I think you will have to share snippets of your source code in order to explain how you initialize and use JPEGLiveSource and explain what quality is in your context. We would like to understand what goes on and perhaps try to reproduce the issue.
Please also test one thing for us. Try to run the MediaLiveSource sample unmodified, when using this sample unmodified in your test setup do you see the issue?
Sharing our code is nearly impossible. But let me assure you that we don’t do anything different than in your samples.
I can’t find a MediaLiveSource sample application in all of the downloaded SDKs.
I tried the MediaLiveViewer sample, which does exactly what we do and it shows the unwanted behavior.
Let me create a case for this issue for further investigation.
Yes, please do so!
Thanks for your effort so far!
Hi again,
I created the case (MSC419739) just before and you will soon receive an email from us. If not, please let us know through this question.
Hello and a happy new year!
As we didn’t receive a notification about the created support case (MSC419739) yet, I kindly ask what the current state is?
BR
Hi Jochen,
Happy new year to you as well.
I checked and the case is created, but it is your colleague (David) that is the registered contact person for you company, so he is probably receiving the updates?
So far we have not been able to reproduce. We have tried using the MediaLiveViewer sample and the video starts streaming almost immediately after selecting a camera. Even when running with FullHD resolution. I assume you reproduced using an unmodified version of MediaLiveViewer sample?
What does the sample report in the “Hardware Decoding Status” field?
The fact that reducing the quality helps on this for you sounds to me like this is a bandwidth issue, but given that Smart Client works fine on the same machine makes this unlikely.
Best regards,
Peter
Hi Peter,
you are right: we didn’t modify your samples!
The “Hardware Decoding Status” field shows the value “hardwareNvidia” in our test environment. The test machine has a Nvidia display adapter.
The customer actually doesn’t have the samples …
BR
An additional question: Are you running your application (and the sample) on the same machine as the server or another one?
The sample runs on another machine than the server.
BR
Hi again
Could you please provide a wireshark of the communication between the MediaLiveViewer sample and the XProtect server machine during one of the instances of this happening. Please also provide us with timestamps of when you select the camera in the item picker of the sample and when the video starts showing.
Also, do you have any antivirus on your machine that might cause file load to be slower? When running hardware decoding we need to load some dll’s, which might take a couple of seconds (but should never be 10-40 seconds).
Could you please also check whether selecting another camera after the first camera starts streaming, is faster than the first one?
Br,
Peter
Hello,
attached find the Wireshark files for the MediaLiveViewer sample (MediaLiveViewer.pcapng):
Camera was selected at No. 500
Video started at No. 605
BR
Thank you for the wireshark, but I am a bit confused. I assume that 10.10.77.124 is the address of the server and 10.10.1.199 is the address of the client (?), as that is source and destination of the video packages, but I don’t see a single package going from client to server?
Hello again.
You’re right about the ip addresses. We saw video on the client and you got the complete wireshark file.
Can we do something else?
Our customer is asking for a time estimation. As I am an engineer too I know its hard to tell, but are we talking about weeks or months roghly?
BR
Is there any chance that multicast is enabled? Or some sort of proxy is in place between client and server?
With the current information we have no way of getting any further, since we cannot reproduce and we cannot really see anything from the wireshark as long as no requests from the client can be seen.
Hi again,
the problem for wireshark was, that the camera and XProtect and the client were in different subnets. That’s why you did not see all the traffic.
Trying to get better files we noticed, that there is no more delay in the sample application and also not in our application.
We don’t know what could have caused this and are trying to figure it out together with out IT department. Maybe they changed something?
I think it might be interesting for you to know what it was if we find something?
BR
Hi Jochen,
We definitely would like to hear about your findings.
Could also be relevant for other users of the forum.
Br,
Peter