Encoder, a decoder and corresponding methods using compact MV storage
The disclosure provides a motion vector compression method, comprising: obtaining a temporal motion vector; determining a compressed motion vector using a binary representation of the temporal motion vector comprising an exponent part and/or a mantissa part, wherein the exponent part comprises N bits, the mantissa part comprises M bits, and wherein N is a non-negative integer and M is a positive integer; and performing a temporal motion vector prediction (TMVP) using the compressed motion vector.
1. A motion vector compression method, comprising:
obtaining an 18 bit temporal motion vector;
determining a compressed motion vector of the 18 bit temporal motion vector when an indicator indicates the 18 bit temporal motion vector need to be compressed, wherein a binary representation of the compressed motion vector comprising an exponent part or a mantissa part, wherein the exponent part comprises N bits, the mantissa part comprises M bits, wherein N is a non-negative integer and M is a positive integer, and wherein the representation of the compressed motion vector is 10 bits; and
performing a temporal motion vector prediction (TMVP) using the compressed motion vector;
wherein the determining a compressed motion vector, comprises:
removing least significant bits (LSB) by clipping the obtained temporal motion vector.
2. The motion vector compression method of claim 1 , wherein
the exponent part corresponds to most significant bit(s) (MSB) of the binary representation and the mantissa part corresponds to least significant bit(s) (LSB) of the binary representation; or,
the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation.
3. The motion vector compression method of claim 2 , further comprising: when the exponent part corresponds to MSB of the binary representation and the mantissa part corresponds to LSB of the binary representation, deriving a value of the compressed motion vector by:
deriving a first shift value by applying a right shift of M bit to the binary representation;
deriving last M bit of the binary representation as a first basic binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the first shift value to the first basic binary representation.
4. The motion vector compression method of claim 2 , further comprising: when the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation, deriving a value of a motion vector component by:
deriving last N bit of the binary representation as a second shift value;
deriving a second basic binary representation by applying a right shift of N bit to the binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the second shift value to the second basic binary representation.
5. The motion vector compression method of claim 1 , wherein the 18 bit temporal motion vector comprises a motion vector horizontal component and a motion vector vertical component.
6. A non-transitory computer-readable storage medium storing executable instructions for execution by a processing circuitry, wherein the executable instructions, when executed by the processing circuitry, configures the processing circuitry to carry out the method according to claim 1 .
7. A video data decoding device, comprising:
a memory storage, configured to store video data in a form of a bitstream; and
a video decoder, configured to perform the method according to claim 1 .
8. A encoder, comprising:
a circuitry configured to perform the method of claim 1 .
9. A video data encoding device, comprising:
a memory storage, configured to store video data in a form of a bitstream; and
a video decoder, configured to perform the method according to claim 1 .
10. A non-transitory storage medium storing a bitstream and one or more instructions executable by at least one processor to perform operations of encoding or decoding of the bitstream, the operations comprising:
obtaining an 18 bit temporal motion vector;
determining a compressed motion vector of the 18 bit temporal motion vector when an indicator indicates the 18 bit temporal motion vector need to be compressed, wherein a binary representation of the compressed motion vector comprising an exponent part or a mantissa part, wherein the exponent part comprises N bits, the mantissa part comprises M bits, wherein N is a non-negative integer and M is a positive integer, and wherein the representation of the compressed motion vector is 10 bits; and
performing a temporal motion vector prediction (TMVP) using the compressed motion vector;
wherein the determining a compressed motion vector, comprises:
removing least significant bits (LSB) by clipping the obtained temporal motion vector.
11. The storage medium of claim 10 , wherein
the exponent part corresponds to most significant bit(s) (MSB) of the binary representation and the mantissa part corresponds to least significant bit(s) (LSB) of the binary representation; or,
the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation.
12. The storage medium of claim 11 , wherein the operations further comprise: when the exponent part corresponds to MSB of the binary representation and the mantissa part corresponds to LSB of the binary representation, deriving a value of the compressed motion vector by:
deriving a first shift value by applying a right shift of M bit to the binary representation;
deriving last M bit of the binary representation as a first basic binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the first shift value to the first basic binary representation.
13. The storage medium of claim 11 , wherein the operations further comprise: when the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation, deriving a value of a motion vector component by:
deriving last N bit of the binary representation as a second shift value;
deriving a second basic binary representation by applying a right shift of N bit to the binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the second shift value to the second basic binary representation.
14. The storage medium of claim 10 , wherein the 18 bit temporal motion vector comprises a motion vector horizontal component and a motion vector vertical component.
15. A decoder, comprising:
a circuitry configured to perform operations of:
obtaining an 18 bit temporal motion vector;
determining a compressed motion vector of the 18 bit temporal motion vector when an indicator indicates the 18 bit temporal motion vector need to be compressed, wherein a binary representation of the compressed motion vector comprising an exponent part or a mantissa part, wherein the exponent part comprises N bits, the mantissa part comprises M bits, wherein N is a non-negative integer and M is a positive integer, and wherein the representation of the compressed motion vector is 10 bits; and
performing a temporal motion vector prediction (TMVP) using the compressed motion vector;
wherein the determining a compressed motion vector, comprises:
removing least significant bits (LSB) by clipping the obtained temporal motion vector.
16. The decoder of claim 15 , wherein
the exponent part corresponds to most significant bit(s) (MSB) of the binary representation and the mantissa part corresponds to least significant bit(s) (LSB) of the binary representation; or,
the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation.
17. The decoder of claim 16 , wherein the operations further comprise: when the exponent part corresponds to MSB of the binary representation and the mantissa part corresponds to LSB of the binary representation, deriving a value of the compressed motion vector by:
deriving a first shift value by applying a right shift of M bit to the binary representation;
deriving last M bit of the binary representation as a first basic binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the first shift value to the first basic binary representation.
18. The decoder of claim 16 , wherein the operations further comprise: when the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation, deriving a value of a motion vector component by:
deriving last N bit of the binary representation as a second shift value;
deriving a second basic binary representation by applying a right shift of N bit to the binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the second shift value to the second basic binary representation.
19. The decoder of claim 15 , wherein the 18 bit temporal motion vector comprises a motion vector horizontal component and a motion vector vertical component.