Stream switching based on gradual decoder refresh

Abstract

A signaling method and device for use in stream switching in which GDR random access points are used. In order to indicate the GDR switching points in the bitstreams, a Sync Sample Information Box, which is contained in a Sync Sample Box, is used to provide information of such GDR switching points. The information also includes which slice group is the isolated region and which slice group is the leftover region. The signaling method can be used in video data transmission using Real-time Transport Protocol (RTP), and a Session Description Protocol (SDP) can be used to convey information indicative of the characteristics of the bitstreams.

Description

FIELD OF THE INVENTION

[0001] The present invention relates generally to video streaming and, more particularly, to stream adaptation in accordance with changing transmission conditions.

BACKGROUND OF THE INVENTION

[0002] In video streaming or video-on-demand services, because of the dynamic network conditions, the end-to-end transmission characteristics between the server and the client may change frequently. For example, the transmission bitrate may be reduced. To maintain the continuity of the streaming session and to maximize the Quality of Service, the server should adapt the transmitted stream to the changing transmission conditions. This process is called stream adaptation.

[0003] Stream adaptation is either multi-encoding based or transcoding based. In multi-encoding based stream adaptation, the server stores the same video content in a plurality of encoded streams of different forms or with different parameters, and the transmitted data in the encoded streams may be switched between different streams. In transcoding based stream adaptation, the server contains a transcoder to transcode a stream to different forms or with different parameters.

[0004] To enable switching from one bitstream to another, the switched-to bitstream must contain switching points, such that the client-side decoder can still receive image data of acceptable decoding quality after switching. Switching points can be random access points or non-random access points. SP/SI pictures can be used for stream switching at non-random access points. Random access points, however, are natural switching points.

[0005] Random access refers to the ability of the decoder to start decoding a stream at a point in the stream other than the beginning of the stream, and to recover an exact or approximate representation of the decoded pictures. Thus, a random access point is a switching point where decoding of any following coded picture can be initiated.

[0006] A random access point and a recovery point characterize a random access operation. All decoded pictures located at or subsequent to a recovery point in the output order are correct or approximately correct in content. If the random access point is the same as the recovery point, the random access operation is Instantaneous Decoder Refresh (IDR), otherwise it is Gradual Decoder Refresh (GDR). IDR points in a video stream can be used in fast forward and random access, but they can also be used for error resiliency and recovery. IDR is also used in bitrate adaptation by stream switching, especially on the server side.

[0007] IDR pictures are pictures that are coded without any reference to other pictures, and all the pictures following and IDR picture in decoding order are coded without reference to any earlier picture than the IDR picture in decoding order, whereas GDR can be implemented using the technique called isolated regions as described later in this document. The picture at a GDR random access point is called a GDR picture.

[0008] Random access points render it possible to seek operations in locally stored video streams. In video on-demand streaming, servers can respond to seek requests by transmitting data starting from the random access point that is closest to the requested destination of the seek operation. Switching between coded streams of different bit-rates is a method that is used commonly in unicast streaming for the Internet to match the transmitted bitrate to the expected network throughput and to avoid congestion in the network. Switching to another stream is possible at a random access point. Furthermore, random access points enable tuning in to a broadcast or multicast. In addition, a random access point can be coded as a response to a scene cut in the source sequence or as a response to an intra picture update request.

[0009] File Format

[0010] MPEG-4 Part 12 specifies ISO (International Organization for Standardization) base media file format. It is designed to contain timed media information for a presentation in a flexible, extensible format that facilitates interchange, management, editing, and presentation of the media. This presentation may be ‘local’ to the system containing the presentation, or may be carried out via a network or other stream delivery mechanism. The file structure is object-oriented in that a file can be decomposed into constituent objects, and the structure of the objects can be inferred directly from their type. The file format is designed to be independent of any particular network protocol while enabling efficient support for them in general. ISO base media file format is used as the basis for MP4 file format (MPEG-4 Part 14) and AVC (Advanced Video Coding) file format (MPEG-4 Part 15). AVC file format specifies how AVC content is stored in an ISO base media file format. It is normally used in the context of a specification, such as the MP4 file format, derived from ISO base media file format that permits the use of AVC video.

