Handling real-time transport protocol (RTP) media packets in voice over internet protocol (VoIP) terminal

BACKGROUND OF THE INVENTION

Field of the Invention

The present invention relates handling Real-time Transport Protocol (RTP) media packets in a Voice over Internet Protocol (VoIP) terminal, and more particularly, to an apparatus and method of handling RTP media packets in a VoIP terminal, allowing network resources to be more efficiently used in handling voice packets and transmitting the voice packets to a correspondent at either a VoIP wired terminal or a Voice over Wireless LAN (VoWLAN) terminal using Wireless LAN (WLAN) and a VoIP technology.

Description of the Related Art

In delivering video, voice, and facsimile messages over the Internet, a Voice over Internet Protocol (VoIP) system transmits real-time media such as voice and video when a user desiring to use the Internet gains access to the Internet by using a Personal Computer (PC), by using any independent device with Internet Protocol applied thereto, or by making a call toward a gateway with an existing Public service Telephone Network (PSTN) phone.

The VoIP system is used because it has the following advantages.

First, integration of a telephone network and a data network reduces investment cost for network equipment. A telephone network for voice communication and a data network for data communication are not disposed separately, such that the investment cost for network equipment is saved. Second, the integrated network reduces management cost and improves efficiency. By handling data and voice with one network, VoIP provides two advantages of management cost reduction and efficiency improvement, unlike an existing network in which data and voice are handled by distinct networks. Third, VoIP is easy to work with Internet-based multimedia services. Using the same network for voice and data makes it possible to provide a number of additional services, such as video conferencing, which are difficult to provide in conventional telephones that adopt circuit switching.

To provide the VoIP service, there is a need for means of discovering and signaling a correspondent to communicate. VoIP signaling includes H.323 of ITU-T and Session Initiation Protocol (SIP) of IETF.

A number of H.323-based VoIP services have been developed. SIP facilitates parsing and compiling tasks as well as provides excellent extensibility. Further, SIP is text-based and thus is easy to implement, unlike H.323.

VoIP end-point devices, e.g., gateways, IP phones, PCs, and the like perform voice communication by continuously transmitting and receiving RTP packets as packetized voice between a sender and a recipient through the RTP protocol on an IP network. However, the consistent RTP packet transmission and reception causes a traffic load on the IP network and impacts overall performance of VoIP equipment by processing RTP packets at a socket interface of a media processor.

The RTP allows an end-to-end transmission service in which real-time data such as audio and video is forwarded using a multicast or unicast network. The RTP has no concept of a connection. Typically, the RTP operates on an upper layer of a User Datagram Protocol (UDP) and utilizes multiplexing and checksum services of the UDP.

In addition to a typical wired VoIP system, Voice over Wireless LAN (VoWLAN) technology, in which voice is forwarded using a widespread WLAN, is recently emerging as a new mobile telephone technology. This is because the VoWLAN realizes lower fees and greater convenience by adding mobility to a wired Internet telephone, i.e., a VoIP telephone.

The VoWLAN forwards voice over a wireless LAN network. In other words, the VoWLAN uses the wireless LAN as a medium, unlike an existing Internet phone working on a wired network.

The VoWLAN provides convenient voice communication by guaranteeing mobility within an area of an Access Point (AP). Further, using an existing network considerably reduces telephony cost as compared with a telephone circuit provider. In particular, the VoWLAN enables a video telephone service desired by customers, and therefore, is advantageous for future customer services.

In VoIP, since voice data is all formed as an RTP packet and is continuously forwarded over a data network, a predetermined network bandwidth is required to perform smooth communication.

To efficiently use network bandwidth in the VoIP system, there is a method with silence that is a characteristic of voice conversation, and an RTP multi-framing method in which voice data is multiplexed into one RTP packet.

The method with silence utilizes a silence processing scheme, such as silence suppression or Voice Activity Detection (VAD)/Comfort Noise Generation (CNG). Typically, a VoIP Digital Signal Processor (DSP) has a VAD/CNG function. When silence is detected by this function, a normal voice payload is not generated but a smaller silence payload indicating the silence is generated. This is sent to the correspondent via the RTP and a local noise is sent to the correspondent during a silence period, such that network bandwidth is saved and smooth communication is maintained.

The multi-framing method is described below.

In VoIP communication, a voice payload periodically generated by a VoIP DSP is formed as an RTP packet and transmitted to a correspondent. To form the RTP packet, protocol header information for transmission, such as an Ethernet header, an IP header, a UDP header, an RTP header, and the like, as well as the voice payload, are added to every RTP packet. This increases the size of actually transmitted data and requires additional bandwidth.

