In order to improve the quality of heterogeneous network video transmission, a linear network video combined source channel encoding method based on lattice quantization is proposed. First, the multi-description independent parallel channel transmission framework for the Gaussian video source transmission is established by multi-describing the combined source channel transmission scheme. Secondly, the lattice scale quantization method is considered to be considered or an boundless information decoder. Realize the decrease in the analog mapping bandwidth, improves heterogeneous network video transmission performance by bandwidth expansion; The comparative test on the frame delay and effective loss rate indicators verifies the effectiveness of the method.
introduction
In recent years, video transmission based on mobile technology (such as live broadcast, sports live, etc.) has become a hot streaming media application, and video transmission traffic has also increased rapid growth [1]. The 2012 video traffic accounts for about 58% of total flow, and will reach 70% in 2017. From 2012 to 2017, the total traffic flow will increase by 14 times [2]. How to ensure that the reliability of ultra-clear video data transmission is the main problem facing the service provider under the existing network.
Although the current network infrastructure construction has provided users with different Internet access methods, but the data transmission capacity of a single network still has limited restrictions, and unable to provide satisfactory mobile video transmission performance [3]. The main problem with the WLAN network is limited by bandwidth, small coverage, and cannot meet a large number of mobile users' mobile video service requirements. WiMAX networks provide a broader coverage and higher peak rate, but the user has a large share amount, multimedia real-time high throughput demand cannot meet. Single network performance is limited to make the heterogeneous network bandwidth integration problem is highly valued by scholars. Classic Joint Source Channel Encoding [4] (Joint Source-Channel Coding, JSCC) The main research content is the optimization of channel coding and source of sources, and video data error correction coding and channel status: (1) channel coding and trusted sources Code ratio optimization, such as document [5, 6]; (2) How to adjust the coding rate in the case of setting the channel parameters and state to achieve the required transmission target, such as the literature [7]; (3) channel encoding Reliability, such as fountain, turbo, etc .; (4) combined encoding optimization algorithm design, realize system performance improvement, such as literature [8]; literature [9] studied the optimization problem of coding rate in bandwidth restrictions, Realizing video ends to end distortion minimization.
However, the above algorithm utilizes an error control method, but does not consider the fault problem in which the channel itself is, it is easy to tend to transmit data video transmission quality, affect the quality of the user experience and video service. In this regard, this paper mainly targets heterogeneous network video transmission issues, using lattice quantitative conversion in the decoder, multi-describing the combined source channel encoding method, with the aim of obtaining a low distortion and low delay video transmission scheme.
1 problem model description
The parallel channel transmission scheme of the signal source integrated network with multi-describing the combined source channel coding (MD-JSCC) is shown in Figure 1.
The ratio is pronounced in a five-tuple group (R1, R2, D0, D1, D2). Among them, the rate distortion region R (D0, D1, D2) of the MD-SC problem is the closing set of rate pairs (R1, R2). There is or no side information (SI) in this area is unknown. The exception is the secondary Gaussian MD-SC case, that is, the Gaussian source and secondary distortion function:
2 simulated MD-JSCC solution based on lattimetricizer
2.1 Multi-descriptive analog mapping
2.2 Lattizer quantizer
The L-weeque lattice is a discrete population of European space, which can be described as:
2.3 Coding - Decoding Process
Where θ is a rotation angle. The maximum correlation of the output can be obtained by the rotation angle θ = π / 2, see [9]. The joined analog source channel MD encoding scheme is shown in Figure 2.
