Nokia Corporation

Morten With Pedersen

Carsten Juncker

Pedro Hojen-Sorensen

Kåre Jean Jensen

Søren Riis

Embodiments of the invention include apparatuses, systems, computer readable media, and methods for processing speech signals in a manner that enhances capacity, efficiency and hardware utilization of a communications network. A method, according to one embodiment, includes receiving speech signals, determining a subchannel power imbalance ratio of at least two subchannels, and selecting a receiver architecture for processing the speech signals in accordance with the determined subchannel power imbalance ratio.

VAMOS&#x2014;DARP receiver switching for mobile receivers

An approach is provided for estimating a channel of a multi-carrier transmission by splitting an interpolation or extrapolation or smoothing of estimates into a frequency domain part and a time domain part, wherein the channel estimation unit is configured to use for both the frequency domain part and the time domain part respective recursive filters, and wherein at least one of the recursive filters is adapted to channel conditions of the multi-carrier transmission.

Method and apparatus of recursive time-frequency channel estimation

There is provided a novel approach for generating multilingual text-to-phoneme mappings for use in multilingual speech recognition systems. The multilingual mappings are based on the weighted output from a neural network text-to-phoneme model, trained on data mixed from several languages. The multilingual mappings used together with a branched grammar decoding scheme is able to capture both inter- and intra-language pronunciation variations which is ideal for multilingual speaker independent recognition systems. A significant improvement in overall system performance is obtained for a multilingual speaker independent name dialing task when applying multilingual instead of language dependent text-to-phoneme mapping.

Growing community of inventors

Frederiksberg, Denmark

Morten With Pedersen