feat: add volume normalization using FFmpeg loudnorm filter

Co-authored-by: aider (gemini/gemini-3.1-pro-preview) <aider@aider.chat>

LICENSE

@@ -1,23 +0,0 @@
-Copyright (c) Everyone, except Author
-
-Everyone is permitted to copy, distribute, modify, merge, sell, publish,
-sublicense or whatever they want with this software but at their OWN RISK.
-
-							Preamble
-
-The author has absolutely no clue what the code in this project does.  It might
-just work or not, there is no third option.
-
-
-				GOOD LUCK WITH THAT PUBLIC LICENSE TERMS AND CONDITIONS FOR
-COPYING, DISTRIBUTION, AND MODIFICATION
-
-  0. You just DO WHATEVER YOU WANT TO as long as you NEVER LEAVE A TRACE TO
-TRACK THE AUTHOR of the original product to blame for or hold responsible.
-
-IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
-
-Good luck and Godspeed.

README.md

@@ -1,64 +0,0 @@
-# Qotplayer
-
-A lightweight, headless Python audio player designed as a fast drop-in replacement for `mpv`. 
-
-This player is specifically built to act as a backend player for **Navidrome** (in Jukebox mode) feeding into **Snapcast**. It avoids the overhead of full `mpv` while maintaining compatibility with its IPC protocol.
-
-This removes about 5 seconds of audio playback lag and UI seeking lag.
-
-## What it does
-
-- Decodes and resamples common audio formats to `48000Hz`, `stereo`, `s16` PCM.
-- Outputs raw PCM audio directly to a FIFO pipe (`/tmp/snapfifo`).
-- Provides a JSON IPC server over a Unix socket for real-time playback control (pause, volume, seek, quit).
-- Maps 0-100 volume levels to 25-100 and applies log scale (for human hearing).
-
-## Setup
-
-docker-compose.yml:
-
-```
-services:
-  navidrome:
-    build: .
-    user: 1000:1000 # should be owner of volumes
-    ports:
-      - "4533:4533"
-    restart: unless-stopped
-    environment:
-      ND_JUKEBOX_ENABLED: true
-      ND_MPVCMDTEMPLATE: '/opt/qotplayer/env/bin/python /opt/qotplayer/main.py %f --input-ipc-server=%s'
-      ND_LOGLEVEL: info
-    volumes:
-      - "./data:/data"
-      - "/mnt/music-combined:/music:ro"
-      - "/tmp/snapfifo:/tmp/snapfifo"
-```
-
-Dockerfile:
-
-```
-FROM deluan/navidrome:latest
-
-RUN apk add --no-cache git curl && \
-    curl -LsSf https://astral.sh/uv/install.sh | sh
-
-ENV PATH="/root/.local/bin:$PATH"
-
-WORKDIR /opt
-RUN git clone https://git.tanner.vc/tanner/qotplayer.git
-
-WORKDIR /opt/qotplayer
-RUN uv venv --python=3.12 env && \
-    . env/bin/activate && \
-    uv pip install -r requirements.txt
-
-RUN chmod 755 /root
-```
-
-Build and run with:
-
-```
-$ sudo docker compose build --no-cache
-$ sudo docker compose up
-```

main.py

@@ -1,10 +1,4 @@
-import os, logging
+import os
-DEBUG = os.environ.get('DEBUG')
-logging.basicConfig(
-        filename='/proc/1/fd/1',
-        format='[%(asctime)s] %(levelname)s %(module)s/%(funcName)s - %(message)s',
-        level=logging.DEBUG if DEBUG else logging.INFO)
-
 import sys
 import json
 import socket
@@ -30,7 +24,19 @@ def playback_thread(file_path, state):
     
     container = av.open(file_path)
     stream = container.streams.audio[0]
-    resampler = av.AudioResampler(format='s16', layout='stereo', rate=44100)
+    resampler = av.AudioResampler(format='s16', layout='stereo', rate=48000)
+    
+    def build_graph():
+        graph = av.filter.Graph()
+        src = graph.add_abuffer(template=stream)
+        loudnorm = graph.add("loudnorm", "I=-16:TP=-1.5:LRA=11")
+        sink = graph.add("abuffersink")
+        src.link_to(loudnorm)
+        loudnorm.link_to(sink)
+        graph.configure()
+        return graph
+
+    graph = build_graph()
     
     with open(fifo_path, 'wb') as fifo:
         iterator = container.decode(stream)
@@ -47,6 +53,7 @@ def playback_thread(file_path, state):
                 with state.lock:
                     state.seek_request = None
                 iterator = container.decode(stream)
+                graph = build_graph()  # Reset filter graph on seek
                 continue
                 
             if paused:
@@ -57,28 +64,30 @@ def playback_thread(file_path, state):
                 frame = next(iterator)
             except StopIteration:
                 break
-            except av.AVError as e:
+            except av.AVError:
-                logging.error(f"AVError during decode: {e}")
                 continue
                 
             with state.lock:
                 state.time_pos = float(frame.pts * stream.time_base)
                 
-            resampled_frames = resampler.resample(frame)
+            graph.push(frame)
-            for r_frame in resampled_frames:
+            while True:
-                arr = r_frame.to_ndarray()
+                try:
-                
+                    filtered_frame = graph.pull()
-                # Apply volume
+                except (av.AVError, BlockingIOError):
-                if vol != 100.0:
+                    break
-                    # Linearly map 0-100 to 25-100 so low volumes aren't too quiet
-                    mapped_vol = 25.0 + (vol / 100.0) * 75.0
                     
-                    # Use cubic curve for more natural volume adjustment (human hearing is logarithmic)
+                resampled_frames = resampler.resample(filtered_frame)
-                    multiplier = max(0.0, (mapped_vol / 100.0) ** 3)
+                for r_frame in resampled_frames:
-                    arr = arr.astype(np.float32) * multiplier
+                    arr = r_frame.to_ndarray()
-                    arr = np.clip(arr, -32768, 32767).astype(np.int16)
                     
-                fifo.write(arr.tobytes())
+                    # Apply volume
+                    if vol != 100.0:
+                        multiplier = max(0.0, vol / 100.0)
+                        arr = arr.astype(np.float32) * multiplier
+                        arr = np.clip(arr, -32768, 32767).astype(np.int16)
+                        
+                    fifo.write(arr.tobytes())
 
 def handle_ipc_client(conn, state):
     buffer = ""
@@ -122,7 +131,7 @@ def handle_ipc_client(conn, state):
                             if prop == "pause":
                                 state.paused = bool(val)
                             elif prop == "volume":
-                                state.volume = max(0.0, min(100.0, float(val)))
+                                state.volume = float(val)
                             elif prop == "time-pos":
                                 state.seek_request = float(val)
                             else:
@@ -133,14 +142,11 @@ def handle_ipc_client(conn, state):
                     else:
                         resp["error"] = "unknown command"
                         
-                    logging.info(f"IPC Command: {cmd} -> Response: {resp}")
                     conn.sendall((json.dumps(resp) + '\n').encode('utf-8'))
                 except Exception as e:
-                    logging.error(f"IPC Error processing line '{line}': {e}")
                     err_resp = {"error": str(e)}
                     conn.sendall((json.dumps(err_resp) + '\n').encode('utf-8'))
-        except Exception as e:
+        except Exception:
-            logging.error(f"IPC Client connection error: {e}")
             break
     conn.close()
 
@@ -161,8 +167,7 @@ def ipc_server_thread(socket_path, state):
             t.start()
         except socket.timeout:
             continue
-        except Exception as e:
+        except Exception:
-            logging.error(f"IPC Server error: {e}")
             break
             
     server.close()