumer usage flags to GRALLOC_USAGE_HW_TEXTURE when it creates BufferQueue instances for external textures to ensure that the data in the buffer is recognizable by GLES.

Because SurfaceTexture instances interact with an EGL context, an app can only call its methods while the EGL context that owns the texture is current on the calling thread. For more information see the SurfaceTexture class documentation.

Timestamps and transformations

SurfaceTexture instances include the getTimeStamp() method, which retrieves a timestamp, and getTransformMatrix() method, which retrieves a transformation matrix. Calling updateTexImage() sets both the timestamp and the transformation matrix. Each buffer that BufferQueue passes includes transformation parameters and a timestamp.

Transformation parameters are useful for efficiency. In some cases, source data might be in the incorrect orientation for the consumer. Instead of rotating the data before sending it to the consumer, send the data in its orientation with a transform that corrects it. The transformation matrix can be merged with other transformations when the data is used, minimizing overhead.

The timestamp is useful for buffer sources that are time dependent. For example, when setPreviewTexture() connects the producer interface to the output of the camera, frames from the camera can be used to create a video. Each frame needs to have a presentation timestamp from when the frame was captured, not from when the app received the frame. The camera code sets the timestamp provided with the buffer, resulting in a more consistent series of timestamps.

Case study: Grafika's continuous capture

Grafika's continuous capture involves recording frames from a device's camera and displaying those frames on screen. To record frames, create a surface with the MediaCodec class's createInputSurface() method and pass the surface to the camera. To display frames, create an instance of SurfaceView and pass the surface to setPreviewDisplay(). Note that recording frames and displaying them at the same time is a more involved process.

The continuous capture activity displays video from the camera as video is being recorded. In this case, encoded video is written to a circular buffer in memory that can be saved to disk at any time.

This flow involves three buffer queues:

  • App — The app uses a SurfaceTexture instance to receive frames from the camera, converting them to an external GLES texture.
  • SurfaceFlinger — The app declares a SurfaceView instance to display the frames.
  • MediaServer — Configure a MediaCodec encoder with an input surface to create the video.

In the figure below, the arrows indicate data propagation from the camera. BufferQueue instances are in color (producers are teal, consumers are green).

Grafika continuous
capture activity

Figure 1. Grafika's continuous capture activity

Encoded H.264 video goes to a circular buffer in RAM in the app process. When a user presses the capture button, the MediaMixer class writes the encoded video to an MP4 file on disk.

All BufferQueue instances are handled with a single EGL context in the app while the GLES operations are performed on the UI thread. The handling of encoded data (managing a circular buffer and writing it to disk) is done on a separate thread.

When using the SurfaceView class, the surfaceCreated() callback creates the EGLContext and EGLSurface instances for the display and the video encoder. When a new frame arrives, SurfaceTexture performs four activities:
  1. Acquires the frame.
  2. Makes the frame available as a GLES texture.
  3. Renders the frame with GLES commands.
  4. Forwards the transform and timestamp for each instance of EGLSurface.

The encoder thread then pulls the encoded output from MediaCodec and stashes it in memory.

Secure texture video playback

Android supports GPU post-processing of protected video content. This lets apps use the GPU for complex, nonlinear video effects (such as warps), mapping protected video content onto textures for use in general graphics scenes (for example, using GLES), and virtual reality (VR).

Secure Texture Video Playback

Figure 2. Secure texture video playback

Support is enabled using the following two extensions:

  • EGL extension — (EGL_EXT_protected_content) Enables the creation of protected GL contexts and surfaces, which can both operate on protected content.
  • GLES extension — (GL_EXT_protected_textures) Enables tagging textures as protected so they can be used as framebuffer texture attachments.

Android enables SurfaceTexture and ACodec (libstagefright.so) to send protected content even if the window's surface doesn't queue to SurfaceFlinger and provides a protected video surface for use within a protected context. This is done by setting the protected consumer bit (GRALLOC_USAGE_PROTECTED) on surfaces created in a protected context (verified by ACodec).

Secure texture video playback sets the foundation for strong DRM implementation in the OpenGL ES environment. Without a strong DRM implementation, such as Widevine Level 1, many content providers don't allow rendering of their high-value content in the OpenGL ES environment, preventing important VR use cases such as watching DRM-protected content in VR.

AOSP includes framework code for secure texture video playback. Driver support is up to OEMs. Device implementers must implement EGL_EXT_protected_content and GL_EXT_protected_textures extensions. When using your own codec library (to replace libstagefright), note the changes in /frameworks/av/media/libstagefright/SurfaceUtils.cpp that allow buffers marked with GRALLOC_USAGE_PROTECTED to be sent to ANativeWindow (even if ANativeWindow doesn't queue directly to the window composer) as long as the consumer usage bits contain GRALLOC_USAGE_PROTECTED. For detailed documentation on implementing the extensions, refer to the Khronos registries (EGL_EXT_protected_content, and GL_EXT_protected_textures).