A four-speaker stereo audio optimization method and system for a tablet

Abstract

The application relates to the technical field of audio processing, and discloses a four-speaker stereo audio optimization method and system of a tablet computer, which comprises the following steps: acquiring stereo audio and collecting current impedance data of four speakers, matching the current impedance data with shielding calibration data, determining target speakers subjected to partial shielding and current acoustic transfer parameters, determining reference sound field contribution and actual sound field contribution, and obtaining to-be-compensated sound field contribution according to amplitude difference and phase difference, selecting at least two compensation speakers, adjusting multi-channel distribution relationship in combination with the to-be-compensated sound field contribution and the acoustic transfer parameters, and performing gain processing and phase processing on the stereo audio through a multi-channel sound effect engine to generate four-way speaker driving signals. The application reduces sound channel imbalance and sound image deviation caused by partial shielding and improves stereo sound field output stability.

Description

Technical Field

[0001] This invention relates to the field of audio processing technology, and more specifically, to a method and system for optimizing four-speaker stereo audio in a tablet computer. Background Technology

[0002] With the increasing use of tablets in audio-visual playback, gaming, and online video scenarios, external audio output has evolved from single-speaker and dual-speaker systems to multi-speaker collaborative output. Speakers positioned at different locations on the device create left and right channels and a spatial sound field to improve sound image width, directionality, and immersion. Due to the slim and lightweight design of tablets, speakers typically radiate sound through sound holes on the edges of the device. However, the user's hand, fingers, and contact surfaces can alter the sound output conditions of some speakers when holding the tablet horizontally, vertically, or on a table or stand. For four-speaker tablets, all speakers work together to form a stereo sound field. A change in the output state of one speaker not only affects the local loudness but can also affect the left and right channel balance, sound image position, and the overall multi-speaker synthesis effect. Therefore, maintaining stable four-speaker stereo audio output under changing usage conditions has become a practical technical requirement in multi-speaker tablet audio processing.

[0003] In the prior art, Chinese invention patent application CN114697855A - Multi-channel Car Audio System, discloses that in a multi-speaker system, when there is obstruction between one speaker and the listener, the surrounding speakers are used for directional compensation, and the playback effect after obstruction can be improved by adjusting the gain, frequency, and phase of the surrounding speakers. However, in the actual holding process of a tablet computer, the speaker outlet is more often partially obstructed by fingers, the edge of the palm, or the supporting surface, and the degree of obstruction changes with the holding position. Different obstruction states have different effects on the sound field formed by the four speakers working together. If a predetermined surrounding speaker compensation method is used only based on whether there is obstruction, it is difficult to fully reflect the actual sound changes of the affected speakers under different degrees of obstruction. This can easily cause the channel balance and spatial sound image to shift before and after compensation, especially when the obstruction state changes continuously, making it more difficult to maintain the stability of the stereo sound field.

[0004] Therefore, it is necessary to design a four-speaker stereo audio optimization method and system for tablet computers to solve the problems existing in the current technology. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for optimizing stereo audio of a tablet computer with four speakers, aiming to solve the problem of channel imbalance and sound image shift caused by changes in the actual sound contribution of the four speakers of a tablet computer under partially obscured conditions.

[0006] This invention proposes a method for optimizing four-speaker stereo audio in a tablet computer, comprising: Acquire stereo audio and collect the current impedance data of the four speakers; The current impedance data is matched with the occlusion calibration data to determine the target loudspeaker that is partially occluded, and the current acoustic transmission parameters of the target loudspeaker are obtained. The occlusion calibration data includes the correlation between impedance data and acoustic transmission parameters under different occlusion states. The reference sound field contribution is determined based on the stereo audio and the reference acoustic transmission parameters of the target loudspeaker in an unobstructed state. The actual sound field contribution is determined based on the stereo audio and the current acoustic transmission parameters. The amplitude difference and phase difference between the reference sound field contribution and the actual sound field contribution are determined. The sound field contribution to be compensated is obtained based on the amplitude difference and phase difference. Select at least two compensation speakers from the three speakers other than the target speaker, and adjust the multi-channel allocation relationship according to the contribution of the sound field to be compensated and the acoustic transfer parameters of the compensation speakers; The stereo audio is processed by a multi-channel sound engine according to the adjusted multi-channel allocation relationship, and the gain and phase are processed to generate four-way speaker drive signals to control the output of the four speakers.

