Techniques for measurement of perceptual audio quality

• US 7,548,855 B2
• Filed: 06/26/2006
• Issued: 06/16/2009
• Est. Priority Date: 12/14/2001
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Lee, Ming-Chieh, Thumpudi, Naveen, Chen, Wei-Ge
• Primary Examiner(s): Armstrong; Angela A
• Application Number:
  US11/475,301
• Publication Number:
  US 20060241941A1
• Time in Patent Office: 1,086 Days
• Field of Search: 704/200.1, 704/230, 375/240
• US Class Current: 704/230
• CPC Class Codes: G10L 25/69 for evaluating synthetic or decoded voice signals

Abstract

An audio processing tool measures the quality of reconstructed audio data. For example, an audio encoder measures the quality of a block of reconstructed frequency coefficient data in a quantization loop. The invention includes several techniques and tools, which can be used in combination or separately. First, before measuring quality, the tool normalizes the block to account for variation in block sizes. Second, for the quality measurement, the tool processes the reconstructed data by critical bands, which can differ from the quantization bands used to compress the data. Third, the tool accounts for the masking effect of the reconstructed data, not just the masking effect of the original data. Fourth, the tool band weights the quality measurement, which can be used to account for noise substitution or band truncation. Finally, the tool changes quality measurement techniques depending on the channel coding mode.

Claims

1. A storage medium having stored therein computer-executable instructions for causing a computer programmed thereby to perform a method of encoding audio, the method comprising:

encoding audio organized as plural blocks of audio data, including measuring quality of the plural blocks of audio data, wherein each of the plural blocks has one of plural available block sizes, and wherein the measuring comprises, for each of the plural blocks of audio data;

normalizing the block, including one or more of;

normalizing amplitude scale of plural values in the block to compensate for changes in the amplitude scale relating to block size of the block; and

normalizing the block size of the block to compensate for block size variation among the plural blocks of audio data; and

computing a quality measure for the normalized block;

and outputting the encoded audio in a bit stream.

2. The storage medium of claim 1, wherein the plural blocks of audio data comprise plural transform blocks of spectral data.

3. The storage medium of claim 2 wherein the measuring further comprises:

before the computing, processing the normalized transform block according to an auditory model that includes temporal smearing.

4. The storage medium of claim 2 wherein the measuring further comprises:

before the computing, processing the normalized transform block as plural critical bands according to an auditory model, thereby normalizing band scale.

5. The storage medium of claim 1 wherein the normalizing the block size of the block includes normalizing to a standard size.

6. The storage medium of claim 5 wherein the standard size is a largest block size of the plural available block sizes.

7. The storage medium of claim 1 wherein the normalizing amplitude scale of the plural values in the block uses a scaling factor that is based at least in part on the block size of the block.

8. The storage medium of claim 1 wherein the normalizing the block size of the block includes:

computing ratio of a maximum block size of the plural available block sizes to the block size of the block; and

setting at least some values in the normalized block based at least in part on the ratio.

9. The storage medium of claim 1 wherein the normalizing the block size of the block includes, for each value in the block, repeating the value by an expansion factor in the normalized block, wherein the expansion factor is proportional to ratio of maximum block size to the block size of the block.

10. An audio encoder comprising:

one or more processors;

memory;

at least one input device, output device or communication connection; and

one or more storage media storing computer-executable instructions for causing the audio encoder to perform a method comprising;

encoding audio, including;

using a frequency transformer to transform a time domain block of audio samples into a transform block of frequency coefficients, wherein the transform block has a transform block size selected from among plural available transform block sizes; and

using a program module to normalize the transform block, wherein the normalizing the transform block comprises;

normalizing amplitude scale of plural coefficient values in the transform block to compensate for changes in the amplitude scale relating to the transform block size of the transform block; and

normalizing the transform block size of the transform block to compensate for transform block size variation; and

outputting the encoded audio in a bit stream.

11. The audio encoder of claim 10 wherein the encoding further includes:

using a measurer to compute a quality measure for the normalized transform block.

12. The audio encoder of claim 10 wherein the normalizing the transform block size includes normalizing to a standard size, and wherein the normalizing the amplitude scale of the coefficient values of the block uses a scaling factor that is based at least in part on the transform block size of the transform block.

13. The audio encoder of claim 12 wherein the standard size is a largest transform block size of the plural available transform block sizes.

14. The audio encoder of claim 10 wherein the frequency transformer applies a modulated lapped transform.

15. The audio encoder of claim 10 wherein the encoding further includes:

using a modeler to process the normalized transform block according to an auditory model that includes temporal smearing.

16. The audio encoder of claim 10 wherein the normalizing the transform block size comprises for each frequency coefficient in the transform block, repeating the frequency coefficient by an expansion factor in the normalized transform block, wherein the expansion factor is proportional to ratio of maximum transform block size to the transform block size of the transform block.

17. An audio encoder comprising:

one or more processors;

memory;

at least one input device, output device or communication connection; and

one or more storage media storing computer-executable instructions for causing the audio encoder to perform a method comprising;

encoding audio, including;

using a frequency transformer to transform a time domain block of audio samples into a transform block of frequency coefficients, wherein the transform block has a transform block size selected from among plural available transform block sizes;

using a program module to normalize the transform block, wherein the normalizing comprises for each frequency coefficient in the transform block, repeating the frequency coefficient by an expansion factor in the normalized transform block, wherein the expansion factor is proportional to ratio of maximum transform block size to the transform block size of the transform block; and

outputting the encoded audio in a bit stream.

18. In an audio encoder, a computer-implemented method comprising:

encoding audio organized as plural blocks of audio data, wherein the encoding includes measuring quality of the plural blocks of audio data, wherein each of the plural blocks has one of plural available block sizes, and wherein the measuring quality comprises, for each of the plural blocks of audio data;

normalizing the block, including one or more of;

normalizing amplitude scale of plural values in the block to compensate for changes in the amplitude scale relating to block size of the block; and

normalizing the block size of the block to compensate for block size variation among the plural blocks of audio data; and

computing a quality measure for the normalized block; and

outputting the encoded audio in a bit stream.

19. The method of claim 18 wherein the plural blocks of audio data comprise plural transform blocks of spectral data.

20. The method of claim 19 wherein the measuring quality further comprises:

before the computing, processing the normalized transform block according to an auditory model that includes temporal smearing.

21. The method of claim 19 wherein the measuring quality further comprises:

before the computing, processing the normalized transform block as plural critical bands according to an auditory model, thereby normalizing band scale.

22. The method of claim 18 wherein the normalizing the block size of the block includes normalizing to a standard size.

23. The method of claim 22 wherein the standard size is a largest block size of the plural available block sizes.

24. The method of claim 18 wherein the normalizing amplitude scale of the plural values in the block uses a scaling factor that is based at least in part on the block size of the block.

25. The method of claim 18 wherein the normalizing the block size of the block includes:

computing ratio of a maximum block size of the plural available block sizes to the block size of the block; and

setting at least some values in the normalized block based at least in part on the ratio.

26. The method of claim 18 wherein the normalizing the block size of the block includes, for each value in the block, repeating the value by an expansion factor in the normalized block, wherein the expansion factor is proportional to ratio of maximum block size to the block size of the block.