Frame-based audio coding with additional filterbank to suppress aliasing artifacts at frame boundaries

• US 5,913,191 A
• Filed: 10/17/1997
• Issued: 06/15/1999
• Est. Priority Date: 10/17/1997
• Status: Expired due to Term

• Current Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Original Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Inventors: Fielder, Louis Dunn
• Primary Examiner(s): Dorvil, Richemond
• Application Number:
  US08/953,121
• Time in Patent Office: 606 Days
• Field of Search: 704/201, 704/229, 704/230, 704/224, 704/225, 704/203, 704/205, 704/206, 704/211, 704/500, 704/504, 704/502, 704/503, 704/265, 704/266, 704/262, 704/216, 704/219
• US Class Current: 704/230
• CPC Class Codes: G10L 19/0204 using subband decomposition
• CPC Class Codes: G10L 25/18 the extracted parameters being spectral information of each sub-band
• CPC Class Codes: G11B 20/10 Digital recording or reproducing

Abstract

Several audio signal processing techniques may be used in various combinations to improve the quality of audio represented by an information stream formed by splice editing two or more other information streams. The techniques are particularly useful in applications that bundle audio information with video information. In one technique, gain-control words conveyed with the audio information stream are used to interpolate playback sound levels across a splice. In another technique, special filterbanks or forms of TDAC transforms are used to suppress aliasing artifacts on either side of a splice. In yet another technique, special filterbanks or crossfade window functions are used to optimize the attenuation of spectral splatter created at a splice. In a further technique, audio sample rates are converted according to frame lengths and rates to allow audio information to be bundled with, for example, video information. In yet a further technique, audio blocks are dynamically aligned so that proper synchronization can be maintained across a splice. An example for 48 kHz audio with NTSC video is discussed.

Claims

1. A method for signal processing comprising:

receiving an input signal comprising a sequence of frames, a respective input signal frame comprising a start block of signal samples, one or more interim blocks of signal samples and an end block of signal samples, said blocks of signal samples representing audio information,

generating, in response to a respective input signal frame, first filtered signals by applying a first filterbank to said start block of signal samples, second filtered signals by applying a second filterbank to said one or more interim blocks of signal samples, and third filtered signals by applying a third filterbank to said end block of signal samples, wherein said first, second and third filterbanks each have a respective length, the respective lengths of said first and second filterbanks being unequal and the respective lengths of said second and third filterbanks being unequal, and

generating an output signal suitable for transmission or storage by assembling said first filtered signals, said second filtered signals and said third filtered signals into a respective output signal frame, whereby a sequence of output signal frames is generated by assembling first, second and third filtered signals generated in response to said plurality of input signal frames.

3. A method according to claim 1 or 2 wherein said signal sample blocks in a respective input signal frame overlap one another by N/2 samples,

said first filterbank is of length 3N/2 and has aliasing characteristics such that, in response to said first filtered signals, a complementary first synthesis filterbank generates a recovered start block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a first N/2 samples,

said second filterbank is of length N and has aliasing characteristics such that, in response to said second filtered signals, a complementary second synthesis filterbank generates one or more recovered interim blocks of signal samples each having aliasing components, and

said third filterbank is of length 3N/2 and has aliasing characteristics such that, in response to said third filtered signals, a complementary third synthesis filterbank generates a recovered end block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a last N/2 samples.

4. A method according to claim 3 wherein

said first filterbank is implemented by a first transform and a first analysis window function of length 3N/2 samples, wherein said first window function has a first portion of length N/2 and a second portion of length N/2 that are substantially not equal to zero, and a third portion of length N/2 that is substantially equal to zero,

said second filterbank is implemented by a second transform and a second analysis window function of length N samples, wherein said second window function has a first portion of length N/2 samples and a second portion of length N/2 samples that is substantially not equal to zero, and

said third filterbank is implemented by a third transform and a third analysis window function of length 3N/2 samples, wherein said third window function has a first portion of length N/2 that is substantially equal to zero, and a second portion of length N/2 and a third portion of length N/2 that are substantially not equal to zero.

