bob_q3tts.tts_node

Classes

ROS node that provides an interface to Qwen3-TTS with streaming text aggregation.

Functions

main([args])

Start the TTS node.

Module Contents

class bob_q3tts.tts_node.TTSnode

Bases: rclpy.node.Node

ROS node that provides an interface to Qwen3-TTS with streaming text aggregation.

text_buffer = ''

last_text_time

text_queue

audio_queue

lock

file_index

inferred_target_sr = None

_is_speaking = False

_speaking_finish_time = 0.0

sub

pub

audio_pub

speaking_pub

model = None

sd = None

sf = None

torch = None

tts_thread

audio_thread

flush_timer

speaking_monitor_timer

_get_sf(): Lazy load soundfile library.

text_callback(msg): Aggregate incoming words/text into sentences based on delimiters.

flush_timer_callback(): Flush the text buffer if the timeout has passed.

_apply_substitutions(text): Apply regex-based substitutions to the text.

tts_loop(): Process the text queue and generate audio.

audio_loop(): Consume the audio queue for playback and storage.

play_audio(audio_data, sr): Play audio using native library or configured executable.

_do_native_play(audio_data, sr): Perform native audio playback using sounddevice.

resample_audio(audio_data, orig_sr, target_sr): Resample audio data using librosa if available, else linear interpolation.

_play_with_executable(executable, audio_data, sr): Play audio using an external command.

speaking_monitor_callback(): Monitor the speaking duration and reset flag after delay.

bob_q3tts.tts_node.main(args=None): Start the TTS node.