Quantification method and device, storage medium and program product

Abstract

A method, apparatus, storage medium and program product for quantification. In the method, a sampling device generates a plurality of quantization results with different precisions based on a signal to be quantized. Furthermore, the sampling device generates quantized data of the signal to be quantized based on the plurality of quantization results. Therefore, according to the embodiment of the invention, a plurality of quantization results with different precisions can be generated through the uniform quantization ADC, and non-uniform quantization is realized through processing the plurality of quantization results. Compared with a scheme based on nonlinear transformation and inverse transformation, the embodiment of the invention can reduce distortion and reduce implementation complexity.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application generally relate to the field of signal processing, and more particularly to a method, an apparatus, a computer readable storage medium and a computer program product for quantization. BACKGROUND

[0002] Analog to digital converter (ADC) is widely used in scenarios such as digital communication, and in some scenarios, a non-uniform quantization method is used. It is necessary to optimize the non-uniform quantization ADC. SUMMARY

[0003] Embodiments of the present application provide a technical solution for quantization, which can not use nonlinear transformation and inverse transformation, but can realize non-uniform quantization ADC through multiple uniform quantization ADCs with different precisions, and can be applied to various scenarios such as digital communication, audio processing, video processing, underwater acoustic communication, etc.

[0004] In a first aspect, a quantization method is provided. The execution subject of the method can be a quantization apparatus or a chip applied to the quantization apparatus. Hereinafter, the execution subject is taken as an example to be described. In the method, the quantization apparatus generates multiple quantization results with different precisions based on a to-be-quantized signal. Then, the quantization apparatus generates quantized data of the to-be-quantized signal based on the multiple quantization results. In this way, multiple quantization results with different precisions can be generated through uniform quantization ADC, and non-uniform quantization can be realized through processing of the multiple quantization results. Compared with nonlinear transformation and inverse transformation, the method can reduce distortion and reduce implementation complexity.

[0005] In some implementations, the quantization apparatus generates multiple quantization results with different precisions in the following manner. The quantization apparatus generates multiple copy signals of the to-be-quantized signal. Then, the quantization apparatus applies different multiple gains to the multiple copy signals to generate multiple amplified signals. Then, the quantization apparatus quantizes the multiple amplified signals to generate multiple quantized signals. Then, the quantization apparatus generates multiple quantization results based on the multiple quantized signals. In this way, multiple quantization signals with different precisions can be generated by using different multiple gains, thereby facilitating the realization of non-uniform quantization.

[0006] In some implementations, the quantization of the multiple amplified signals by the quantization apparatus includes that the quantization apparatus quantizes the multiple amplified signals based on the same analog to digital converter (ADC) in time division. In this way, multiple quantized signals with different precisions can be generated by using one ADC, reducing the number of ADCs and saving cost and power consumption.

[0007] In some implementations, a dynamic range of a first amplified signal in the plurality of amplified signals is equal to a dynamic range of the ADC, and a dynamic range of a second amplified signal in the plurality of amplified signals is greater than the dynamic range of the ADC. In this way, some of the amplified signals will overflow the dynamic range of the ADC, thereby achieving different quantization precisions for large and small signals, and achieving non-uniform quantization.

[0008] In some implementations, the quantizing, by the quantizing apparatus, the plurality of amplified signals comprises: the quantizing apparatus quantizing the plurality of amplified signals in parallel based on the plurality of ADCs. In this way, the same hardware design structure can be used for the multi-path quantization, thereby reducing hardware implementation complexity.

[0009] In some implementations, the plurality of ADCs have different precisions. In this way, in combination with the multi-path different gain amplification, multi-precision quantization can be achieved more flexibly and finely, thereby achieving non-linear quantization more flexibly and finely.

[0010] In some implementations, the plurality of ADCs have the same precision. In this way, implementation complexity of the non-linear quantization can be reduced.

[0011] In some implementations, a dynamic range of a first amplified signal in the plurality of amplified signals is equal to a dynamic range of a first ADC used to quantize the first amplified signal. And a dynamic range of a second amplified signal in the plurality of amplified signals is greater than a dynamic range of a second ADC used to quantize the second amplified signal. In this way, some of the amplified signals will overflow the dynamic range of the corresponding ADC, thereby achieving different quantization precisions for large and small signals, and achieving non-uniform quantization.

[0012] In some implementations, the same ADC or the plurality of ADCs comprises an n-bit quantization ADC, where n is an integer. In this way, by selecting an ADC with an appropriate number of quantization bits, the quantization precision can be flexibly controlled.

[0013] In some embodiments, the quantization apparatus generates the plurality of quantization results by: if the quantization apparatus determines that an amplified amplitude corresponding to a first sampling point of the signal to be quantized in a first amplified signal of the plurality of amplified signals does not exceed a dynamic range of an ADC used to quantize the first amplified signal, generating a first quantization result corresponding to the first amplified signal based on the first sampling point; and if the quantization apparatus determines that an amplified amplitude corresponding to a second sampling point of the signal to be quantized in a second amplified signal of the plurality of amplified signals exceeds the dynamic range of the ADC used to quantize the second amplified signal, not generating a second quantization result corresponding to the second amplified signal based on the second sampling point. In this way, under the condition that the amplitude of the signal to be quantized is large and some of the ADCs overflow, the quantization result exceeding the dynamic range of the ADC can be discarded, and only the quantization result not exceeding the dynamic range of the ADC is used, so that a large quantization step is achieved under the condition of a large signal. Under the condition that the amplitude of the signal to be quantized is small and none of the ADCs overflow, all the quantization results can be used, so that a small quantization step is achieved under the condition of a small signal. Thus, non-uniform quantization is achieved.

[0014] In some embodiments, the quantization apparatus generates the plurality of quantization results based on the plurality of quantized signals by: normalizing, by the quantization apparatus, the plurality of quantized signals based on the plurality of normalization factors to generate a plurality of normalized signals; and synchronizing, by the quantization apparatus, the plurality of normalized signals to generate the plurality of quantization results. In this way, the influence of gain amplification is compensated by normalization processing, and different time delays in different precision quantization are compensated by synchronization, so that non-uniform quantization is accurately achieved.

[0015] In some embodiments, the plurality of normalization factors are equal to inverses of the plurality of gains, respectively. In this way, the influence of gain amplification can be accurately compensated, and non-uniform quantization can be accurately achieved.

[0016] In some embodiments, the quantization apparatus generates the quantized data of the signal to be quantized by: determining, by the quantization apparatus, a plurality of quantization intervals corresponding to a plurality of intermediate quantization values corresponding to sampling points of the signal to be quantized in the plurality of quantization results; and determining, by the quantization apparatus, a quantization value corresponding to the sampling points in the quantized data based on the plurality of quantization intervals. In this way, the plurality of quantization intervals of the plurality of quantizations are processed, and the overall non-uniform quantization is achieved by the way of uniform quantization of the plurality of amplified signals. Compared with the scheme based on nonlinear transformation and inverse transformation, the quantization apparatus can reduce distortion and reduce complexity.

