Pepper: Define PPB_VideoEncoder API.

ppapi/api/pp_codecs.idl

Index: ppapi/api/pp_codecs.idl
diff --git a/ppapi/api/pp_codecs.idl b/ppapi/api/pp_codecs.idl
index 5a93ff1063f054a130ee0a797280c491138f1d5a..d3ac262be2b33fcae22eca81214dcced0f5cf79a 100644
--- a/ppapi/api/pp_codecs.idl
+++ b/ppapi/api/pp_codecs.idl
@@ -115,3 +115,56 @@ struct PP_VideoPicture_0_1 {
    */
   PP_Size texture_size;
 };
+
+/**
+ * Supported video profile information. See the PPB_VideoEncoder function
+ * GetSupportedProfiles() for more details.
+ */
+struct PP_SupportedVideoProfile {
+  /**
+   * The codec profile.
+   */
+  PP_VideoProfile profile;
+
+  /**
+   * Dimensions of the maximum resolution of video frames, in pixels.
+   */
+  PP_Size max_resolution;
+
+  /**
+   * The numerator of the maximum frame rate.
+   */
+  uint32_t max_framerate_numerator;

[dmichael (off chromium), 2015/02/03 23:26:07]

Why a fraction? Would a float or double suffice?

I notice that Android uses a float or integer...
http://developer.android.com/reference/android/media/MediaFormat.html#KEY_FRA...

[bbudge, 2015/02/04 13:56:40]

This is copied from the VEA interface. I think it's the correct way to specify
frame rate because the encoder uses exact arithmetic internally, and floats or
even doubles can't represent these ratios well.
See: http://avisynth.nl/index.php/FPS

+
+  /**
+   * The denominator of the maximum frame rate.
+   */
+  uint32_t max_framerate_denominator;
+
+  /**
+   * A value indicating if the profile is available in hardware, software, or
+   * both.
+   */
+  PP_HardwareAcceleration acceleration;
+};
+
+/**
+ * Struct describing a bitstream buffer.
+ */
+struct PP_BitstreamBuffer {
+  /**
+   * The size, in bytes, of the bitstream data.
+   */
+  uint32_t size;
+
+  /**
+   * The base address of the bitstream data.
+   */
+  mem_t buffer;
+
+  /**
+   * Whether the buffer represents a key frame.
+   */
+  PP_Bool key_frame;
+};
+