TH2501002000A - Quantization of non-constant parameters for advanced connectivity. - Google Patents

Abstract

DEPCT68 This disclosure provides methods, devices, and computer software products for the quantization of non-constant parameters associated with parameter-dependent spatial audio encoding. This disclosure further relates to methods and toolkits for restoring audio objects in audio decoding systems that consider non-constantly quantized parameters. According to the disclosure, such manipulation makes it possible to reduce bit usage without significantly degrading the quality of the restored audio object.

Claims (18)

DEPCT68 1. Methods in audio encoders for quantization of parameters related to parameter-dependent spatial audio encoding include: taking at least the first and second parameters to be quantized; quantizing the first parameter based on a scalar first-map quantization scheme with a magnitude of the first non-constant step to obtain the quantized first parameter, where the magnitude of the first non-constant step is chosen such that a smaller step is used for the range of first parameters where human hearing is most sensitive, and a larger step is used for the range of first parameters where human hearing is less sensitive; dequantizing the quantized first parameter using a scalar first-map quantization scheme to obtain a dequantized first parameter that is an approximation of the first parameter; determining a scaling factor based on the dequantized first parameter. And the quantization of the second parameter based on scaling factors and a scalar quantization plan of the second non-constant step size to obtain the quantized second parameter. 2. The method of claim 1, where the quantization procedure of the second parameter is based on a scaling factor and a second scalar quantization plan is incorporated by dividing the second parameter by the scaling factor before quantizing the second parameter according to the second scalar quantization plan. 3. The method of claim 1, where the magnitude of the second non-constant step of the plan, the second scalar plan quantization is scaled by a pre-scaling factor of the second parameter quantization. 4. The method of claim 1, where the magnitude of the second non-constant step of the second scalar quantization of the plan increases with the value of the second parameter. 5. A method of claim 1 in which the first scalar quantization plan incorporates more quantization steps than the second scalar quantization plan. 6. The method of claim 1, in which the first scalar quantization plan is constructed by offsetting, reflecting, and combining the second scalar quantization plan. 7. The method of claim 1, where the magnitude of the largest step of the first and/or second scalar quantization plan is approximately four times greater than the magnitude of the smallest step of the first and/or second scalar quantization plan. 8. A quantized audio encoder that quantizes parameters related to spatial audio signal encoding relies on parameters and consists of: a receiver that accepts at least the first and second parameters to be quantized; a first quantizer provided at the receiver end that quantizes the first parameter based on a scalar first quantization plan with a magnitude of the first non-constant step to obtain the quantized first parameter, where the magnitude of the first non-constant step is chosen such that a smaller step magnitude is used for the range of first parameters where human hearing is most sensitive, and a larger step magnitude is used for the range of first parameters where human hearing is less sensitive; and a dequantizer that takes the quantized first parameter from the first quantizer and dequantizes the quantized first parameter using the scalar first quantization plan to obtain the dequantized first parameter. The first quantizer takes the dequantized first parameter and its scaling factor; the second quantizer takes the second parameter and its scaling factor, and quantizes the second parameter based on the scaling factor and a second scalar quantization plan with a non-constant second step size to obtain the dequantized second parameter. 9. Methods in audio decoders for dequantized parameter dequantization relating to parameter-based spatial audio signal encoding include: taking at least the first and second quantized parameters; dequantizing the first quantized parameter according to a scalar quantization scheme of the first non-constant step size to obtain the first dequantized parameter, where the size of the first non-constant step is chosen such that a smaller step size is used for the range of first parameters where human hearing is most sensitive, and a larger step size is used for the range of first parameters where human hearing is less sensitive; determining a scaling factor based on the first dequantized parameter; and dequantizing the second quantized parameter based on a scaling factor. In scaling and scalar quantization plans, a second scalar plan with a non-constant step size is used to obtain a second, dequantized parameter value. 10. The method of claim 9, in which the procedure for dequantizing the second quantized parameter based on a scaling factor and a second scalar quantization plan, involves dequantizing the second quantized parameter according to the second scalar quantization plan and multiplying its result by a scaling factor. 11. The method of claim 9, where the magnitude of the second non-constant step of the second scalar quantization plan is scaled by a pre-scaling factor, involves the dequantization of the second quantized parameter. 12. The method of claim number 9, where the magnitude of the second non-constant step of the second scalar quantization of the plan increases with the value of the second parameter. 13. The method of claim 9, in which the first scalar quantization plan incorporates more quantization steps than the second scalar quantization plan. 14. The method of claim 9, in which the first scalar quantization plan is constructed by offsetting, reflecting, and combining the second scalar quantization plan. 15. The method of claim 9, where the magnitude of the largest step of the first and/or second scalar quantization plot is approximately four times greater than the magnitude of the smallest step of the first and/or second scalar quantization plot. 16. The method of claim 1 where at least one of the magnitudes of the first non-constant step or the magnitude of the second non-constant step is in the range from 0.0594 to 0.1906. 17. An audio encoder of claim number 8 in which at least one of the magnitudes of the first non-constant step or the magnitude of the second non-constant step is in the range from 0.0594 to 0.1906. 18. The method of claim 9 where at least one of the magnitudes of the first non-constant step or the magnitude of the second non-constant step is in the range from 0.0594 to 0.1906.