Adaptive Multi-Rate audio codec

About: Adaptive Multi-Rate audio codec is a research topic. Over the lifetime, 1467 publications have been published within this topic receiving 19736 citations. The topic is also known as: AMR & Adaptive Multi-Rate.

Uncoded bit error rate of eCall modem through AMR codec and AWGN channel

Abstract: Data communications through the voice channel of the cellular network are important for telematics applications such as automatic vehicle crash notification, an implementation of which is the pan-European eCall system scheduled for deployment in 2015. The cellular voice channel includes the vocoder which is designed for compressing speech waveforms and is a hindrance for data communications. No mathematical model exists for the vocoder as a data communications channel. Thus, in this paper, the uncoded bit error rate (BER) is found by simulation using the bit-exact C-code implementation of the vocoder. The adaptive multi-rate (AMR) vocoder is used in GSM, the network on which eCall runs, and is the one used in these simulations. The BER is first presented for the vocoder-only channel at different AMR compression rates. Then it is presented for the vocoder and additive white Gaussian noise (AWGN) channel. Finally, an attempt is made to find the optimum receiver filter for the vocoder channel.

Using MPEG-4 audio for DRM digital narrowband broadcasting

Abstract: DRM (Digital Radio Mondiale) is an international consortium, working on the definition of a world-wide standard for digital narrowband broadcasting in short- and medium-wave bands. One major requirement for a digital system in these bands is the bandwidth compatibility with existing analogue services. This means that the system needs to operate with a channel bandwidth of 9 or 10 kHz. Therefore DRM has a need for highly efficient audio and speech coding. A further requirement is error robustness. Due to the fading characteristics of the channel, signal errors may occur in spite of digital modulation and channel coding. Last but not least DRM wants to use an international standard to make sure that widespread, well-tested and well-supported algorithms are being used. Following these requirements, DRM has chosen parts of MPEG-4 audio for the DRM system. Based on listening tests performed in cooperation with the NADIB project, the European predecessor of DRM, MPEG has proposed audio and speech coding algorithms suitable for use within DRM. Since then, more work on these algorithms has been done within Version 2 of MPEG-4. MPEG-4 offers highly efficient audio coding by the AAC algorithm as well as state-of-the-art CELP speech coding. Both algorithms are very flexible in terms of codec configuration and bit rate. In addition to that, they support special error robust modes for use in error prone channels. DRM has chosen MPEG-4 AAC for audio coding. The MPEG-4 CELP is a very good candidate for the speech coding part of the system. The DRM specification will be finalized in 2000.

130 Mbit/s (H4 rate) HDTV codec based on the DCT algorithm

Abstract: The hardware implementation of a 130 Mbit/s (H4 rate) HDTV CODEC based on the DCT (discrete cosine transform) algorithm is studied for transmission in broadband ISDN (BISDN). The intrafield 2-dimensional DCT, nonlinear fixed bit quantisation of eight classes with switchable operation between y and Pr/Pb, and the four parallel signal processing possible by DCT are also implemented in the CODEC. Performance evaluation confirms that the developed CODEC has sufficient picture quality and is practical for HDTV distribution use.

Synchronizing a channel codec and vocoder of a mobile station

Abstract: In one embodiment, the present invention includes a method for maintaining a vocoder and channel codec in substantial synchronization. The method may include receiving a configuration message that includes rate information and an effective radio block identifier at a mobile station, coding a current radio block via a vocoder and channel codec, configuring an encoding portion of the vocoder and channel codec with the rate information after performing the coding, and then coding the effective radio block using the rate information. Other embodiments are described and claimed.

Joint encoding of the waveform and speech recognition features using a transform codec

Abstract: We propose a new transform speech codec that jointly encodes a wideband waveform and its corresponding wideband and narrowband speech recognition features. For distributed speech recognition, wideband features are compressed and transmitted as side information. The waveform is then encoded in a manner that exploits the information already captured by the speech features. Narrowband speech acoustic features can be synthesized at the server by applying a transformation to the decoded wideband features. An evaluation conducted on an in-car speech recognition task show that at 16 kbps our new system typically shows essentially no impact in word error rate compared to uncompressed audio, whereas the standard transform codec produces up to a 20% increase in word error rate. In addition, good quality speech is obtained for playback and transcription, with PESQ scores ranging from 3.2 to 3.4.