Methods for frequency domain packet loss concealment and related decoder
A method, decoder, and program code for filling an analysis window length with a time domain signal for concealing a lost audio frame associated with a received audio signal. A first segment of a previously received part of a received audio signal is copied from a prototype buffer. A second segment of the previously received part of the received audio signal is overlap added from the prototype buffer to an initial portion of a reconstructed part of the received audio signal followed by a remaining portion of the reconstructed part of the received audio signal.
1 . A method of filling an analysis window length for concealing a lost audio frame associated with a received audio signal, the method comprising:
copying a first segment of the received audio signal from a prototype buffer, wherein the first segment comprises a first portion of a time domain signal of a most recently synthesized audio frame, wherein the most recently synthesized audio frame includes a last decoded good frame prior to a lost audio frame associated with the received audio signal;
overlap adding a second segment of the received audio signal from the prototype buffer with an initial portion of a reconstructed audio signal, wherein the second segment comprises a second portion, directly following the first portion, of the time domain signal of the most recently synthesized audio frame, wherein the reconstructed audio signal includes a recreation of the lost audio frame generated using an inverse fast Fourier transform (IFFT); and
generating a temporal signal to fill an analysis window length of a lapped transform modified discrete cosine transform (MDCT) analysis window, the temporal signal comprising (i) the copied first segment of the received audio signal followed by (ii) the overlap added second segment of the received audio signal with the initial portion followed by (iii) a remaining portion of the reconstructed audio signal.
2 . The method of claim 1 , wherein copying the first segment from the prototype buffer comprises copying the first segment into a processing buffer, and wherein overlap adding the second segment from the prototype buffer with the initial portion of the reconstructed audio signal comprises overlap adding the second segment from the prototype buffer with the initial portion of the reconstructed audio signal into the processing buffer.
3 . The method of claim 2 , wherein the processing buffer is a lapped transform MDCT buffer.
4 . The method of claim 1 , wherein the reconstructed audio signal comprises a time evolved transform signal.
5 . The method of claim 1 , wherein the MDCT analysis window is asymmetric.
6 . The method of claim 1 , further comprising copying and overlap adding for consecutively lost frames, comprising:
copying a first segment of a recreated audio signal from a time continuity buffer into a processing buffer;
overlap adding a second segment of the recreated audio signal from the time continuity buffer with an initial portion of another reconstructed audio signal into the processing buffer; and
generating another temporal signal to fill another analysis window length comprising (i) the copied first segment of the recreated audio signal followed by (ii) the overlap added second segment of the recreated audio signal with the initial portion of the another reconstructed audio signal followed by (iii) a remaining portion of the another reconstructed audio signal.
7 . The method of claim 6 , wherein the time continuity buffer is updated with a newly recreated signal after the overlap adding.
8 . The method of claim 1 , wherein overlap adding the second segment of the received audio signal from the prototype buffer with the initial portion of the reconstructed audio signal comprises:
applying a first window to obtain a first scaled sample of the previously received first segment of the received audio signal from the prototype buffer;
applying a second window to obtain a second scaled sample of the reconstructed audio signal; and
summing the first and second scaled samples.
9 . The method of claim 1 , wherein a length of the copied first segment depends on a sampling frequency.
10 . The method of claim 1 , wherein a length of the overlap added second segment with the initial portion depends on a sampling frequency.
11 . The method of claim 1 , further comprising:
dynamically adapting lengths of the first and second segments based on an analysis of a previously synthesized time domain signal from a filled analysis window length.
12 . A decoding device for filling an analysis window length for concealing a lost audio frame associated with a received audio signal, the decoding device comprising:
a processor; and
memory coupled with the processor, wherein the memory includes instructions that when executed by the processor causes the decoding device to perform operations comprising:
copying a first segment of the received audio signal from a prototype buffer, wherein the first segment comprises a first portion of a time domain signal of a most recently synthesized audio frame, wherein the most recently synthesized audio frame includes a last decoded good frame prior to a lost audio frame associated with the received audio signal;
overlap adding a second segment of the received audio signal from the prototype buffer with an initial portion of a reconstructed audio signal, wherein the second segment comprises a second portion, directly following the first portion, of the time domain signal of the most recently synthesized audio frame, wherein the reconstructed audio signal includes a recreation of the lost audio frame generated using an inverse fast Fourier transform (IFFT); and
generating a temporal signal to fill an analysis window length of a lapped transform modified discrete cosine transform (MDCT) analysis window, the temporal signal comprising (i) the copied first segment of the received audio signal followed by (ii) the overlap added second segment of the received audio signal with the initial portion followed by (iii) a remaining portion of the reconstructed audio signal.
