Method subband of coding and decoding audio signals using variable length windows

• US 5,848,391 A
• Filed: 07/11/1996
• Issued: 12/08/1998
• Est. Priority Date: 07/11/1996
• Status: Expired due to Term

• Current Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated), Fraunhofer Gesellschaft Zur Foerderung Der Angewandten Forsching E.V.
• Original Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated), Fraunhofer Gesellschaft Zur Foerderung Der Angewandten Forsching E.V.
• Inventors: Davidson, Grant, Robinson, Charles, Dietz, Martin, Gbur, Uwe, Bosi, Marina, Kunz, Oliver, Brandenburg, Karlheinz
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): ZINTEL, HAROLD ALBERT
• Application Number:
  US08/678,666
• Time in Patent Office: 880 Days
• Field of Search: 704/229, 704/500
• US Class Current: 704/200.1
• CPC Class Codes: G10L 19/022 Blocking, i.e. grouping of samples in time; Choice of analysis windows; Overlap factoring
• CPC Class Codes: H04B 1/665 using psychoacoustic properties of the ear, e.g. masking effect

Abstract

A method of encoding time-discrete audio signals comprises the steps of weighting the time-discrete audio signal by means of window functions overlapping each other so as to form blocks, the window functions producing blocks of a first length for signals varying weakly with time and blocks of a second length for signals varying strongly with time. A start window sequence is selected for the transition from windowing with blocks of the first length to windowing with blocks of the second length, whereas a stop window sequence is selected for the opposite transition. The start window sequence is selected from at least two different start window sequences having different lengths, whereas the stop window sequence is selected from at least two different stop window sequences having different lengths. A method of decoding blocks of encoded audio signals selects a suitable inverse transformation as well as a suitable synthesis window as a reaction to side information associated with each block.

Claims

1. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function including a first half and a second half for producing blocks of a first length for signals varying weakly with time and a second window function including a first half and a second half for producing blocks of a second length for signals varying strongly with time, said method comprising:

applying said first window function, to said time-discrete audio signals to produce blocks of said first length;

upon detecting signals varying strongly with time, applying one start window sequence from a plurality of different start window sequences to effect a transition from producing blocks of said first length to producing blocks of said second length, one of said start window sequences including a first part corresponding to said first half of said first window function, a second non-overlapping part having a length, a third part corresponding to said second half of said second window function, and a fourth part having at least one said overlapping second window function;

upon detecting signals varying weakly with time, applying one stop window sequence from a plurality of different stop window sequences to effect a transition producing blocks of said second length to producing blocks of said first length, one of said stop window sequences comprising a first part having at least one overlapping second window function, a second part corresponding to said first half of said second window function;

a third non-overlapping part having a length; and

a fourth part corresponding to said second half of said first window function; and

wherein, said length of said non-overlapping part and said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said start window sequence and said stop window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.

2. A method according to claim 1, wherein said start window sequence and said stop window sequence are not a mirror image of each other.

3. A method according to claim 1, wherein the first window function has the second half overlapping with the first half of the subsequent first window function and the second window function has the second half overlapping with the first half of the subsequent second window function, thereby said window functions windowing twice as many time-discrete signal values as there are frequency coefficients.

4. A method according to claim 1, wherein said applying includes said start window sequence including said fourth part having three said overlapping second window functions.

5. A method according to claim 1, wherein said applying includes said stop window sequence including said first part having three said overlapping second window functions.

8. A method according to claim 1, wherein said applying said start window sequence includes using said start window sequence including fourth part consisting of three said overlapping second window functions, and said stop window sequence including first part with consisting of three said overlapping second window functions.

9. A method according to claim 1, wherein said applying said start window sequence includes using said start window sequence including said fourth part consisting of three said overlapping second window functions, at least one window sequence including eight said overlapping second window functions, and said stop window sequence including said first part with three said overlapping second window functions.

12. A method according to claim 1, wherein said applying said start window sequence includes using said start window sequence including said fourth part consisting of three said overlapping second window functions, and said applying said stop window sequence includes using said stop window sequence including said first part which corresponds to said first half of said second window function.

