Sound Tools eXtended (STx) 5.0 — A Powerful Sound Analysis Tool Optimized for Speech

Anton Noll, Jonathan Stuefer, Nicola Klingler, Hannah Leykum, Carina Lozo, Jan Luttenberger, Michael Pucher, Carolin SchmidInterspeech 2019

TL;DR

This paper introduces Sound Tools eXtended (STx) 5.0, a free, powerful, and user-friendly software for acoustic speech and sound analysis. It's designed to simplify complex signal processing for researchers in humanities and related fields, offering an intuitive interface, robust analysis tools, and seamless integration with existing workflows.

What Problem Does This Paper Address?

Many researchers in fields like linguistics, philology, and psychology need to conduct data-driven research involving speech analysis. However, they often lack the specialized technical training required to use complex signal processing software. Existing tools can be cumbersome, especially when dealing with field recordings that are not acoustically pristine or when analyzing data from multiple speakers. Furthermore, automatic analysis tools can produce inaccurate results without careful parameter tuning, and integrating segmented data into statistical analyses can be a significant hurdle. STx 5.0 aims to bridge this gap by providing an accessible yet powerful solution.

Methods

STx 5.0 is presented as a significant update to an existing software suite, focusing on an improved user experience. The core innovation highlighted is the new workspace mode, which simplifies the interface by hiding features not directly relevant to speech analysis. This mode is designed to be intuitive for users without extensive technical backgrounds. The software offers a range of signal processing features, including frequency analysis (FFT, DFT, wavelet, filter banks), spectral smoothing, de-noising, and signal enhancement. For speech analysis specifically, STx 5.0 includes a built-in recorder, supports various import/export formats (including TextGrid), and offers tools for manual segmentation and correction of acoustic measurements. It also incorporates a call function for WebMAUS, a web application for automatic speech segmentation.

Datasets

Not discussed in the paper. The paper focuses on the software's capabilities rather than specific datasets used for training or evaluation.

Key Insights and Ideas

  • Democratizing Speech Analysis: The primary insight is that advanced acoustic analysis tools can be made accessible to a broader range of researchers, not just those with deep technical expertise. The simplified GUI and workspace mode are key to this.
  • Bridging the Gap Between Humanities and Technical Fields: STx 5.0 acts as a crucial intermediary, allowing scholars in humanities to leverage sophisticated signal processing without needing to become software engineers.
  • Workflow Integration: The software's compatibility with established tools like TextGrid and its ability to export data in common formats (CSV, XLS, XLSX, TXT) are vital for integrating acoustic analysis into larger research projects and statistical workflows.
  • Flexibility for Real-World Data: The ability to manually correct measurements and handle non-ideal recording conditions is a significant advantage for researchers working with field data.
  • Data Protection Features: The anonymization tool for marking and replacing sensitive data is a practical and important feature for ethical research.

Key Results

The paper doesn't present quantitative results in terms of accuracy or performance benchmarks. Instead, it emphasizes the qualitative improvements and new features of STx 5.0. The key "results" are the enhanced usability and expanded functionality that enable:

  • Simplified speech analysis for new users and non-technical researchers.
  • Efficient handling of large quantities of audio data through project organization and grouping.
  • Streamlined segmentation and analysis of long sound files.
  • Improved workflow efficiency with customizable hotkeys and a user-friendly interface.
  • Seamless integration with existing speech analysis tools and workflows.

Limitations and Future Work

The authors acknowledge that while STx 5.0 aims to simplify analysis, some advanced features might still require programming knowledge or advanced computer skills. They also mention that the software is "aimed to be constantly improved" and explicitly welcome "all feedback and improvement ideas." Specific future work directions are not detailed, but the emphasis on continuous improvement suggests ongoing development based on user needs.

Summary updated

BibTeX
@inproceedings{noll2019sound,
  author = {Anton Noll and Jonathan Stuefer and Nicola Klingler and Hannah Leykum and Carina Lozo and Jan Luttenberger and Michael Pucher and Carolin Schmid},
  title = {{Sound Tools eXtended (STx) 5.0 — A Powerful Sound Analysis Tool Optimized for Speech}},
  booktitle = {Interspeech},
  year = {2019},
}
Figure