Re: [Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Ross Finlayson
> On Jul 30, 2020, at 1:50 PM, Zhang Qian(张倩) wrote: > > > Hi Ross, > > Yes, for function RTPInterface::sendDataOverTCP , I found it will not resend > the header data if sent failed for TCP, then it will lose the packet with > this header. That’s correct. > We need to check why data sen

[Live-devel] 答复: [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Zhang Qian(张倩)
Hi Ross, Yes, for function RTPInterface::sendDataOverTCP , I found it will not resend the header data if sent failed for TCP, then it will lose the packet with this header. We need to check why data sent failed for our OS. Thanks for your response. :) Thanks. /Qian -邮件原件- 发件人: live

Re: [Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Alexander Tumarov
Thank you for clarification, Alexander On Wednesday, July 29, 2020 2:33:14 PM IDT Ross Finlayson wrote: > > On Jul 29, 2020, at 11:12 PM, Alexander Tumarov > > wrote: > > > > Hi Ross, > > > > In the case of TCP should it suspend sending up to the point there is a > > space available instead of d

Re: [Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Ross Finlayson
> On Jul 29, 2020, at 11:12 PM, Alexander Tumarov wrote: > > Hi Ross, > > In the case of TCP should it suspend sending up to the point there is a space > available instead of dropping RTP packets? This would not be real-time streaming - the service that our RTSP/RTP server provides. If yo

Re: [Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Alexander Tumarov
Hi Ross, In the case of TCP should it suspend sending up to the point there is a space available instead of dropping RTP packets? or close connection in such a case? or at least should it be up to implementation/user code? RTSP can be user for streaming stored data - not only real time strea

Re: [Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Ross Finlayson
> On Jul 29, 2020, at 9:05 PM, Zhang Qian(张倩) wrote: > > > Hi Ross, > > > Sorry for misunderstanding. I catch the tcpdump log for rtsp server, and I > use the TCP socket. Seems that these FU packets are not sent to TCP protocol > stack. So I want to check whether these packets are lost in

[Live-devel] 答复: [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Zhang Qian(张倩)
Hi Ross, Sorry for misunderstanding. I catch the tcpdump log for rtsp server, and I use the TCP socket. Seems that these FU packets are not sent to TCP protocol stack. So I want to check whether these packets are lost in rtsp server for large bitrate. Thanks. /Qian -邮件原件- 发件人: live

Re: [Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Ross Finlayson
> On Jul 29, 2020, at 8:34 PM, Zhang Qian(张倩) wrote: > > > Hi Ross, > > I found sometimes RTP sequence is not continuous when video bitrate is very > high such as 8Mbps. Seems that it will lose some FU packets. Congratulations! You have discovered datagrams! UDP (and RTP, which uses UDP

[Live-devel] [live555] RTP sequence number is not continuous when video bitrate is high (for example: 8Mbps)

2020-07-29 Thread Zhang Qian(张倩)
Hi Ross, I found sometimes RTP sequence is not continuous when video bitrate is very high such as 8Mbps. Seems that it will lose some FU packets. Have you ever come across this issue? How to avoid it. Thanks. :) tcpdump log: 36300 15:34:54.70596210.*.**.** 10.**.**.** H265