ComfyUI Node: Egregora Resample Audio (HQ)

Resample Audio (HQ):

The Resample Audio (HQ) node is designed to change the sample rate of audio data with high quality, ensuring that the audio maintains its integrity and clarity during the conversion process. This node is particularly useful when you need to match the sample rate of different audio files or when preparing audio for specific playback or processing requirements. By utilizing advanced resampling techniques, such as those provided by librosa, resample_poly, or torchaudio, the node offers flexibility and precision in handling audio data. The primary goal of this node is to facilitate seamless audio sample rate conversion, making it an essential tool for audio processing tasks where maintaining audio quality is crucial.

Resample Audio (HQ) Input Parameters:

audio

This parameter represents the audio data to be resampled. It can be provided as a NumPy array or a PyTorch tensor, with the shape either being (channel, num_samples) for multi-channel audio or (num_samples) for single-channel audio. The audio data serves as the primary input for the resampling process, and its quality and format will directly impact the output.

origin_sr

The origin_sr parameter specifies the original sample rate of the input audio. It is an integer value that indicates how many samples per second the audio was originally recorded or stored at. This parameter is crucial for accurately converting the audio to the desired target sample rate.

target_sr

This parameter defines the target sample rate to which the audio should be resampled. Like origin_sr, it is an integer value representing the number of samples per second for the output audio. The target_sr determines the new resolution of the audio data, affecting its playback speed and quality.

resample_module

The resample_module parameter allows you to choose the resampling library or method to be used. Options include 'librosa', 'resample_poly', and 'torchaudio', with 'librosa' being the default. Each option offers different algorithms and performance characteristics, allowing you to select the most suitable method for your specific needs.

resample_type

This parameter specifies the type of resampling algorithm to be used when librosa is selected as the resample_module. The default value is "kaiser_fast", which is a commonly used algorithm for its balance between speed and quality. This parameter is only applicable when using librosa and allows for further customization of the resampling process.

audio_path

The audio_path parameter is optional and can be used to specify the file path of the audio data. This is useful if you want to keep track of the source file or if additional metadata is required for processing. However, it is not necessary for the resampling operation itself.

Resample Audio (HQ) Output Parameters:

resampled_audio

The resampled_audio is the primary output of the node, representing the audio data after it has been resampled to the target sample rate. This output maintains the same format as the input, either as a NumPy array or a PyTorch tensor, and reflects the changes made during the resampling process. The quality and characteristics of this output are directly influenced by the input parameters and the chosen resampling method.

Resample Audio (HQ) Usage Tips:

• When working with high-quality audio, consider using the torchaudio resample module for efficient processing, especially if you are resampling multiple audio files with the same parameters.
• If you need a balance between speed and quality, the librosa module with the kaiser_fast resample type is a good choice for most applications.
• Always ensure that the origin_sr and target_sr parameters are set correctly to avoid unexpected changes in audio speed or pitch.

Resample Audio (HQ) Common Errors and Solutions:

"Sample rate mismatch"

• Explanation: This error occurs when the origin_sr and target_sr parameters are set to the same value, making resampling unnecessary.
• Solution: Double-check the sample rates and ensure they are different if resampling is intended. If not, you can skip the resampling process.

"Invalid resample module"

• Explanation: This error indicates that the specified resample_module is not recognized or supported.
• Solution: Verify that the resample_module is set to one of the supported options: 'librosa', 'resample_poly', or 'torchaudio'.

"Audio data format error"

• Explanation: This error suggests that the input audio data is not in the expected format or shape.
• Solution: Ensure that the audio data is provided as a NumPy array or PyTorch tensor with the correct shape, either (channel, num_samples) or (num_samples).