13 . The decoding device of claim 12 , wherein copying the first segment from the prototype buffer comprises copying the first segment into a processing buffer, and wherein overlap adding the second segment from the prototype buffer with the initial portion of the reconstructed audio signal comprises overlap adding the second segment from the prototype buffer with the initial portion of the reconstructed audio signal into the processing buffer.
14 . The decoding device of claim 13 , wherein the processing buffer is a lapped transform MDCT buffer.
15 . The decoding device of claim 12 , wherein the reconstructed audio signal comprises a time evolved transform signal.
16 . The decoding device of claim 12 , wherein the MDCT analysis window is asymmetric.
17 . The decoding device of claim 12 , wherein the operations further comprise copying and overlap adding for consecutively lost frames, comprising:
copying a first segment of a recreated audio signal from a time continuity buffer into the processing buffer;
overlap adding a second segment of the recreated audio signal from the time continuity buffer with an initial portion of another reconstructed audio signal into a processing buffer; and
generating another temporal signal to fill another analysis window length comprising (i) the copied first segment of the recreated audio signal followed by (ii) the overlap added second segment of the recreated audio signal with the initial portion of the another reconstructed audio signal followed by (iii) a remaining portion of the another reconstructed audio signal.
18 . The decoding device of claim 17 , wherein the operations further comprise updating the time continuity buffer with a newly recreated signal after the overlap add.
19 . The decoding device of claim 12 , wherein overlap adding the second segment of the received audio signal from the prototype buffer with the initial portion of the reconstructed audio signal comprises:
applying a first window to obtain a first scaled sample of the received audio signal from the prototype buffer;
applying a second window to obtain a second scaled sample of the reconstructed audio signal; and
summing the first and second scaled samples.
20 . The decoding device of claim 12 , wherein a length of the first segment depends on a sampling frequency.
21 . The decoding device of claim 12 , wherein a length of the overlap added second segment with the initial portion depends on a sampling frequency.
22 . The decoding device of claim 12 , wherein the operations further comprise:
dynamically adapting lengths of the first and second segments based on an analysis of a previously synthesized time domain signal from a filled analysis window length.
23 . A computer program product comprising a non-transitory storage medium including program code to be executed by at least one processor of a decoding device for filling an analysis window length for concealing a lost audio frame associated with a received audio signal, whereby execution of the program code causes the decoding device to perform operations comprising:
copying a first segment of the received audio signal from a prototype buffer, wherein the first segment comprises a first portion of a time domain signal of a most recently synthesized audio frame, wherein the most recently synthesized audio frame includes a last decoded good frame prior to a lost audio frame associated with the received audio signal;
overlap adding a second segment of the received audio signal from the prototype buffer to an initial portion of a reconstructed audio signal, wherein the second segment comprises a second portion, directly following the first portion, of the time domain signal of the most recently synthesized audio frame, wherein the reconstructed audio signal includes a recreation of the lost audio frame generated using an inverse fast Fourier transform (IFFT); and
generating a temporal signal to fill an analysis window length of a lapped transform modified discrete cosine transform (MDCT) analysis window, the temporal signal comprising (i) the copied first segment of the received audio signal followed by (ii) the overlap added second segment of the received audio signal with the initial portion followed by (iii) a remaining portion of the reconstructed audio signal.
24 . The computer program product of claim 23 , wherein the operations further comprise:
copying and overlap adding for consecutively lost frames, comprising:
copying a first segment of a recreated audio signal from a time continuity buffer into a processing buffer;
overlap adding a second segment of the recreated audio signal from the time continuity buffer with an initial portion of another reconstructed audio signal into a processing buffer; and
generating another temporal signal to fill another analysis window length comprising (i) the copied first segment of the recreated audio signal followed by (ii) the overlap added second segment of the recreated audio signal with the initial portion of the another reconstructed audio signal followed by (iii) a remaining portion of the another reconstructed audio signal.