[0007] Furthermore, establishing the occlusion calibration data includes: Each speaker is sequentially placed in an unobstructed state and multiple partially obstructed states arranged according to the degree of obstruction; impedance data is collected in the unobstructed state and each partially obstructed state, the same calibration audio is played, and amplitude and phase are collected at two reference listening positions to form acoustic transfer parameters; the correlation between impedance data and acoustic transfer parameters in each state is established.

[0008] Furthermore, obtaining the current acoustic transmission parameters includes: The current impedance data of the target loudspeaker is compared with the impedance data of each part of the occlusion state to determine the first matching occlusion state with the smallest impedance difference. Determine the second matching blockage state with the smallest impedance difference from the current impedance data from the adjacent partial blockage states of the first matching blockage state; Based on the impedance difference between the current impedance data and the first and second matched blocking states, the interpolation ratio is determined, and the acoustic transmission parameters of the first and second matched blocking states are interpolated to obtain the current acoustic transmission parameters.

[0009] Furthermore, obtaining the current acoustic transmission parameters also includes: When the first and second matching occlusion states used for interpolation switch as the current impedance data changes, the acoustic transfer parameters of the common occlusion states before and after the switch are retained; the interpolation ratio is re-determined based on the first and second matching occlusion states after the switch to obtain the continuously changing current acoustic transfer parameters.

[0010] Furthermore, when obtaining the contribution of the sound field to be compensated, it includes: The amplitude and phase of the reference acoustic transfer parameters and the current acoustic transfer parameters are obtained according to the same processing frequency band; the amplitude difference and phase difference between the reference sound field contribution and the actual sound field contribution in each processing frequency band are determined respectively, and the sound field contribution to be compensated in the current processing frequency band is determined according to the amplitude difference and phase difference.

[0011] Furthermore, when obtaining the contribution of the sound field to be compensated, it also includes: when the phase difference does not exceed a preset phase threshold, maintaining the current output of the target loudspeaker in the current processing frequency band; When the phase difference exceeds the preset phase threshold, reduce the current output of the target loudspeaker in the current processing frequency band and update the sound field contribution to be compensated in the current processing frequency band.

[0012] Furthermore, when updating the sound field contribution to be compensated, the following is included: After the target loudspeaker reduces its output in the current processing frequency band, the actual sound field contribution is updated based on the reduced target loudspeaker output and the current acoustic transfer parameters. The amplitude and phase differences between the reference sound field contribution and the updated actual sound field contribution are redefined, and the sound field contribution to be compensated in the current processing frequency band is updated based on the redefined amplitude and phase differences.

[0013] Furthermore, adjusting the multi-channel allocation relationship includes: decomposing the contribution of the sound field to be compensated into a compensation component undertaken by each compensation loudspeaker based on the acoustic transfer parameters of each compensation loudspeaker at two reference listening positions; determining the gain adjustment amount and phase adjustment amount of each compensation loudspeaker based on the amplitude and phase of each compensation component; and correcting the multi-channel allocation relationship using the gain adjustment amount and phase adjustment amount.

[0014] Furthermore, determining the gain adjustment amount and the phase adjustment amount includes: Using multiple sets of candidate gain adjustment and candidate phase adjustment, the contribution of the synthetic sound field formed by the compensation loudspeaker at two reference listening positions is calculated, and the amplitude error and phase error of the synthetic sound field contribution relative to the contribution of the sound field to be compensated are determined. A set of candidate gain adjustment and candidate phase adjustment values ​​is selected from those whose amplitude error does not exceed a preset amplitude error threshold and whose phase error does not exceed a preset phase error threshold at both reference listening positions, and these are used as the gain adjustment and phase adjustment values.

[0015] Compared with existing technologies, the advantages of this invention are as follows: By collecting the current impedance data of four loudspeakers to identify the partial occlusion state, and using occlusion calibration data to obtain the acoustic transmission parameters of the target loudspeaker under the current occlusion state, the impact of occlusion is transformed from a simple "whether occlusion occurs" judgment into acoustic parameters that can characterize the actual sound production state. Based on this, using the reference sound field contribution under the unoccluded state as a reference, the actual sound field contribution that can still be formed after the target loudspeaker is partially occluded is retained. The missing sound field contribution to be compensated is determined based on the amplitude and phase differences between the two. Then, at least two compensation loudspeakers are used to adjust the gain and phase through multi-channel allocation relationships. This avoids the repetitive compensation, channel imbalance, and sound image shift caused by directly switching loudspeakers or increasing the output of other loudspeakers in traditional occlusion compensation. The compensation amount is adjusted specifically according to the actual sound production changes of the target loudspeaker. Even with varying degrees of partial occlusion from fingers, palms, or supporting surfaces, a relatively stable amplitude and phase relationship between the four loudspeakers can be maintained, improving the continuity and spatial positioning stability of the stereo sound field during changes in occlusion state.

