)]}'
{
  "commit": "d460e60e1955be72f3479a944fabde401fe848a2",
  "tree": "665048fcbdf49e0cd7f3ab5de1f2c8e9dd0c2376",
  "parents": [
    "5460d74631c9df16c9f0d160fb872c9834f8bb59"
  ],
  "author": {
    "name": "Sam Zackrisson",
    "email": "saza@webrtc.org",
    "time": "Fri Mar 13 14:10:34 2026"
  },
  "committer": {
    "name": "WebRTC LUCI CQ",
    "email": "webrtc-scoped@luci-project-accounts.iam.gserviceaccount.com",
    "time": "Fri Mar 13 15:40:14 2026"
  },
  "message": "AEC3: Fix render buffer headroom calculation on underruns\n\nThe read and write pointers can meet. Currently, this causes an overflow in the headroom calculation (headroom \u003d\u003d buffer size) which offsets the buffers needlessly. This causes a loss of transparency when using the neural residual echo estimation module.\n\nTested:\n  1. audioproc_f is bitexact on a large set of aecdumps, without ML-REE\n  2. I see improved transparency on a dump with underruns, with ML-REE\n\nBug: b:487988676, webrtc:442444736\nChange-Id: I902a20f28c603f85a1c10ae5b525788e3a54c285\nReviewed-on: https://webrtc-review.googlesource.com/c/src/+/456301\nCommit-Queue: Sam Zackrisson \u003csaza@webrtc.org\u003e\nReviewed-by: Jesus de Vicente Pena \u003cdevicentepena@webrtc.org\u003e\nCr-Commit-Position: refs/heads/main@{#47153}\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "f25416f828dae7ba5acd24bec269876e40f1b9e0",
      "old_mode": 33188,
      "old_path": "modules/audio_processing/aec3/render_buffer.h",
      "new_id": "6fc0317c6b2ab0974085599616521ad74e00ed45",
      "new_mode": 33188,
      "new_path": "modules/audio_processing/aec3/render_buffer.h"
    }
  ]
}