5. A method according to claim 4 wherein the first and second portions of said first analysis window function have been derived from a Kaiser-Bessel window function, said second analysis window function has been derived from said Kaiser-Bessel window function, and the second and third portions of said third analysis window function have been derived from said Kaiser-Bessel window function.

2. A method for signal processing comprising:

receiving an input signal comprising a sequence of frames, a respective input signal frame comprising a start block of signal samples, one or more interim blocks of signal samples and an end block of signal samples, said blocks of signal samples representing audio information,

generating, in response to a respective input signal frame, a first filtered signal block by applying a first filterbank to said start block of signal samples, one or more second filtered signal blocks by applying a second filterbank to said one or more interim blocks of signal samples, and a third filtered signal block by applying a third filterbank to said end block of signal samples, wherein said second filterbank generates said second filtered signal blocks having aliasing artifacts, said first filterbank generates said first filtered signal block having aliasing artifacts that cancel aliasing artifacts in a respective second filtered signal block but having substantially no other aliasing artifacts, and said third filterbank generates said third filtered signal block having aliasing artifacts that cancel aliasing artifacts in a respective second filtered signal block but having substantially no other aliasing artifacts, and

generating an output signal suitable for transmission or storage by assembling said first filtered signal block, said one or more second filtered signal blocks and said third filtered signal block into a respective output signal frame, whereby a sequence of output signal frames is generated by assembling first, second and third filtered signal blocks generated in response to said plurality of input signal frames.

6. A method for signal processing comprising:

receiving an input signal comprising a sequence of frames, a respective input signal frame comprising first filtered signals, second filtered signals and third filtered signals,

generating, in response to a respective input signal frame, a start block of signal samples by applying a first synthesis filterbank to said first filtered signals, one or more interim blocks of signal samples by applying a second synthesis filterbank to said second filtered signals, and an end block of signal samples by applying a third synthesis filterbank to said third filtered signals, wherein said first, second and third synthesis filterbanks each have a respective length, the respective lengths of said first and second synthesis filterbanks being unequal and the respective lengths of said second and third synthesis filterbanks being unequal, and

generating a sequence of output signal frames, a respective output signal frame comprising a plurality of signal sample blocks, wherein a respective frame comprises said start block, said one or more interim blocks and said end block, wherein said signal sample blocks represents audio information.

8. A method according to claim 6 or 7 wherein said signal sample blocks in a respective output signal frame overlap one another by N/2 samples,

said first synthesis filterbank is of length 3N/2 generates said start block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a first N/2 samples,

said second synthesis filterbank is of length N and generates said one or more interim blocks of signal samples each having aliasing components, and

said third synthesis filterbank is of length 3N/2 and generates said end block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a last N/2 samples.

9. A method according to claim 8 wherein

said first synthesis filterbank is implemented by a first transform and a first synthesis window function of length 3N/2 samples, wherein said first window function has a first portion of length N/2 and a second portion of length N/2 that are substantially not equal to zero, and a third portion of length N/2 that is substantially equal to zero,

said second synthesis filterbank is implemented by a second transform and a second synthesis window function of length N samples, wherein said second window function has a first portion of length N/2 samples and a second portion of length N/2 samples that is substantially not equal to zero, and

said third synthesis filterbank is implemented by a third transform and a third synthesis window unction of length 3N/2 samples, wherein said third window function has a first portion of length N/2 that is substantially equal to zero, and a second portion of length N/2 and a third portion of length N/2 that are substantially not equal to zero.

10. A method according to claim 9 wherein the first and second portions of said first synthesis window function have been derived from a Kaiser-Bessel window function, said second synthesis window function ha s been derived from said Kaiser-Bessel window function, and the second and third portions of said third synthesis window function have been derived from said Kaiser-Bessel window function.