[0011] In the current design of AVC file format, SP/SI pictures are stored in switching picture tracks, which are tracks separate from the track that is being switched from and the track being switched to. Switching picture tracks can be identified by the existence of a specific required track reference in that track. A switching picture is an alternative to the sample in the destination track that has exactly the same decoding time.

[0012] Each IDR random access point corresponds to a sync sample indicated in the Sync Sample Box. The design of Sync Sample Box is specified in the ISO base media file format as follows: 1 Definition Box Type: ‘stss’ Container: Sample Table Box (‘stbl’) Mandatory: No Quantity: Zero or one

[0013] This box provides a compact marking of the random access points within the stream. The table is arranged in strictly increasing order of sample number. If the sync sample box is not present, every sample is a random access point. 2 Syntax aligned(8) class SyncSampleBox  extends FullBox(‘stss’, version = 0, 0) {  unsigned int(32) entry_count;  int i;  for (i=0; i < entry_count; i++) {   unsigned int(32) sample_number;  } }

[0014] Semantics

[0015] version is an integer that specifies the version of the box.

[0016] entry_count is an integer that gives the number of entries in the following table. If entry_count is zero, there are no random access points within the stream and the following table is empty.

[0017] sample_number gives the numbers of the samples that are random access points in the stream.

[0018] Isolated Regions

[0019] The isolated regions technique provides an elegant solution for many applications, such as GDR (gradual decoder refresh) (JVT-C074), error resiliency and recovery (JVT-C073), region-of-interest coding and prioritization, picture-in-picture functionality, and coding of masked video scene transitions (JVT-C075). With GDR being based on isolated regions, media channel switching for receivers, bitstream switching for the server, and allowing newcomers for multicast streaming will be as easy as instantaneous random access with smoother bitrate.

[0020] An isolated region in a picture can contain any macroblock and a picture can contain zero or one isolated region, or more isolated regions that do not overlap. A leftover region is the area of the picture that is not covered by any isolated region of a picture. When coding an isolated region, all predictive coding within the same coded or decoded picture, herein referred to as in-picture prediction, is disabled across its boundaries. A leftover region may be predicted from isolated regions of the same picture.

[0021] A coded isolated region can be decoded without the presence of any other isolated or leftover region of the same coded picture. It may be necessary to decode all isolated regions of a picture before the leftover region. An isolated region or a leftover region contains at least one slice.

[0022] Pictures, whose isolated regions are predicted from each other, are grouped into an isolated-region picture group. An isolated region can be coupled with a corresponding isolated region in each earlier picture within the same isolated-region picture group. An isolated region can be inter-predicted from the corresponding isolated region within the same isolated-region picture group. However, inter prediction of an isolated region from other isolated regions is disallowed. In contrast, a leftover region may be inter-predicted from any isolated region. The shape, location, and size of coupled isolated regions may evolve from picture to picture in an isolated-region picture group.

[0023] Coding of isolated regions can be realized in the JVT codec based on slice groups. Each GDR random access point is characterized by a recovery point Supplemental Enhancement Information (SEI) message.

[0024] SP/SI Pictures

[0025] The JVT coding standard supports SP/SI pictures. It is known that in stream switching involving only P-slices, the decoder will not have the correct decoded reference frames required in image reconstruction. By inserting an I-slice at regular intervals in the coded sequence to create switching points can solve this problem. However, an I-splice is likely to contain much more coded data than a P-slice. As such, a peak in the coded birate is resulted at each switching point. SP-slices and SI-slices are designed to support switching without the increased bitrate penalty of I-slices.

[0026] An SP/SI picture is encoded in such a way that another SP/SI picture using different reference pictures can have exactly the same reconstructed picture. SP/SI pictures can be applied for bitstream switching, splicing, random access, fast forward, fast backward and error resilience/recovery. For example, let us assume that there are two bitstreams, bs1 and bs2, of different bitrates, originated from the same video sequence. In bs1, an SP picture (s1) is coded, and another SP picture (s2) is coded at the same location in bs2. In bs1, an additional SP picture (s12) is coded having exactly the same reconstructed picture as s2. s12 and s2 use different reference pictures (from bs1 and bs2, respectively). Thus, switching from bs1 to bs2 can be carried out by transmitting s12 instead of s1 in the switching location. Since s12 has exactly the same reconstruction as s2, reconstructed pictures after switching are error-free.

