Method, device, and computer program for optimizing transmission of motion vector related information when transmitting a video stream from an encoder to a decoder

A method of decoding encoded video data, the method comprising:

• selecting one prediction mode among a skip mode, a merge mode or a motion prediction mode;
• and
• decoding an image portion using the selected prediction mode;
• wherein during the decoding, an adaptive motion vector resolution is used only if the selected prediction mode is the motion prediction mode.

A device for decoding encoded video data, the device comprising at least one microprocessor configured for carrying out the steps of:

• determining a motion vector for decoding a first image portion, the decoding of the first image portion is performed according to one of a skip mode, a merge mode or a motion prediction mode;
• locating a reference image portion for the first image portion using the determined motion vector;
• and
• decoding the first image portion based on the reference image portion;
• wherein the motion vector has a sub-pixel resolution if the first image portion is decoded according to skip or merge modes;
• and
• wherein the motion vector has a resolution dependent on a coding mode flag, specifying the resolution of the motion vector for the first image portion, if the first image portion is decoded according to the motion prediction mode.

A device for decoding encoded video data, the device comprising at least one microprocessor configured for carrying out the steps of:

• selecting one prediction mode among a skip mode, a merge mode or a motion prediction mode;
• and
• decoding an image portion using the selected prediction mode;
• wherein during the decoding, an adaptive motion vector resolution is used only if the selected prediction mode is the motion prediction mode.