EP0376553A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Perceptual coding of audio signals
Assignee: AT&T Inc
•
Filing: 1989-12-14
•
Granted: —
•
Expires: 2009-12-14
Abstract
Abstract A technique for the masking of quantizing noise in the coding of audio signals is adapted to include geometric interpolation between the thresholds for a tone masking noise and for noise masking a tone, in order to reduce use of bit-rate capability where it is not necessary for transparent or high quality. The technique is usable with the types of channel coding known as "noiseless" or Huffman coding and with variable radix packing. The stereophonic embodiment eliminates redundancies in the sum and difference signals, so that the stereo coding uses significantly less than twice the bit rate of the comparable monaural signal. The technique can be used both in transmission of signals and in recording for reproduction, particularly recording and reproduction of music. Compatibility with the ISDN transmission rates known as 1B, 2B and 3B rates has been achieved.
Inventors
- 👤 Joseph Lindley Hall Ii
- 👤 James David Johnston
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-04 14:24
CPC Classifications
H; H03; H03M; H03M1/00; H03M1/12; H04; H04S; H04S1/00; H04S1/007; H03M7/00; H03M7/30; H03M7/40; H03M7/42; H04B; H04B1/00; H04B1/66; H04B1/665
Claims Analysis
Total Claims: 10
Claim 1
Independent Claim
153 words
1 . A method for the coding of audio signals, of the type comprising the steps of receiving an audio signal; analyzing said signal into components for which the perceptual properties of human hearing can be determined; determining a representation of a noise threshold for each of said components; quantizing that portion of each component which exceeds its noise threshold representation; and coding each quantized portion of each component and the representations of said thresholds, whereby the coded components and threshold representation may be utilized, SAID METHOD BEING CHARACTERIZED IN THAT the determining step includes taking into account the mixed tone and noise qualities of each of the components so that said threshold representation more accurately represents the level below which added noise would be just imperceptible; and the quantizing step includes calculating a bit rate such that the resulting quantizing noise is less than, but an appreciable fraction of, said threshold representations.
Claim 2
Dependent Claim
153 words
1 . A method for the coding of audio signals, of the type comprising the steps of receiving an audio signal; analyzing said signal into components for which the perceptual properties of human hearing can be determined; determining a representation of a noise threshold for each of said components; quantizing that portion of each component which exceeds its noise threshold representation; and coding each quantized portion of each component and the representations of said thresholds, whereby the coded components and threshold representation may be utilized, SAID METHOD BEING CHARACTERIZED IN THAT the determining step includes taking into account the mixed tone and noise qualities of each of the components so that said threshold representation more accurately represents the level below which added noise would be just imperceptible; and the quantizing step includes calculating a bit rate such that the resulting quantizing noise is less than, but an appreciable fraction of, said threshold representations.
Claim 3
Dependent Claim
46 words
2 . A method of the type claimed in claim 1, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the quantizing step includes selectively adjusting the bit rate and the quantizing levels permitted by the bit rate such that the bit rate is kept within predetermined bounds.
Claim 4
Dependent Claim
125 words
3 . A method of the type claimed in claim 2, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the analyzing step comprises performing an invertible frequency analysis to produce frequency bands appropriate for a Bark power spectrum; and the threshold determining step includes creating the Bark power spectrum from the analyzed signal; convolving the Bark power spectrum with a spreading function representing mutual masking properties of signal components not within the same frequency band; renormalizing the convolved Bark power spectrum; interpolating geometrically between threshold offsets for tones masking noise and noise masking tones to create a composite threshold offset; and combining each portion of the renormalized Bark power spectrum and its composite threshold offset to yield the threshold representation for each portion of said spectrum.
Claim 5
Dependent Claim
44 words
4 . A method of the type claimed in claim 3, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the selectively adjusting step comprises changing the lowest bit rates to a bit rate at least equal to the highest bound predetermined for the bit rate.
Claim 6
Dependent Claim
27 words
5 . A method of the type claimed in claim 4, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the coding step includes employing entropy-type coding with compression.
Claim 7
Dependent Claim
71 words
6 . A method of the type claimed in any of claims 1, 4 or 5, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the receiving and analyzing steps include receiving two stereophonic audio signals and the analyzing step includes identifying redundancies in the two signals; and the determining and quantizing steps are combined for the two signals to yield a lower bit rate for said redundant portions without perceptible quantizing noise.
Claim 8
Dependent Claim
68 words
7 . A method of the type claimed in any of claims 1, 4 or 5, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the receiving and analyzing steps include receiving left and right stereophonic audio signals and analyzing signals that are derived from the sum and difference of said left and right signals; and the threshold determining step acts upon the summed power spectra of the analyzed signals.
Claim 9
Dependent Claim
85 words
8 . A method of the type claimed in claim 7, SAID METHOD BEING FURTHER CHARACTERIZED BY generating a statistical measure of each of a plurality of groups of the analyzed components of the sum and difference signals; and quantizing each statistical measure; the coding step including selecting a set of entropy-type codes suitable for the encoding of each group of the quantized components of the sum and difference signals in response to each quantized statistical measure, and coding an identification for the selected set.
Claim 10
Dependent Claim
70 words
9 . A method of the type claimed in claim 8, SAID METHOD BEING FURTHER CHARACTERIZED IN THAT the steps of selecting a set of entropy-type codes and coding an identification therefore are performed for related ones of the respective sum and difference signal set identifications as a pair, said identifications being related in that the component groups from which they are derived are related, whereby coding bits are saved.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.