Skip to main content
TCP Streamer provides extensive configuration options for audio streaming. These options are passed to the start_stream() command.

​
Configuration Schema

All configuration is provided when calling start_stream(). There is no persistent configuration file; settings are managed by the frontend application.

​
Device Configuration

​
device_name

​
device_name
string
required
Name of the audio input device to capture from.Valid values: Any device name returned by get_input_devices()Platform notes:
  • On Windows, loopback devices are prefixed with [Loopback]
  • On Linux, ALSA virtual devices are automatically filtered out
  • Device must support 2-channel (stereo) audio

​
is_loopback

​
is_loopback
boolean
default:"false"
Whether the selected device is a loopback device for system audio capture.Valid values: true, falseEffect:
  • When true, searches output devices instead of input devices
  • Uses BufferSize::Default instead of fixed buffer size
  • Increases minimum ring buffer to 8000ms (vs 5000ms for standard input)
  • Adjusts adaptive buffer range to 4000-12000ms
Platform support: Windows only (WASAPI loopback)

​
Network Configuration

​
ip

​
ip
string
required
Target server IPv4 address.Valid values: Any valid IPv4 address (e.g., "192.168.1.100", "10.0.0.5")Examples:
  • Local network: "192.168.1.100"
  • Localhost: "127.0.0.1"

​
port

​
port
u16
required
Target server port number.Valid values: 1-65535Common values: 5000-5010 (convention for audio streaming)

​
auto_reconnect

​
auto_reconnect
boolean
default:"false"
Enable automatic reconnection on connection loss.Valid values: true, falseBehavior:
  • Uses exponential backoff: 2s → 4s → 8s → … → 60s (max)
  • Adds ±500ms jitter to prevent thundering herd
  • Minimum retry delay: 2 seconds
  • Drains stale audio from buffer before reconnecting

​
dscp_strategy

​
dscp_strategy
string
default:"voip"
DSCP/QoS marking strategy for network packets.Valid values:
  • "voip" - Expedited Forwarding (EF, DSCP 46, TOS 0xB8) - Recommended for audio
  • "lowdelay" - Low Delay (IPTOS_LOWDELAY, TOS 0x10)
  • "throughput" - Throughput (IPTOS_THROUGHPUT, TOS 0x08)
  • "besteffort" - Best Effort (TOS 0x00)
Default: Falls back to "voip" if invalid value providedPlatform notes:
  • Requires administrator/root privileges on some systems
  • May be ignored by some network equipment
  • Most effective on managed networks with QoS policies

​
Audio Configuration

​
sample_rate

​
sample_rate
u32
default:"48000"
Audio sample rate in Hz.Valid values: Device-dependent, typically:
  • 44100 (CD quality)
  • 48000 (professional audio, recommended)
  • 96000 (high-resolution)
Effect on bitrate (stereo, 16-bit):
  • 44100 Hz = ~1411 kbps
  • 48000 Hz = ~1536 kbps
  • 96000 Hz = ~3072 kbps

​
buffer_size

​
buffer_size
u32
default:"480"
Hardware audio buffer size in frames.Valid values: Device-dependent, typically 128-2048Common values:
  • 128 frames (~2.7ms at 48kHz) - Ultra-low latency
  • 480 frames (~10ms at 48kHz) - Balanced
  • 1024 frames (~21ms at 48kHz) - Stable for older hardware
Platform notes:
  • Ignored for loopback devices (uses system default)
  • Lower values = lower latency but higher CPU usage
  • Too low may cause audio glitches on slower systems

​
chunk_size

​
chunk_size
u32
default:"960"
Number of samples (frames) per network transmission packet.Valid values: Typically 240-4800Effect:
  • Smaller values = lower latency, more CPU/network overhead
  • Larger values = higher latency, more efficient
Recommended values:
  • 480 samples (~10ms at 48kHz) - Low latency
  • 960 samples (~20ms at 48kHz) - Balanced
  • 2400 samples (~50ms at 48kHz) - Efficient for high-bandwidth