[0027] Streaming System

[0028] As mentioned earlier, in multi-encoding based stream adaptation, the server stores in a plurality of encoded streams the same video content, but only one of the encoded streams is selected for transmission. FIG. 1 depicts a transmitting system 10, which includes a server 20 capable of receiving a plurality of streams from a transcoder or multi-stream generator or storage device 12. As shown, the streaming server 20 comprises a stream selector 22 to select one of the encoded streams 1 to n. The selected encoded stream is divided into packets by a packetizer 24 and coded in a channel coder 26 for transmission. To maintain continuity of the streaming session and to maximize the Quality of Service, the server generally selects the best possible encoded stream for transmission. When the transmission condition changes, the server may have to increase or reduce the bitrate, for example. Accordingly, the stream selector switches streams by selecting a different encoded stream at a switching point. At the client side, however, the decoder can simply decode whatever transmission data it receives. Basically, a streaming client device 40 comprises a channel decoder 42, a de-packetizer 44 and a decoder 46 for providing decoded video signals to a display 48 for display, as shown in FIG. 2. However, in a streaming system that supports client-driven stream adaptation, the streaming client device can send a request signal to the server to request switching of the stream. The streaming system is shown in FIG. 3, which shows the connection between a streaming server 20 and a streaming client 40 through a network 60.

[0029] Instantaneous/Gradual Decoder Refresh

[0030] As mentioned earlier, a random access point is any picture from which decoding can be initiated. At such an access point, all decoded pictures at, or subsequent to, a recovery point are correct or approximately correct in content. It should be noted that the phrase “correct in content” as used in this disclosure means that the decoded slice or picture is exactly the same as when the decoding is started from the beginning of the stream, and the phrase “approximately correct in content” means that the decoded slice or picture is approximately the same as when the decoding is started from the beginning of the bitstream. As shown in FIG. 4a, the recovery point is the same as the switching point, and the pictures with correct or approximately correct in content start at the switching point. As such, the random access operation is referred to as Instantaneous Decoder Refresh (IDR). With IDR random access points, only an I slice or an SI slice can be used for stream switching.

[0031] In contrast, a Gradual Decoder Refresh (GDR) random access point can contain any kind of slices (I, P, SI, SP). As shown in FIG. 4b, however, the content in the picture is correct or approximately correct starting from a picture following the switching point in the output order. The pictures between the recovery point and the switching point may be visually annoying or otherwise unacceptable for viewing.

[0032] Currently, an efficient method to signal GDR switching points to be used in file format is lacking. An example of the file format is AVC file format, which is important for a server file containing streaming content with GDR based video coding to support stream switching. For AVC contents stored in the AVC file format, a GDR switching point can only be identified when an access unit contains a recovery point SEI message, and the syntax element changing_slice_group_idc is equal to 1 or 2, as specified in the JVT coding standard. This method requires that each AVC access unit is checked to see whether there is a recovery point SEI message and whether changing_slice_group_idc is equal to 1 or2.

SUMMARY OF THE INVENTION

[0033] The present invention provides an efficient signal method and device for GDR switching points in file format. Furthermore, information on how the GDR is encoded using isolated regions is also signaled so as to achieve faster stream switching. With the signaling method of present invention, GDR switching points can be identified as easily as other switching points, such as IDR and SP/SI switching points. In addition, the server can select to transmit only the isolated region for the access units from the GDR switching point to the recover point, inclusive, to achieve faster GDR switching and reduced bitrate.

[0034] Thus, according to the first aspect of the present invention, there is provided a signaling method for use in stream switching among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching. The method is characterized by

[0035] providing in the bitstreams information indicative of the switching point so that said stream switching can be carried out based on the provided information, wherein

[0036] the recovery point is different from the switching point.

[0037] Furthermore, the video frames contain at least one isolated region associated with said one or more slices in the second bitstream decoded subsequent to said stream switching, and the provided information is further indicative of the isolated region.