[0016] On the other hand, this application also provides a four-speaker stereo audio optimization system for a tablet computer, used to perform the aforementioned four-speaker stereo audio optimization method for a tablet computer, including: The data acquisition module is used to acquire stereo audio and collect the current impedance data of the four speakers; The occlusion analysis module is used to determine the target loudspeaker based on the current impedance data and occlusion calibration data, and to obtain the current acoustic transmission parameters of the target loudspeaker. The sound field contribution processing module is used to determine the reference sound field contribution, the actual sound field contribution, and the sound field contribution to be compensated based on the stereo audio, the reference acoustic transfer parameters, and the current acoustic transfer parameters. A multi-channel adjustment module is used to adjust the multi-channel allocation relationship according to the contribution of the sound field to be compensated and the acoustic transmission parameters of the compensation loudspeaker; The audio output module includes a multi-channel sound engine, which performs gain and phase processing on the stereo audio according to the adjusted multi-channel allocation relationship, generates four-speaker drive signals, and controls the output of the four speakers.

[0017] It is understandable that the aforementioned tablet computer's four-speaker stereo audio optimization method and system have the same beneficial effects, and will not be elaborated upon here. Attached Figure Description

[0018] The accompanying drawings are used to illustrate the technical solutions of the present invention and do not constitute a limitation on the scope of protection of the present invention. In the drawings: Figure 1 A flowchart of a four-speaker stereo audio optimization method for a tablet computer provided in an embodiment of the present invention; Figure 2 This is a flowchart of sound field contribution compensation under partially obstructed conditions provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the arrangement of four speakers and a reference listening position provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the speaker partial obstruction calibration state provided in an embodiment of the present invention; Figure 5 A functional block diagram of a four-speaker stereo audio optimization system for a tablet computer provided in an embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will now be described with reference to the accompanying drawings. These embodiments are used to illustrate the technical solutions of the present invention and do not limit the scope of protection of the present invention. Unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0020] In some embodiments of this application, see Figure 1-4 As shown, this application proposes a four-speaker stereo audio optimization method for a tablet computer, comprising: S100: Acquire stereo audio and collect current impedance data of the four speakers.

[0021] S200: Match the current impedance data with the occlusion calibration data to determine the target loudspeaker that is partially occluded, and obtain the current acoustic transmission parameters of the target loudspeaker. The occlusion calibration data includes the correlation between impedance data and acoustic transmission parameters under different occlusion states.

[0022] S300: Determines the reference sound field contribution based on the stereo audio and the reference acoustic transmission parameters of the target loudspeaker in an unobstructed state; determines the actual sound field contribution based on the stereo audio and the current acoustic transmission parameters; determines the amplitude difference and phase difference between the reference sound field contribution and the actual sound field contribution; and obtains the sound field contribution to be compensated based on the amplitude difference and phase difference.

[0023] S400: Select at least two compensation speakers from the three speakers other than the target speaker, and adjust the multi-channel allocation relationship according to the contribution of the sound field to be compensated and the acoustic transmission parameters of the compensation speakers.

[0024] S500: Through a multi-channel audio engine, the stereo audio is processed for gain and phase according to the adjusted multi-channel distribution relationship, generating four-speaker drive signals to control the output of the four speakers.

[0025] Specifically, the tablet features four independently driveable speakers, each positioned in a different location on the tablet. When playing stereo audio, the multi-channel audio engine processes the stereo audio into four speaker drive signals.

[0026] Each loudspeaker employs an audio amplifier capable of feeding back the drive voltage and current at the loudspeaker end. The drive voltage and current are synchronously acquired within the same sampling period, and the current impedance data is obtained based on the processing frequency band with effective audio excitation. In this embodiment, the current impedance data is constructed using the impedance amplitudes at several fixed frequency points within each processing frequency band. The same fixed frequency points are used in the obstruction calibration phase and the normal playback phase to ensure that the current impedance data and the obstruction calibration data have the same data caliber. When the loudspeaker's sound output area is obstructed, the acoustic load on the sound output side changes, and this change is reflected in the loudspeaker impedance through the diaphragm's mechanical load. Therefore, different obstruction states can create distinguishable impedance changes.