13. A method according to claim 1, wherein said applying said start window sequence includes using said start window sequence including said fourth part consisting of three overlapping second window functions, at least one window sequence consisting of eight overlapping said second window functions, and said applying said stop window sequence includes using said stop window sequence including said first part which corresponds to said first half of said second window function.

14. A method according to claim 1, wherein said applying said start window sequence includes using said start window sequence including a first part corresponding to said first half of said first window function, a constant second part, and a fourth part which corresponds to said second half of said second window function, and said applying said stop window sequence includes using said stop window sequence including said first part having three said overlapping second window functions.

15. A method according to claim 1, wherein said applying said start window sequence includes using said start window sequence including a fourth part which corresponds to said second half of said second window function, at least one window sequence including eight overlapping second window functions, and said applying said stop window sequence includes using said stop window sequence including said first part having three said overlapping second window functions.

6. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function, including a first half and a second half, for producing blocks of a first length for signals varying weakly with time and a second window function, including a first half and a second half, for producing blocks of a second length for signals varying strongly with time, said method comprising:

applying said first window function to said time-discrete audio signals to produce blocks of said first length; and

upon detecting signals varying strongly with time, applying a start window sequence to effect a transition from producing blocks of said first length to producing blocks of said second length;

said start window sequence including a first part corresponding to said first half of said first window function, a second non-overlapping part having a length, a third part corresponding to said second half of said second window function, and a fourth part having at least one said overlapping second window function;

wherein said length of said non-overlapping part and said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said start window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.

10. A method according to claim 6, wherein said applying said start window sequence includes using said start window sequence including said fourth part having three said overlapping second window functions.

7. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function, including a first half and a second half, for producing blocks of a first length for signals varying weakly with time and a second window function, including a first half and a second half, for producing blocks of a second length for signals varying strongly with time, said method comprising:

applying said second window function to said time-discrete audio signals to produce blocks of said second length;

upon detecting signals varying weakly with time, applying a stop window sequence to effect a transition from producing blocks of said second length to producing blocks of said first length, said stop window sequence including a first part having at least one said overlapping second window function, a second part corresponding to said first half of said second window function, a third non-overlapping part having a length, and a fourth part corresponding to said second half of said first window function;

wherein said length of said non-overlapping part and said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said stop window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.

11. A method according to claim 7, wherein said applying said stop window sequence includes using said stop window sequence including said first part having three said overlapping second window functions.

16. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function including a first half and a second half for producing blocks of a first length for signals varying weakly with time and a second window function including a first half and a second half for producing blocks of a second length for signals varying strongly with time, said method comprising:

windowing said time-discrete audio signals by employing said first window function producing blocks of said first length;

upon detecting transient signals applying a start window sequence to achieve a transition from windowing with blocks of said first length to windowing with blocks of said second length and

subsequently applying a stop window sequence for an opposite transition,

wherein said start or stop window sequence includes a segment having a non-overlapping length and another segment having a number of said second window functions being overlapped, said non-overlapping length and said number of said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said start window sequence and said stop window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.

17. A method according to claim 16, wherein said applying includes said start window sequence including a first part corresponding to said first half of said first window function and a last part which corresponds to said second half of said second window function said stop start window sequence includes a first part having said number of overlapping second window functions.

18. A method according to claim 16, wherein said applying includes said start window sequence including a first part which corresponds to said first half of said second window function and a last part having said number of overlapping second window functions and said stop window sequence including a first part having said first half of said second window function.

19. A method of decoding time-discrete audio signals encoded in blocks, said method comprising the following steps:

reading of side information with which said encoded blocks are provided, said side information referring to a first or second window function, a start window sequence or a stop window sequence which has been used in connection with the block in question and to the transform which has been used with the block in question, wherein said start or stop window sequence includes a segment having a non-overlapping length and another segment having a number of said second window functions being overlapped, said non-overlapping length and said number of said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said start window sequence and said stop window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function;

selecting an inverse transform and a synthesis window function as a reaction to the side information read; and

re-transforming and windowing with the selected inverse transform and the selected synthesis window function.

