Obviousness

To analyze the obviousness of US patent 6795805 under 35 U.S.C. § 103, we must consider whether the differences between the claimed invention and the prior art would have been obvious at the time the invention was made to a person having ordinary skill in the art. This involves determining the scope and content of the prior art, ascertaining the differences between the claimed invention and the prior art, and resolving the level of ordinary skill in the pertinent art. A conclusion of obviousness must be supported by articulated reasoning with a rational underpinning.

The patent US6795805B1, titled "Periodicity enhancement in decoding wideband signals," aims to enhance the periodicity of an excitation signal in a wideband speech decoder. This is achieved by filtering an innovative codevector with an innovation filter, reducing low-frequency energy of the innovative codevector and enhancing low-frequency periodicity of the excitation signal.

A person having ordinary skill in the art (POSITA) in this field would likely have a strong understanding of digital signal processing, speech coding techniques (particularly CELP), and psychoacoustics. They would be familiar with common filters and their effects on frequency components, as well as methods for analyzing and synthesizing speech.

Prior Art References and their Teachings:

The patent itself acknowledges existing techniques for periodicity enhancement:

• Prior Art Periodicity Enhancement: The patent states that "Enhancing the periodicity of the excitation signal improves the quality in case of voiced segments." It further notes that in the past, this was done by "filtering the innovative codevector from the fixed codebook through a filter having a transfer function of the form 1/(1−εbz−T) where ε is a factor below 0.5 which controls the amount of introduced periodicity." However, it critiques this approach for wideband signals, stating it is "less efficient... since it introduces periodicity over the entire spectrum." This implies that a POSITA would be aware of and seeking improvements upon this existing method.

The patent also references several U.S. Patents by Adoul et al. (U.S. Pat. No. 5,444,816; U.S. Pat. No. 5,699,482; U.S. Pat. No. 5,754,976; and U.S. Pat. No. 5,701,392) in the context of the innovative codebook search, specifically mentioning the use of an algebraic codebook. These patents, therefore, establish the state of the art for innovative codebook generation within CELP systems.

Furthermore, the patent cites "B. S. Atal and M. R. Schroeder in “Predictive coding of speech and subjective error criteria”, IEEE Transaction ASSP, vol. 27, no. 3, pp. 247-254, June 1979" when discussing perceptual weighting in analysis-by-synthesis encoders. This work, and related publications such as "Adaptive Predictive Coding of - Speech Signals" by Atal and Schroeder, are foundational to Code-Excited Linear Prediction (CELP) and the concept of exploiting the masking properties of the human ear to shape quantization error.

Obviousness Analysis:

The core invention of US6795805B1 lies in calculating a periodicity factor and using it to filter the innovative codevector such that low-frequency energy is reduced, and low-frequency periodicity of the excitation signal is enhanced. This is specifically applied to wideband signals, addressing a perceived inefficiency of prior art methods that applied periodicity enhancement across the entire spectrum.

Consider a combination of:

• Prior Art Periodicity Enhancement (as described in US6795805B1's background): This reference teaches the general concept of enhancing periodicity in voiced segments by filtering the innovative codevector using a filter like 1/(1−εbz−T).
• B. S. Atal and M. R. Schroeder, "Predictive coding of speech and subjective error criteria" and related CELP literature: This body of work establishes the principles of CELP coding, including the use of adaptive and innovative codebooks, perceptual weighting, and the goal of optimizing speech quality. It also highlights the importance of understanding the human auditory system and its masking properties for efficient speech coding.
• General knowledge of digital signal processing and filters: A POSITA would be well aware of various types of filters (e.g., low-pass, high-pass, band-pass) and their frequency-shaping characteristics. They would also understand how to adapt filter coefficients based on signal characteristics.

Motivation to Combine:

A POSITA, confronted with the acknowledged inefficiency of prior art periodicity enhancement (which applies uniformly across the spectrum) when dealing with wideband signals, would be motivated to find a more spectrally selective approach. The goal is to improve the quality of wideband voiced segments by more effectively enhancing periodicity.

