ComfyUI Node: RunningHub VoxCPM Multi-Speaker (Dynamic Audio)

RunningHub VoxCPM Multi-Speaker (Dynamic Audio):

The RunningHub_VoxCPM_MultiSpeaker_ListReference node is designed to facilitate dynamic audio processing for multi-speaker scenarios using the VoxCPM model. This node allows you to input multiple audio references, enabling the generation of speech that can mimic different speakers based on the provided audio samples. It is particularly beneficial for applications requiring diverse voice outputs, such as virtual assistants, audiobooks, or any AI-driven voice synthesis tasks. By leveraging this node, you can achieve a more natural and varied speech synthesis, enhancing the overall user experience with dynamic and contextually appropriate audio outputs.

RunningHub VoxCPM Multi-Speaker (Dynamic Audio) Input Parameters:

model

The model parameter specifies the VoxCPM model to be used for speech synthesis. This is a required input and determines the underlying architecture and capabilities of the speech generation process. The model influences the quality and characteristics of the generated speech, making it crucial to select a model that aligns with your specific needs.

script

The script parameter is a string input that contains the text to be converted into speech. It supports dynamic speaker tags, such as [spk5], allowing you to specify which speaker's voice should be used for different parts of the text. This flexibility enables the creation of multi-speaker dialogues or narratives within a single script. The default value is set to a sample text, but you can customize it to suit your content requirements.

normalize_text

The normalize_text parameter is a boolean option that, when enabled, normalizes the input text to ensure consistent formatting and pronunciation. This can be particularly useful for handling variations in text input, such as abbreviations or numbers, to produce a more natural-sounding speech output. The default value is False.

denoise_reference

The denoise_reference parameter is a boolean option that, when enabled, applies noise reduction to the reference audio inputs. This can improve the clarity and quality of the generated speech, especially when the reference audio contains background noise or other distortions. The default value is False.

max_len

The max_len parameter is an integer that defines the maximum length of the generated speech in terms of text characters. It helps manage the processing load and ensures that the output remains within a manageable size. The default value is 4096, with a minimum of 64 and a maximum of 8192, adjustable in steps of 64.

retry_badcase

The retry_badcase parameter is a boolean option that, when enabled, allows the node to retry generating speech if the initial attempt results in poor quality or errors. This can enhance the reliability of the output by providing additional opportunities to achieve a satisfactory result. The default value is True.

audio_1, audio_2, ..., audio_n

These parameters represent optional audio inputs that serve as reference samples for different speakers. Each audio_n input is expected to be an audio file that the node uses to mimic the corresponding speaker's voice. The number of audio inputs can be dynamically adjusted based on your requirements, allowing for flexible multi-speaker configurations.

RunningHub VoxCPM Multi-Speaker (Dynamic Audio) Output Parameters:

synthesized_audio

The synthesized_audio output is the generated speech audio file that results from processing the input script and reference audio samples. This output is the primary deliverable of the node, providing a synthesized voice that reflects the characteristics of the specified speakers. The quality and accuracy of this output depend on the input parameters and the selected model.

RunningHub VoxCPM Multi-Speaker (Dynamic Audio) Usage Tips:

• Ensure that your reference audio samples are clear and representative of the desired speaker's voice to achieve the best synthesis results.
• Use the normalize_text option to handle diverse text inputs consistently, especially when dealing with complex scripts or multiple languages.
• Experiment with different max_len settings to balance between processing time and output length, particularly for longer scripts.

RunningHub VoxCPM Multi-Speaker (Dynamic Audio) Common Errors and Solutions:

"Model not found"

• Explanation: This error occurs when the specified VoxCPM model is not available or incorrectly referenced.
• Solution: Verify that the model name is correct and that the model is properly installed and accessible in your environment.

"Invalid audio input"

• Explanation: This error indicates that one or more of the audio inputs are not in the expected format or are corrupted.
• Solution: Check the format and integrity of your audio files, ensuring they are compatible with the node's requirements.

"Text length exceeds maximum limit"

• Explanation: This error arises when the input script exceeds the specified max_len parameter.
• Solution: Reduce the length of your script or increase the max_len parameter to accommodate longer texts.