Conclusion
Even though we did not get much time to experiment with method 3, between methods 2 and 3, we concluded that method 2 is more ideal for our scenario since there is only one receiver allowed and multiple senders. Syncing all the devices is much more easier and hassle free.
Also a good plugin for Gstreamer needs to be created for measuring relative latencies.
As discussed in the result section the scenario involving one camera connected to local interface and another one connected to WiFi network is the extreme boundary condition that could possibly ever be, adding the fact that the one of the devices have a codec problem which leads to lesser frame rates for one of the cameras. As our system worked despite such a large degree of variation we can conclude that video synchronizations can be achieved with large accuracies making way for real time applications.
Also a good plugin for Gstreamer needs to be created for measuring relative latencies.
As discussed in the result section the scenario involving one camera connected to local interface and another one connected to WiFi network is the extreme boundary condition that could possibly ever be, adding the fact that the one of the devices have a codec problem which leads to lesser frame rates for one of the cameras. As our system worked despite such a large degree of variation we can conclude that video synchronizations can be achieved with large accuracies making way for real time applications.