ComfyUI Node: Whisper Transcribe Batch

Whisper Transcribe Batch:

The TT-WhisperTranscriptionBatch node is designed to facilitate the batch processing of audio files for transcription purposes. It leverages a transcription pipeline to convert audio data into text, making it an invaluable tool for users who need to transcribe multiple audio files efficiently. This node is particularly beneficial for AI artists and other users who work with large volumes of audio data and require a streamlined method to generate text transcriptions. The node not only transcribes the audio but also offers options to save the transcriptions and their corresponding time-stamped chunks, providing a comprehensive solution for managing and utilizing transcribed data. Its ability to handle errors gracefully and provide detailed feedback on any issues encountered during the transcription process further enhances its utility and reliability.

Whisper Transcribe Batch Input Parameters:

pipeline

This parameter specifies the transcription pipeline to be used for processing the audio files. It is crucial as it determines the method and accuracy of the transcription process. The pipeline is expected to be a pre-configured transcription model that can handle the audio data provided. There are no specific minimum or maximum values, but it must be a valid transcription pipeline.

wav_bytes_batch

This parameter is a batch of audio files in byte format that the node will process. It is essential for the node's operation as it provides the raw audio data that needs to be transcribed. The batch can contain multiple audio files, allowing for efficient processing of large datasets. There are no specific minimum or maximum values, but it should be a valid batch of audio bytes.

format_newlines_on_punctuation

This boolean parameter determines whether newlines should be inserted after punctuation marks in the transcribed text. When set to true, it enhances the readability of the transcription by formatting it into more natural, sentence-like structures. The default value is true, and it can be set to either true or false.

save_transcription

This boolean parameter indicates whether the transcribed text should be saved to a file. If set to true, the node will save the transcription as a text file, allowing for easy access and storage of the transcribed data. The default value is false, and it can be set to either true or false.

save_chunks

This boolean parameter specifies whether the time-stamped chunks of the transcription should be saved to a CSV file. This is useful for users who need detailed information about the timing of each segment of the transcription. The default value is false, and it can be set to either true or false.

filename_prefix

This string parameter allows you to specify a prefix for the filenames of the saved transcription and chunk files. It is useful for organizing and categorizing the output files. The default value is an empty string, and it can be set to any valid string.

filename_suffix

This string parameter allows you to specify a suffix for the filenames of the saved transcription and chunk files. Similar to the prefix, it helps in organizing the output files. The default value is an empty string, and it can be set to any valid string.

overwrite_existing

This boolean parameter determines whether existing files with the same name should be overwritten. If set to true, the node will overwrite any existing files, ensuring that the latest transcription is saved. If set to false, the node will increment the filename to avoid overwriting. The default value is true, and it can be set to either true or false.

Whisper Transcribe Batch Output Parameters:

None

The TT-WhisperTranscriptionBatch node does not produce any direct output parameters. Instead, its primary function is to process and save transcriptions and their corresponding chunks to files, based on the input parameters provided.

Whisper Transcribe Batch Usage Tips:

• Ensure that the pipeline parameter is set to a valid and well-configured transcription model to achieve accurate results.
• Use the format_newlines_on_punctuation parameter to improve the readability of the transcriptions, especially if the text will be used for presentations or publications.
• Consider setting save_transcription and save_chunks to true if you need to keep a record of the transcriptions and their time-stamped segments for future reference.
• Utilize the filename_prefix and filename_suffix parameters to organize your output files systematically, which can be particularly helpful when dealing with large batches of audio files.

Whisper Transcribe Batch Common Errors and Solutions:

Some files failed to transcribe

• Explanation: This error occurs when one or more audio files in the batch could not be processed successfully. The error message will include details about the specific files and the nature of the error.
• Solution: Check the error details provided in the message to identify the cause. Common issues might include corrupted audio files or unsupported formats. Ensure that all audio files are in a compatible format and try running the node again.

File already exists and cannot be overwritten

• Explanation: This error occurs when the node attempts to save a transcription or chunk file, but a file with the same name already exists and overwrite_existing is set to false.
• Solution: Either set overwrite_existing to true to allow the node to overwrite existing files, or ensure that the filename_prefix and filename_suffix parameters are set to create unique filenames for each output.