The RTP multi-framing method multiplexes a number of voice payloads into one RTP packet within a predetermined limit and transmits the RTP packet, instead of forming and transmitting the RTP packet directly after a voice payload is generated. This reduces the quantity of additional protocol header information for transmission, such that an entire required network bandwidth is decreased.

The silence-using method and the multi-framing method consume considerable processing time in a terminal or cause a delay while two or three voice packets are being sequentially accumulated, thus deteriorating voice quality. In addition, if a multi-framed RTP packet is lost, a multi-framed amount of the voice packet can be lost all at once, further deteriorating voice quality.

The above problem associated with the VoWLAN terminal is described below in greater detail.

In the VoWLAN phone, battery duration is critical because the VoWLAN phone is a wireless communication terminal.

The silence method requires a consistent processing time because an RTP packet is formed and transmitted over a network by the VoIP technology directly after packetized media data is generated during communication. This causes battery power of the VoWLAN phone to be more rapidly consumed during communication.

In the multi-framing method, generated voice packets are sequentially accumulated by a few frames and then transmitted. This causes voice packet delivery to be delayed, as well as losing a large amount of voice data at once when the packet is lost, resulting in voice quality degradation.

<patcit id="pcit0001" dnum="US6621834B1"><text>US 6,621,834 B1</text></patcit> discloses a system and method for voice transmission over high level network protocols. On the Internet and the World Wide Web, such high level protocols are HTTP/TCP. The restrictions imposed by firewalls and proxy servers are avoided by using HTTP level connections to transmit voice data. In addition, packet delivery guarantees are obtained by using TCP instead of UDP. Variable compression based on silence detection takes advantage of the natural silences and pauses in human speech, thus reducing the delays in transmission caused by using HTTP/TCP. The silence detection includes the ability to bookend the voice data sent with small portions of silence to insure that the voice sounds natural. Finally, the voice data is transmitted to each client computer independently from circular list of voice data, thus insuring that all clients wall stay current with the most recent voice data. The combination of these features enables simple, seamless, and interactive Internet conferencing.

<patcit id="pcit0002" dnum="WO03052747A1"><text>WO 03/052747 A1</text></patcit> addresses the need for reducing audio overhang in wireless communication systems. It provides for the deletion of silent frames before they' are converted to audio by the listening devices. It only provides for the deletion of a portion of the silent frames that make up a period of silence or low voice activity in the speaker's audio. Voice frames that make up periods of silence less than a given length of time are not deleted.

SUMMARY OF THE INVENTION

The invention is set out in the appended set of claims. The embodiments and/or examples of the following description which are not covered by the appended claims are considered as not being part of the present invention.

It is an object of the present invention to provide an apparatus and method of handling RTP media packets in a VoIP terminal, capable of reducing network bandwidth required for VoIP communication while minimizing the impact of multi-framing on voice quality by enhancing RTP packet multi-framing.

According to one aspect of the present invention, a method of handling Voice over Internet Protocol (VoIP) voice packets communicated between terminals in real time is provided, in accordance with claim 1.

Checking whether a communication payload generated at an RTP transmitting side is a silence payload preferably comprises: determining that an entire communication payload is the voice payload upon the number of voice data having more than a silence reference value within a periodically generated communication payload exceeds a predetermined number; and determining that the entire communication payload is the silence payload upon the number of the voice data having more than the silence reference value does not exceed the predetermined number.

Accumulating the silence payload preferably comprises continuously accumulating the silence payload without transmission until the number of accumulated silence payloads reaches the multi-frame limit and then transmitting the silence payloads without further accumulation.

The multi-frame limit is preferably set such that a correspondent terminal receiving packets from a transmitting terminal does not cause a delay in replaying the voice.

The RTP packet preferably comprises the accumulated silence payloads and a succeeding voice payload.

The communication payload is preferably generated using at least one of a G.711 CODEC, a G.723.1 CODEC, and a G.729 CODEC.

Each terminal is preferably at least one of a VoIP wired terminal and a Voice over Wireless LAN (VoWLAN) terminal.

According to yet another aspect of the present invention, a terminal communicating using a Voice over Internet Protocol (VoIP) system is provided, in accordance with claim 8.

The terminal preferably further comprises a wired/wireless transmission processor adapted to receive an RTP packet from the media processor, add a User Datagram Protocol (UDP) header and an IP header to the RTP packet, and to transmit the resultant RTP packet to a correspondent terminal.

The multi-frame limit is preferably set such that a correspondent terminal receiving packets from a transmitting terminal does not cause a delay in replaying the voice.