[0038] The stream switching can be initiated by a server device or requested by a client device in a streaming network based on transmission conditions between the server device and the client device.

[0039] The signaling method is used in a transmission utilizing Real-time Transport Protocol (RTP), and wherein a Session Description Protocol (SDP) is used to convey information indicative of characteristics of the first and second bitstreams.

[0040] According to the second aspect of the present invention, there is provided a streaming server device capable of switching streams among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching. The streaming server device is characterized by

[0041] a stream selector for selecting the first bitstream for transmission; and

[0042] means for providing in the bitstreams information indicative of the switching point, so as to allow the stream selector to select the second bitstream for transmission based on the provided information, wherein the recovery point is different from the switching point.

[0043] According to the third aspect of the present invention, there is provided a streaming system capable of switching stream among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching. The streaming system is characterized by

[0044] at least one streaming client; and

[0045] at least one streaming server for transmitting one of the bitstreams to the streaming client so as to allow the streaming client to reconstruct the video frames based on the transmitted bitstream, wherein the streaming server comprises:

[0046] a stream selector for selecting the first bitstream for transmission and for further selecting the second bitstream, and

[0047] means for providing in the bitstreams information indicative of the switching point so as to allow the stream selector to select the second bitstream based on the provided information, wherein the recovery point is different from the switching point.

[0048] The streaming system is further characterized by

[0049] a video encoder to convert a video input signal into the video data; and

[0050] means, responsive to the video data, for encoding the video data into the plurality of bitstreams.

[0051] According to the fourth aspect of the present invention, there is provided a software program for use in a streaming system for stream switching among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching. The computer program is characterized by

[0052] a code for determining said switching point; and

[0053] a code for indicating said switching point in information provided in the bitstreams, so as to allow a streaming server to carrying out the stream switching based on the provided information, wherein the recovery point is different from the switching point.

[0054] The present invention will become apparent upon reading the description taken in conjunction with FIGS. 5 to 7.

BRIEF DESCRIPTION OF THE DRAWINGS

[0055] FIG. 1 is a block diagram illustrating a streaming server that supports stream switching.

[0056] FIG. 2 is a block diagram illustrating a streaming client.

[0057] FIG. 3 is a schematic representation of a streaming system.

[0058] FIG. 4a is a schematic representation illustrating stream switching using an instantaneous decoder refresh picture.

[0059] FIG. 4b is a schematic representation illustrating stream switching using a gradual decoder refresh picture.

[0060] FIG. 5 is a block diagram illustrating a sync sample box.

[0061] FIG. 6 is a block diagram illustrating a sync sample information box, according to the present invention.

[0062] FIG. 7 is a schematic representation illustrating a streaming system, according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION

[0063] According to the present invention, information on the switchable GDR pictures is included in a sync sample information box (ssif) that is contained in the sync sample box so as to indicate the access points. Furthermore, the slice groups need to be associated to the isolated region and to the leftover region in the ssif. Using this information, the decoder can use the GDR picture to correctly switch streams. Using GDR pictures in switching, the information of pictures in the switching points can be transmitted faster than that for IDR pictures, because the leftover region in a GDR picture does not need to be sent. Though using GDR pictures for switching the users may see only part of the picture area at beginning, they could be happier if they can see something as soon as possible. In addition, the leftover region in a picture from the GDR switching point to the recovery point, inclusively, does not need to be sent. As such, reduced transmission rate is achieved.

[0064] The implementation of the present invention in AVC file format is characterized in that each random access point is a switching point. It should be noted that all random access points, including both IDR random access points (IDR access units) and GDR random access points (access units containing recovery point SEI messages with the syntax element changing_slice_group_idc equal to 1 or 2), are marked in Sync Sample Box. In addition, a Sync Sample Information Box (contained in sync sample box) is defined as follows: 3 Definition Box Type: ‘ssif’ Container: Sync Sample Box (‘stss’) Mandatory: No Quantity: Zero or one

[0065] This box provides information of the random access points within the stream. The information includes whether a random access point is a GDR or an IDR random access point. If the random access point is a GDR point, the information also includes which slice group is the isolated region and which slice group is the leftover region. If the sync sample box does not contain a sync sample information box, all the sync samples marked by the sync sample box are IDR random access points. 4 Syntax aligned(8) class SyncSampleInformationBox  extends FullBox(‘ssif’, version = 0, 0) {  int i;  for (i=0; i < entry_count; i++) {   unsigned int(2) random_access_point_idc;   bit (6) reserved = ‘111111’b;  } }

[0066] Semantics

[0067] version is an integer that specifies the version of this box.

