media::VideoEncodeAccelerator::Initialize() [1] is defined to be synchronous
(this forces the IPC message [2] and Mojo message [3] to be Sync). 

This is probably due to webrtc's encoder interface [4] method definition
originating long time ago in software encoders, which are easy and 
synchronous.

[1] https://cs.chromium.org/chromium/src/media/video/video_encode_accelerator.h?l=122
[2] https://cs.chromium.org/chromium/src/media/gpu/ipc/common/media_messages.h?l=151&gsn=GpuCommandBufferMsg_CreateVideoEncoder
[3] https://cs.chromium.org/chromium/src/media/mojo/interfaces/video_encode_accelerator.mojom?q=video_encode_accelerator.mojom&sq=package:chromium&dr&l=50
[4] https://cs.chromium.org/chromium/src/third_party/webrtc/api/video_codecs/video_encoder.h?l=106

This forces the Chrome implementation to have to jump through hoops [5] to
reconcile this interface with the fact that VEA::Initialize() has a
synchronous true/false success indication [6] and a deferred indication 
via either a RequireBitstreamBuffers() (good) [7] or NotifiyError (bad) [8].

[5] https://cs.chromium.org/chromium/src/content/renderer/media/gpu/rtc_video_encoder.cc?l=844
[6] https://cs.chromium.org/chromium/src/content/renderer/media/gpu/rtc_video_encoder.cc?l=345
[7] https://cs.chromium.org/chromium/src/content/renderer/media/gpu/rtc_video_encoder.cc?l=493
[8] https://cs.chromium.org/chromium/src/content/renderer/media/gpu/rtc_video_encoder.cc?l=584

The point of this bug is to isolate the adaptation of synch interface vs
async Mojo in just one file: rtc_video_encoder.cc.