[0017] In some embodiments, the quantization apparatus determines the quantization value corresponding to the sampling points in the quantized data by:

[0018] If the quantization device determines that the multiple quantization intervals have an intersection, the quantization value is determined based on the intersection. Alternatively or additionally, if the quantization device determines that the multiple quantization intervals do not have an intersection, the quantization value is determined based on a mean value of the multiple quantization intervals. In this way, the cases where the multiple quantization intervals have an intersection and do not have an intersection can be flexibly handled, so that accurate non-uniform quantization is performed, and the precision is improved.

[0019] In some implementations, the quantization device determining the quantization value based on the intersection includes the quantization device determining a midpoint value of the intersection as the quantization value. In this way, accurate non-uniform quantization can be performed, and the precision is improved.

[0020] In a second aspect, a quantization device is provided. The quantization device can be a quantization module that implements a quantization function, or can be a chip in the quantization module. The quantization device can be implemented entirely in hardware, or can be implemented in a combination of hardware, software, firmware, or other manners, which are not limited in the present disclosure. The quantization device includes a quantization result generation module configured to generate multiple quantization results with different precisions based on a signal to be quantized. The quantization device also includes a post-quantization data generation module configured to generate post-quantization data of the signal to be quantized based on the multiple quantization results. In this way, multiple quantization results with different precisions can be generated by a uniform quantization ADC, and non-uniform quantization can be achieved by processing the multiple quantization results. Compared with a scheme based on a non-linear transformation and an inverse transformation, the quantization device can reduce distortion and reduce implementation complexity.

[0021] In some implementations, the quantization result generation module includes a replication sub-module configured to generate multiple replicated signals of the signal to be quantized based on the signal to be quantized. The quantization result generation module also includes a gain amplification sub-module configured to apply different multiple gains to the multiple replicated signals to generate multiple amplified signals. The quantization result generation module also includes a quantization sub-module configured to quantize the multiple amplified signals to generate multiple quantized signals. The quantization result generation module also includes a normalization sub-module configured to normalize the multiple quantized signals based on multiple normalization factors to generate multiple normalized signals, the multiple normalization factors of the normalization sub-module respectively equaling inverses of the multiple gains. The quantization result generation module also includes a synchronization sub-module configured to synchronize the multiple normalized signals to generate the multiple quantization results. In this way, multiple quantization results with different precisions can be generated by a uniform quantization ADC, and the influence of different gains in the gain amplification sub-module can be compensated by the normalization sub-module, and different time delays caused by quantization with different precisions can be compensated by the synchronization sub-module, so that timing-synchronized and accurate quantization results are generated.

[0022] In some implementations, the quantization sub-module can include a same analog-to-digital converter (ADC) configured to quantize the multiple amplified signals in time division to generate the multiple quantized signals. In this way, the number of ADCs can be reduced, and power consumption and cost can be saved.

[0023] In some implementations, the dynamic range of the first amplified signal in the plurality of amplified signals is equal to the dynamic range of the ADC, and the dynamic range of the second amplified signal in the plurality of amplified signals is greater than the dynamic range of the ADC. In this way, some of the amplified signals will exceed the dynamic range of the ADC and cause overflow, thereby achieving different quantization precisions for large and small signals and achieving non-uniform quantization.

[0024] In some implementations, the quantization sub-module can include a plurality of ADCs for quantizing the plurality of amplified signals in parallel to generate a plurality of quantized signals. In this way, the same hardware design structure can be used for multi-channel quantization, reducing hardware implementation complexity.

[0025] In some implementations, the plurality of ADCs can have different precisions. In this way, in combination with multi-channel different gain amplification, multi-precision quantization can be more flexible and fine, thereby achieving more flexible and fine non-linear quantization.

[0026] In some implementations, the plurality of ADCs can have the same precision. In this way, the implementation complexity of non-linear quantization can be reduced.

[0027] In some implementations, the dynamic range of the first amplified signal in the plurality of amplified signals can be equal to the dynamic range of the first ADC used to quantize the first amplified signal. And the dynamic range of the second amplified signal in the plurality of amplified signals can be greater than the dynamic range of the second ADC used to quantize the second amplified signal. In this way, some of the amplified signals will exceed the dynamic range of the corresponding ADC and cause overflow, thereby achieving different quantization precisions for large and small signals and achieving non-uniform quantization.

[0028] In some implementations, the quantized data generation module includes a quantization interval determination sub-module for determining a plurality of quantization intervals corresponding to a plurality of intermediate quantization values corresponding to the sampling points of the signal to be quantized based on the plurality of intermediate quantization values in the plurality of quantization results. The quantized data generation module also includes a quantization interval processing sub-module for determining quantization values corresponding to the sampling points in the quantized data based on the plurality of quantization intervals. In this way, by processing the plurality of quantization intervals of the multi-channel quantization, the overall non-uniform quantization is achieved by the way of uniform quantization through the multi-channel amplified signals. Compared with non-uniform quantization based on non-linear transformation and inverse transformation, distortion is reduced and complexity is reduced.

[0029] In some embodiments, the quantization interval processing submodule determining the quantization value corresponding to the sampling point in the quantized data can include the following steps. If the quantization interval processing submodule determines that the multiple quantization intervals have an intersection, the quantization value is determined based on the intersection. Alternatively or additionally, if the quantization interval processing submodule determines that the multiple quantization intervals do not have an intersection, the quantization value is determined based on the mean value of the multiple quantization intervals. In this way, the cases where the multiple quantization intervals have an intersection and do not have an intersection can be flexibly processed, thereby performing accurate non-uniform quantization and improving the precision.

[0030] In some embodiments, the quantization interval processing submodule determining the quantization value based on the intersection includes the quantization interval processing submodule determining the midpoint value of the intersection as the quantization value. In this way, accurate non-uniform quantization can be performed, and the precision can be improved.

[0031] In a third aspect, a quantization device is provided. The quantization device includes the quantization apparatus in the second aspect.

[0032] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program. When the computer program is executed, the method performed by the quantization apparatus in the first aspect is implemented.

[0033] In a fifth aspect, a computer readable storage medium is provided, which stores a computer program. When the computer program is executed, the method performed by the quantization apparatus in the first aspect is implemented.

[0034] In a sixth aspect, a chip is provided, which implements the method performed by the quantization apparatus in the first aspect. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1A A sampling system in which embodiments of the present application can be implemented is shown.

[0036] Figure 1A Another sampling system in which embodiments of the present application can be implemented is shown.

[0037] Figure 2 A schematic diagram of non-uniform quantization implemented using non-linear transformation and inverse transformation is shown.

[0038] Figure 3 A flowchart of a non-uniform quantization method implemented at the quantization apparatus in embodiments of the present application is shown.

[0039] Figure 4 A schematic diagram of non-uniform quantization implemented using multi-path multi-precision simultaneous quantization in embodiments of the present application is shown.

[0040] Figure 5A schematic diagram of non-uniform quantization implemented by using multi-path multi-precision time-sharing quantization in the embodiment of the application is shown.

[0041] Figure 6 A schematic diagram of data splicing in the embodiment of the application is shown.

