Toward the long-term evolution of the xkl speech analysis software

A legacy software, “xkl,” developed in the 1980’s by Dennis Klatt in Kenneth N. Stevens’ Speech Communication Group at the Massachusetts Institute of Technology (MIT) has been recently revamped, thanks to a joint effort between MIT and Sapienza University of Rome (https://www.sciencedirect.com/science/article/pii/S2352711023001887). This collaboration led to the development of a modern xkl, which is now compatible with modern computing platforms (Windows, macOS, Linux) (more on current software developments can be found at http://acts.ing.uniroma1.it/xkl.php). The original development was based on Motif libraries, but these libraries currently are not being updated. To ensure long-term compatibility, the xkl graphical user interface is currently being redesigned using GTK. Moreover, the software will include the possibility of computing the reassigned spectrogram, enriching the available time-frequency representations, and facilitating the process of performing detailed analyses of the properties of speech, such as formant estimation. The xkl software has superior qualities for acoustic parameter estimation, in particular, determination of spectral peaks and their evolution in time, thanks to a flexible interface that allows the investigator to conveniently sweep through successive spectral slices. This new-xkl is currently under use in preliminary analyses of vowels in the LaMIT dataset, a set of 100 sentences in Italian (see https://www.sciencedirect.com/science/article/pii/S2352340922004772).