fix(engine): mux uppercase WebM exports with Opus audio (#6579)

* fix(engine): mux uppercase WebM exports with Opus audio

* test(engine): terminate ffprobe options before input

docs/packages/engine.mdx

@@ -95,6 +95,10 @@ Paths passed to `encodeFramesChunkedConcat()` and `concatVideoFiles()` may be
 relative or absolute. Relative paths resolve from the calling process's working
 directory.
 
+`muxVideoWithAudio()` recognizes output filename extensions regardless of case.
+WebM output, including `.WEBM` and `.WeBm`, converts audio to Opus while copying
+the video stream.
+
 These are rendering internals. Their result and error conventions differ by
 operation: browser and capture orchestration throws, FFmpeg wrappers generally
 return result objects, and teardown helpers avoid masking the original failure.

packages/engine/src/services/chunkEncoder.muxContainer.test.ts

@@ -0,0 +1,80 @@
+import { spawnSync } from "node:child_process";
+import { mkdtempSync, rmSync } from "node:fs";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { afterAll, beforeAll, describe, expect, it } from "vitest";
+import { getFfmpegBinary, getFfprobeBinary } from "../utils/ffmpegBinaries.js";
+import { muxVideoWithAudio } from "./chunkEncoder.js";
+
+const HAS_FFMPEG = spawnSync(getFfmpegBinary(), ["-version"]).status === 0;
+
+function ffmpeg(args: string[]): Buffer {
+  const result = spawnSync(getFfmpegBinary(), ["-nostdin", "-v", "error", "-y", ...args]);
+  if (result.status !== 0) throw new Error(result.stderr.toString());
+  return result.stdout;
+}
+
+describe.skipIf(!HAS_FFMPEG)("WebM audio muxing", () => {
+  let dir: string;
+  let video: string;
+  let audio: string;
+  let sourceFrames: Buffer;
+
+  beforeAll(() => {
+    dir = mkdtempSync(join(tmpdir(), "hf-mux-container-"));
+    video = join(dir, "video.webm");
+    audio = join(dir, "audio.m4a");
+    ffmpeg([
+      "-f",
+      "lavfi",
+      "-i",
+      "testsrc2=size=64x36:rate=10:duration=1",
+      "-c:v",
+      "libvpx-vp9",
+      "-deadline",
+      "realtime",
+      video,
+    ]);
+    ffmpeg([
+      "-f",
+      "lavfi",
+      "-i",
+      "sine=frequency=440:sample_rate=48000:duration=1",
+      "-c:a",
+      "aac",
+      audio,
+    ]);
+    sourceFrames = ffmpeg(["-i", video, "-f", "rawvideo", "-pix_fmt", "rgb24", "-"]);
+  });
+
+  afterAll(() => {
+    if (dir) rmSync(dir, { recursive: true, force: true });
+  });
+
+  it.each(["webm", "WEBM", "WeBm"])(
+    "muxes AAC as Opus into .%s without changing the video frames",
+    async (extension) => {
+      const output = join(dir, `output.${extension}`);
+      const result = await muxVideoWithAudio(video, audio, output, undefined, {
+        audioCodec: "aac",
+      });
+      expect(result.success, result.error).toBe(true);
+
+      const probe = spawnSync(
+        getFfprobeBinary(),
+        ["-v", "error", "-show_entries", "stream=codec_name", "-of", "csv=p=0", "--", output],
+        { encoding: "utf8" },
+      );
+      expect(probe.status, probe.stderr).toBe(0);
+      expect(probe.stdout.trim().split(/\r?\n/)).toEqual(["vp9", "opus"]);
+      expect(ffmpeg(["-i", output, "-an", "-f", "rawvideo", "-pix_fmt", "rgb24", "-"])).toEqual(
+        sourceFrames,
+      );
+
+      const pcm = ffmpeg(["-i", output, "-vn", "-t", "0.5", "-ac", "1", "-f", "s16le", "-"]);
+      let energy = 0;
+      for (let i = 0; i < pcm.length; i += 2) energy += pcm.readInt16LE(i) ** 2;
+      expect(Math.sqrt(energy / (pcm.length / 2))).toBeGreaterThan(100);
+    },
+  );
+});

packages/engine/src/services/chunkEncoder.ts

@@ -857,8 +857,9 @@ export async function muxVideoWithAudio(
   const outputDir = dirname(outputPath);
   if (!existsSync(outputDir)) mkdirSync(outputDir, { recursive: true });
 
-  const isWebm = outputPath.endsWith(".webm");
-  const isMov = outputPath.endsWith(".mov");
+  const outputExtension = extname(outputPath).toLowerCase();
+  const isWebm = outputExtension === ".webm";
+  const isMov = outputExtension === ".mov";
   const shouldCopyAudio = isWebm ? false : await shouldCopyAacSidecar(audioPath, config);
   const args = [...(config?.videoInputArgs ?? ["-i", videoPath]), "-i", audioPath, "-c:v", "copy"];