For some reason we've ended up with three different sender-side APIs to generate a keyframe:

1. https://w3c.github.io/webrtc-extensions/#rtcrtpsender-setparameters-keyframe

   // main.html
   const params = sender.getParameters();
   await sender.setParameters(params, {encodingOptions: params.encodings.map(() => {keyFrame: true}});

2. https://w3c.github.io/webrtc-encoded-transform/#rtcrtpsender-extension

   // main.html
   await sender.generateKeyFrame(sender.getParameters().encodings.map(({rid} => rid));

3. https://w3c.github.io/webrtc-encoded-transform/#dom-rtcrtpscripttransformer-generatekeyframe

   // worker.js
   onrtctransform = async ({transformer}) => await transformer.generateKeyFrame(rid);

Observations:

• 1 is implemented in Chrome https://wpt.fyi/results/webrtc/RTCRtpSender-setParameters-keyFrame.html
• 3 is implemented in Safari and Firefox https://wpt.fyi/results/webrtc-encoded-transform?q=keyframe
• 2 is not implemented by anyone and seems redundant with 1.

Let's remove 2.