Digital Speech Processing

Services & Products

Corpus Collection

The Digital Speech Processing Group at ARCON Corporation is available to generate speech corpora for use in speech science or speech communication research and test & evaluation purposes. Over several years, ARCON has collected a number of speech data bases. These have included:

LL/FBI - a custom corpus of speech in reverberant and babble environments using specialized sensors.
LL/ASE - an extensive multi-sensor corpus in extreme acoustic noise environments with acoustic and non-acoustic sensors.
DDVPC - a test corpus using resident sensors in military acoustic noise environments.
In-house - a test corpus recorded in typical mobile phone environments.

The generation of a speech corpus requires a diverse set of experience and capabilities. These include:

Corpus design,
Multi-channel capability,
Storage and distribution,
Acoustic and non-acoustic sensors,
Human Factors,
Noise field simulation,
Audio, computer, human factors expertise,
Scripted and conversational material,
Resident and calibration sensors.

The Digital Speech Processing Group at ARCON Corporation has these capabilities and the required experience.