7. A method for signal processing comprising:

receiving an input signal comprising a sequence of frames, a respective input signal frame comprising a first filtered signal block, one or more second filtered signal blocks and a third filtered signal block,

generating, in response to a respective input signal frame, a start block of signal samples by applying a first synthesis filterbank to said first filtered signal block, one or more interim blocks of signal samples by applying a second synthesis filterbank to said one or more second filtered signal blocks, and an end block of signal samples by applying a third synthesis filterbank to said third filtered signal block, wherein said second synthesis filterbank generates said interim blocks having aliasing artifacts, said first synthesis filterbank generates said start block having aliasing artifacts that cancel aliasing artifacts in a respective interim block but having substantially no other aliasing artifacts, and said third synthesis filterbank generates said end block having aliasing artifacts that cancel aliasing artifacts in a respective interim block but having substantially no other aliasing artifacts, and

generating a sequence of output signal frames, a respective output signal frame comprising a plurality of signal sample blocks, wherein a respective frame comprises said start block, said one or more interim blocks and said end block, wherein said signal sample blocks represents audio information.

11. A device for signal processing comprising:

means for receiving an input signal comprising a sequence of frames, a respective input signal frame comprising a start block of signal samples, one or more interim blocks of signal samples and an end block of signal samples, said blocks of signal samples representing audio information,

means for generating, in response to a respective input signal frame, first filtered signals by applying a first filterbank to said start block of signal samples, second filtered signals by applying a second filterbank to said one or more interim blocks of signal samples, and third filtered signals by applying a third filterbank to said end block of signal samples, wherein said first, second and third filterbanks each have a respective length, the respective lengths of said first and second filterbanks being unequal and the respective lengths of said second and third filterbanks being unequal, and

means for generating an output signal suitable for transmission or storage by assembling said first filtered signals, said second filtered signals and said third filtered signals into a respective output signal frame, whereby a sequence of output signal frames is generated by assembling first, second and third filtered signals generated in response to said plurality of input signal frames.

13. A device according to claim 11 or 12 wherein said signal sample blocks in a respective input signal frame overlap one another by N/2 samples,

said first filterbank is of length 3N/2 and has aliasing characteristics such that, in response to said first filtered signals, a complementary first synthesis filterbank generates a recovered start block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a first N/2 samples,

said second filterbank is of length N and has aliasing characteristics such that, in response to said second filtered signals, a complementary second synthesis filterbank generates one or more recovered interim blocks of signal samples each having aliasing components, and

said third filterbank is of length 3N/2 and has aliasing characteristics such that, in response to said third filtered signals, a complementary third synthesis filterbank generates a recovered end block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a last N/2 samples.

14. A device according to claim 13 wherein

said first filterbank is implemented by a first transform and a first analysis window function of length 3N/2 samples, wherein said first window function has a first portion of length N/2 and a second portion of length N/2 that are substantially not equal to zero, and a third portion of length N/2 that is substantially equal to zero,

said second filterbank is implemented by a second transform and a second analysis window function of length N samples, wherein said second window unction has a first portion of length N/2 samples and a second portion of length N/2 samples that is substantially not equal to zero, and

said third filterbank is implemented by a third transform and a third analysis window function of length 3N/2 samples, wherein said third window function has a first portion of length N/2 that is substantially equal to zero, and a second portion of length N/2 and a third portion of length N/2 that are substantially not equal to zero.

15. A device according to claim 14 wherein the first and second portions of said first analysis window function have been derived from a Kaiser-Bessel window function, said second analysis window function has been derived from said Kaiser-Bessel window function, and the second and third portions of said third analysis window function have been derived from said Kaiser-Bessel window function.

12. A device for signal processing comprising:

means for receiving an input signal comprising a sequence of frames, a respective input signal frame comprising a start block of signal samples, one or more interim blocks of signal samples and an end block of signal samples, said blocks of signal samples representing audio information,

means for generating, in response to a respective input signal frame, a first filtered signal block by applying a first filterbank to said start block of signal samples, one or more second filtered signal blocks by applying a second filterbank to said one or more interim blocks of signal samples, and a third filtered signal block by applying a third filterbank to said end block of signal samples, wherein said second filterbank generates said second filtered signal blocks having aliasing artifacts, said first filterbank generates said first filtered signal block having aliasing artifacts that cancel aliasing artifacts in a respective second filtered signal block but having substantially no other aliasing artifacts, and said third filterbank generates said third filtered signal block having aliasing artifacts that cancel aliasing artifacts in a respective second filtered signal block but having substantially no other aliasing artifacts, and

means for generating an output signal suitable for transmission or storage by assembling said first filtered signal block, said one or more second filtered signal blocks and said third filtered signal block into a respective output signal frame, whereby a sequence of output signal frames is generated by assembling first, second and third filtered signal blocks generated in response to said plurality of input signal frames.