[0068] random_access_point_idc :

[0069] 0 indicates that the random access point is not a IDR random access point;

[0070] 1 indicates that the isolated region is covered by slice group 0 while the leftover region is covered by slice group 1;

[0071] 2 indicates that the isolated region is covered by slice group 1 while the leftover region is covered by slice group 0;

[0072] 3 is not allowed.

[0073] With the signaling method, according to the present invention, all switching points can be explicitly marked so that the stream server does not need to parse each picture to find the switching points. If there are no GDR switching points, the Sync Sample Information Box (contained in the Sync Sample Box) does not need to be used.

[0074] An exemplary Sync Sample Box is shown in FIG. 5 and an exemplary Sync Sample Information Box is shown in FIG. 6.

[0075] According to the present invention, a computer program is used in the streaming system to provide information on the switchable GDR pictures in a Sync Sample Information Box that is contained in a Sync Sample Box. The information includes the switching points. In addition, the computer program also specifies the slice groups that are associated to the isolation region and to the leftover region. Such a computer program is denoted by reference numeral 16, as shown in FIG. 7. The computer program 16 is part of a video coder 14, which provides encoded video input signal and GDR related information to the multi-stream transcoder/generator 12. The stream server 20 is capable of selecting one of the encoded streams for transmission, based on the dynamic network conditions in the network 60. If the end-to-end transmission characteristics between the streaming server 20 and the streaming client 40 have changed, the streaming server 20 may initiate stream switching in that the streaming server chooses another encoded stream, according with the GDR related information provided in the Sync Sample Information Box. Alternatively, the streaming client 40 may send a request signal to the streaming server 20, requesting a different transmitted stream if the streaming client 40 detects a change in the transmission conditions in the network 60.

[0076] The GDR signaling method, according to the present invention, can be used in video data transmission using Real-time Transport Protocol (RTP), and a Session Description Protocol (SDP) can be used to convey information indicative of the characteristics of bitstreams in stream switching. As it is known, RTP provides end-to-end network transport functions suitable for applications transmitting real-time data, such as audio, video or simulation data, over multicast or unicast network services. RTP does not address resource reservation and does not guarantee quality-of-service (QoS) for real-time services. The data transport is augmented by a control protocol (RTCP) to allow monitoring of the data delivery in a manner scalable to large multicast networks, and to provide minimal control and identification functionality. RTP and RTCP are designed to be independent of the underlying transport and network layers. The protocol supports the use of RTP-level translators and mixers. The Session Description Protocol is intended for describing multimedia sessions for the purposes of session announcement, session invitation, and other forms of multimedia session initiation. SDP can be used, for example, by the server to notify the client what bitrate alternatives of a bistream is available.

[0077] The GDR signaling method, according to the present invention, is applicable to the video coding standard ITU-T H.264 (also known as MPEG-4 Part 10 or AVC)developed by Joint Video Team (JVT). However, the application of the present invention is not limited to the above-mentioned JVT coding standard. The present invention may also be applied to other video coding standards and devices.

[0078] Thus, although the invention has been described with respect to a preferred embodiment thereof, it will be understood by those skilled in the art that the foregoing and various other changes, omissions and deviations in the form and detail thereof may be made without departing from the scope of this invention.

Claims

1. A signaling method for use in stream switching among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching, said method characterized by

providing in the bitstreams information indicative of the switching point so that said stream switching can be carried out based on the provided information, wherein

the recovery point is different from the switching point.

2. The signaling method of claim 1, wherein each video frame comprises one or more slices and the video frames contain at least one isolated region associated with said one or more slices in the second bitstream decoded subsequent to said stream switching, said method characterized in that the provided information is further indicative of the isolated region.

3. The signaling method of claim 1, wherein the bitstreams are conveyed from a server device to a client device in a streaming network, said method characterized in that said stream switching is initiated by the server device.

