An adaptive control method, device and equipment for network audio cross-clock domain transmission

Abstract

The application discloses a kind of self-adapting control network audio cross-clock domain transmission method, device and equipment.The application estimates long-term frequency deviation by network time information and local I2S sampling count, so that ASRC conversion ratio can follow actual clock deviation change, avoid long-term one-way deviation of buffer occupancy;Through buffer occupancy feedback control, short-term deviation caused by network jitter, scheduling delay and burst data arrival change is corrected, so that the number of valid samples in network audio receiving buffer, ASRC FIFO or I2S DMA buffer is maintained within the target range, the probability of underrun and overrun is reduced, thereby improving system stability;ASRC parameters are updated under preset boundary conditions, reducing the pop, crack and transient distortion caused by sudden change of conversion ratio or filter coefficient;Through multi-channel same group coherence control, the phase consistency between multiple channels in the same audio group is maintained.

Description

Technical Field

[0001] This invention relates to the fields of network audio transmission, digital audio processing, asynchronous sampling rate conversion, and embedded audio systems, and particularly to an adaptive control method, apparatus, and computer device for cross-clock domain transmission of network audio based on a hardware ASRC module (Asynchronous Sample Rate Converter). Background Technology

[0002] In network audio transmission systems like Dante, audio data is typically transmitted between multiple devices via Ethernet. The network audio transmitter, network audio receiver, and local audio codec operate in different clock domains. For the network audio receiver, the sampling clock for the network-side audio data usually originates from the network audio clock or a synchronization clock based on PTP (Precision Time Protocol), while the local I2S audio interface is typically driven by a local crystal oscillator, a phase-locked loop (PLL), or an internal clock module of the SoC. Because the network audio sampling clock domain and the local I2S sampling clock domain are not strictly from the same source, even if they have the same nominal sampling rate, such as 48 kHz or 96 kHz, small but persistent frequency and phase deviations can occur due to factors such as crystal oscillator errors, PTP synchronization errors, network transmission jitter, Linux scheduling delays, and DMA transfer delays. These deviations can cause the number of valid samples in the network audio receiver buffer to continuously increase or decrease, eventually leading to problems such as buffer underrun, buffer overrun, audio dropouts, popping sounds, distortion, or abnormal playback speed.

[0003] To address the aforementioned issues, existing technologies offer two typical solutions: First, reproducing the network audio clock using an external clock chip and using this external clock to drive the local I2S audio interface. This solution requires additional external clock chips, DAC control circuits, feedback detection circuits, or VCXO adjustment circuits, resulting in high hardware costs, high system complexity, and the clock adjustment process is prone to introducing problems such as frequency abrupt changes, phase abrupt changes, clock glitches, or codec lockout. Second, asynchronously converting the PCM audio data using software ASRC. This method requires the CPU to perform interpolation, filtering, and resampling operations on the audio sampling points. In multi-channel, high-sampling-rate, and low-latency scenarios, this significantly increases the CPU load and introduces additional software processing latency. When the CPU load increases or real-time scheduling becomes unstable, the continuity of audio transmission cannot be guaranteed.

[0004] Furthermore, while existing hardware ASRC can achieve cross-clock domain conversion, its fixed conversion ratio design cannot adapt to dynamic scenarios such as network jitter, sampling rate switching, and multi-channel phase consistency. Standalone deployment still struggles to solve problems like transient distortion and buffer offset accumulation. Therefore, there is an urgent need for a hardware ASRC adaptive control scheme for network audio scenarios to achieve low-cost, highly stable cross-clock domain audio transmission. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an adaptive control method, apparatus and computer device for cross-clock domain transmission of network audio, which can realize stable conversion between the network audio sampling clock domain and the local I2S sampling clock domain without adding an external clock reproduction chip.

[0006] Firstly, an adaptive control method for cross-clock domain transmission of network audio is provided, comprising the following steps: S10: receiving network audio data packets, unpacking them to obtain PCM audio samples, packet sequence number, sampling sequence number, timestamp, and audio channel information, and writing the PCM audio samples into the network audio receiving buffer; S20: periodically detecting the number of valid samples in the network audio receiving buffer, and generating a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving buffer and the target buffer occupancy; S30: collecting network time information, local I2S sampling count, and local reference time information, and estimating the network audio sampling clock domain and the local I2S clock domain. S40: The long-term frequency deviation between sampling clock domains is used to generate a feedforward estimate of the ASRC conversion ratio; S50: The feedforward estimate is combined with the feedback correction to generate the target ASRC conversion ratio; S60: The target ASRC conversion ratio is smoothed, limited, and speed-limited to obtain the actual ASRC conversion ratio that can be directly written to the hardware ASRC module; S70: The ASRC parameters of the hardware ASRC module are updated under preset boundary conditions, and the ASRC parameters include the actual ASRC conversion ratio; S80: Steps S20 to S60 are repeated periodically to dynamically maintain the buffer occupancy within the target range by adaptively correcting the ASRC conversion ratio.

[0007] Secondly, an adaptive control network audio cross-clock domain transmission device is provided, comprising: a network audio receiving and unpacking module for receiving network audio data packets, unpacking them to obtain PCM audio samples, packet sequence numbers, sampling sequence numbers, timestamps, and audio channel information, and writing the PCM audio samples into the network audio receiving buffer; a buffer feedback module for periodically detecting the number of valid samples in the network audio receiving buffer, and generating a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving buffer and the target buffer occupancy; and an ASRC feedforward estimation module for collecting network time information, local I2S sampling counts, and local reference time information, estimating the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generating a feedforward estimate of the ASRC conversion ratio; ASRC The conversion ratio control module combines the feedforward estimate with the feedback correction to generate the ASRC target conversion ratio; the ASRC parameter smoothing and limiting module performs smoothing, amplitude limiting, and speed limiting on the ASRC target conversion ratio to obtain the actual ASRC conversion ratio that can be directly written into the hardware ASRC module; the ASRC parameter update module updates the ASRC parameters of the hardware ASRC module under preset boundary conditions, wherein the ASRC parameters include the actual ASRC conversion ratio.