16. A device for signal processing comprising:

means for receiving an input signal comprising a sequence of frames, a respective input signal frame comprising first filtered signals, second filtered signals and third filtered signals,

means for generating, in response to a respective input signal frame, a start block of signal samples by applying a first synthesis filterbank to said first filtered signals, one or more interim blocks of signal samples by applying a second synthesis filterbank to said second filtered signals, and an end block of signal samples by applying a third synthesis filterbank to said third filtered signals, wherein said first, second and third synthesis filterbanks each have a respective length, the respective lengths of said first and second synthesis filterbanks being unequal and the respective lengths of said second and third synthesis filterbanks being unequal, and

means for generating a sequence of output signal frames, a respective output signal frame comprising a plurality of signal sample blocks, wherein a respective frame comprises said start block, said one or more interim blocks and said end block, wherein said signal sample blocks represents audio information.

18. A device according to claim 16 or 17 wherein said signal sample blocks in a respective output signal frame overlap one another by N/2 samples,

said first synthesis filterbank is of length 3N/2 generates said start block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a first N/2 samples,

said second synthesis filterbank is of length N and generates said one or more interim blocks of signal samples each having aliasing components, and

said third synthesis filterbank is of length 3N/2 and generates said end block of signal samples having aliasing components in a middle N/2 samples and having no aliasing components in a last N/2 samples.

19. A device according to claim 18 wherein

said first synthesis filterbank is implemented by a first transform and a first synthesis window function of length 3N/2 samples, wherein said first window function has a first portion of length N/2 and a second portion of length N/2 that are substantially not equal to zero, and a third portion of length N/2 that is substantially equal to zero,

said second synthesis filterbank is implemented by a second transform and a second synthesis window function of length N samples, wherein said second window function has a first portion of length N/2 samples and a second portion of length N/2 samples that is substantially not equal to zero, and

said third synthesis filterbank is implemented by a third transform and a third synthesis window function of length 3N/2 samples, wherein said third window function has a first portion of length N/2 that is substantially equal to zero, and a second portion of length N/2 and a third portion of length N/2 that are substantially not equal to zero.

20. A device according to claim 19 wherein the first and second portions of said first synthesis window function have been derived from a Kaiser-Bessel window function, said second synthesis window function has been derived from said Kaiser-Bessel window function, and the second and third portions of said third synthesis window function have been derived from said Kaiser-Bessel window function.

17. A device for signal processing comprising:

means for receiving an input signal comprising a sequence of frames, a respective input signal frame comprising a first filtered signal block, one or more second filtered signal blocks and a third filtered signal block,

means for generating, in response to a respective input signal frame, a start block of signal samples by applying a first synthesis filterbank to said first filtered signal block, one or more interim blocks of signal samples by applying a second synthesis filterbank to said one or more second filtered signal blocks, and an end block of signal samples by applying a third synthesis filterbank to said third filtered signal block, wherein said second synthesis filterbank generates said interim blocks having aliasing artifacts, said first synthesis filterbank generates said start block having aliasing artifacts that cancel aliasing artifacts in a respective interim block but having substantially no other aliasing artifacts, and said third synthesis filterbank generates said end block having aliasing artifacts that cancel aliasing artifacts in a respective interim block but having substantially no other aliasing artifacts, and

means for generating a sequence of output signal frames, a respective output signal frame comprising a plurality of signal sample blocks, wherein a respective frame comprises said start block, said one or more interim blocks and said end block, wherein said signal sample blocks represents audio information.