In Figure 2, (S1, S2) can be encoded by the following steps:
(2) Situations 2: (Si-available encoder) will follow the case 1 encoding, in addition to the initial pair (Z1, Z2) 0 of the auxiliary variable, and its alternative form is:
3 experimental analysis
3.1 experiment settings
Select EXATA as a network emulator, set the following: The simulation platform is EXATA 2.1, which is the advanced simulation version developed by QualNet, which can be used to perform experimental simulation in semi-real estate. In order to achieve the purpose of obtaining the H.264 real-time video stream, the use of EXATA 2.1 is integrated with this algorithm, and the specific development details can be referred to the Exita User Manual. In the network structure design, the wired network access port is reserved, and the end-collected wireless network interface has a WiMAX interface, WLAN interface, and HSDPA interface. The connection to the server can be established through the IP address binding. The relevant parameter settings for the heterogeneous network are shown in Table 1.
This article compares the SCLQ-JSCC algorithm to the following multi-path / heterogeneous video transmission scheme: (1) Virtual Path System, VPS. The method uses the fountain code to build a heterogeneous video transmission path, in the algorithm implementation, the parameter update period is 0.5 s. The fountain code packet size is 8 b, the symbol length is 512 B. (2) Media Flow Rate Allocation, MFRA). The policy uses the maximum use of the maximum algorithm for multi-path video transmission, and the number of video layers is set to the value 1, since SVC / H.264 encoded scalability.
3.2 Analysis
To verify the performance of the SCLQ-JSCC algorithm, select the tolerable transmission loss rate to compare the forward error correction, FEC) redundancy and video coding rates.
Figure 4 shows the comparative case of three comparative algorithm redundancy and video coding rates. As can be seen from Figure 4 (a), the SCLQ-JSCC algorithm is obviously better than VPS and MFRA algorithm, and the MFRA policy is considered to be redundant. Supervisory than VPS strategy. According to Fig. 4 (b), the algorithm is superior to the two comparative strategies selected in the video coding rate indicator, and the VPS policy is higher than the MFRA algorithm because of the virtual path transmission problem.
During the experiment, the PSNR indicator (peak signal-to-noise ratio) standard deviation, mean and instantaneous value of the video received are shown in Table 2. Table 2 Experimental results show that in the PSNR index mean, the algorithm is always higher than the VPS and MFRA algorithm on 4 sets of video transmission in CITY, indicating that the distortion of the transmission signal is relatively smaller, the algorithm of this algorithm Relatively better. At the PSNR index standard, the standard deviation of this algorithm is the smallest, indicating that the stability of the video transmission of the specified algorithm is better.
The results shown in FIG. 5 are frame delay accumulation distributions during video transmission. According to the experimental results of Figure 5, the frame delay of the SCLQ-JSCC algorithm as mentioned herein is significantly lower than the VPS and the MFRA algorithm, which reflects the low delay characteristics of the algorithm, although the virtual path transmission problem is considered in the VPS algorithm. However, after the video frame is lost, you need to recall the virtual path, so it will affect the transmission delay of the video frame.
Figure 6 shows the comparative case of the effective loss rate index in the [30, 80] S period, pay attention to the PSNR value in the video transmission is not only associated with the loss rate, but also related to the loss of the video frame, so the indicator The quality of the video transmission process can be reflected to some extent.
According to Fig. 6, in the effective loss rate indicator, the algorithm is less than the selected VPS and MFRA algorithm, MFRA, because of the video split transmission technology, resulting in the loss rate than VPS and this article SCLQ-JSCC algorithm . The above experimental results verify the advantages of the proposed algorithm in video data transmission quality and transmission speed.
4 Conclusion
This paper proposes a multi-description independent parallel channel transmission framework based on Jing Jing quantified heterogeneous network video, establishes a multi-description independent parallel channel transmission framework for Gaussian video source transmission, and uses the lattice scale quantization method to reduce the decrease in the analog mapping bandwidth. Structure network video transmission performance. In the future, it will focus on the multi-access relay channel or multi-hop network MD-JSCC scheme to simulate the network topology. , Read full text, original text title: [Academic papers] based on lattice quantitative heterogeneous network video joint source channel coding
Article Source: [Micro Signal: Chinaaet, WeChat Public Number: Electronic Technology Application Chinaaet] Welcome to add attention! Please indicate the source of the article.
Our other product:
