We are developing a gstreamer based pipeline to read a video
source, rtp payload the video frame and then udpsink the rtp packets. We are finding that the pipeline bursts the rtp data unto
the network, creating network spikes. In one example, the interpacket network delay
is about 15micro seconds for approx 150 packets, meaning the whole video frame
is delivered unto the network in 2.25ms. Is there some way in which to ‘pace’ the rtp
packets so that it takes 20-30ms to send the same frame? Keith
Tyco Safety Products/CEM Systems Ltd.
Registered in Northern Ireland: NI 25728. Registered Office: 195 Airport Road West, Belfast, BT3 9ED.
Please note that any views or opinions presented in this email are solely those of the author and do not necessarily represent those of the company. This email and any attachments are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender immediately and delete any copies in your possession.
------------------------------------------------------------------------------ Protect Your Site and Customers from Malware Attacks Learn about various malware tactics and how to avoid them. Understand malware threats, the impact they can have on your business, and how you can protect your company and customers by using code signing. http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Hi,
On Thu, Jan 20, 2011 at 11:54 AM, Redfern, Keith <[hidden email]> wrote: > We are developing a gstreamer based pipeline to read a video source, rtp > payload the video frame and then udpsink the rtp packets. > > We are finding that the pipeline bursts the rtp data unto the network, > creating network spikes. In one example, the interpacket network delay is > about 15micro seconds for approx 150 packets, meaning the whole video frame > is delivered unto the network in 2.25ms. It's conceptually right that fragmented packets, having all the same timestamp, are sent as fast as possible over the network, the bottleneck being the network and physical layer setting. It's usually bad to try and artificially delay packets, as you will accrue network latencies over the time. If your problem is the number of packets (not the size of them) it's good to try with an higher MTU, you can usually set it from the payloader. If the MTU is already set appropriately (it really depends on the network you're using) or you're actually concerned about the # of bytes sent across the network, then you're encoding with too high parameters / quality. Try one or more of the following: - Reduce frame resolution. - Reduce frame rate. - Reduce encoder output bitrate. - Reduce encoder output quality. - Use a better encoder (e.g. h264 is usually better than mpeg-2). - Increase the number of P frames between I frames (or, that is the same, decrease the frequency of I frames). - Already said, increase the MTU (it will reduce the overhead caused by RTP, UDP and IP on the payload). > > Is there some way in which to ‘pace’ the rtp packets so that it takes > 20-30ms to send the same frame? Probably, but you're going through a lot of problems doing that ;) . Regards, Marco > > > > Keith > > > > > ________________________________ > Tyco Safety Products/CEM Systems Ltd. > Registered in Northern Ireland: NI 25728. Registered Office: 195 Airport > Road West, Belfast, BT3 9ED. > > Please note that any views or opinions presented in this email are solely > those of the author and do not necessarily represent those of the company. > This email and any attachments are confidential and intended solely for the > use of the individual or entity to whom they are addressed. If you have > received this email in error please notify the sender immediately and delete > any copies in your possession. > > > ------------------------------------------------------------------------------ > Protect Your Site and Customers from Malware Attacks > Learn about various malware tactics and how to avoid them. Understand > malware threats, the impact they can have on your business, and how you > can protect your company and customers by using code signing. > http://p.sf.net/sfu/oracle-sfdevnl > _______________________________________________ > gstreamer-devel mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/gstreamer-devel > > ------------------------------------------------------------------------------ Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)! Finally, a world-class log management solution at an even better price-free! Download using promo code Free_Logger_4_Dev2Dev. Offer expires February 28th, so secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsight-sfd2d _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Administrator
|
In reply to this post by Redfern, Keith
On Thu, 2011-01-20 at 09:54 +0000, Redfern, Keith wrote:
> We are developing a gstreamer based pipeline to read a video source, > rtp payload the video frame and then udpsink the rtp packets. > > We are finding that the pipeline bursts the rtp data unto the network, > creating network spikes. In one example, the interpacket network delay > is about 15micro seconds for approx 150 packets, meaning the whole > video frame is delivered unto the network in 2.25ms. > > Is there some way in which to ‘pace’ the rtp packets so that it takes > 20-30ms to send the same frame? You can pace (or throttle) the time at which the rtp packets are sent by adjusting the outgoing timestamps of all buffers (i.e. the ones coming out of the payloaders and going into the udpsinks). The packets without timestamps will be sent as soon as they are received. Wim is also working on a way to throttle that based on buffer(list?) duration also fwiw. Edward > > > > Keith > > > > > > > ______________________________________________________________________ > Tyco Safety Products/CEM Systems Ltd. > Registered in Northern Ireland: NI 25728. Registered Office: 195 > Airport Road West, Belfast, BT3 9ED. > > Please note that any views or opinions presented in this email are > solely those of the author and do not necessarily represent those of > the company. This email and any attachments are confidential and > intended solely for the use of the individual or entity to whom they > are addressed. If you have received this email in error please notify > the sender immediately and delete any copies in your possession. > > > > ------------------------------------------------------------------------------ > Protect Your Site and Customers from Malware Attacks > Learn about various malware tactics and how to avoid them. Understand > malware threats, the impact they can have on your business, and how you > can protect your company and customers by using code signing. > http://p.sf.net/sfu/oracle-sfdevnl > _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel ------------------------------------------------------------------------------ Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)! Finally, a world-class log management solution at an even better price-free! Download using promo code Free_Logger_4_Dev2Dev. Offer expires February 28th, so secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsight-sfd2d _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Free forum by Nabble | Edit this page |