[0042] Figure 7 A schematic diagram of non-uniform quantization implemented by using single 3-bit ADC time-sharing quantization in the embodiment of the application is shown.

[0043] Figure 8 A schematic diagram of non-uniform quantization implemented by using two 2-bit and 3-bit ADCs for simultaneous quantization in the embodiment of the application is shown.

[0044] Figure 9 A block diagram of a sampling device in the embodiment of the application is shown.

[0045] Figure 10 A structural schematic diagram of a sampling device in the embodiment of the application is shown. DETAILED DESCRIPTION

[0046] To make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The specific operation methods, function descriptions and the like in the method embodiments can also be applied to the device embodiments or system embodiments.

[0047] As described above, an analog-to-digital converter (ADC) is a bridge for signal conversion between an analog radio frequency circuit and a digital circuit, and its performance will directly affect the performance of the entire digital system. Limited by manufacturing processes, the ADC used in a baseband digital system with a large bandwidth has low quantization precision and large quantization error, and the system performance is severely limited. For example, the bandwidth of a terahertz (THz) system can usually reach several tens of GHz, and the Nyquist sampling frequency is 2 times the bandwidth. At this time, the sampling frequency of the ADC will reach several tens of Gsps. However, with the current manufacturing process, the sampling rate and quantization precision of the ADC cannot be considered at the same time. The precision of the ADC with a sampling rate of several tens of Gsps is below 5 bits. The reduction of quantization precision leads to an exponential decline in the performance of the THz system. In the scenario of low quantization precision, non-uniform quantization is an effective solution to reduce quantization error and improve performance. However, a general ADC is a uniform quantizer. Therefore, the design of a non-uniform quantization scheme based on a uniform quantization ADC has become a research hotspot.

[0048] In the embodiments of the present disclosure, non-uniform quantization can be used in various scenarios.

[0049] Figure 1A A sampling system in which the embodiment of the present application can be implemented is shown, specifically a digital communication scenario.

[0050] In the scenario 100 of the wireless communication system, the transmitting end 110 transmits a radio frequency analog signal. The analog signal passes through the wireless channel 115 and enters the receiving end. The receiving end can be the receiving end of a network device such as a base station in the wireless communication system, a receiving end of a terminal device in the wireless communication system, a receiving end of a router, a receiving end of a vehicle-mounted communication system, a receiving end of a satellite communication system, and the like. In the receiving end, the radio frequency analog signal is received at 120, and the sampling and quantization of the analog signal are performed at the sampling device 105, and the received digital signal is obtained at 125. The wireless channel 115 can also be a wired channel, and the corresponding scenario 100 is a scenario in a wired communication system.

[0051] The wireless communication system in the embodiments of the present application includes, but is not limited to, a narrowband Internet of Things (NB-IoT) system, a Long Term Evolution (LTE) system, and three application scenarios of a 5G mobile communication system, eMBB, URLLC, and eMTC, and the like.

[0052] It should be understood that the above wireless communication system can be applied to both high frequency scenarios (above 6G) such as millimeter waves and low frequency scenarios (sub-6G). The application scenarios of the wireless communication system include, but are not limited to, existing communication systems such as the fifth generation system (5G), the new radio (NR) communication system, and the like, and the future evolved public land mobile network (PLMN) system and the like.

[0053] In the embodiments of the present application, Figure 1A The receiving radio frequency analog signal 120, the sampling device 105, and the received digital signal 125 can be used for the receiving end of a terminal device, or for the receiving end of a network device.

[0054] The above terminal device can be a user equipment (UE), a terminal, an access terminal, a terminal unit, a terminal station, a mobile station (MS), a remote station, a remote terminal, a mobile terminal, a wireless communication device, a terminal agent, or a terminal device, and the like. The terminal device can also be a communication chip with a communication module, or a vehicle with a communication function, or a vehicle-mounted device (such as a vehicle-mounted communication device, a vehicle-mounted communication chip), and the like. The terminal device 101 can have a wireless transceiver function, which can communicate (such as wireless communication) with one or more network devices of one or more communication systems and accept network services provided by the network device.

[0055] The terminal device can be a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA) device, a handheld device having wireless communication function, a computing device, or other processing device connected to a wireless modem, an in-vehicle device, a wearable device, a terminal device in a future 5G network, or a terminal device in a future evolved PLMN network, and the like.

[0056] The terminal device can be a mobile phone, a pad, a computer with wireless transceiver function, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wireless terminal in industrial control, a wireless terminal in self driving, a wireless terminal in remote medical treatment, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, and the like.

[0057] In addition, the terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; the terminal device can also be deployed on the water surface (such as a ship, etc.); the terminal device 101 can also be deployed in the air (such as an airplane, a balloon, and a satellite, etc.). The network device can be an access network device (or an access site). The access network device refers to a device having a network access function, such as a radio access network (RAN) base station, and the like. The network device can specifically include a base station (BS), or include a base station and a radio resource management device for controlling the base station, and the like. The network device can also include a relay station (relay device), an access point, a base station in a 5G network or an NR base station, a base station in a future evolved PLMN network, and the like. The network device can be a wearable device or a vehicle-mounted device. The network device (103) can also be a communication chip with a communication module.

[0058] For example, the network device includes but is not limited to: a base station (g nodeB, gNB) in 5G, an evolved node B (eNB) in a long term evolution (LTE) system, a radio network controller (RNC), a wireless controller under a cloud radio access network (CRAN) system, a base station controller (BSC), a home base station (for example, a home evolved nodeB, or a home node B, HNB), a baseband unit (BBU), a transmitting and receiving point (TRP), a transmitting point (TP), a mobile switching center, and can also be an evolved NB (eNB or eNodeB) in LTE, and can also be a base station device in a future 5G network or an access network device in a future evolved PLMN network, and can also be a wearable device or a vehicle-mounted device.

[0059] In some deployments, a network device can include a centralized unit (CU) and a distributed unit (DU). The network device can also include an active antenna unit (AAU). The CU implements part of the functionality of the network device, and the DU implements part of the functionality of the network device. For example, the CU is responsible for handling non-real-time protocols and services, implements the radio resource control (RRC), and the functionality of the packet data convergence protocol (PDCP) layer. The DU is responsible for handling the physical layer protocols and real-time services, implements the functionality of the radio link control (RLC) layer, the media access control (MAC) layer, and the physical (PHY) layer. The AAU implements part of the physical layer processing functionality, the radio frequency processing, and the related functionality of the active antenna. Since the information of the RRC layer eventually becomes the information of the PHY layer, or is transformed from the information of the PHY layer, under this architecture, high layer signaling, such as RRC layer signaling, can also be considered as being sent by the DU, or by the DU+AAU. It can be understood that the network device can be a device including one or more of a CU node, a DU node, and an AAU node. In addition, the CU can be divided into a network device in a radio access network (RAN), or can be divided into a network device in a core network (CN), which is not limited in the present application. Examples of the network device include, but are not limited to, a Node B (NodeB or NB), an evolved Node B (eNodeB or eNB), a next generation Node B (gNB), a transmission reception point (TRP), a remote radio unit (RRU), a radio head (RH), a remote radio head (RRH), an IAB node, a low power node such as a femto node, a pico node, a reconfigurable intelligent surface (RIS), a network-controlled relay, and the like.

