Hi Sugata,
Firstly as you may know already, traffic marking are done to classify traffic into different classes according to the priority so that they are meted with a different behaviour. It ensures smaller and critical traffic does not get stuck behind bulkier traffic which may not need priority.
Among the real time traffic, audio is the most critical. The Cisco recommended markings are according to the priority of the traffic type. Voice has a higher priority than video. Video is further separated into two categories - Interactive (Video conferencing etc.) and Streaming (playing a unicast or multicast video). The order of priority and reason for that is below:
1) Audio - Voice traffic is very sensitive to packet loss, latency. It requires no packet loss and latency of about 150 ms.
2) Interactive Video - Sensitive but can tolerate packet loss of about 1% and latency almost same as voice
3) Streaming Video - Less sensitive - can tolerate about 4-5% packet loss and latency of about 4-5 seconds
So it’s a general guideline\standard from Cisco to mark voice with a dscp 46 or ef, Interactive video 34 or af41, Streaming video as dscp 32 or cs4. Since this is Cisco recommended best practise, Cisco's most UC apps by default mark the traffic like this.
You can change it but strongly not advised unless you have a very good reason. Another reason not to change is some of the auto templates etc. you can normally use at access level catalyst switches use these markings, which otherwise you will have to manually change the mappings etc.
Hope this clarifies - if you have any further questions please feel free to ask.
Terry
