Browse Prior Art Database

Generic Multimedia Codec Payload Format for Unequal Error Protection

IP.com Disclosure Number: IPCOM000020126D
Published in the IP.com Journal: Volume 3 Issue 11 (2003-11-25)
Included in the Prior Art Database: 2003-Nov-25
Document File: 6 page(s) / 143K

Publishing Venue

Siemens

Related People

Juergen Carstens: CONTACT

Abstract

Today's multimedia codecs generate frames consisting of bit groups with different grades of sensitivity for transmission errors, also known as "priority classes". For instance, an AMR (Adaptive Multirate Codec) generates media frames consisting of 3 bit groups: class A bits, class B bits and class C bits. Class A bits are very sensitive to transmission errors: if there is an error in the class A bits received at the decoder side, this may cause a very annoying sound after decoding. Thus a very tight error protection and detection is required for class A bits. On the contrary, class C bit errors can at most affect the background noise. "Unequal Error Protection" (UEP) allows protecting each priority class differently with an appropriate grade of error protection. Possible error protections can range from simply adding a checksum for error detection, adding different grades of redundancy to allow "forward error correction" (FEC), retransmission of frames with errors (ARQ, Automatic ReQuest of Retransmission) etc. The use of UEP means that the media frames need to be split in groups of bits belonging to the same priority class, and then a different error protection mechanism is applied to each such group. This frame splitting as well as the application of error protections is most efficient when applied at the lower OSI (Open System Interconnection) layers (physical and/or link layer). E.g. for UEP of the UMTS (Universal Mobile Telecommunications System) radio interface this means splitting the media stream into multiple Radio Bearers (RB), whereby each RB has different protection mechanisms. The equipment that performs the frame splitting task for UEP must know the first and last bit of each priority class in the media frames received from the encoder. For the more complex codecs such as H.263, MPEG-4, etc these boundaries are varying from frame to frame.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 16% of the total text.

Page 1 of 6

S

© SIEMENS AG 2003 file: 2003J13393.doc page: 1

Generic Multimedia Codec Payload Format for Unequal Error Protection

Idea: Eddy Troch, BE-Herentals

Today's multimedia codecs generate frames consisting of bit groups with different grades of sensitivity for transmission errors, also known as "priority classes". For instance, an AMR (Adaptive Multirate Codec) generates media frames consisting of 3 bit groups: class A bits, class B bits and class C bits. Class A bits are very sensitive to transmission errors: if there is an error in the class A bits received at the decoder side, this may cause a very annoying sound after decoding. Thus a very tight error protection and detection is required for class A bits. On the contrary, class C bit errors can at most affect the background noise. "Unequal Error Protection" (UEP) allows protecting each priority class differently with an appropriate grade of error protection. Possible error protections can range from simply adding a checksum for error detection, adding different grades of redundancy to allow "forward error correction" (FEC), retransmission of frames with errors (ARQ, Automatic ReQuest of Retransmission) etc. The use of UEP means that the media frames need to be split in groups of bits belonging to the same priority class, and then a different error protection mechanism is applied to each such group. This frame splitting as well as the application of error protections is most efficient when applied at the lower OSI (Open System Interconnection) layers (physical and/or link layer). E.g. for UEP of the UMTS (Universal Mobile Telecommunications System) radio interface this means splitting the media stream into multiple Radio Bearers (RB), whereby each RB has different protection mechanisms. The equipment that performs the frame splitting task for UEP must know the first and last bit of each priority class in the media frames received from the encoder. For the more complex codecs such as H.263, MPEG-4, etc these boundaries are varying from frame to frame.

There exist several possibilities to perform UEP up to now. UEP is not yet widely deployed in commercial products. One exception is the usage of UEP for the transmission of AMR encoded speech media over circuit switched radio bearers in the GSM or UMTS network. AMR codecs have a small number of "codec modes", and for each codec mode the boundaries between class A, B and C bits are fixed. This codec mode is also included in the header of each media frame. At either side of the radio link, the terminal resp. RNC (Radio Network Controller) are performing the splitting of radio stream. The encoder at the network side is located at MSC (Mobile Switching Center). At call establishment, the terminal signals to the MSC a list of possible frame formats (boundary table), i.e. specifying the class A,B and C bit length for each format, together with the QoS (Quality of Service) parameters to be applied for each RB. The error protection mechanism t...