[0060] In addition, the network device can be connected to a core network (CN) device, and the core network device can be used to provide core network services for the access network device and the terminal device. The core network device can correspond to different devices under different systems. For example, in 3G, the core network device can correspond to a serving GPRS support node (SGSN) and / or a gateway GPRS support node (GGSN) of a general packet radio service (GPRS). In 4G, the core network device can correspond to a mobility management entity (MME) and / or a serving gateway (S-GW). In 5G, the core network device can correspond to an access and mobility management function (AMF), a session management function (SMF), or a user plane function (UPF).

[0061] Figure 1B Another sampling system in which embodiments of the present application can be implemented is shown, specifically a quantization scenario for an audio signal.

[0062] In scenario 130, the analog audio signal 140 is sampled and quantized at the sampling device 135 to obtain a digital audio signal 145 for subsequent processing or transmission at 150.

[0063] In the sampling device 105 or the sampling device 135, a non-uniform quantization method can be used. Those skilled in the art can understand that, in addition to the communication scenario or audio signal processing in 100 and 130, non-uniform quantization can also be used in underwater acoustic communication, video signal processing, and other scenarios, which are not limited by the present disclosure.

[0064] One way to implement non-uniform quantization based on uniform quantization ADC is to perform nonlinear transformation on the quantized signal. The combination of the uniform quantizer and the signal nonlinear transformation is ultimately equivalent to non-uniform quantization, such as Figure 2quantization input signal 205 first undergoes a non-linear transformation 210, then the non-linear transformed signal is uniformly quantized by a uniform quantizer ADC 215, and the uniformly quantized signal undergoes an inverse non-linear transformation 220 to finally obtain a quantization output signal 225. Due to the existence of the non-linear transformation, the quantization input-output is finally equivalent to a non-uniform quantization effect. This scheme needs to consider the selection of the non-linear transformation 210 and the inverse non-linear transformation 220 and the hardware implementation. The non-linear transformation function plays an important role in the final equivalent non-uniform quantization effect, and the hardware implementation of the non-linear transformation and its inverse transformation determines the hardware implementation complexity of the scheme. Common non-linear transformations 210 are A-law and mu-law compression functions.

[0065] The non-uniform quantization in the scheme 200 is implemented by the non-linear transformation 210 and the inverse non-linear transformation 220, which has the disadvantages of high hardware implementation complexity and cost and poor accuracy. The scheme 200 needs to increase customized analog devices to implement the non-linear transformation 200 before the ADC quantization, which significantly increases the cost and complexity of the hardware implementation. Moreover, the non-linear transformation 210 and the inverse non-linear transformation 220 need to be matched with each other, which increases the difficulty of the hardware implementation. Figure 1A If the non-linear transformation is performed at the transmitting end 110, the inverse non-linear transformation is performed at the receiving end 120, and the quantization is performed at 105, the channel distortion between the transmitting end and the receiving end will cause the inverse non-linear transformation at the receiving end 120 to fail to restore the transmitted signal. In this scenario, the robustness of the non-linear transformation 210 and the inverse non-linear transformation 220 to resist channel distortion is poor.

[0066] Therefore, in view of the problems such as high ADC sampling rate, low quantization precision, and large quantization error in a large-bandwidth communication system, it is necessary to provide a non-uniform quantization scheme based on a uniform quantization ADC to reduce the quantization error and improve the system performance. It can be understood by those skilled in the art that in the fields of audio processing, video processing, underwater acoustic processing, etc., when using non-uniform quantization, using non-linear transformation and inverse transformation also has the above problems and needs to be solved.

[0067] In view of this, the present disclosure provides a multi-precision quantization splicing non-uniform quantization method, which quantizes the quantization input signal multiple times with different precisions, splices the multiple quantization results based on a criterion such as an intersection verification criterion, and realizes the non-uniform quantization effect of the quantization input signal.

[0068] Figure 3 A flowchart of a non-uniform quantization method implemented at a quantization device in an embodiment of the present application is shown. The sampling device implementing the flowchart 300 can be implemented in the sampling device 105 in Figure 1A the sampling device 135 in Figure 1B other scenarios such as video processing, underwater acoustic processing, etc., and the present disclosure does not limit this.

[0069] In the flow 300, at 310, the quantization device generates a plurality of quantization results with different precisions based on the signal to be quantized. At 320, the quantization device generates quantized data of the signal to be quantized based on the plurality of quantization results. In this way, a plurality of quantization results with different precisions can be generated by the uniform quantization ADC, and the non-uniform quantization can be realized by processing the plurality of quantization results. Compared with the scheme based on the nonlinear transformation and the inverse transformation, the flow 300 can reduce distortion and reduce implementation complexity.

[0070] Figure 4 A schematic diagram of the non-uniform quantization realized by the multi-path multi-precision simultaneous quantization in the embodiments of the present application is shown. Specifically, 400 shows a schematic diagram of the non-uniform quantization realized by a plurality of ADCs. The sampling device for realizing the embodiment 400 can be realized in the sampling device 105 in Figure 1A , or the sampling device 135 in Figure 1B , or a sampling device in other scenarios such as video processing, underwater acoustic processing, etc., and the present disclosure does not limit this.

[0071] In the embodiments of the present application, the received signal is quantized with different precisions at the receiving end, which specifically includes the following steps. The received signal output by the radio frequency module 400, i.e., the signal to be quantized 406, is decomposed into K replica signals such as 407, 408. The decomposition of the signal to be quantized 406 into the replica signals 407, 408 can be realized by a power divider, or can be realized by a circuit on a circuit board, and the present disclosure does not limit this. For the K replica signals such as 407, 408, K gain amplifiers with K different gains g1-g k are used for gain amplification, and K amplified signals such as 415, 417 are obtained. In this way, different K gains are used to facilitate the quantization of the replica signals 407, 408 derived from the signal to be quantized 406 with different quantization precisions, so as to realize different quantization steps for large signals and small signals, and realize non-uniform quantization.

[0072] In this embodiment of the present application, K ADCs, such as 420 and 423, are used to simultaneously sample and quantize K amplified signals, such as 415 and 417, to obtain K quantized signals, such as 425 and 427. The K ADCs can use different or equal precision to quantize the amplified signals 415 and 417, which have been amplified by different gain factors. Among the K amplified signals, when the amplitude of the signal to be quantized 506 is large, the amplitude of at least one amplified signal, such as 415, is equal to or reaches the dynamic range of the corresponding ADC 420. This equalization can be approximately equal, for example, within ±10% of the dynamic range, or within another range. However, if the amplitude of at least one other amplified signal, such as 417, among the K amplified signals exceeds the dynamic range of the corresponding ADC 423, resulting in overflow, the quantized signal 425 is retained, while the quantized signal 427 is discarded due to the overflow, achieving a large quantization step size for the large-amplitude signal to be quantized 506. When the amplitude of the signal to be quantized 506 is small, for example, the amplitudes of the K amplified signals 415 and 417 are all within the dynamic range of the corresponding ADCs 420 and 423, thus achieving a small quantization step size for the small-amplitude signal to be quantized 506. In this way, different quantization step sizes are achieved for signals to be quantized 506 with different amplitudes, thus achieving non-uniform quantization. Figure 2 Compared with the non-uniform quantization using nonlinear transformation and inverse transformation, the distortion is reduced and the implementation complexity is lowered.

