> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/curator/_mcp/server. # Curate Audio ## Docs - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/tutorials.md): Collection of tutorials for audio curation workflows including beginner guides and advanced quality assessment techniques - [Beginner Tutorial](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/tutorials/beginner.md): Beginner tutorial for audio processing using the FLEURS dataset - [ALM Tutorial](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/tutorials/alm.md): Tutorial for curating audio language model training data using the ALM pipeline with window construction and overlap filtering - [ReadSpeech Tutorial](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/tutorials/readspeech.md): Tutorial for processing the DNS Challenge Read Speech dataset through AudioDataFilterStage with automatic download and configurable quality filters - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/load-data.md): Load audio datasets from various sources including FLEURS, custom manifests, and local files - [Custom Manifests](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/load-data/custom-manifests.md): Create and load custom audio manifests in JSONL format for NeMo Curator audio processing - [FLEURS Dataset](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/load-data/fleurs-dataset.md): Load and process the multilingual FLEURS speech dataset with automated download and manifest creation - [Local Files](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/load-data/local-files.md): Load audio files from local directories by creating custom manifests - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data.md): Process audio data using ASR inference, quality assessment, audio analysis, and text integration for high-quality speech datasets - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/asr-inference.md): Perform automatic speech recognition using NeMo Framework models with GPU acceleration and batch processing - [NeMo ASR Models](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/asr-inference/nemo-models.md): Guide to using NeMo Framework's pretrained ASR models for speech recognition in audio curation pipelines - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-assessment.md): Assess and filter audio quality using transcription accuracy metrics, duration analysis, and custom quality measures - [WER Filtering](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-assessment/wer-filtering.md): Filter audio samples based on Word Error Rate thresholds to ensure high-quality transcription accuracy - [Duration Filtering](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-assessment/duration-filtering.md): Filter audio samples by duration ranges, speech rate metrics, and temporal characteristics for optimal dataset quality - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering.md): Compose preprocessing, VAD, band, MOS scoring, and speaker separation stages to extract clean single-speaker training segments from raw audio - [Preprocessing Stages](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/preprocessing.md): Audio preprocessing stages — mono conversion, segment concatenation, and timestamp mapping for downstream filtering pipelines - [VAD Segmentation](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/vad.md): Voice activity detection with VADSegmentationStage to split audio into speech segments using Silero VAD - [Band Filter](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/band-filter.md): Filter audio by spectral bandwidth — full-band vs narrow-band — using BandFilterStage and a scikit-learn classifier - [UTMOS Filter](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/utmos.md): Filter audio segments by predicted Mean Opinion Score using UTMOSFilterStage and the utmos22_strong model - [SIGMOS Filter](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/sigmos.md): Filter audio across seven independent perceptual quality dimensions — noise, overall, signal, coloration, discontinuity, loudness, reverb — using SIGMOSFilterStage - [Speaker Separation](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/speaker-separation.md): Diarize audio with offline SortFormer or streaming Sortformer and fan out one task per speaker for downstream per-speaker filtering - [AudioDataFilterStage Composite](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/quality-filtering/audio-data-filter-stage.md): AudioDataFilterStage composite — end-to-end audio curation pipeline that decomposes into VAD, band, UTMOS, SIGMOS, and speaker separation sub-stages from a single YAML config - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/audio-analysis.md): Extract and analyze audio file characteristics including duration calculation, format validation, and metadata extraction - [Duration Calculation](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/audio-analysis/duration-calculation.md): Calculate precise audio duration using soundfile library for quality assessment and metadata generation - [Format Validation](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/audio-analysis/format-validation.md): Audio format support and error handling in NeMo Curator audio processing stages - [Overview](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/alm.md): ALM data curation stages for constructing and filtering training windows from diarized audio segments - [Data Builder](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/alm/data-builder.md): ALMDataBuilderStage reference for constructing training windows from diarized audio segments with quality filtering - [Overlap Filtering](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/alm/overlap-filtering.md): ALMDataOverlapStage reference for removing overlapping training windows based on configurable thresholds - [Text Integration](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/process-data/text-integration.md): Convert processed audio data to DocumentBatch format for downstream processing - [Save and Export](https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/save-export.md): Export processed audio data and transcriptions in formats optimized for ASR training and multimodal applications > **Note:** This page contains both a page directory (above) and the landing page content (below). The page directory is generated for agent use and does not appear on the landing page. > For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.nvidia.com/nemo/curator/v26.04/curate-audio/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/curator/_mcp/server. > Comprehensive audio curation capabilities for speech data processing including ASR inference, quality assessment, and text integration workflows # About Audio Curation NeMo Curator provides comprehensive audio curation capabilities to prepare high-quality speech data for automatic speech recognition (ASR) and multi-modal model training. The toolkit includes processors for loading audio datasets, performing ASR inference, assessing transcription quality, and integrating with text curation workflows. ## Use Cases * Process and curate large-scale speech datasets for ASR model training * Perform quality assessment and filtering based on transcription accuracy metrics * Generate transcriptions using state-of-the-art NVIDIA NeMo ASR models * Integrate audio processing with text curation pipelines for multi-modal workflows * Scale audio processing across GPU clusters efficiently --- ## Introduction Master the fundamentals of NeMo Curator and set up your audio processing environment. #### [Concepts](/about/concepts/audio) Learn about AudioTask, ASR pipelines, and other core data structures for efficient audio curation data-structures asr-pipeline quality-metrics #### [Get Started](/get-started/audio) Learn prerequisites, setup instructions, and initial configuration for audio curation setup configuration quickstart ## Curation Tasks ### Load Data Import your audio data from various sources into NeMo Curator's processing pipeline. #### [Local Files](/curate-audio/load-data/local-files) Load audio files from local directories and file systems local-storage file-discovery batch-processing #### [Custom Manifests](/curate-audio/load-data/custom-manifests) Create and load custom audio dataset manifests with metadata manifests metadata custom-formats #### [FLEURS Dataset](/curate-audio/load-data/fleurs-dataset) Load and process the multilingual FLEURS speech dataset fleurs multilingual benchmarks ### Process Data Transform and enhance your audio data through ASR inference, quality assessment, and analysis. #### [ASR Inference](/curate-audio/process-data/asr-inference) Generate transcriptions using NVIDIA NeMo ASR models nemo-models transcription gpu-accelerated #### [Quality Assessment](/curate-audio/process-data/quality-assessment) Assess transcription quality using WER and CER wer-filtering duration-filtering #### [Audio Analysis](/curate-audio/process-data/audio-analysis) Analyze audio characteristics including duration and format validation duration-calculation format-validation metadata-extraction #### [Text Integration](/curate-audio/process-data/text-integration) Integrate audio processing results with text curation workflows multimodal text-filtering pipeline-integration ### Save & Export Save processed audio data and transcriptions in formats suitable for downstream training and analysis. #### [Save & Export](/curate-audio/save-export) Export curated audio datasets with transcriptions and quality metrics manifests parquet metadata --- ## Tutorials Build practical experience with step-by-step guides for common audio curation workflows. #### [Beginner Tutorial](/curate-audio/tutorials/beginner) Learn the basics of audio loading, ASR inference, and quality filtering asr-inference quality-filtering basic-workflow