The media processor is preferably adapted to check whether a communication payload generated at an RTP transmitting side is a silence payload, and to accumulate the silence payload instead of transmitting it upon the communication payload being the silence payload, and to form an RTP packet upon the communication payload being a voice payload or upon the number of accumulated silence payloads reaching a predetermined multi-frame limit.

The communication payload is preferably generated using at least one of a G.711 CODEC, a G.723.1 CODEC, and a G.729 CODEC.

The media processor is preferably adapted to determine that an entire communication payload is the voice payload upon the number of voice data having more than a silence reference value within a periodically generated communication payload exceeding a predetermined number and to determine that the entire communication payload is the silence payload upon the number of the voice data having more than the silence reference value not exceeding the predetermined number.

The media processor preferably comprises a VoIP Digital Signal Processor (DSP) adapted to receive voice data from a speaker and to periodically generate a communication payload.

The terminal is preferably at least one of a VoIP wired terminal and a Voice over Wireless LAN (VoWLAN) terminal.

BRIEF DESCRIPTION OF THE DRAWINGS

A more complete appreciation of the present invention, and many of the attendant advantages thereof, will be readily apparent as the present invention becomes better understood by reference to the following detailed description when considered in conjunction with the accompanying drawings, in which like reference symbols indicate the same or similar components, wherein: <ul id="ul0001" list-style="none"><li><figref idref="f0001">FIG. 1</figref> is a block diagram of a VoIP processing configuration in a VoIP end-point device according to an embodiment of the present invention;</li><li><figref idref="f0002">FIG. 2</figref> is a diagram of VoIP processing flow in a VoIP end-point device according to an embodiment of the present invention;</li><li><figref idref="f0003">FIG. 3</figref> is a view of a method of handling RTP transmission according to an embodiment of the present invention; and</li><li><figref idref="f0004">FIG. 4</figref> is a flowchart of RTP selective multi-framing according to an embodiment of the present invention.</li></ul>

DETAILED DESCRIPTION OF THE INVENTION

Hereinafter, exemplary embodiments of the present invention are described in greater detail with reference to the accompanying drawings. The present invention is described with respect to a VoWLAN terminal with which advantages of the present invention are maximized. However, the present invention is not limited to the VoWLAN terminal. It will be apparent to those skilled in the art that the present invention can be applied to systems communicating using a VoIP system, e.g., a typical wired VoIP system, and to all communication systems to which RTP multi-framing transmission and reception is applicable.

<figref idref="f0001">FIG. 1</figref> is a block diagram of a VoIP processing configuration in a VoIP end-point device according to an embodiment of the present invention.

A Voice over Wireless LAN (VoWLAN) terminal wirelessly communicates with an Access Point (AP) using IEEE 802.11-A, B and G standards. The VoWLAN terminal is able to gain access to a system via the AP and communicate with a correspondent terminal. A Session Initiation Protocol (SIP) is used as a signal processing protocol for VoIP call setup, and a Real-time Transport Protocol (RTP) is used for voice communication. As a voice CODEC, G.711, G.729 or G.712 is mainly used. The CODEC serves to convert an analog signal to a digital signal, and vise versa. One of two VoWLAN phones performing VoIP communication continuously transmits and receives RTP packets to and from the other to deliver voice packets. To perform VoIP processing, a VoIP end-point device of <figref idref="f0001">FIG. 1</figref> includes a wired/wireless transmission processor 130 that transmits and receives IP packets, a signaling protocol processor 110 for controlling a VoIP communication session, and a media processor 120 for processing voice data.

The wired/wireless transmission processor 130 uses a wired/wireless transmission scheme, such as an IEEE 802.3 wired LAN or an IEEE 802.11 wireless LAN, to transmit and receive IP packets. The signaling protocol processor 110 uses a VoIP signaling protocol, such as H.323 or SIP, to set up and terminate a VoIP call by controlling a communication session with a correspondent. After the signaling protocol processor 110 initiates the VoIP communication session, the media processor 120 continuously transmits and receives packetized voice data by a VoIP DSP 121 to and from the correspondent using the RTP protocol so that communication therebetween is achieved.

<figref idref="f0002">FIG. 2</figref> is a diagram of VoIP processing flow in a VoIP end-point device according to an embodiment of the present invention.

When a communication session is initiated with a communication correspondent by the signaling protocol processor 110, a VoIP end-point device continuously packetizes voice data at an encoder of the VoIP DSP121 and replays voice at a decoder, as well as transmits and receives the packetized voice data using the RTP protocol until the session is terminated.