[0008] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described adaptive control network audio cross-clock domain transmission method.

[0009] The aforementioned adaptive control network audio cross-clock domain transmission method, apparatus, and computer equipment achieve stable conversion between the network audio sampling clock domain and the local I2S sampling clock domain without adding external clock reproduction chips, VCXO adjustment circuits, or DAC control circuits, reducing system complexity and hardware costs. Hardware ASRC performs resampling operations, reducing the CPU burden of software interpolation and filtering of multi-channel audio samples, thus lowering system CPU utilization. Long-term frequency deviation is estimated using network time information and local I2S sampling counts, allowing the ASRC conversion ratio to follow actual clock deviation changes, avoiding long-term unidirectional shifts in buffer occupancy. Buffer occupancy feedback control corrects short-term offsets caused by network jitter, scheduling delays, and sudden data arrival changes, maintaining the number of effective samples in the network audio receiving buffer, ASRCFIFO, or I2S DMA buffer within the target range, reducing the probability of underrun and overrun, thereby improving system stability. ASRC parameters are updated under preset boundary conditions, reducing pops, distortions, and transient distortions caused by sudden changes in the conversion ratio or filtering coefficients. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating an adaptive control network audio cross-clock domain transmission method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of an adaptive control network audio cross-clock domain transmission device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] Terminology Explanation: ASRC: Asynchronous Sample Rate Converter, used to convert audio sample rates between different sampling clock domains; I2S: Inter-Integrated Circuit Audio Interface, commonly used for transmitting PCM audio data between SoC and audio codec; PTP: Precision Time Protocol, used for time synchronization between network devices; Dante: A professional network audio transmission technology based on Ethernet; PCM: Pulse Code Modulation Audio Data; FIFO: First-In-First-Out (FIFO) cache structure; DMA: Direct Memory Access, used for data transfer between peripherals and memory; flow: An audio stream transmission unit in a network audio system, which may contain one or more audio channels; underrun: The data consumption rate is greater than the data supply rate, resulting in no valid data to be read on the output side; Overrun: The data supply rate exceeds the data consumption rate, causing the input-side buffer to overflow.

[0014] This invention discloses an adaptive control network audio transmission method, apparatus, and computer device across clock domains, aiming to solve the following technical problems: 1. When there is a long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, the number of valid samples in the network audio receiving buffer is prone to continuous shift, resulting in underrun or overrun. 2. Network transmission jitter, PTP synchronization fluctuations, Linux scheduling delays, and DMA transmission delays can cause uncertainty in the arrival time of audio data, affecting the stability of real-time playback; 3. Hardware ASRC with a fixed conversion ratio cannot adaptively adjust according to the network audio status and local I2S status, making it difficult to balance low latency and high stability. 4. When the ASRC conversion ratio or filter coefficients are updated abruptly at any sampling point, transient distortion, popping sounds, crackling sounds, or codec lockout may be introduced. 5. When multi-channel audio undergoes ASRC conversion, if different conversion ratios or different parameter update times are used for each channel, inconsistencies in the phase relationship between channels may occur, affecting the spatial imaging and synchronization performance of the multi-channel audio system.

[0015] The adaptive control method for cross-clock domain transmission of network audio provided by this invention is applied to a network audio receiving device. This device employs an RK3506B or other embedded processor platform with a hardware ASRC module, I2S, DMA, and Ethernet interface, and is suitable for transmitting audio streams such as Dante, AES67, AVB, and RTP. After the network audio data is received via the Ethernet interface, the software module performs unpacking and buffer management, the hardware ASRC module performs sampling rate conversion, and the local I2S interface outputs the converted PCM audio data.

[0016] Figure 1 This is a flowchart illustrating the adaptive control network audio cross-clock domain transmission method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the adaptive control network audio cross-clock domain transmission method includes steps S10 to S70: S10: Receive network audio data packets, unpack them to obtain PCM audio samples, packet sequence number, sampling sequence number, timestamp, and audio channel information, and write the PCM audio samples into the network audio receiving buffer.

[0017] The software control layer of the network audio receiving device is configured with a network audio receiving module, a network audio unpacking module, and a network audio receiving buffer module. The network audio receiving module receives audio data packets transmitted via an Ethernet interface using Dante or other network audio protocols. The network audio unpacking module parses the received network audio data packets to obtain information such as PCM audio samples, packet sequence numbers, sample sequence numbers and timestamps, audio channel information, and PTP timestamps. The unpacked PCM audio data (a set of PCM audio samples) is temporarily stored in the network audio receiving buffer module.

[0018] S20: Periodically detect the number of valid samples in the network audio receiving buffer, and generate a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving buffer and the target buffer occupancy.

[0019] 2.1 Periodically detect the number of valid samples in the network audio receiving buffer to obtain the number of valid samples in the network audio receiving buffer in the k-th control period.

[0020] This invention periodically detects the number of valid samples in the network audio receiving buffer and uses this number of valid samples as the feedback quantity for ASRC control.

[0021] Let the number of valid samples in the network audio reception buffer during the k-th control period be: in, This represents the cumulative number of samples written to the network audio receive buffer up to the kth control cycle; This represents the cumulative number of samples read from the network audio receive buffer up to the k-th control cycle; This indicates the number of valid samples in the network audio reception buffer.

[0022] 2.2 Calculate the cache usage deviation.

[0023] Set the target cache usage to The cache usage deviation is defined as follows: when When the current cache usage is higher than the target value, it means that the long-term equivalent data write rate on the network audio input side is higher than the data consumption rate on the local I2S output side. The system needs to adjust the ASRC conversion mapping relationship appropriately so that the cache usage gradually returns to the target range. when When the cache usage is below the target value, it indicates that the long-term data consumption rate on the local I2S output side is higher than the data write rate on the network audio input side. The system needs to adjust the ASRC conversion mapping relationship appropriately to prevent the cache usage from continuing to decrease.

[0024] The target buffer occupancy in this invention is not a fixed value, but is dynamically calculated based on network status, system scheduling status, sampling rate, number of channels, and latency configuration. This design aims to achieve an adaptive balance between low latency and high stability.

[0025] Let the target cache usage be: in, Indicates the current audio sampling rate, such as 48 kHz or 96 kHz; Indicates the current target cache time.

[0026] The current target cache time can be calculated using the following formula: in, Based on the base cache time, To estimate the jitter of network audio packets, This is the estimated system scheduling delay. This refers to the time corresponding to the I2S DMA period or the audio period. The compensation time is related to the current number of Dante flows, the number of channels, or the number of multicast receptions. , , , These are the weighting coefficients.

[0027] Calculate target cache usage: .

[0028] Network audio packet jitter estimation value The arrival time deviation of packets within a sliding window can be used for calculation. For example, for the i-th network audio data packet, its actual arrival time is... The theoretical arrival interval is The package arrival time deviation is: The estimated jitter value for network audio packets can be taken within the sliding window. The average, root mean square, or high percentile value. For example: in, This represents the 99th percentile statistic.

[0029] To prevent underruns caused by insufficient target cache usage or excessive system latency caused by excessive target cache usage, this invention further sets a target cache usage constraint: in, To be the minimum allowed target cache size, The maximum allowed target cache size.

[0030] When network jitter is low, system scheduling latency is low, and the number of Dante flows is small, the system reduces... To reduce end-to-end audio latency; when network jitter increases, system scheduling latency increases, or the number of Dante flows increases, the system improves... This enhances the system's tolerance to sudden delays and scheduling jitter.

[0031] S23: Based on the buffer usage deviation, a PID control algorithm is used to generate a feedback correction amount: in, This is the proportional control coefficient. The integral control coefficient, This is the differential control coefficient.

[0032] In practical implementation, a PID control algorithm with only proportional terms, a PID control algorithm with only proportional-integral terms, or a PID control algorithm with only proportional-integral-derivative terms can be used to calculate the feedback correction. For real-time audio systems, a PID control algorithm with amplitude limiting and integral suppression is preferred for calculating the feedback correction to avoid excessive oscillation of the control quantity.

[0033] S30: Collect network time information, local I2S sampling count and local reference time information, estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generate a feedforward estimate of the ASRC conversion ratio.

[0034] To avoid being affected by short-term network jitter by relying solely on changes in cache occupancy, this invention further introduces network time information and local I2S sampling counts to estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain.

[0035] 3.1 Collect network time information, local I2S sampling count, and local reference time information.

[0036] The software control layer of the network audio receiving device is configured with a network time information extraction module, a local I2S sampling counting module, and a local reference time module. The network time information extraction module obtains network time information such as the timestamp, packet sequence number, sample sequence number, and PTP timestamp of the network audio data packets. The local I2S sampling counting module counts local I2S sampling count data such as the local I2S LRCLK frame clock count, I2S DMA transmission cycle count, or local cumulative playback sample count. The local reference time module provides local reference time in real time.

[0037] 3.2 Calculate the estimated sampling frequency on the network audio side.

[0038] Suppose that the change in sample index of the network audio data within the k-th estimation window is: in, The sequence number of the network audio sample at the end of the k-th estimation window; The sequence number of the network audio sample at the start time of the k-th estimation window.

[0039] The change in network audio timestamps is as follows: in, Let be the network audio timestamp at the end of the k-th estimation window, and The network audio timestamp represents the start time of the k-th estimation window.

[0040] The estimated sampling frequency on the network audio side is: in, To estimate the window length, , It can be obtained from the network audio data packet sequence number, sample sequence number, or the cumulative number of PCM audio samples after unpacking. , It can be calculated from PTP timestamps, network audio timestamps, or local received timestamps.

[0041] 3.3 Calculate the estimated sampling frequency on the local I2S side.

[0042] In this invention, the local I2S sampling and counting module counts the cumulative number of samples in the local LRCLK frame clock or I2S DMA cycle as the local I2S sample count.

[0043] Within the k-th estimation window, the change in local I2S sample count is: in, This represents the cumulative local I2S audio frame count read at the end of the k-th estimation window; This represents the local I2S cumulative audio frame count read at the start of the k-th estimation window.

[0044] The change in local reference time is: in, Represents the local reference clock timestamp at the end of the k-th estimation window; This represents the local reference clock timestamp at the start of the k-th estimation window.

[0045] The estimated sampling frequency on the local I2S side is: The relative frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain is defined as: 3.4 Feedforward estimates of ASRC conversion ratios.

[0046] Based on the long-term frequency deviation, the feedforward estimate of the ASRC conversion ratio is obtained: .

[0047] S40: Combine the feedforward estimate with the feedback correction to generate the ASRC target conversion ratio.

[0048] This invention uses a combination of feedforward estimation and cache usage feedback to generate the ASRC target conversion ratio: This method can simultaneously utilize network audio timing information, local I2S sample counts, and buffer occupancy status. Long-term frequency deviations are mainly caused by... Compensation, short-term cache offset is mainly caused by This correction improves control stability and resistance to network jitter.

[0049] S50: Performs smoothing, amplitude limiting, and speed limiting on the target ASRC conversion ratio to obtain the actual ASRC conversion ratio that can be directly written to the hardware ASRC module.

[0050] First, the ASRC target conversion ratio Smoothing is performed to obtain the actual ASRC conversion ratio that can be directly written to the hardware ASRC module. : in, The smoothing coefficient is , and satisfies The larger the value, the smoother the change in the ASRC conversion ratio; The smaller the value, the faster the ASRC conversion ratio response.

[0051] Secondly, regarding the actual conversion ratio of ASRC Limit the single change amount: in, This represents the maximum allowable change in the conversion ratio within a single control cycle.

[0052] Meanwhile, the actual conversion ratio of ASRC Set the valid range: in, and These are the minimum and maximum conversion ratios supported by the ASRC hardware module, or the safety range set by the system based on audio quality requirements.

[0053] S60: Update the ASRC parameters of the hardware ASRC module under preset boundary conditions, wherein the ASRC parameters include the actual ASRC conversion ratio.

[0054] In network audio systems, if the ASRC conversion ratio, interpolation filter coefficients, or related control registers are suddenly updated at any sampling point, it may cause discontinuities in the output audio waveform, resulting in pops, distortions, or transient distortions. To avoid these issues, this invention does not immediately apply new ASRC parameters after generation. Instead, it first writes the ASRC parameters to be updated into the pending parameter area or the spare parameter buffer, and completes the parameter switching when preset boundary conditions are met.

[0055] Let the set of ASRC parameters that are currently in effect be... The set of ASRC parameters to be effective is as follows .in, At least include ASRC actual conversion ratio It can also include interpolation filter coefficients, decimation filter coefficients, initial values ​​of the phase accumulator, FIFO thresholds, or control register configurations.

[0056] After generating new ASRC parameters, the system first sets the parameter update request flag: But not immediately This applies to the real-time data path of the hardware ASRC module. It is executed only when preset boundary conditions are met. Otherwise, keep: The ASRC parameter update method described above ensures that ASRC parameter updates occur at stable boundaries of audio data structures, DMA transfer structures, or multi-channel synchronization structures, thereby reducing the impact of parameter mutations on the continuity of output audio.

[0057] In this invention, the preset boundary conditions include I2S period boundaries, I2S DMA buffer boundaries, ASRCFIFO safe intervals, network audio frame boundaries, audio sampling frame boundaries, near zero-crossing, and unified parameter switching points for multi-channel groups. The activation mechanism of each boundary condition is explained in detail below: (6.1) I2S period boundary activation mechanism An I2S period is a basic block of data processed periodically in an audio drive or DMA transfer. Let each I2S period contain... Number of sampled frames, local I2S sampled frame count is Then the I2S period boundary conditions are as follows: This indicates that the current sampling frame is located at the start boundary of the I2S period.

[0058] When the system detects that the current sampling frame is at the start boundary of the I2S period and that there are parameters to be applied, it performs ASRC parameter updates: The purpose of this mechanism is to align ASRC parameter updates with the periodic scheduling boundaries driven by I2S, avoiding changes in the sampling rate conversion relationship within a period, thereby reducing the problem of inconsistent audio sample time mapping within the same period.

[0059] (6.2) I2S DMA buffer boundary activation mechanism An I2S DMA buffer typically consists of one or more periods and operates as a circular buffer. Let the total sampling frame length of the I2S DMA buffer be... The current read / write position of the DMA is The conditions for the DMA buffer start boundary to take effect are: It can also be expressed as: This indicates that the current read / write position of the DMA is located at the beginning boundary of the DMA buffer.

[0060] When the system detects that the current DMA read / write position is at the start boundary of the DMA buffer and that there are parameters to be applied, it performs ASRC parameter update: In one implementation, the DMA semi-buffer boundary can also be used as a safe boundary for allowing updates. Let the set of DMA boundaries that allow updates be... ,but: When it is detected that the current read / write position of the DMA belongs to the set of DMA boundaries that are allowed to be updated, and there are parameters to be applied, perform ASRC parameter update: This mechanism ensures that parameter updates are aligned with the DMA data transfer boundary, avoiding changes to ASRC parameters in the middle of consecutive sample blocks being transferred or played by the DMA, thereby reducing the risk of waveform discontinuity within DMA data blocks.

[0061] (6.3) ASRC FIFO safety zone activation mechanism Hardware ASRC modules typically include input FIFOs and output FIFOs. If ASRC parameter updates occur when the FIFOs are too empty or too full, it may increase the risk of underrun or overrun. Therefore, this invention sets a safe range for the ASRC FIFOs, allowing parameters to take effect only when the FIFO occupancy is within the safe range.

[0062] Let the number of valid samples in the ASRC input FIFO be... The number of valid samples in the output FIFO is The input FIFO safety range is: The safe range for the output FIFO is: in, , Input the minimum and maximum thresholds for the effective number of samples in the FIFO for ASRC. , Set the minimum and maximum thresholds for the number of valid samples in the ASRC output FIFO.

[0063] When both the input FIFO and output FIFO meet the safety range and there are parameters pending activation, perform ASRC parameter update: This mechanism can avoid parameter updates when there is insufficient data at the ASRC input, insufficient data at the output, or when the FIFO is close to overflow, thereby improving data continuity and system stability during parameter switching.

[0064] (6.4) Network audio frame boundary activation mechanism Network audio data is typically transmitted in network audio frames, audio packets, or protocol frames. For Dante, RTP, or similar network audio protocols, a network audio frame usually contains a fixed number of PCM sample frames. If ASRC parameters take effect within a network audio frame, samples within the same network audio frame may use different time mappings. Therefore, this invention sets network audio frame boundaries, allowing ASRC parameter updates to take effect only at the network audio frame boundaries.

[0065] Let the network audio frame number be Each network audio frame contains Given a sampled frame and a sampling offset of Onet(n) within the current network audio frame, the network audio frame boundary condition is as follows: This indicates that the current sampled frame is located at the beginning of the network audio frame.

[0066] When the system detects the start of a new network audio frame and there are parameters to be applied, it performs ASRC parameter updates: This mechanism ensures that samples within the same network audio frame use the same set of ASRC parameters, avoiding changes in the conversion ratio within a single network audio frame, which helps reduce transient distortion caused by discontinuous intra-frame sampling time mapping.

[0067] (6.5) Audio sampling frame boundary activation mechanism In multichannel PCM audio, an audio sampling frame typically contains samples from multiple channels at the same sampling time. For example, for two-channel audio, an audio sampling frame includes samples from the left and right channels; for multichannel audio, an audio sampling frame includes samples from multiple channels at the same sampling time.

[0068] Assume the system has a total of There are 1 audio channel, and the PCM audio sample number is 1. The current channel offset of the PCM audio sample within the audio sampling frame is The channel offset is then defined as: The conditions for audio sampling frame boundaries to take effect are: This indicates that the current PCM audio sample is located in the starting channel of the audio sampling frame.

[0069] When it is detected that the current PCM audio sample is at the beginning of an audio sampling frame and there are parameters to be applied, perform ASRC parameter update: This mechanism ensures that multiple channel samples at the same sampling time use the same set of ASRC parameters, avoiding different conversion ratios for different channels within the same audio sampling frame, thereby maintaining the synchronization relationship and phase consistency between channels.

[0070] (6.6) Mechanism for activation near zero crossover When using zero-crossing proximity updates, the system determines whether the current output audio sample is in the zero-crossing region based on the changes in the amplitude and sign of the output audio sample. Let the output audio sample at the current sampling point n be... The output audio sample of the previous sampling point n-1 is The zero-crossing amplitude threshold is The condition for the update to take effect near the zero crossover is: and: When the above conditions are met and there are parameters yet to be applied, perform ASRC parameter update: in, For use in representing The sign function for positive and negative signs.

[0071] In practical digital audio systems, due to PCM sample quantization errors, noise, and DC bias, a sign function with a threshold can also be used. in, A threshold is set for the sign determination. Using a sign function with a threshold can avoid frequent sign changes caused by noise jitter when the sample is close to zero.

[0072] The advantage of the zero-crossing proximity update mechanism is that ASRC parameter switching is performed at positions where the audio waveform amplitude is low. Even if the parameter switching causes a small waveform disturbance, it is not easy to form obvious audible pops or distortions.

[0073] (6.7) Mechanism for the activation of unified parameter switching points for multi-channel groups For multiple audio channels belonging to the same Dante flow, the same PTP clock domain, the same sampling rate, or having a fixed phase relationship, this invention divides them into the same ASRC control group. Let the m-th ASRC control group be: in, This indicates multiple audio channels within the same control group.

[0074] For the same control group The system generates a uniform ASRC conversion ratio for all channels: And set a unified parameter switching time: in, The ASRC conversion ratio for the m-th control group is... This is the unified parameter update time for this control group.

[0075] When all channel sampling points within the m-th ASRC control group reach the commonly allowed audio sampling frame boundary, I2S DMA buffer boundary, or network audio frame boundary, the current sampling point n is used as the unified parameter switching point for the control group. And set the sampling time corresponding to the current sampling point n to the unified parameter update time of the control group. .

[0076] Let the set of unified parameter switching points for the m-th ASRC control group be . When the sampling points of all channels within the same control group reach the commonly allowed audio sampling frame boundary, I2S DMA buffer boundary, or network audio frame boundary, a unified parameter switch is performed: in, This represents the set of ASRC parameters for the m-th ASRC control group that is currently in effect. This represents the set of ASRC parameters for the m-th ASRC control group that is yet to take effect.

[0077] This mechanism ensures that multiple channels within the same control group use the same set of ASRC parameters at the same sampling time, avoiding phase drift or image shift caused by different channels updating parameters at different times. For stereo, multi-channel sound reinforcement, audio matrix, and spatial audio systems, this mechanism can improve the consistency and synchronization of multi-channel audio.

[0078] (6.8) Method for determining the effectiveness of boundary conditions The above seven activation mechanisms can be used individually or in combination. To improve system stability, this invention adopts a combined determination method for boundary conditions to take effect, meaning that ASRC parameter updates are only performed when multiple boundary conditions are simultaneously met. For example, when both the I2S period boundary and the ASRC FIFO safe interval are required to be met, the activation determination condition is: When all of the above conditions are met, execute: In another implementation, the system can also dynamically configure the priority of different boundary conditions based on the current network status, audio sampling rate, number of channels, CPU scheduling latency, and cache usage. For example, in low-latency scenarios, the I2S period boundary or audio sampling frame boundary can be prioritized; in multi-channel phase-sensitive scenarios, the unified parameter switching point of the multi-channel group can be prioritized; in audio-sensory-sensitive scenarios, the vicinity of zero crossover can be prioritized; and when the system load is high or the FIFO state is unstable, the ASRC FIFO safe zone can be prioritized.

[0079] S70: Periodically repeat steps S20~S60, dynamically maintaining cache usage within the target range by adaptively adjusting the ASRC conversion ratio.

[0080] The hardware ASRC module converts the PCM data in the network audio sampling clock domain into PCM data in the local I2S sampling clock domain based on the updated ASRC conversion ratio and filtering parameters, and outputs it to the audio codec through the local I2S audio interface.

[0081] The system periodically repeats steps S20~S60, continuously adjusting the ASRC conversion ratio based on the network audio status, local I2S status, and buffer usage status, so as to keep the buffer usage within the target range and keep the audio output continuous and stable.

[0082] In Dante network audio systems, multi-channel audio is often transmitted with the same flow, sampling rate, PTP clock domain, or audio routing group. If different ASRC conversion ratios or different parameter update times are used for multiple channels within the same audio group, it may cause changes in the phase relationship between channels, affecting the synchronization performance of stereo, multi-channel sound reinforcement, matrix audio, and spatial audio.

[0083] Therefore, in some embodiments of the present invention, the adaptive control network audio cross-clock domain transmission method further includes a coherent control step for multiple audio channels in the same group: dividing multiple audio channels that meet preset association conditions into the same ASRC control group; applying a uniform ASRC conversion ratio to all channels in the same control group, and synchronously updating the hardware ASRC module parameters of all channels when the same preset boundary conditions are met.

[0084] Specifically, multiple audio channels that meet preset association conditions are grouped into the same ASRC control group. The preset conditions include at least one of the following: belonging to the same Dante flow, belonging to the same network audio subscription relationship, belonging to the same PTP clock domain, having the same sampling rate, having the same network delay configuration, having a fixed channel phase relationship, or belonging to the same physical audio interface or the same audio playback path.

[0085] Let the m-th ASRC control group be: in, This indicates multiple audio channels within the same control group.

[0086] For the same control group The system uses a uniform ASRC conversion ratio: And adopt a unified parameter update time: in, The ASRC conversion ratio for the m-th control group is... This is the unified parameter update time for this control group.

[0087] In this way, multiple channels within the same control group maintain a consistent time mapping relationship during the sampling rate conversion process, avoiding inter-channel phase drift caused by independent adjustment of ASRC parameters for each channel.

[0088] For audio channels with different sampling rates, different PTP clock domains, different Dante flows, or different audio routing groups, the system can establish different ASRC control groups and calculate the corresponding conversion ratios and target buffer occupancy.

[0089] The adaptive control network audio cross-clock domain transmission method provided in this embodiment of the invention has at least the following advantages over the prior art: First, this invention does not simply use hardware ASRC to convert the sampling rate between different clock domains. Instead, it establishes an adaptive control relationship between network audio timing information, local I2S sampling count, buffer usage, and hardware ASRC conversion ratio. Through this control relationship, the system can dynamically adjust the ASRC conversion ratio according to the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain. Second, this invention combines long-term frequency deviation estimation with short-term buffer occupancy feedback. Long-term frequency deviation estimation is used to compensate for the stable frequency error between the network audio sampling clock and the local I2S sampling clock, while buffer occupancy feedback is used to correct short-term offsets caused by network jitter, scheduling delays, and sudden changes in data arrival, thereby improving system stability. Third, this invention proposes an adaptive target buffer occupancy calculation method, which dynamically adjusts the target buffer occupancy based on network jitter, system scheduling delay, DMA cycle, sampling rate, flow count, and channel count. This method can reduce audio latency when the network is stable and improve anti-jitter capability when network jitter increases or system load rises. Fourth, this invention proposes a mutation-free update mechanism for ASRC parameters. By smoothing the conversion ratio, limiting the amount of a single change, limiting the effective range of parameters, double-buffered parameter updates (preparing two sets of ASRC parameter sets, one set already in effect and the other set to be in effect), boundary synchronous updates (synchronous updates of all channels within the same control group), or updates near zero crossover, the transient distortion, popping sounds, crackling sounds, and codec lockout risks caused by changes in ASRC parameters are reduced. Fifth, this invention proposes a coherent control method for multiple audio channels in the same group: for multiple audio channels with the same Dante flow, the same PTP clock domain, the same sampling rate, or a fixed phase relationship, the system adopts a uniform ASRC conversion ratio and a uniform parameter update time to maintain phase consistency among multiple channels.

[0090] The adaptive control network audio cross-clock domain transmission method provided in this embodiment of the invention has the following beneficial technical effects compared with the prior art: (1) Without adding an external clock reproduction chip, VCXO adjustment circuit or DAC control circuit, a stable conversion between the network audio sampling clock domain and the local I2S sampling clock domain is achieved, reducing the complexity of the system structure and the hardware cost. (2) Resampling operations are performed by hardware ASRC, which reduces the burden on the CPU to perform software interpolation and filtering on multi-channel audio samples and reduces the CPU utilization rate of the system. (3) By estimating the long-term frequency deviation through network time information and local I2S sampling count, the ASRC conversion ratio can follow the actual clock deviation and avoid long-term unidirectional shift of buffer occupancy; (4) By controlling the buffer occupancy, the system corrects the short-term offset caused by network jitter, scheduling delay and sudden data arrival changes, so that the number of effective samples in the network audio receiving buffer, ASRC FIFO or I2S DMA buffer is kept within the target range, reducing the probability of underrun and overrun, thereby improving system stability. (5) By adaptively calculating the target cache usage, the system latency is reduced when the network condition is good, and the playback stability is improved when the network jitter or system load increases. (6) Through the ASRC parameter update mechanism without sudden change, the ASRC parameters are updated under preset boundary conditions, reducing popping, crackling and transient distortion caused by sudden changes in conversion ratio or filter coefficient; (7) By using multi-channel coherent control, the phase consistency between multiple channels in the same audio group is maintained. It is suitable for multi-channel Dante audio equipment, audio matrix, conference system, sound reinforcement system and professional audio transmission equipment.

[0091] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0092] In one embodiment, an adaptive control network audio cross-clock domain transmission device is provided, wherein each functional module of the adaptive control network audio cross-clock domain transmission device corresponds one-to-one with steps S10 to S60 of the adaptive control network audio cross-clock domain transmission method in the above embodiment. For example... Figure 2 As shown, the adaptive control network audio cross-clock domain transmission device includes a network audio receiving and unpacking module 10, a buffer feedback module 20, an ASRC feedforward estimation module 30, an ASRC conversion ratio control module 40, an ASRC parameter smoothing and limiting module 50, and an ASRC parameter update module 60. Detailed descriptions of each functional module are as follows: The network audio receiving and unpacking module 10 is used to receive network audio data packets, unpack them to obtain PCM audio samples, packet sequence number, sampling sequence number, timestamp and audio channel information, and write the PCM audio samples into the network audio receiving buffer. The cache feedback module 20 is used to periodically detect the number of valid samples in the network audio receiving cache and generate a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving cache and the target cache occupancy. ASRC feedforward estimation module 30 is used to collect network time information, local I2S sampling count and local reference time information, estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generate a feedforward estimate of the ASRC conversion ratio. ASRC conversion ratio control module 40 is used to combine the feedforward estimate with the feedback correction to generate the ASRC target conversion ratio. ASRC parameter smoothing and limiting module 50 is used to smooth, limit and speed up the target ASRC conversion ratio to obtain the actual ASRC conversion ratio that can be directly written into the hardware ASRC module. ASRC parameter update module 60 is used to update the ASRC parameters of the hardware ASRC module under preset boundary conditions. The ASRC parameters include the actual ASRC conversion ratio.

[0093] In one embodiment, the network audio receiving and unpacking module 10 specifically includes: A network audio receiving module, used to receive audio data packets transmitted via Dante or other network audio protocols; The network audio unpacking module is used to parse the received network audio data packets and extract PCM audio samples, packet numbers, sampling numbers, timestamps, and audio channel information. The network audio receiving buffer module is used to temporarily store unpacked PCM audio data.

[0094] In one embodiment, the cache feedback module 20 specifically includes: The buffer occupancy detection module is used to periodically detect the number of valid samples in the network audio receiving buffer; The feedback correction calculation module is used to calculate and generate the feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving buffer and the target buffer occupancy.

[0095] In one embodiment, the ASRC feedforward estimation module 30 specifically includes: The network time information extraction module is used to obtain network time information; The local I2S sampling and counting module is used to count local I2S samplings. The local reference time module is used to provide local reference time. The clock skew estimation module is used to estimate the long-term frequency skew between the network audio sampling clock domain and the local I2S sampling clock domain; The feedforward estimation module is used to generate feedforward estimates of the ASRC conversion ratio based on the long-term frequency deviation.

[0096] The adaptive control network audio cross-clock domain transmission device provided in this embodiment of the invention has the following beneficial technical effects compared with the prior art: (1) Without adding an external clock reproduction chip, VCXO adjustment circuit or DAC control circuit, a stable conversion between the network audio sampling clock domain and the local I2S sampling clock domain is achieved, reducing the complexity of the system structure and the hardware cost. (2) Resampling operations are performed by hardware ASRC, which reduces the burden on the CPU to perform software interpolation and filtering on multi-channel audio samples and reduces the CPU utilization rate of the system. (3) By estimating the long-term frequency deviation through network time information and local I2S sampling count, the ASRC conversion ratio can follow the actual clock deviation and avoid long-term unidirectional shift of buffer occupancy; (4) By controlling the buffer occupancy, the system corrects the short-term offset caused by network jitter, scheduling delay and sudden data arrival changes, so that the number of effective samples in the network audio receiving buffer, ASRC FIFO or I2S DMA buffer is kept within the target range, reducing the probability of underrun and overrun, thereby improving system stability. (5) By adaptively calculating the target cache usage, the system latency is reduced when the network condition is good, and the playback stability is improved when the network jitter or system load increases. (6) Through the ASRC parameter update mechanism without sudden change, the ASRC parameters are updated under preset boundary conditions, reducing popping, crackling and transient distortion caused by sudden changes in conversion ratio or filter coefficient; (7) By using multi-channel coherent control, the phase consistency between multiple channels in the same audio group is maintained. It is suitable for multi-channel Dante audio equipment, audio matrix, conference system, sound reinforcement system and professional audio transmission equipment.

[0097] Specific limitations regarding the adaptive control network audio cross-clock domain transmission device can be found in the limitations of the adaptive control network audio cross-clock domain transmission method described above, and will not be repeated here. Each module in the aforementioned adaptive control network audio cross-clock domain transmission device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in the computer device, or stored in software in the memory of the computer device, so that the processor can call and execute the corresponding operations of each module.

[0098] In one embodiment, a computer device is provided, the internal structure of which can be shown as follows: Figure 3 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with other external electronic devices via a network connection. When the computer program is executed by the processor, it implements the functions or steps of the aforementioned adaptive control network audio cross-clock domain transmission method.

[0099] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0100] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0101] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for audio transmission across clock domains in an adaptive control network, characterized in that, Includes the following steps: S10: Receive network audio data packets, unpack them to obtain PCM audio samples, packet number, sampling number, timestamp and audio channel information, and write the PCM audio samples into the network audio receiving buffer; S20: Periodically detect the number of valid samples in the network audio receiving buffer, and generate a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving buffer and the target buffer occupancy. S30: Collect network time information, local I2S sampling count and local reference time information, estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generate a feedforward estimate of the ASRC conversion ratio; S40: Combine the feedforward estimate with the feedback correction to generate the ASRC target conversion ratio; S50: Smooths, limits, and speeds up the target ASRC conversion ratio to obtain the actual ASRC conversion ratio that can be directly written to the hardware ASRC module. S60: Update the ASRC parameters of the hardware ASRC module under preset boundary conditions, wherein the ASRC parameters include the actual ASRC conversion ratio. S70: Periodically repeat steps S20~S60, and dynamically maintain the cache usage within the target range by adaptively adjusting the ASRC conversion ratio; The process of acquiring network time information, local I2S sampling counts, and local reference time information to estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generating a feedforward estimate of the ASRC conversion ratio, specifically includes the following steps: Collect network time information, local I2S sampling count, and local reference time information; Within the k-th estimation window, the change in sample sequence number of the statistical network audio data is calculated. Changes in network audio timestamps And calculate the estimated sampling frequency on the network audio side: ; Within the k-th estimation window, count the change in local I2S sample counts. Change from local reference time And calculate the estimated sampling frequency on the local I2S side: ; Feedforward estimates of the ASRC conversion ratio: .

2. The method for audio cross-clock domain transmission in an adaptive control network according to claim 1, characterized in that, The process of periodically detecting the number of valid samples in the network audio receiver buffer and generating a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiver buffer and the target buffer occupancy includes the following steps: The number of valid samples in the network audio receiving buffer is periodically detected to obtain the number of valid samples in the network audio receiving buffer within the k-th control period. ; Calculate cache usage deviation: ,in, Target cache usage; The PID control algorithm is used to generate feedback correction values: ,in, This is the proportional control coefficient. The integral control coefficient, This is the differential control coefficient.

3. The method for audio cross-clock domain transmission in an adaptive control network according to claim 2, characterized in that, The target cache usage The result is dynamically calculated using an adaptive algorithm, and includes the following steps: Get the current audio sampling rate ; Calculate the current target cache time: ,in, Based on the base cache time, To estimate the jitter of network audio packets, This is the estimated system scheduling delay. This refers to the I2S DMA period or the time corresponding to the audio period. The compensation time is related to the current number of Dante flows, the number of channels, or the number of multicast receptions. , , , These are the weighting coefficients; Calculate target cache usage: .

4. The method for audio cross-clock domain transmission in an adaptive control network according to claim 1, characterized in that, The process of smoothing, limiting, and speed-limiting the ASRC target conversion ratio specifically includes the following steps: ASRC target conversion ratio Smoothing is performed to obtain the actual ASRC conversion ratio. : ,in, The smoothing coefficient is , and satisfies ; ASRC actual conversion ratio Limit the single change amount: ,in, This represents the maximum allowable change in the conversion ratio within a single control cycle. ASRC actual conversion ratio Set the valid range: , and These are the minimum and maximum conversion ratios supported by the ASRC hardware module, respectively.

5. The method for audio cross-clock domain transmission in an adaptive control network according to claim 1, characterized in that, The preset boundary conditions include any one or any combination of the following seven boundary conditions: I2S Period Boundary: Local I2S Sampling Frame Count satisfy ,in The number of sampled frames contained in a single I2S Period; I2S DMA buffer boundary: Current DMA read / write position satisfy ,in , This is the total sampling frame length of the I2S DMA buffer; ASRC FIFO safe zone: The number of valid samples in the ASRC input FIFO satisfy And the number of valid samples in the ASRC output FIFO satisfy ,in, , Input the minimum and maximum thresholds for the effective number of samples in the FIFO for ASRC. , Set the minimum and maximum thresholds for the effective number of samples in the ASRC output FIFO; Network audio frame boundaries; the current sampling frame is located at the beginning of the network audio frame, and the sampling offset within the network audio frame. satisfy ; Audio sampling frame boundary: The current PCM audio sample is located at the beginning of the audio sampling frame, and the channel offset of the current PCM audio sample within the audio sampling frame. satisfy , , This represents the total number of audio channels. The PCM audio sample number; Near zero crossover: Current output audio sample satisfy and ,in, For use in representing The sign function for positive and negative signs. Zero crossover amplitude threshold; Multi-channel group unified parameter switching point: The sampling points of all channels within the same control group reach the common allowed audio sampling frame boundary, I2S DMA buffer boundary, or network audio frame boundary.

6. The method for audio cross-clock domain transmission in an adaptive control network according to claim 5, characterized in that, Dynamically configure the priority of different boundary conditions based on the current scenario: In low-latency scenarios, I2S period boundaries or audio sampling frame boundaries should be preferred. In multi-channel phase-sensitive scenarios, the unified parameter switching point of the multi-channel group should be selected first; In audio-sensitive scenarios, prioritize selections near zero crossover. When the system load is high or the FIFO state is unstable, the ASRC FIFO safe zone should be selected first.

7. The method for audio cross-clock domain transmission in an adaptive control network according to claim 1, characterized in that, It also includes coherent control steps for multi-channel audio: Multiple audio channels that meet the preset association conditions are grouped into the same ASRC control group; A uniform ASRC conversion ratio is used for all channels within the same control group, and the hardware ASRC module parameters of all channels are updated synchronously when the preset boundary conditions are met. The preset association conditions include at least one of the following: belonging to the same Dante flow, belonging to the same network audio subscription relationship, belonging to the same PTP clock domain, having the same sampling rate, having the same network delay configuration, having a fixed channel phase relationship, belonging to the same physical audio interface or the same audio playback path.

8. An adaptive control network audio transmission device across clock domains, characterized in that, include: The network audio receiving and unpacking module is used to receive network audio data packets, unpack them to obtain PCM audio samples, packet sequence number, sampling sequence number, timestamp and audio channel information, and write the PCM audio samples into the network audio receiving buffer; The buffer feedback module is used to periodically detect the number of valid samples in the network audio receiving buffer and generate a feedback correction amount for the ASRC conversion ratio based on the deviation between the number of valid samples in the network audio receiving buffer and the target buffer occupancy. The ASRC feedforward estimation module is used to collect network time information, local I2S sampling count and local reference time information, estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generate a feedforward estimate of the ASRC conversion ratio. The ASRC conversion ratio control module is used to combine the feedforward estimate with the feedback correction to generate the ASRC target conversion ratio. The ASRC parameter smoothing and limiting module is used to smooth, limit and speed up the target ASRC conversion ratio to obtain the actual ASRC conversion ratio that can be directly written into the hardware ASRC module. The ASRC parameter update module is used to update the ASRC parameters of the hardware ASRC module under preset boundary conditions. The ASRC parameters include the actual ASRC conversion ratio. The process of acquiring network time information, local I2S sampling counts, and local reference time information to estimate the long-term frequency deviation between the network audio sampling clock domain and the local I2S sampling clock domain, and generating a feedforward estimate of the ASRC conversion ratio, specifically includes the following steps: Collect network time information, local I2S sampling count, and local reference time information; Within the k-th estimation window, the change in sample sequence number of the statistical network audio data is calculated. Changes in network audio timestamps And calculate the estimated sampling frequency on the network audio side: ; Within the k-th estimation window, count the change in local I2S sample counts. Change from local reference time And calculate the estimated sampling frequency on the local I2S side: ; Feedforward estimates of the ASRC conversion ratio: .

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the adaptive control network audio cross-clock domain transmission method as described in any one of claims 1 to 7.

Images