The prior art of Atal and Schroeder teaches the importance of subjective error criteria and exploiting the masking properties of the human ear, implying that perceived quality improvements are paramount. Given that wideband signals have a broader frequency range (50-7000 Hz compared to 200-3400 Hz for telephone band), a POSITA would recognize that uniformly enhancing periodicity across the entire wideband spectrum might not be optimal or perceptually desirable. The stated problem in US6795805B1 is that the prior art "introduces periodicity over the entire spectrum," which is "less efficient in case of wideband signals." This explicitly points to a problem a POSITA would seek to overcome.

Therefore, a POSITA would be motivated to combine the concept of periodicity enhancement (as known in the prior art) with their understanding of frequency-selective filtering and the perceptual considerations for wideband speech. The idea of selectively enhancing periodicity in certain frequency bands (e.g., low frequencies, where fundamental pitch harmonics are strongest) to achieve a more efficient and perceptually pleasing wideband signal would be a logical next step.

Specific Obviousness Argument for Independent Claims:

• Claim 1 (Device for enhancing periodicity) & Claim 11 (Method for enhancing periodicity): These claims center on calculating a periodicity factor and filtering the innovative codevector to "reduce energy of a low frequency portion of the innovative codevector and enhance periodicity of a low frequency portion of the excitation signal."

  • Motivation: Given the prior art's general teaching of periodicity enhancement and its noted inefficiency for wideband signals (due to non-selective application across the spectrum), a POSITA would seek to make this enhancement frequency-dependent. Knowing that the perception of periodicity (pitch) is primarily linked to lower frequencies, a POSITA would be motivated to focus the enhancement there.
  • Means: The prior art provides the basis for deriving a "periodicity factor" (e.g., related to pitch gain 'b' or energy ratios as described in US6795805B1 itself). The general knowledge of filtering techniques would suggest using a filter whose characteristics are controlled by this factor to achieve the desired frequency-selective effect (i.e., reducing low-frequency energy of the innovative codevector to allow the adaptive codebook's periodicity to dominate, thus enhancing periodicity in the low-frequency portion of the combined excitation). The specific filter forms (F(z)=−αz+1−αz−1 or F(z)=1−σz−1) might represent optimization, but the underlying concept of frequency-dependent filtering based on a periodicity measure to selectively enhance periodicity at lower frequencies would be obvious.

• Claim 21 (Decoder for producing a synthesized wideband speech signal) & Claim 31 (Improvement in a decoder for producing a synthesized wideband speech signal): These claims integrate the periodicity enhancing device/method into a broader CELP decoder architecture.

  • Motivation: Once the concept of spectrally-selective periodicity enhancement for wideband signals is conceived (as argued above for Claims 1 and 11), a POSITA would be motivated to incorporate this improvement into existing CELP decoder designs to achieve the noted benefits in wideband speech quality. The existing CELP framework, as taught by Adoul et al. and Atal and Schroeder, already includes components like signal fragmenting devices, pitch and innovative codebooks, combiner circuits, and signal synthesis filters.
  • Means: Integrating a periodicity enhancing device (as defined in claim 1) into a standard CELP decoder would be a straightforward engineering task for a POSITA. The placement of the innovation filter before the combiner circuit (as specified in Claim 31) is a logical design choice to affect the innovative codevector before it contributes to the overall excitation signal. The existing decoder components would function in their established roles, with the added periodicity enhancement improving the overall synthesized wideband speech quality.

In summary, the motivation to combine known periodicity enhancement techniques with frequency-selective processing, guided by an understanding of wideband speech characteristics and perceptual considerations, would lead a POSITA to the claimed invention. The components and their general functions would be known in the art, and their combination to address the identified inefficiency in wideband periodicity enhancement would be predictable.