[0073] exist Figure 4 In the embodiment shown, the 1st to the Kth ADCs use different M1 to M2. k Thus, when the quantization accuracy of K ADCs is different, the combination of g1-g k Different gains of the K ADCs can achieve accurate and flexible control of multiple quantization accuracies of the replica signals 407 and 408, thereby optimizing the quantization results. When the quantization accuracies of the K ADCs are the same, the circuit implementation can be simplified.

[0074] In the embodiment of the present application, multiple quantized signals 425 and 427 of different precisions can also be amplitude normalized and synchronously processed, and then spliced ​​to achieve overall non-uniform quantization. Specifically, for example, the gain coefficients of the K normalization modules of 430 and 433 are is the gain coefficient g1-g of the K gain amplification modules such as 410 and 413 k The reciprocal of g1-g kCompensation is performed to obtain K normalized signals, such as 435 and 437. Multiple quantized signals 425 and 427 with different quantization precisions have different delays. K synchronization processes 440 and 443, for example, using a digital synchronization algorithm, can be used to eliminate the delay differences between the quantized signals 425 and 427, resulting in K synchronized quantization results, such as 445 and 447. This increases the delay robustness of the multi-precision quantization and facilitates subsequent data splicing 450.

[0075] In an embodiment of the present application, in data splicing 450, K synchronized quantization results, such as 445 and 447, can be spliced. The quantized data 455 finally obtained has a larger quantization step when the amplitude of the signal to be quantized 406 is larger, and has a smaller quantization step when the amplitude of the signal to be quantized 406 is smaller, thereby realizing non-uniform quantization.

[0076] The specific method of using the intersection verification method to achieve data splicing 450 is as follows Figure 6 As shown in 600.

[0077] In the embodiment of the present application, Figure 4 In the multi-precision quantization process shown in FIG, each sampling point of the signal to be quantized 406 corresponds to multiple quantization intervals. Figure 4 As shown, K different gain amplifiers such as 410 and 413 are used. After quantization by K ADCs, each sampling point of the signal to be quantized 406 can generate K quantized signals such as 425 and 427, with K quantization intervals, which are expressed as When calculating the quantization interval, only the quantized signal that does not exceed the ADC dynamic range is used. The quantized signal that exceeds the ADC dynamic range is discarded and its quantization interval is not calculated.

[0078] In the embodiment of the present application, the minimum common interval can be determined by finding the intersection of the aforementioned K quantization intervals. The intersection of the K quantization intervals can be expressed as If the intersection of K quantization intervals is not an empty set, that is, As shown in 610, the final quantized splicing value of the current sampling point is the midpoint of the intersection, that is, As the output quantized data 455. If due to noise, interference, small fluctuation of the signal, etc., the intersection of the K quantization intervals is an empty set, that is, As shown in 620, the final quantized splicing value of the current sampling point is the average of the K quantization intervals, that is, the average of the upper and lower limits of the K quantization intervals. The quantized data 455 is outputted. In this way, the situations where multiple quantization intervals have intersections or do not have intersections can be handled flexibly, thereby performing accurate non-uniform quantization, improving quantization accuracy, and enhancing robustness in actual use.

[0079] Figure 5 A schematic diagram of the non-uniform quantization implemented by the multi-path multi-precision time-division quantization in the embodiments of the present application is shown. The sampling device of the implementation embodiment 500 can be implemented in the sampling device 105 in Figure 1A , or the sampling device 135 in Figure 1B , or a sampling device in other scenarios such as video processing, underwater acoustic processing, etc., which are not limited by the present disclosure.

[0080] Compared with the non-uniform quantization implementation embodiment 500 of time-division processing and the non-uniform quantization implementation embodiment 400 of multi-path parallel processing, the roles of the radio frequency 505 and the gain amplification 510, 513 are the same as the roles of the radio frequency 405 and the gain amplification 410, 413, that is, to generate a signal to be quantized and to perform gain amplification with different gains. The difference lies in that, in 500, K parallel ADCs such as 420, 423 in 400 are not used, but a time-division module 517 and a single ADC 520 are used. The time-division module 517 obtains K amplified signals such as 515, 516 in a time-division manner and inputs them into a single M-bit ADC 520 in a time-division manner. The ADC 520 samples and quantizes the K amplified signals in a time-division manner and outputs K quantized signals such as 525, 527 in a time-division manner. The normalization processing such as 530, 533 and the Figure 4 in 430, 433 have the same effect, using gain to compensate for the gain amplification 510, 513. The K synchronization algorithms such as 540, 543 perform synchronization processing on the different time delays of the time-division signals, obtain K synchronized quantization results such as 545, 547, and perform data splicing at 550. The data splicing process in 550 is the same as that in Figure 4 450, which can use the intersection verification method in Figure 6 , which is not repeated here. Through the single-ADC time-division processing mode in the embodiment 500, non-uniform quantization of the signal to be quantized 506 is also implemented. The single-ADC mode reduces the number of ADCs used, reducing cost and power consumption.

[0081] In the embodiments of the present disclosure, in Figure 4 , Figure 5 , K-ADC parallel processing and single-ADC time-division processing are respectively shown to implement non-uniform quantization, and the intersection verification method of data splicing is shown in Figure 6 . In subsequent Figure 7 , Figure 8 , embodiments using specific gain coefficients and specific quantization precision ADCs will be shown.

[0082] Figure 7 A schematic diagram of the non-uniform quantization implemented by the single 3-bit ADC time-division quantization in the embodiments of the present application is shown, and corresponds toFigure 5 Specifically, in embodiment 700, a 3-bit precision ADC is used to perform two quantizations on the two replica signals in a time-division manner, and the quantization results are spliced ​​based on the intersection verification method, thereby achieving the effect of non-uniform quantization. The sampling device for implementing embodiment 700 can be implemented in Figure 1A The sampling device 105, or Figure 1B The sampling device 135 in the embodiment, or other sampling devices in scenarios such as video processing and underwater acoustic processing, is not limited in this disclosure.

