"Moreover I would like to know the difference between priority percent and bandwidth percent."
"priority percent" sets a cap on bandwidth usage. When there's congestion any traffic that exceeds the cap will be discarded. All traffic within/under the cap will be sent next.
"bandwidth percent" sets a minimum amount of bandwidth. The class may obtain more if it's available. When more traffic is offered then the class minimum, and excess bandwidth is not available for it, it is queued.
The norm for VoIP is priority classification. However, don't forget you might need to also treat voice signaling as special too.
For video conferencing, I prefer priority for it too, but it sometimes will work, not too badly, within a normal class. Amount of bandwidth is dependent on the VC device. Video is often very bursty in bandwidth demands. The latter addressed by either setting more bandwidth for the class than the "average" and/or increase queue depth. The former, setting more bandwidth then "required", also avoids transient queuing delays. (Remember unused bandwidth, regardless of what's configured, is still available to other traffic. What's important is to avoid delays or drops of your real-time traffic.)