Depending on the VoIP voice CODEC used, the voice data, which is generated by the VoIP DSP 121, is continuously produced in a shorter period of 10 msec (in G.711) or in a longer period of 30 msec (in G.723.1) (typically, using interruption). The voice data is formed as an RTP packet by the RTP processor directly after it is generated, and transmitted to the correspondent in real time via a socket interface 201 over a communication network.

RTP transmission and reception of the voice data continuously produced in real time is typically performed by application software in the RTP processor, on which overall performance of the VoIP end-point device depends.

<figref idref="f0003">FIG. 3</figref> is a view of a method of handling RTP transmission according to an embodiment of the present invention.

After a call is set up through SIP signaling at the signaling protocol processor 110 as described with reference to <figref idref="f0001">FIG. 1</figref>, respective VoWLAN phones transmit and receive RTP packets therebetween and replay voice data present in the RTP packets, such that voice communication is achieved.

<figref idref="f0003">FIG. 3</figref> shows accumulation of the silence payload and processing of the voice payload of the RTP packet with time in processing the RTP packet using the RTP selective multi-framing method according to an embodiment of the present invention. The embodiment of <figref idref="f0003">FIG. 3</figref> will be described in conjunction with a VoIP environment using a G.71 CODEC.

It will be appreciated that, in the RTP selective multi-framing according to an embodiment of the present invention, only silence payloads are multi-framed within a multi-frame limit when a silence period is long, and the accumulated silence payloads and a succeeding voice payload are multi-framed together as one RTP packet when the succeeding voice payload is generated.

The operation in respective periods is described in more detail as follows.

It can be seen that in a period 31, only silence payloads are generated.

In other words, when a G.711 CODEC is used, voice data is generated at 10 msec intervals by the VoIP DSP of the media processor 120. In conventional art, an RTP packet is formed and transmitted to a correspondent over a network directly after voice data is generated. On the other hand, based on the fact that "actual conversation does not fill successive time periods during voice communication, and silence periods in which there is no conversation occupy about 40 % of the overall communication", the present invention checks a level of periodically generated voice data, and if the level is lower than a predetermined level, i.e., if the voice data is not substantial conversation data but silence data that is generated during conversation, stores the voice data in an internal buffer instead of immediately transmitting it to the correspondent.

An actual experiment showed that 36.3 % of entire voice data corresponds to the silence data, wherein several voice samples are inspected and determined to be silence data if the voice samples have voice level values less than a reference value for which the silence data occupies 30-40 % of the entire voice data.

As can be seen in a period 32, the buffering proceeds until voice data having more than a predetermined level, that is, substantial voice data, is generated, or until the number of frames reaches a maximum buffering limit (a multi-framing limit). If the number of the frames reaches the multi-frame limit, an RTP packet corresponding to the silence data is formed and forwarded.

Normally, a multi-frame range that does not impact communication voice quality is 60 to 80 ms. The present invention obtained satisfactory communication voice quality by setting the range value as 40 ms.

A voice payload appears in period 33. The generated voice data is immediately forwarded together with the previous buffered silence data.

The voice payloads continue to be generated in a period 34 and are forwarded immediately.

In a period 35, the silence payload appears again and is buffered.

<figref idref="f0004">FIG. 4</figref> is a flowchart of RTP selective multi-framing according to an embodiment of the present invention.

In addition to existing RTP processing, the RTP selective multi-framing of <figref idref="f0004">FIG. 4</figref> further includes (i) checking whether communication payloads generated at an RTP transmitting side are silence payloads, (ii) accumulating the silence payloads within such a limit that voice quality deterioration due to voice delay does not arise, and (iii) forming an RTP packet if the voice payload rather than the silence payload is generated, or if the number of accumulated silence payloads reaches the multi-frame limit, and transmitting the RTP packet to a correspondent.

When a communication payload is periodically generated by the VoIP DSP 121 (S401), a voice level of the generated communication payload is checked (S402). In this step, a determination is made as to whether the communication payload is regarded as voice data or silence data.

Detecting and processing silence in the present invention is described below.

If the number of voice data having more than the silence reference value within the voice payload periodically generated by the VoIP DSP 121 exceeds a predetermined number, an entire payload is determined to be a voice payload. If the number of the data having more than the silence reference value is less than the predetermined number, it is handled as the silence payload.

That is, if the level of the voice constituting the communication payload is lower than or equal to the silence data level, the communication payload is determined to be the silence payload. Otherwise, the communication payload is determined to be the voice payload.

If the communication payload is determined to be the silence data, it is necessary to check whether the number of the accumulated communication payloads reaches the multi-frame limit (S403). If the number of the accumulated communication payloads does not yet reach the multi-frame limit, the communication payload continues to be accumulated (or, buffered) (S404). If the number of the accumulated communication payloads reaches the multi-frame limit, a current communication payload is combined with the previous buffered communication payloads to form the RTP packet (S405), and the RTP packet is transmitted to a wireless LAN phone of a correspondent (S406).

