Transient detection with hangover indicator for encoding an audio signal

• US 10,311,883 B2
• Filed: 10/18/2016
• Issued: 06/04/2019
• Est. Priority Date: 08/27/2007
• Status: Active Grant

• Current Assignee: Telefonaktiebolaget LM Ericsson
• Original Assignee: Telefonaktiebolaget LM Ericsson
• Inventors: Taleb, Anisse, Ullberg, Gustaf
• Primary Examiner(s): Lerner, Martin
• Application Number:
  US15/296,600
• Publication Number:
  US 20170040024A1
• Time in Patent Office: 959 Days
• Field of Search: 7042001, 704205, 704211, 704500, 704501, 704206
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• CPC Class Codes: G10L 19/025 Detection of transients or attacks for time/frequency resolution switching

Abstract

A transient detector (100) analyzes (110) a given frame n of the input audio signal to determine, based on audio signal characteristics of the given frame n, a transient hangover indicator for a following frame n+1, and signals (120) the determined transient hangover indicator to an associated audio encoder (10) to enable proper encoding of the following frame n+1.

Claims

1. An audio encoder apparatus comprising transient detector circuitry operating on an audio signal, wherein said transient detector circuitry is configured to:

analyze a given frame n of said audio signal to determine whether a transient is present in frame n of the audio signal, wherein frame n of the audio signal comprises at least a first block and a second block, and frame n of the audio signal is immediately followed by frame n+1 of the audio signal; and

set an istransient flag for frame n+1 to a logical value of TRUE as a result of determining that frame n is a transient frame, to enable proper encoding of said following frame n+1, wherein

the audio encoder is configured to determine whether frame n of the audio signal is a transient frame by performing a transient detector process comprising;

calculating a short term energy value for the first block of frame n (E_ST(1)), calculating a long term energy value for the first block of frame n (E_LT(1)), and determining whether E_ST (1)/E_LT(1) is greater than or equal to a threshold (ρ

), and

the audio encoder is configured such that the audio encoder determines that a transient is not present in frame n of the audio signal as a result of determining that E_ST (1)/E_LT(1) is less than ρ

, wherein

the audio encoder apparatus is further configured to;

determine whether the frame n of the audio signal is a transient frame in dependence on a predetermined window function;

scale said given frame n by said window function to produce a first scaled frame;

determine a transient indicator for said given frame n based on the first scaled frame;

scale said given frame n by said window function shifted one frame forward in time to produce a second scaled frame; and

determine a transient hangover indicator for said following frame n+1 based on the second scaled frame.

2. An audio encoding method, the method comprising:

analyzing a given frame n of an audio signal comprising a plurality of frames including frame n and a frame n+1, wherein frame n of the audio signal comprises at least a first block and a second block, and frame n of the audio signal is immediately followed by frame n+1 of the audio signal, to determine whether a transient hangover condition is satisfied for frame n+1; and

triggering a transient for frame n+1 as a result of determining that the transient hangover condition for frame n+1 is satisfied, thereby enabling proper encoding of frame n+1, wherein

determining that the transient hangover condition is satisfied comprises determining whether a transient is present in frame n of the audio signal,

determining whether a transient is present in frame n of the audio signal comprises performing a transient detector process comprising;

i) calculating a short term energy value for the first block of frame n (E(1)),

ii) calculating a long term energy value for the first block of frame n (E_LT(1)), and

iii) determining whether a ratio of E(1) to E_LT(1) satisfies a first condition, and

a transient is determined not to be present in frame n of the audio signal as a result of determining that the ratio of E(1) to E_LT(1) does not satisfy the first condition.

3. The method of claim 2, wherein determining whether the transient hangover condition is satisfied in comprises determining whether audio signal characteristics representative of a transient in said given frame n is not suppressed after a windowing operation based on a window function.

4. The method of claim 3, wherein said window function corresponds to a window function used for transform coding of frame n of said audio signal in said audio encoder apparatus, but shifted one frame forward in time.

5. The method of claim 4, further comprising encoding the audio signal using a lapped transform.

6. The method of claim 3, further comprising:

scaling said given frame n by said window function to produce a first scaled frame;

determining a transient indicator for said given frame n based on the first scaled frame;

scaling said given frame n by said window function shifted one frame forward in time to produce a second scaled frame; and

determining a transient hangover indicator for said following frame n+1 based on the second scaled frame.

7. The method of claim 2, wherein determining that the transient hangover condition is satisfied further comprises determining a location of the transient in said given frame n.

8. The method of claim 7, wherein determining that the transient hangover condition is satisfied further comprises determining that the transient in said given frame n is located at the center or end of frame n.

9. The method of claim 8, wherein determining that the transient hangover condition is satisfied further comprises determining whether a transient that is present in frame n is located at the beginning of frame n.

10. The method of claim 2, further comprising encoding frame n+1 based on the triggering of the transient for frame n+1.