SurfaceTexture

Construct a new SurfaceTexture to stream images to a given OpenGL texture.

Construct a new SurfaceTexture to stream images to a given OpenGL texture. In single buffered mode the application is responsible for serializing access to the image content buffer. Each time the image content is to be updated, the releaseTexImage() method must be called before the image content producer takes ownership of the buffer. For example, when producing image content with the NDK ANativeWindow_lock and ANativeWindow_unlockAndPost functions, releaseTexImage() must be called before each ANativeWindow_lock, or that call will fail. When producing image content with OpenGL ES, releaseTexImage() must be called before the first OpenGL ES function call each frame.

Attach the SurfaceTexture to the OpenGL ES context that is current on the calling thread. A new OpenGL ES texture object is created and populated with the SurfaceTexture image frame that was current at the time of the last call to detachFromGLContext(). This new texture is bound to the GL_TEXTURE_EXTERNAL_OES texture target. This can be used to access the SurfaceTexture image contents from multiple OpenGL ES contexts. Note, however, that the image contents are only accessible from one OpenGL ES context at a time.

Detach the SurfaceTexture from the OpenGL ES context that owns the OpenGL ES texture object. This call must be made with the OpenGL ES context current on the calling thread. The OpenGL ES texture object will be deleted as a result of this call. After calling this method all calls to updateTexImage() will throw an IllegalStateException until a successful call to attachToGLContext(int) is made. This can be used to access the SurfaceTexture image contents from multiple OpenGL ES contexts. Note, however, that the image contents are only accessible from one OpenGL ES context at a time.

Retrieve the timestamp associated with the texture image set by the most recent call to updateTexImage. This timestamp is in nanoseconds, and is normally monotonically increasing. The timestamp should be unaffected by time-of-day adjustments, and for a camera should be strictly monotonic but for a MediaPlayer may be reset when the position is set. The specific meaning and zero point of the timestamp depends on the source providing images to the SurfaceTexture. Unless otherwise specified by the image source, timestamps cannot generally be compared across SurfaceTexture instances, or across multiple program invocations. It is mostly useful for determining time offsets between subsequent frames.

Retrieve the 4x4 texture coordinate transform matrix associated with the texture image set by the most recent call to updateTexImage. This transform matrix maps 2D homogeneous texture coordinates of the form (s, t, 0, 1) with s and t in the inclusive range [0, 1] to the texture coordinate that should be used to sample that location from the texture. Sampling the texture outside of the range of this transform is undefined. The matrix is stored in column-major order so that it may be passed directly to OpenGL ES via the glLoadMatrixf or glUniformMatrix4fv functions.

release() frees all the buffers and puts the SurfaceTexture into the 'abandoned' state. Once put in this state the SurfaceTexture can never leave it. When in the 'abandoned' state, all methods of the IGraphicBufferProducer interface will fail with the NO_INIT error. Note that while calling this method causes all the buffers to be freed from the perspective of the the SurfaceTexture, if there are additional references on the buffers (e.g. if a buffer is referenced by a client or by OpenGL ES as a texture) then those buffer will remain allocated. Always call this method when you are done with SurfaceTexture. Failing to do so may delay resource deallocation for a significant amount of time.

Releases the the texture content. This is needed in single buffered mode to allow the image content producer to take ownership of the image buffer. For more information see SurfaceTexture(int, boolean).

Set the default size of the image buffers. The image producer may override the buffer size, in which case the producer-set buffer size will be used, not the default size set by this method. Both video and camera based image producers do override the size. This method may be used to set the image size when producing images with Canvas (via lockCanvas(Rect)), or OpenGL ES (via an EGLSurface). The new default buffer size will take effect the next time the image producer requests a buffer to fill. For Canvas this will be the next time lockCanvas(Rect) is called. For OpenGL ES, the EGLSurface should be destroyed (via eglDestroySurface), made not-current (via eglMakeCurrent), and then recreated (via eglCreateWindowSurface) to ensure that the new default size has taken effect. The width and height parameters must be no greater than the minimum of GL_MAX_VIEWPORT_DIMS and GL_MAX_TEXTURE_SIZE (see glGetIntegerv). An error due to invalid dimensions might not be reported until updateTexImage() is called.

Register a callback to be invoked when a new image frame becomes available to the SurfaceTexture. Note that this callback may be called on an arbitrary thread, so it is not safe to call updateTexImage() without first binding the OpenGL ES context to the thread invoking the callback.

Update the texture image to the most recent frame from the image stream. This may only be called while the OpenGL ES context that owns the texture is current on the calling thread. It will implicitly bind its texture to the GL_TEXTURE_EXTERNAL_OES texture target.

Invoked when the garbage collector has detected that this instance is no longer reachable. The default implementation does nothing, but this method can be overridden to free resources.

Note that objects that override finalize are significantly more expensive than objects that don't. Finalizers may be run a long time after the object is no longer reachable, depending on memory pressure, so it's a bad idea to rely on them for cleanup. Note also that finalizers are run on a single VM-wide finalizer thread, so doing blocking work in a finalizer is a bad idea. A finalizer is usually only necessary for a class that has a native peer and needs to call a native method to destroy that peer. Even then, it's better to provide an explicit close method (and implement Closeable), and insist that callers manually dispose of instances. This works well for something like files, but less well for something like a BigInteger where typical calling code would have to deal with lots of temporaries. Unfortunately, code that creates lots of temporaries is the worst kind of code from the point of view of the single finalizer thread.

If you must use finalizers, consider at least providing your own ReferenceQueue and having your own thread process that queue.

Unlike constructors, finalizers are not automatically chained. You are responsible for calling super.finalize() yourself.

Uncaught exceptions thrown by finalizers are ignored and do not terminate the finalizer thread. See Effective Java Item 7, "Avoid finalizers" for more.