Method for transcoding video streams with reduced number of predictions

A method for transcoding video streams by a server, comprising:

• receiving a first request from a first client terminal;
• receiving a second request from a second client terminal;
• predicting macroblocks of frames of a first video stream to generate a first set of predictions in response to the first request and the second request;
• transforming the first set of predictions into a first set of quantized transform coefficients in response to the first request;
• transforming the first set of predictions into a second set of quantized transform coefficients in response to the second request;
• coding the first set of quantized transform coefficients into a first coded video stream;
• coding the second set of quantized transform coefficients into a second coded video stream having a resolution identical to a resolution of the first coded video stream;
• transmitting the first coded video stream from the server to the first client terminal, and transmitting the second coded video stream from the server to the second client terminal;
• receiving a third request from a third client terminal;
• predicting macroblocks of frames of a second video stream to generate a second set of predictions in response to the third request, wherein the second video stream has a resolution different from a resolution of the first video stream;
• transforming the second set of predictions into a third set of quantized transform coefficients;
• coding the third set of quantized transform coefficients into a third coded video stream;
• and
• transmitting the third coded video stream from the server to the third client terminal.