[0027] To eliminate impedance measurement errors caused by insufficient audio energy, a preset sampling energy threshold is determined when the tablet is in silent mode. Specifically, a 20-millisecond sampling period is used, and the noise floor is continuously sampled for 1 second. The 95th percentile value of the audio energy in each sampling period is used as the noise floor benchmark, and the preset sampling energy threshold is set by increasing the noise floor benchmark by 6 dB. If the audio energy of the currently processed frequency band is lower than the preset sampling energy threshold, the impedance data of the currently processed frequency band is not updated, and the most recently obtained valid impedance data that meets the sampling conditions continues to be used.

[0028] See Figure 3 As shown, the four speakers are positioned at different locations on the tablet's body, forming four separate sound output locations. To illustrate the sound field compensation process under partial obstruction conditions, Figure 3 The first speaker is designated as the target speaker, the second and third speakers as compensation speakers, and the fourth speaker remains in normal operation for stereo audio output. The identities of the first, second, third, and fourth speakers are only used to indicate one operating state. In actual operation, the target speaker is determined based on the current impedance data, and at least two compensation speakers are selected from the remaining three speakers. This embodiment addresses the issue of partial obstruction of one speaker. Since the other three speakers do not meet the target speaker determination criteria, the compensation speakers use pre-calibrated acoustic transmission parameters under unobstructed conditions. Figure 3 In the operating state shown, the second and third speakers are used as compensation speakers.

[0029] The tablet computer has a left and a right reference listening position on the front, and four speakers form an acoustic transmission relationship with the two reference listening positions respectively. Figure 3 The lines connecting each loudspeaker to the two reference listening positions represent the acoustic transmission relationship between each loudspeaker and the two reference listening positions, and do not represent fixed sound wave propagation paths. The left and right reference listening positions are used to characterize the left and right listening areas under normal viewing conditions, so that the subsequent reference sound field contribution, actual sound field contribution, and sound field contribution to be compensated are all determined using the same spatial reference.

[0030] The occlusion calibration data is pre-established during the acoustic calibration of the tablet computer. Specifically, four speakers are calibrated separately, while the other three speakers are excluded from the calibration playback for each speaker. First, the unobstructed state with the speaker apertures fully open is recorded. Then, an occlusion fixture is used to progressively cover the speaker apertures, creating multiple partially occluded states. These partially occluded states are arranged in order of decreasing effective open area of ​​the apertures. For example, unobstructed, approximately one-quarter occluded, approximately one-half occluded, and approximately three-quarter occluded states can be used as a set of calibration states. The occlusion ratio is only used to form calibration states with stable differences and does not require actual calculation of the occlusion area of ​​the apertures during operation.

[0031] See Figure 4 As shown, the same loudspeaker is configured with an unobstructed state, a first partially obstructed state, a second partially obstructed state, and a third partially obstructed state. These multiple partially obstructed states are set in ascending order of the degree of obstruction of the loudspeaker's sound-emitting area. In the unobstructed state, the loudspeaker's sound-emitting area remains open. The first, second, and third partially obstructed states respectively increase the obstruction range of the sound-emitting area, creating distinguishable acoustic load variations between each calibration state. Figure 4 The occlusion range in the figure represents different degrees of occlusion and does not limit the specific shape of the occlusion. During actual calibration, the effective open area of ​​the sound output region can be changed by using an occlusion fixture. Impedance data and acoustic transfer parameters are collected for each calibration state, so that impedance changes and acoustic changes at two reference listening positions under the same calibration state have a unified state basis. Subsequently, the current acoustic transfer parameters can be determined among multiple calibration states based on the current impedance data obtained during operation.

[0032] In each calibration state, the speaker being calibrated is driven using the same calibration audio, and the input amplitude and phase of the calibration audio in each processing frequency band are recorded, while impedance data is acquired synchronously. Calibration measurement microphones are set up at the left and right reference listening positions, respectively. The calibration audio input signal and the sampling of the two measurement microphones use the same clock reference, and the same calibration audio playback start point is used as the phase reference. The measurement microphones are only used to establish occlusion calibration data; sound field data is not required during normal tablet playback.