[0083] In the embodiment of the present disclosure, the sampling device uses a power divider 705 to divide the received RF signal x(t) to be quantized output by the RF module 703 into two replica signals: x1(t) 708 and x2(t) 707. The first gain amplification module 713 is used to amplify the first replica signal x1(t) 708. Times the gain adjustment to obtain the first amplified signal And make the first amplified signal The dynamic range of ADC 720 is fully utilized under large signal conditions. The specific value of may not be a fixed value, but may be variable. The second gain amplification module 710 is for the second replica signal x2(t), by Times the gain adjustment to obtain the second amplified signal And after passing through the delay device 716 with a delay value of T, the delayed second amplified signal is obtained Where T is the total duration of a sampling period of the RF signal to be quantized x(t). Under large signal conditions, the second amplified signal The second amplified signal exceeds the dynamic range of the ADC 720, causing the ADC 720 to overflow. The subsequent sampling and quantization signals are discarded, and no data splicing is performed. Only the first amplified signal is used. Under small signal conditions, the first amplified signal and the second amplified signal The quantization steps are all within the ADC dynamic range, and the subsequent quantization signals can be quantized and spliced. In this way, a large quantization step size can be achieved under large signal conditions and a small quantization step size can be achieved under small signal conditions, thus achieving non-uniform quantization as a whole.

[0084] Through the different gain amplification processing of g1 and 2g1, and the second amplified signal Delay is performed to achieve the first amplified signal and the delayed second amplified signal The 3-bit ADC720 samples and quantizes the two amplified signals in a time-division manner, thus saving cost and power consumption. After sampling and quantization, the first quantized signal 727 is normalized by the amplitude of 733 to obtain the first normalized signal 3-bit ADC720 amplifies the first After the sampling and quantization is completed, the delayed second amplified signal is After sampling and quantization, the second quantized signal 725 is normalized by the amplitude of 730 to obtain the second normalized signal In synchronization modules 740 and 743, the first normalized signal is eliminated by a digital synchronization algorithm such as the coarse synchronization and fine synchronization algorithms of the OFDM system. and the second normalized signal The delay difference between them is used to obtain synchronization signals 745 and 747, which provide good synchronization data for subsequent data splicing and improve quantization accuracy. The values ​​of synchronization signals 745 and 747 are and Finally, follow Figure 6 The dataset splicing method based on intersection verification method is used for a given n and That is, the intermediate quantization value is used to calculate the quantization interval, and the quantization interval is spliced ​​to obtain the final non-uniform quantization result 755. In this way, a single ADC can be used to perform sampling and quantization of multiple amplified signals with different precisions in a time-division manner, and different quantization step sizes under different signal amplitudes are achieved by discarding the quantization signal due to the fact that some amplified signals exceed the dynamic range of the ADC. Compared with the method based on nonlinear transformation and inverse transformation, the quantization accuracy is improved and the complexity is reduced. In addition, a single ADC reduces cost and power consumption. It can be understood by those skilled in the art that in addition to using an ADC with 2-bit quantization, any n-bit quantization can also be used, such as quantization less than 8 bits. In addition to using the 2-path copy signal processing method in embodiment 700, any K paths greater than 2 can also be used, and the present disclosure does not limit this.

[0085] Figure 8 A schematic diagram showing the embodiment of the present application using two 2-bit and 3-bit ADCs to simultaneously quantize and implement non-uniform quantization, corresponding to Figure 4 .

[0086] In embodiment 800, two ADCs with 2-bit and 3-bit precision are used to quantize two replica signals simultaneously, and the quantization results are spliced ​​based on the intersection verification method to achieve the effect of non-uniform quantization. The sampling device for implementing embodiment 800 can be implemented in Figure 1A The sampling device 105, orFigure 1B The sampling device 135 in the embodiment, or other sampling devices in scenarios such as video processing and underwater acoustic processing, is not limited in this disclosure.

[0087] In the embodiment of the present disclosure, the function is the same as that of embodiment 700. The power divider 805 in the sampling device divides the RF signal to be quantized 804 output by the RF module 803 into two replica signals x1(t) 808 and x2(t) 807. The first gain amplifier 813 performs a gain control on the first replica signal x1(t) 808. times the gain adjustment so that the first amplified signal Under large signal conditions, the dynamic range of the 3-bit ADC 823 is fully utilized. The second gain amplifier 810 performs a gain adjustment on the second replica signal x2(t) 805. Multiplication gain adjustment makes the second amplified signal Under large signal conditions, the dynamic range of the 2-bit ADC 820 is exceeded. Under large signal conditions, the second amplified signal The subsequent quantization results are discarded, no data splicing is performed, and only the first amplified signal is used Under small signal conditions, the first amplified signal and the second amplified signal The quantization results generated subsequently can all be quantized and spliced. In this way, a large quantization step size can be achieved under large signal conditions and a small quantization step size can be achieved under small signal conditions, thus achieving non-uniform quantization as a whole.

[0088] In the embodiment of the present disclosure, the first amplified signal is amplified using the ADC 823 with 3-bit precision. Perform sampling quantization, and normalize the amplitude of the quantization result to obtain the first normalized signal Another 2-bit precision ADC820 is used to simultaneously amplify the second signal Perform sampling quantization, and normalize the amplitude of the quantization result to obtain the second normalized signal In this way, two ADCs with different precisions are used to sample and quantize the amplified signals after amplification with different gains, thereby improving the flexibility of system implementation and the sampling performance.

[0089] In the embodiment of the present disclosure, the function is the same as that of embodiment 700, and the normalization signal is eliminated by the digital synchronization algorithm 840, 843 such as the coarse synchronization and fine synchronization algorithms of the OFDM system. and The delay difference between them is the same as that of embodiment 700. Figure 6 The data splicing method 750 based on the intersection verification method is shown, which performs the normalized signal and For a given n and The quantization intervals of the intermediate quantized values are spliced to obtain non-uniformly quantized quantized data 755. In this way, multiple ADCs can be used to simultaneously sample and quantize multiple amplified signals with different precisions, and by means of some amplified signals exceeding the dynamic range of the ADCs, the quantized signals are discarded, thereby achieving different quantization steps at different signal amplitudes. Compared with the method based on nonlinear transformation and inverse transformation, the quantization accuracy is improved, and the complexity is reduced. Moreover, multiple ADCs improve the design flexibility and improve the quantization performance. Those skilled in the art can understand that, in addition to using 2-bit and 3-bit quantization ADCs, other arbitrary n-bit quantization, such as less than 8-bit quantization, can also be used. In addition to using the 2-way replicated signal processing method in embodiment 800, any K-way greater than 2 can also be used, which is not limited in the present disclosure.

[0090] In the above embodiments provided by the present application, the method provided by the embodiments of the present application is introduced from the perspective of single time division ADC and multi-channel parallel ADC respectively. In order to realize each function in the above method provided by the embodiments of the present application, the sampling device in the network device and the terminal device can include hardware structure and / or software module, and the above functions are realized in the form of hardware structure, software module, or hardware structure plus software module. Whether a certain function in the above functions is executed in the form of hardware structure, software module, or hardware structure plus software module depends on the specific application of the technical solution and the design constraint conditions. The sampling device can also be used in audio processing, video processing, underwater acoustic processing, and other various scenes, which are not limited in the present disclosure.

