Google
 

Home | Product | Solution |Purchase | contract-us

กก
กก
Video server network bandwidth analysis and solutions

    

Perhaps you will encounter this confusion, you have built a 2 M-bandwidth networks, at the very beginning you may just want to build an intranet, web browser for some of the business, you may increase after an internal download business, then After you may in your business network, an increase of the business-demand streaming media, and also increased the network, such as video conferencing business, finally, you decided to increase the network video surveillance business. All of this business you run in 2 M-bandwidth networks, sometimes you find that video is smooth, sometimes you also found that video card can not be very normal here. At this time, even if you artificially lower the bit rate video servers, still unable to get smooth picture. This is because if the bandwidth of only 2 M running a business bandwidth is sufficient, but when a business to run all the time, all operations of the total demand for bandwidth will more than 2 M bandwidth, thus every business in the actual use of isolated On the bandwidth is insufficient. At this time you will find that some time video image is very smooth, some very time video images on the card.

     Perhaps you used early MJPEG encoding video server, you may now have to upgrade to H.264 or MPEG4 video encoding server, and if your network was not sufficient, as described above in the case. Well, you will find that, H.264 or MPEG4 video encoding video server instead of fluency than MJPEG encoding video server fluency. Despite all the manufacturers have to tell you H.264 or MPEG4 occupy less bandwidth, with better image quality, but in your situation the performance of networks is not the case.


     All of the above questions from the network, network bandwidth, or network bandwidth is too low and not clearly destabilizing effects of the video here, we will return to the phenomenon of the above two categories are as follows:

     1, leading to low network bandwidth network video server does not fluent

     The case occurred in the use of ADSL equipment such as video transmission via the Internet network monitoring business. Most users of ADSL upstream bandwidth may be only 512 K or less, the actual transmission bandwidth on the Internet are often significantly less than 512 K, in the use of network video server network video surveillance video business is not very fluent, users have to constantly lower the video server bitrate values, the actual transmission bandwidth to meet the requirements, at the same time ensure that there are relatively good image quality, and often these adjustments work for the customers as it is very troublesome operation.

     2, network bandwidth is not clear and lead to instability video server does not fluent

     As noted in the second case, for example, the surface, there are 2 M sufficient bandwidth, but the actual use in business due to too much excess bandwidth capacity, resulting in insufficient bandwidth video. This situation is unclear and unstable bandwidth, suitable for the construction of some of the earlier networks. Many early 2 M network is in accordance with the construction of optical fiber, optical fiber in 2 M on a variety of business operations, such as visit the company website, internal download, internal e-mail, internal video-on-demand business in this 2 M optical fiber business. An increase in the network video surveillance business, still in the original 2 M fiber all the business, then we will find that video images can not be here very fluent, or even a few seconds before moving to Shiji Miao, unable to meet the requirements of normal monitoring operations. At this time users try to start the video server to reduce the bit rate, eventually found unable to find a suitable value to the image to achieve fluency. Because of too many and miscellaneous business, no one can know the current real-time bandwidth in the end is how much » Video server set up in the end the bit rate in the number of suitable » In particular, the phenomenon of the Internet online video surveillance business is also very prominent.

     We can not help but ask, it's clear the H.264 or MPEG4 compression algorithm advanced a lot, the greater the lower stream, but why in the low-bandwidth performance is not satisfactory but it »

     This series from the MPEG compression algorithm of the core. MPEG compression algorithm from the early series of MPEG-1, MPEG-2 to the current development of MPEG-4 and the latest H.264, the core compression principle has not changed. MPEG compression algorithm in the series of time-domain on the compressed video stream. And MJPEG compression of each frame are separate plans in different ways JPEG, MPEG compression algorithm using a series of I-frame, P and B frame frame model, most of the current compression chip or compression algorithm did not materialize B frame. We do not have too much concern I frame, P-frame the details need to know in these three models in the frame, I frame is the key frame, P-frame is forecast before the frame, B is the frame after frame to the forecast. When I received a decoder time frame, you can decode immediately, without the need for other conditions, but when the decoder received P-frame or B frames, can not be immediately decode, for example, P-frame decoder Decoders needed to decode the frame in front of the P-frame or I frame. If the front of the P-frame or frame I lost in the network, even behind a row received a lot of P-frame, the decoder can not be normal. For example, in the case of insufficient bandwidth network lost an I-frame, even if the subsequent P frames received, still unable to correctly decode the image. This is why low-bandwidth H.264 or MPEG4 compression algorithm performance than MJEPG reasons. MJPEG frame are independent of each frame, as long as a transmission network, we can decode to a picture. In the case of insufficient bandwidth, MJPEG per second could only send three or four frames, but users have in this one can see three or four seconds of screen; on H.264, because their own stream Itself is relatively low, probably in the low-bandwidth under one second to be able to send the 15th and 16th frames, but if I frame on the network lost, even if received 15 consecutive P-frame can not be normal to decode a picture, then to the user The intuitive feeling of the screen long stagnation, and even low-bandwidth H.264 than MJPEG.

     Third, solutions

     Appear above the cause of the problem is not as H.264 compression algorithm as MJPEG, which is totally against this because we do not have the network environment to optimize, any good algorithm, in practical application is carried out according to the application of the corresponding scene Optimized, this is the most optimized for transmission to the network. We can from the following three aspects to solve the problem.

     1, the process of encoding increase the risk of I-frame.

     MPEG encoding algorithm in the series, usually with GOP (GROUP OF PICTURE) to set the value of the I-frame interval, such as when the GOP is set to 30, says that every 30 I have a frame. I frames are usually occupy large volume of data, many GOP encoder is set to be 120 or higher, so I can lower the probability of a frame, thereby reducing the stream, but the consequences when the network bandwidth is not enough, lost a I frame it within a few seconds can not be correct decoding data frame until I receive the next frame, leading to prolonged stagnation screen. After our long-term tests, in the case of low bandwidth GOP's value adjustment of 5 more appropriate, such a matter of seconds I can have five frames, even if I occasionally lose a frame, the next I-frame can quickly make up for the , The image will not long stagnation.

     2, real-time detection network bandwidth

     The intelligent video server to the bandwidth for real-time detection and control, we can respond through the mechanism of control methods such as the current network bandwidth is the actual number, this is completely embedded software can be achieved. As long as we can grasp the real-time bandwidth, the encoder can be the strategy and send real-time strategy adjustment. For different bandwidth adopt a different strategy and sent coded strategy, thus maximizing the utilization of bandwidth for video transmission.

     3, sent optimization strategy, I increase the risk of sending frame

     When embedded software testing to the current lack of bandwidth when the transfer module of the adjustment should be sent smart strategy, I increase the risk of sending the frame. Apart from the GOP by reducing the value of output increased risk of I-frame, the network sent frame strategies to ensure that I can send successful, can be used as a mechanism for re-issued to ensure that I frame agreement on the network will not be lost. In this way, can increase the effective frame of bandwidth usage, reduce the bandwidth invalid frame-consuming.


กก

Recent Post:

กก

IP/Network camera server/Wireless Network Camera/Knowledge Articles: