News

---

Paper

"Oral Billiards (A Physical Model Of Speech)" (arxiv.org/2002.00791).

--TEReisler, 31-Dec-2020

---

Software updates

Version 1.2.2 of Libalinga, java native interface included.

--TEReisler, 15-Sep-2016

---

Software updates

Version 1.2.1 of Libogg++, 0.2.1 of Libneuro and 1.2.1 of Libalinga.

--TEReisler, 06-Nov-2013

---

Paper

"A 10-dimensional Phonetic-prosodic Space and its Stochastic Structure (A framework for probabilistic modeling of spoken languages and their phonology) " (arxiv.org/1306.2593).

--TEReisler, 08-Jul-2013

---

Paper

"Maximally Informative Observables and Categorical Perception" (arxiv.org/1212.5091) on the theory underlying the computational modeling.

--TEReisler, 08-Jul-2013

---

FreeCLAS will be in transition to a new presentation form. In the meantime, the content will be made static, and may contain dead links. All downloadables are available as before.

--TEReisler, 13-Jul-2011

---

Release 1.2.0 of Libogg++

--TEReisler, 24-Jun-2011

---

New releases of both the AnnotationTools and Monowave80 are being loaded. You will need the latest version of everything to use the latest version of Shva to see the addional views available for Monowave80, which will be explained in additional screen movies.

--TEReisler, 04-Jun-2009

---

Version 1.1.0 of Libogg++, 0.2.0 of Libneuro and 1.0.0 of Libalinga have been released.

--TEReisler, 09-Jun-2009

---

Version 0.2.0 of Shva has been released.

--TEReisler, 09-Jun-2009

---

First utterance with a 2D derived signal viewable in shva 0.2.0.

--TEReisler, 10-Jun-2009

---

A new release of Monowave80 with cochleargrams interleaved is now available.

--TEReisler, 11-Jun-2009

---

Using FreeCLAS

FreeCLAS stands for "Free Commons of Linguistically Annotated Speech".

The motivation for FreeCLAS is to build a data base of high-quality speech data. That means annotated data that have been validated by humans. Building such a data base has been expensive because it requires substantial investment of people's attention. As a result, high-quality speech data is not generally available.

This problem is similar to the one that the wikipedia project tackled. This project aims to use the same method, a wiki, to make it possible for people to build FreeCLAS via collaborative contributions over the web. In return, everybody has free and open access to the results.

Although this wiki[1] is in English at this point, the speech data will be multilingual. As a first step, the phonetic annotations will use the International Phonetic Alphabet.

In addition to the data base, this project will also develop the formats and tools needed to build it. One of the tools that we will use is Shva, a graphical editor that opens from your browser to let you hear, view and annotate any utterance in FreeCLAS. To begin, you can use shva to review or edit Monowave80, a small collection of annotated utterances. In the course of improving this seed data base, we hope to improve the current AnnotationTools, annotate more data, and build more tools. That is why all the software here is Free Software.

FreeCLAS is an iheartm project.

• You are free to download any data or software presented here.