[0091] Figure 9is a block diagram of a device 900 that can be used to implement apparatuses according to some embodiments of the application. Sampling apparatuses can be implemented in device 900, e.g., as part of device 900. Sampling apparatuses can be implemented as a single chip, or a combination of several chips, or as hardware circuitry, or partially as hardware circuitry and partially as software, firmware or other means, which the present disclosure does not limit. In some embodiments, device 900 can be an element of a communication network infrastructure, such as a base station (e.g., a NodeB, an evolved NodeB (eNodeB or eNB), a next generation NodeB (sometimes referred to as gNodeB or gNB), a home subscriber server (HSS), a gateway (GW), such as a packet gateway (PGW) or a serving gateway (SGW), or various other nodes or functions within a core network (CN) or a public land mobile network (PLMN). In other embodiments, device 900 can be a device that connects to network infrastructure over a wireless interface, such as a mobile phone, a smartphone, or other such device that can be classified as user equipment (UE). In some embodiments, device 900 can be a machine-type communication (MTC) device (also known as a machine-to-machine (M2M) device), or another such device that can be classified as a UE, although not providing direct services to a user. In some embodiments, device 900 can be a road-side unit (RSU), a vehicle UE (V-UE), a pedestrian UE (P-UE), or an infrastructure UE (I-UE). In some scenarios, device 900 can also be referred to as a mobile device, a term intended to reflect a device that connects to a mobile network, regardless of whether the device itself is designed for or capable of mobility. Particular devices can utilize all or only a subset of the components shown, and levels of integration can vary from device to device. Furthermore, device 900 can contain multiple instances of a component, such as multiple processors, memories, transmitters, receivers, etc.

[0092] Device 900 typically includes a processor 902, such as a central processing unit (CPU) and, in some embodiments, specialized processors such as a graphics processing unit (GPU) or other such processors, a memory 904, a network interface 906, and a bus 908 to connect the components of device 900. Optionally, device 900 can also include components such as mass storage devices 910, a video adapter 912, and I / O interfaces 916 (shown in dashed lines).

[0093] The memory 904 can include any type of non-transitory system memory that is readable by the processor 902, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), read-only memory (ROM), or a combination thereof. In one embodiment, the memory 904 can include more than one type of memory, such as ROM for use at boot-up and DRAM for program and data storage while executing programs. The bus 908 can be one or more of several types of bus architectures including a memory bus or memory controller, a peripheral bus, or a video bus.

[0094] The device 900 can also include one or more network interfaces 906, which can include at least one of a wired network interface and a wireless network interface. As shown, the network interfaces 906 can include a wired network interface for connecting to a network 922, and can also include a wireless access network interface 920 for connecting to other devices through a wireless link. The wireless access network interface 920 can be omitted for nodes or functions that are elements of a PLMN, but not elements at the wireless edge (e.g., eNB) when the device 900 is a network infrastructure element. When the device 900 is infrastructure at the wireless edge of a network, both wired and wireless network interfaces can be included. When the device 900 is a wirelessly connected device, such as a user equipment, the wireless access network interface 920 can be present and can be supplemented by other wireless interfaces, such as a WiFi network interface. The network interfaces 906 allow the device 900 to communicate with remote entities such as those connected to the network 922. Figure 9

[0095] The mass storage 910 can include any type of non-transitory storage device configured to store data, programs, and other information and make the data, programs, and other information accessible via the bus 908. The mass storage 910 can include, for example, one or more of a solid state drive, a hard disk drive, a magnetic disk drive, or an optical disk drive. In some embodiments, the mass storage 910 can be remote from the device 900 and can be accessed through the use of a network interface such as the interface 906. In the illustrated embodiment, the mass storage 910 is distinct from the memory 904 that includes it, and the mass storage 910 can generally perform storage tasks that are compatible with higher latencies, but can generally provide less or no volatility. In some embodiments, the mass storage 910 can be integrated with the heterogeneous memory 904.

[0096] ​Optional video adapter 912 and I / O interface 916 (shown in dotted lines) provide interfaces for coupling device 900 to external input and output devices. Examples of input and output devices include display 66 coupled to video adapter 912 and I / O device 918 such as a touch screen coupled to I / O interface 916. Other devices can be coupled to device 900, and additional or fewer interfaces can be utilized. For example, a serial interface such as a universal serial bus (USB) (not shown) can be used to provide an interface for external devices. Those skilled in the art will appreciate that in embodiments where device 900 is part of a data center, I / O interface 916 and video adapter 912 can be virtualized and provided through network interface 906.

[0097] Figure 10 FIG. 1 is a schematic diagram of the structure of a sampling device 1000 according to some embodiments of the present application. Figure 10 As shown, the apparatus 1000 includes a first generating unit 1002 and a second generating unit 1004. The apparatus 1000 can be applied to Figure 1A The communication system shown and Figure 1B The audio processing system shown in the figure can implement any of the methods provided in the above embodiments. Optionally, the physical manifestation of the device 1000 can be a communication device, such as a network device or UE, and can also be implemented as an audio processing device, a video processing device, an underwater acoustic processing device, etc. Alternatively, the device 1000 can be other devices that can implement the functions of a communication device, such as a processor or chip inside a communication device. Specifically, the device 1000 can be a programmable chip, such as a field programmable gate array (FPGA), a complex programmable logic device (CPLD), an application specific integrated circuit (ASIC), or a system on a chip (SOC).

[0098] In some embodiments, the first generating unit 1002 may be configured to generate multiple quantization results of different precisions based on the signal to be quantized. The second generating unit 1004 may be configured to generate quantized data of the signal to be quantized based on the multiple quantization results.

[0099] In some other embodiments, the apparatus 1000 may include various other units or modules, which may be configured to perform various operations or functions described in relation to the aforementioned method embodiments. Specific details may be obtained by referring to the detailed description of the aforementioned method embodiments, which will not be repeated herein.

[0100] It should be noted that the division of the modules in the above embodiments of the present application is illustrative, and is only a logical function division. In actual implementation, another division manner can be used. In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or can be physically separated, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0101] When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application can be embodied in the form of a software product in essence or all or part of the technical solutions. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the method of each embodiment of the present application. The storage medium described above includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various storage program codes.

[0102] Based on the above embodiments, the embodiments of the present application further provide a computer program, which, when running on a computer, causes the computer to execute any of the methods provided in the above embodiments.

[0103] Based on the above embodiments, the embodiments of the present application further provide a computer readable storage medium, which stores a computer program. The computer program is executed by a computer to cause the computer to execute any of the methods provided in the above embodiments. The storage medium can be any available medium accessible by a computer. By way of example, and not limitation, the computer readable medium can include a RAM, a ROM, an EEPROM, a CD-ROM or other optical disk storage, a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and can be accessed by a computer.

[0104] Based on the above embodiments, the embodiments of the present application further provide a chip for reading a computer program stored in a memory, and implementing any of the methods provided in the above embodiments.

[0105] Based on the above embodiments, the embodiments of the present application provide a chip system, which comprises a processor for supporting a computer device to implement the functions related to the communication devices in the above embodiments. In a possible design, the chip system further comprises a memory for storing the necessary programs and data of the computer device. The chip system can be composed of a chip, or can include the chip and other discrete devices.

[0106] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. In addition, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate means for implementing the functions specified in one or more flows in the flowcharts and / or one or more blocks in the block diagrams.