4. The signaling method of claim 1, wherein the bitstreams are conveyed from a server device to a client device in a streaming network, said method characterized in that said stream switching is requested by the client device.

5. The signaling method of claim 1, wherein the signaling method is used in a transmission utilizing Real-time Transport Protocol (RTP).

6. The signaling method of claim 5, wherein a Session Description Protocol (SDP) is used to convey information indicative of characteristics of the first and second bitstreams.

7. The signaling method of claim 1, wherein said stream switching is carried out in transmission of the video data based on transmission conditions between a server device and a client device in a streaming network.

8. A streaming server device capable of switching streams among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching, said streaming server device characterized by

a stream selector for selecting the first bitstream for transmission; and

means for providing in the bitstreams information indicative of the switching point, so as to allow the stream selector to select the second bitstream for transmission based on the provided information, wherein the recovery point is different from the switching point.

9. The streaming server device of claim 8, wherein each video frame comprises one or more slices and the video frames contain at least one isolated region associated with said one or more slices in the second bitstream decoded subsequent to said stream switching, and wherein the provided information is further indicative of the isolated region.

10. The streaming server device of claim 8, wherein the provided information is used in data transmission utilizing Real-time Transport Protocol (RTP).

11. The streaming server device of claim 10, wherein a Session Description Protocol (SDP) is used to convey information indicative of characteristics of the first and second bitstreams.

12. The streaming server device of claim 8, wherein said stream selector selects the second bitstream for stream switching based on transmission conditions between the streaming server device and a client device in a streaming network.

13. A streaming system capable of switching stream among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching, said streaming system characterized by

at least one streaming client; and

at least one streaming server for transmitting one of the bitstreams to the streaming client so as to allow the streaming client to reconstruct the video frames based on the transmitted bitstream, wherein the streaming server comprises:

a stream selector for selecting the first bitstream for transmission and for further selecting the second bitstream, and

means for providing in the bitstreams information indicative of the switching point so as to allow the stream selector to select the second bitstream based on the provided information, wherein the recovery point is different from the switching point.

14. The streaming system of claim 13, wherein each video frame comprises one or more slices and the video frames contain at least one isolated region associated with said one or more slices in the second bitstream decoded subsequent to said stream switching, and wherein the provided information is further indicative of the isolated region.

15. The streaming system of claim 13, wherein said stream switching is initiated by the streaming server.

16. The streaming system of claim 13, wherein said stream switching is requested by the streaming client.

17. The streaming system of claim 13, wherein the provided information is used in data transmission utilizing Real-time Transport Protocol (RTP).

18. The streaming system of claim 17, wherein a Session Description Protocol (SDP) is used to convey information indicative of characteristics of the first and second bitstreams.

19. The streaming system of claim 13, wherein said stream selects the second bitstream for stream switching based on transmission conditions between the streaming server and the streaming client.

20. The streaming system of claim 13, further characterized by

a video encoder to convert a video input signal into the video data; and

means, responsive to the video data, for encoding the video data into the plurality of bitstreams.

21. A software program for use in a streaming system for stream switching among a plurality of bitstreams, the bitstreams containing video data indicative of a plurality of video frames for each bitstream, wherein the bitstreams comprise at least one switching point so as to allow switching from a first bitstream to a second bitstream at said switching point, and at least one recovery point which defines a first correct or approximately correct picture in output order in the second bitstream decoded subsequent to said stream switching, said computer program characterized by

a code for determining said switching point; and

a code for indicating said switching point in information provided in the bitstreams, so as to allow a streaming server to carrying out the stream switching based on the provided information, wherein the recovery point is different from the switching point.

22. The software program of claim 21, wherein each video frame comprises one or more slices and the video frames contain at least one isolated region associated with said one or more slices in the second bitstream decoded subsequent to said stream switching, and wherein the provided information is further indicative of the isolated region.

23. The software program of claim 21, wherein the provided information is used in data transmission utilizing Real-time Transport Protocol (RTP).

24. The software program of claim 23, wherein a Session Description Protocol (SDP) is used to convey information indicative of characteristics of the first and second bitstreams.