ComfyUI Node: Audio Plotter

Audio Plotter:

The Audio Plotter node is designed to visually represent audio data through waveforms and spectrograms, providing a comprehensive view of audio signals for analysis and comparison. This node is particularly useful for audio engineers and AI artists who need to evaluate the differences between original and processed audio files. By plotting both waveforms and spectrograms, the Audio Plotter allows users to visually assess the impact of audio processing techniques, such as super-resolution or noise reduction, on the audio signal. The node supports toggling between different types of plots, enabling users to focus on specific aspects of the audio data. This visualization capability is essential for understanding the nuances of audio processing and ensuring that the desired audio quality is achieved.

Audio Plotter Input Parameters:

name

The name parameter is a string that serves as an identifier for the test case or audio plot. It allows you to organize and structure your plots by using a hierarchical naming convention, such as taskcase_1/test_set_1. This helps in categorizing and retrieving plots easily, especially when dealing with multiple audio files or test cases.

audio_dict

The audio_dict parameter is a dictionary where each key is a string representing the audio name, and each value is a 1D audio array (numpy ndarray). This parameter provides the audio data that will be plotted. The dictionary format allows for multiple audio files to be processed and visualized simultaneously, facilitating comparison between different audio samples.

global_step

The global_step parameter is an integer that represents the current step or iteration in a process, such as training or evaluation. It is used to track the progress of the process and is often included in the plot's metadata to provide context about when the plot was generated.

sample_rate

The sample_rate parameter is an integer that specifies the number of samples per second in the audio data. It is crucial for accurately plotting the audio waveforms and spectrograms, as it determines the time resolution of the audio signal. The default value is 16000, which is a common sample rate for audio processing.

is_plot_spec

The is_plot_spec parameter is a boolean that indicates whether to plot the spectrogram of the audio data. When set to True, the node will generate a visual representation of the audio's frequency content over time. This is useful for analyzing the spectral characteristics of the audio signal.

is_plot_mel

The is_plot_mel parameter is a boolean that determines whether to plot the mel spectrogram of the audio data. The mel spectrogram provides a perceptually meaningful representation of the audio's frequency content, making it easier to analyze and interpret. The default value is True.

mel_spec_args

The mel_spec_args parameter is an optional dictionary that contains configuration settings for generating the mel spectrogram. If not provided, default settings will be used. This parameter allows for customization of the mel spectrogram, such as adjusting the number of mel bands or the window size, to suit specific analysis needs.

Audio Plotter Output Parameters:

img_waves

The img_waves output parameter is an image that visually represents the waveforms of the original, processed, and null audio signals. This output is crucial for visually comparing the amplitude and time-domain characteristics of different audio signals, helping users to identify any discrepancies or changes introduced by audio processing.

img_spec

The img_spec output parameter is an image of the spectrograms for the original, processed, and null audio signals. Spectrograms provide a time-frequency representation of the audio, allowing users to analyze the spectral content and identify any changes in frequency components due to processing.

img_diff

The img_diff output parameter is an image that highlights the differences between the spectrograms of the original and processed audio signals. This output is particularly useful for identifying specific frequency ranges where processing has had a significant impact, aiding in the fine-tuning of audio processing techniques.

Audio Plotter Usage Tips:

• Use hierarchical naming conventions for the name parameter to organize and easily retrieve plots, especially when dealing with multiple test cases or audio files.
• Adjust the sample_rate parameter to match the sample rate of your audio data for accurate time-domain and frequency-domain representations.
• Enable is_plot_mel to generate mel spectrograms, which provide a more perceptually relevant analysis of the audio's frequency content.
• Customize the mel_spec_args to fine-tune the mel spectrogram settings, such as the number of mel bands or window size, to better suit your analysis needs.

Audio Plotter Common Errors and Solutions:

"Audio dict missing samples/waveform"

• Explanation: This error occurs when the audio_dict parameter does not contain valid audio data or is missing the required samples.
• Solution: Ensure that the audio_dict is correctly populated with audio names as keys and 1D audio arrays as values. Verify that the audio data is properly loaded and formatted before passing it to the node.

"Unsupported AUDIO object for this node"

• Explanation: This error indicates that the audio data provided is not in a supported format or structure.
• Solution: Check that the audio data is provided as a dictionary with the correct keys and values. Ensure that the audio arrays are in the expected 1D format and that the sample rate is specified.