Stream realtime TTS audio as it is synthesized

2026-09-16 14:55:12 +02:00
parent 97c8072b3d
commit 3f684b7325
3 changed files with 60 additions and 31 deletions
@@ -116,12 +116,19 @@ class VoiceSmokeTest(unittest.IsolatedAsyncioTestCase):
        async def tts(request):
            body = await request.json()
            self.assertEqual(body["input"], "Hallo zurück")
-            return web.Response(body=make_wav((6000).to_bytes(2, "little", signed=True) * SAMPLE_RATE),
-                                content_type="audio/wav")
+            self.assertEqual(body["chunk_size"], 4)
+            stream = web.StreamResponse(headers={"Content-Type": "application/octet-stream"})
+            await stream.prepare(request)
+            pcm = (6000).to_bytes(2, "little", signed=True) * SAMPLE_RATE
+            await stream.write(pcm[:SAMPLE_RATE])
+            await asyncio.wait_for(audio_received.wait(), timeout=3)
+            await stream.write(pcm[SAMPLE_RATE:])
+            await stream.write_eof()
+            return stream

        fake = web.Application()
        fake.router.add_post("/audio/transcriptions", stt)
-        fake.router.add_post("/audio/speech", tts)
+        fake.router.add_post("/audio/speech/pcm-stream", tts)
        fake_runner, fake_url = await serve(fake)
        os.environ.update({
            "ATHENA_TALK_REALTIME_SECRET": "test-secret-" * 4,