[0108] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction means, which implements the functions specified in one or more flows in the flowcharts and / or one or more blocks in the block diagrams.

[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to generate a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide steps for implementing the functions specified in one or more flows in the flowcharts and / or one or more blocks in the block diagrams.

[0110] Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A method for quantization, comprising: generating a plurality of quantization results with different precisions based on a signal to be quantized; and generating quantized data of the signal to be quantized based on the plurality of quantization results. The generating the plurality of quantization results with different precisions comprises:

2. The method of claim 1, wherein, generating a plurality of replica signals of the signal to be quantized; applying a plurality of different gains to the plurality of replica signals to generate a plurality of amplified signals; quantizing the plurality of amplified signals to generate a plurality of quantized signals; and generating the plurality of quantization results based on the plurality of quantized signals. The quantizing the plurality of amplified signals comprises:

3. The method of claim 2, wherein, quantizing the plurality of amplified signals time-divisionally based on a same analog-to-digital converter (ADC). 4.The method of claim 3, wherein: a dynamic range of a first amplified signal of the plurality of amplified signals is equal to a dynamic range of the ADC, and a dynamic range of a second amplified signal of the plurality of amplified signals is greater than the dynamic range of the ADC. The quantizing the plurality of amplified signals comprises:

5. The method of claim 2, wherein, quantizing the plurality of amplified signals in parallel based on a plurality of ADCs. The plurality of ADCs have different precisions.

6. The method of claim 5, wherein, The plurality of ADCs have the same precision.

7. The method of claim 5, wherein, 8.The method of any one of claims 5-7, wherein: a dynamic range of a first amplified signal of the plurality of amplified signals is equal to a dynamic range of a first ADC used to quantize the first amplified signal, and a dynamic range of a second amplified signal of the plurality of amplified signals is greater than a dynamic range of a second ADC used to quantize the second amplified signal. The same ADC or the plurality of ADCs comprise n-bit quantization ADCs, where n is an integer.

9. The method according to any one of claims 3-8, characterized in that, The generating the plurality of quantization results comprises:

10. The method according to any one of claims 2 to 9, characterized in that if it is determined that an amplified amplitude corresponding to a first sample point of the signal to be quantized in a first amplified signal of the plurality of amplified signals does not exceed a dynamic range of an ADC used to quantize the first amplified signal, generating a first quantization result corresponding to the first amplified signal based on the first sample point; and if it is determined that an amplified amplitude corresponding to a second sample point of the signal to be quantized in a second amplified signal of the plurality of amplified signals exceeds a dynamic range of an ADC used to quantize the second amplified signal, not generating a second quantization result corresponding to the second amplified signal based on the second sample point. The generating the plurality of quantization results based on the plurality of quantized signals comprises:

11. The method according to any one of claims 2-10, characterized in that, normalizing the plurality of quantized signals based on a plurality of normalization factors to generate a plurality of normalized signals; and synchronizing the plurality of normalized signals to generate the plurality of quantization results. The plurality of normalization factors are equal to inverses of the plurality of gains, respectively.

12. The method of claim 11, wherein, The generating the quantized data of the signal to be quantized comprises:

13. The method according to any one of claims 1-12, characterized in that, determining a plurality of quantization intervals corresponding to a plurality of intermediate quantization values of the plurality of quantization results corresponding to sample points of the signal to be quantized; and determining quantization values corresponding to the sample points in the quantized data based on the plurality of quantization intervals. ​ 14. The method of claim 13, wherein, The determining the quantized value corresponding to the sampling point in the quantized data comprises: if it is determined that the multiple quantization intervals have an intersection, determining the quantized value based on the intersection; and / or if it is determined that the multiple quantization intervals do not have an intersection, determining the quantized value based on a mean value of the multiple quantization intervals.

15. The method of claim 14, wherein, The determining the quantized value based on the intersection comprises: determining a midpoint value of the intersection as the quantized value.

16. A quantization apparatus comprising: a quantization result generation module configured to generate multiple quantization results with different precisions based on a signal to be quantized; and a quantized data generation module configured to generate quantized data of the signal to be quantized based on the multiple quantization results. The quantization result generation module comprises:

17. The apparatus of claim 16, wherein, a replication sub-module configured to generate multiple replicated signals of the signal to be quantized based on the signal to be quantized; a gain amplification sub-module configured to apply multiple different gains to the multiple replicated signals to generate multiple amplified signals; a quantization sub-module configured to quantize the multiple amplified signals to generate multiple quantized signals; a normalization sub-module configured to normalize the multiple quantized signals based on multiple normalization factors to generate multiple normalized signals, wherein the multiple normalization factors of the normalization sub-module are inverses of the multiple gains, respectively; and a synchronization sub-module configured to synchronize the multiple normalized signals to generate the multiple quantization results. The quantization sub-module comprises:

18. The apparatus of claim 17, wherein, a same analog-to-digital converter (ADC) configured to quantize the multiple amplified signals time-divisionally to generate the multiple quantized signals.

19. The apparatus of claim 18, wherein: a dynamic range of a first amplified signal of the multiple amplified signals is equal to a dynamic range of the ADC, and a dynamic range of a second amplified signal of the multiple amplified signals is greater than the dynamic range of the ADC. The quantization sub-module comprises:

20. The apparatus of claim 17, wherein, multiple ADCs configured to quantize the multiple amplified signals in parallel to generate the multiple quantized signals. The multiple ADCs have different precisions.

21. The apparatus of claim 20, wherein, The multiple ADCs have the same precision.

22. The apparatus of claim 20, wherein, 23. The apparatus of any one of claims 20-22, wherein: a dynamic range of a first amplified signal of the multiple amplified signals is equal to a dynamic range of a first ADC used to quantize the first amplified signal, and a dynamic range of a second amplified signal of the multiple amplified signals is greater than a dynamic range of a second ADC used to quantize the second amplified signal. The quantized data generation module comprises:

24. The apparatus of claim 16, wherein, a quantization interval determination sub-module configured to determine multiple quantization intervals corresponding to multiple intermediate quantized values of the multiple quantization results corresponding to sampling points of the signal to be quantized; and a quantization interval processing sub-module configured to determine a quantized value corresponding to the sampling point in the quantized data based on the multiple quantization intervals. The determining the quantized value corresponding to the sampling point in the quantized data comprises:

25. The apparatus of claim 24, wherein, ​ determining the quantization value based on the intersection if it is determined that the plurality of quantization intervals have an intersection; and / or determining the quantization value based on a mean of the plurality of quantization intervals if it is determined that the plurality of quantization intervals do not have an intersection.

26. The apparatus of claim 25, wherein, the determining the quantization value based on the intersection comprises: determining a midpoint value of the intersection as the quantization value.

27. A quantization device comprising the quantization apparatus of any one of claims 16-26.

28. A computer-readable storage medium storing instructions which, when executed by an electronic device, cause the electronic device to perform the method of any one of claims 1-15.

29. A computer program product comprising instructions which, when executed by an electronic device, cause the electronic device to perform the method of any one of claims 1-15.

Images