If the multi-frame limit is set as 40 ms as in the above-mentioned experiment according to the present invention and a G.711 CODEC is used, a maximum of four frames are allowed to be accumulated since one communication payload occupies 10 ms. This limit can be set to be greater or smaller by a system manager in view of a wireless environment, performance of devices used in the system, etc.

The multi-frame limit can be modified depending on whether priority is placed on wireless resources/duration of a wireless phone battery or communication quality, the two having a trade-off relationship.

On the other hand, if the communication payload is determined to be the voice data, an RTP packet is formed using only a current communication payload when there is no previous accumulated silence communication payload and transmitted, irrespective of the multi-frame limit. If there is a previous accumulated silence communication payload, the current communication payload is combined with the previous communication payloads to form the RTP packet (S405), and the RTP packet is transmitted to the correspondent (S406).

In this case, the RTP packet includes voice data generated at a transmitting side of the VoIP equipment, and an RTP header attached to the voice data. The RTP packet is further added to a UDP header and an IP header at the wired/wireless transmission processor 130 described in <figref idref="f0001">FIG. 1</figref> and transmitted to the correspondent.

In the embodiment of <figref idref="f0004">FIG. 4</figref>, the method by which the RTP processor is capable of easily determining from the voice data itself whether the silence data is contained in the voice data, has been mainly described as including RTP selective multi-framing through silence detection with a G.711 CODEC performed on the generated voice data. This processing with a G.711 CODEC similarly works on a compressed CODEC such as G.723.1 or G.729, only if information on the voice level is fed back to the RTP processor together with the voice data from the VoIP DSP.

With the RTP selective multi-framing described above, it is possible to reduce the number of times an RTP packet is transmitted by a rate of silence generation during communication, compared to transmitting the RTP packet directly after voice data is generated irrespective of speech or silence. This reduces processing power consumption in the VoWLAN phone and increases network transmission efficiency since the packets are buffered and then transmitted all at once. As a result, it is possible to reduce battery consumption in the VoWLAN phone and therefore increase battery lifetime.

To test the method according to the present invention, an environment was actually prepared in which a connection is made to an IEEE 802.11b wireless LAN network via an IEEE 802.3 wired LAN switching hub by using one IEEE 802.11b wireless LAN AP for VoIP communication via a wireless LAN and two wireless LAN phones, and by installing software for packet loss simulation in a network into a Linux PC.

An experiment was performed on the RTP selective multi-framing method according to the present invention by using the two wireless LAN phones having an RFC 3261 SIP protocol processor for VoIP signaling. Two network cards were mounted in the Linux PC, and IP addresses were set so that two VoIP WLAN phones belong to a different network. The WLAN phones used for the test had a Central Processing Unit (CPU) with a VoWLAN chipset, and a 1000 mA L-Ion battery. IEEE 802.11 was used as a wireless LAN standard, and SIP and RTP were used as VoIP protocols.

To test a VoIP communication state, a test voice containing normal conversation content recorded for one minute was repeatedly replayed at a PC to be input to a microphone of the wireless LAN phone, such that several tests were conducted in the same VoIP communication environment.

The experimental results confirmed that the RTP selective multi-framing method according to the present invention reduces the number of times an RTP packet is forwarded by a maximum of 28.1 % when a multi-frame limit is set as 4, and saves a corresponding network bandwidth by a maximum of 14.4 %, compared to a case where multi-framing is not used.

The network bandwidth savings of 14.4% may not be great for one wireless LAN phone, but is considerable for an AP communicating with a number of wireless LAN phones, because the bandwidth is associated with quantity of data arid the saved bandwidth allows an increase in capacity of a wireless LAN AP. This effect appears in a VoIP media gateway having a similar function, as well as in the wireless LAN AP.

Furthermore, the experiment confirmed that the method according to the present invention increases battery duration during communication to a maximum of 270 minutes, i.e., by 42.1%, compared to 190 minutes when multi-framing is not used.

With the present invention, it is possible to reduce a network bandwidth required for VoIP communication while minimizing the impact of multi-framing on voice quality, by applying a multi-framing scheme to silence data generated during voice communication while applying an immediate transmitting scheme to generated voice data. In particular, it is possible to reduce battery consumption in a VoWLAN terminal by reducing processing overhead.

While the present invention has been described with reference to exemplary embodiments thereof, it will be understood by those skilled in the art that various modifications in form and detail can be made therein without departing from the scope of the present invention as defined by the following claims.