20. A method of decoding time-discrete audio signals encoded in blocks according to claim 19, wherein, during coding, the time-discrete audio signals were weighted by means of said window functions overlapping one another so as to form blocks of a first length for signals varying weakly with time and blocks of a second length for signals varying strongly with time, said start window sequence was selected for a transition from windowing with blocks of said first length to windowing with blocks of said second length and said stop window sequence was selected for an opposite transition.

21. A method of decoding time-discrete audio signals encoded in blocks according to claim 19, wherein, during coding, the time-discrete audio signals were weighted by means of said window functions overlapping one another so as to form the blocks, the window functions produced blocks of a first length for signals varying weakly with time and blocks of a second length for signals varying strongly with time, said start window sequence was selected for a transition from windowing with blocks of said first length to windowing with blocks of said second length and said stop window sequence was a selected for an opposite transition, and a combined stop start window sequence which produces a block of the first length was situated between said start and stop window sequences and between window functions producing blocks of the second length.

22. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function including a first half and a second half for producing blocks of a first length for signals varying weakly with time and a second window function including a first half and a second half for producing blocks of a second length for signals varying strongly with time, said method comprising the steps of:

windowing said time-discrete audio signals by employing said first window function producing blocks of said first length;

upon detecting signals varying strongly with time, selecting a start window sequence from at least two different start window sequences for transition from windowing with blocks of said first length to windowing with blocks of said second length, one of said start window sequences including a first part corresponding to said first half of said first window function, a second non-overlapping part having a length, a third part corresponding to said second half of said second window function, and a fourth part having a number of overlapped said second window functions;

employing said selected start window sequence for a transition from windowing with blocks of said first length to windowing with blocks of said second length; and

upon detecting signals varying weakly with time, selecting a stop window sequence from at least two different stop window sequences for transition from windowing with blocks of said second length to windowing with blocks of said first length, one of said stop window sequences including a first part having a number of overlapped said second window functions, a second part corresponding to said first half of said second window function, a third non-overlapping part having a length, and a fourth part corresponding to said second half of said first window function;

wherein, said length of said non-overlapping part and said number of said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said start window sequence and said stop window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.

23. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function, including a first half and a second half, for producing blocks of a first length for signals varying weakly with time and a second window function, including a first half and a second half, for producing blocks of a second length for signals varying strongly with time, said method comprising:

windowing said time-discrete audio signals by employing said first window function producing blocks of said first length; and

upon detecting signals varying strongly with time, employing a start window sequence for a transition from windowing with blocks of said first length to windowing with blocks of said second length;

said start window sequence including a first part corresponding to said first half of said first window function, a second non-overlapping part having a length, a third part corresponding to said second half of said second window function, and a fourth part having a number of said second window functions being overlapped;

wherein said length of said non-overlapping part and said number of said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said start window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.

24. A method of encoding time-discrete audio signals including weighting the time-discrete audio signals by means of window functions overlapping each other so as to form blocks, said window functions comprising a first window function, including a first half and a second half, for producing blocks of a first length for signals varying weakly with time and a second window function, including a first half and a second half, for producing blocks of a second length for signals varying strongly with time, said method comprising:

windowing said time-discrete audio signals by employing said second window function producing blocks of said second length; and

upon detecting signals varying weakly with time, employing a stop window sequence for transition from windowing with blocks of said second length to windowing with blocks of said first length, said stop window sequence including a first part having a number of said second window functions being overlapped, a second part corresponding to said first half of said second window function, a third non-overlapping part having a length and a fourth part corresponding to said second half of said first window function;

wherein, said length of said non-overlapping part and said number of said overlapping second window functions are selected in a manner that a number of frequency coefficients obtained by a frequency transform of said time-discrete audio signals windowed by said stop window sequence is equal to a number of frequency coefficients obtained by said frequency transform of time-discrete audio signals windowed by said first window function.