​
Buffer Configuration

​
ring_buffer_duration_ms

​
ring_buffer_duration_ms
u32
default:"5000"
Duration of the internal ring buffer in milliseconds.Valid values: Any positive value, but automatically adjusted:
  • Standard input: Minimum 5000ms
  • Loopback device: Minimum 8000ms
Purpose: Absorbs network jitter and timing variationsMemory usage (stereo, at 48kHz):
  • 5000ms = ~1.8 MB
  • 8000ms = ~2.9 MB
  • 10000ms = ~3.7 MB
Recommendation: Use defaults unless experiencing buffer underruns

​
enable_adaptive_buffer

​
enable_adaptive_buffer
boolean
default:"false"
Enable dynamic buffer resizing based on network conditions.Valid values: true, falseBehavior:
  • Monitors jitter every 10 seconds
  • Increases buffer size if jitter >15ms
  • Decreases buffer size if jitter <5ms
  • Only resizes if change exceeds 10% of current size
  • Emits buffer-resize-event on changes
Note: Currently tracks target size but does not perform actual ring buffer reallocation (future enhancement)

​
min_buffer_ms

​
min_buffer_ms
u32
default:"2000"
Minimum buffer size for adaptive buffering.Valid values: Any positive value, but automatically adjusted:
  • Standard input: Minimum 2000ms
  • Loopback device: Minimum 4000ms
Only used when enable_adaptive_buffer is true

​
max_buffer_ms

​
max_buffer_ms
u32
default:"6000"
Maximum buffer size for adaptive buffering.Valid values: Any positive value, but automatically adjusted:
  • Standard input: Maximum 6000ms
  • Loopback device: Maximum 12000ms
Only used when enable_adaptive_buffer is true

​
Performance Configuration

​
high_priority

​
high_priority
boolean
default:"false"
Set network thread to maximum priority.Valid values: true, falseEffect:
  • Reduces network thread scheduling latency
  • May improve stability under heavy system load
  • Requires elevated privileges on some systems
Warning: May impact overall system responsiveness if system is under heavy load

​
Audio Format

TCP Streamer automatically negotiates audio format with the device:

​
Supported Formats

  1. F32 (32-bit float) - Preferred, native on PipeWire/Linux
  2. I16 (16-bit signed integer) - Common, good compatibility
  3. U16 (16-bit unsigned integer) - Fallback for older devices

​
Transmission Format

All audio is converted to 16-bit signed little-endian PCM before transmission over the network, regardless of capture format.

​
Channel Configuration

  • Required: 2 channels (stereo)
  • Mono or multi-channel devices are not supported

​
Configuration Examples

​
Low Latency Local Streaming

await invoke('start_stream', {
  device_name: 'Microphone (USB Audio)',
  ip: '192.168.1.100',
  port: 5000,
  sample_rate: 48000,
  buffer_size: 128,
  ring_buffer_duration_ms: 5000,
  auto_reconnect: true,
  high_priority: true,
  dscp_strategy: 'voip',
  chunk_size: 480,
  is_loopback: false,
  enable_adaptive_buffer: false,
  min_buffer_ms: 2000,
  max_buffer_ms: 6000,
});

​
System Audio Capture (Windows)

await invoke('start_stream', {
  device_name: '[Loopback] Speakers (Realtek Audio)',
  ip: '192.168.1.100',
  port: 5000,
  sample_rate: 48000,
  buffer_size: 480, // Ignored for loopback
  ring_buffer_duration_ms: 8000,
  auto_reconnect: true,
  high_priority: false,
  dscp_strategy: 'voip',
  chunk_size: 960,
  is_loopback: true,
  enable_adaptive_buffer: true,
  min_buffer_ms: 4000,
  max_buffer_ms: 12000,
});

​
WiFi/Unstable Network

await invoke('start_stream', {
  device_name: 'Microphone (USB Audio)',
  ip: '192.168.1.100',
  port: 5000,
  sample_rate: 48000,
  buffer_size: 480,
  ring_buffer_duration_ms: 10000, // Larger buffer
  auto_reconnect: true,
  high_priority: true,
  dscp_strategy: 'voip',
  chunk_size: 2400, // Larger chunks
  is_loopback: false,
  enable_adaptive_buffer: true,
  min_buffer_ms: 5000,
  max_buffer_ms: 15000,
});

Build docs developers (and LLMs) love

Get started for freeTalk to us