[0033] Two reference listening positions are set in front of the center of the tablet screen. These two positions represent the left and right listening areas under approximate normal viewing conditions. In one embodiment, the forward distance between the two reference listening positions and the center of the screen is set to approximately 400 mm, and the distance between the two reference listening positions is set to approximately 180 mm. The sound pressure amplitude and phase are collected at each of the two reference listening positions in each processing frequency band. The change in sound pressure amplitude relative to the input amplitude is used as the amplitude transfer parameter, and the change in sound pressure phase relative to the input phase is used as the phase transfer parameter. The amplitude transfer parameter and the phase transfer parameter together constitute the acoustic transfer parameters. The amplitude transfer parameter is represented by a linear amplitude transfer value in the sound field contribution calculation, and is converted to decibels when gain adjustment and amplitude error are involved. The phase transfer parameter is uniformly represented by angle. Therefore, for each speaker, a correlation between impedance data and acoustic transfer parameters under the same obstruction state can be established, forming obstruction calibration data.

[0034] During the unobstructed calibration process, at least 30 sets of effective impedance data were continuously acquired for each loudspeaker. The impedance amplitude at each fixed frequency point was normalized according to the average impedance amplitude under unobstructed conditions, and the difference between each set of impedance data and the average impedance data under unobstructed conditions was calculated. The impedance difference was represented by the average of the absolute values ​​of the normalized impedance amplitude differences at each fixed frequency point, and the 95th percentile of the 30 sets of impedance differences was used as the normal fluctuation threshold under unobstructed conditions.

[0035] When the tablet is playing normally, the impedance difference between the current impedance data of each of the four speakers and the average impedance data of the unobstructed state is calculated. If the impedance difference of a speaker does not exceed the normal fluctuation threshold for three consecutive valid sampling periods, the speaker is determined to remain in an unobstructed state. If the impedance difference exceeds the normal fluctuation threshold for three consecutive valid sampling periods, and the impedance difference between the current impedance data and at least one partially obstructed state is less than the impedance difference with the unobstructed state, the speaker is identified as the target speaker. If none of the four speakers meet the target speaker determination criteria, the original multi-channel allocation relationship is maintained. This embodiment will be explained below using the example of a speaker being partially obstructed.

[0036] After identifying the target loudspeaker, the impedance difference between the current impedance data of the target loudspeaker and the impedance data of each partial obstruction state is calculated based on the average of the absolute values ​​of the normalized impedance amplitude differences at each fixed frequency point. The partial obstruction state with the smallest impedance difference is determined as the first matched obstruction state. From the partial obstruction states adjacent to the first matched obstruction state, the partial obstruction state with the smallest impedance difference from the current impedance data is determined as the second matched obstruction state.

[0037] After determining the first and second matched blocking states, the interpolation ratio is determined based on the impedance differences between the current impedance data and the two matched blocking states. The smaller the impedance difference between the current impedance data and the first matched blocking state, the greater the interpolation weight of the first matched blocking state. The smaller the impedance difference between the current impedance data and the second matched blocking state, the greater the interpolation weight of the second matched blocking state, and the sum of the two interpolation weights is one. Specifically, the proportion of the impedance difference of the second matched blocking state to the sum of the two impedance differences can be used as the interpolation weight of the first matched blocking state, and the proportion of the impedance difference of the first matched blocking state to the sum of the two impedance differences can be used as the interpolation weight of the second matched blocking state. When the impedance difference between the current impedance data and a matched blocking state is zero, the acoustic transfer parameters of the matched blocking state are directly used. When the change in the current impedance data exceeds the calibration range of the heaviest part of the blocking state, the acoustic transfer parameters of the heaviest part of the blocking state are used, and no extrapolation beyond the calibration range is performed. The amplitude and phase transfer relationships of the first and second matched occlusion states are interpolated using two interpolation weights to obtain the current acoustic transfer parameters of the target loudspeaker. Before interpolating the phase transfer relationship, the phase spanning ±180 degrees is processed to ensure continuity and avoid phase value jumps that could lead to incorrect interpolation direction.

[0038] After obtaining the current acoustic transfer parameters, before the current sound field adjustment, the stereo audio components assigned to the target loudspeaker are determined according to the pre-adjustment multi-channel allocation. For each processed frequency band and each reference listening position, the stereo audio components are combined with the reference acoustic transfer parameters of the target loudspeaker in an unobstructed state to obtain the amplitude and phase in the unobstructed state, which serve as the reference sound field contribution. The same stereo audio components are combined with the current acoustic transfer parameters to obtain the amplitude and phase in the current partially obstructed state, which serve as the actual sound field contribution. Both the reference sound field contribution and the actual sound field contribution include the amplitude and phase at the left and right reference listening positions.

[0039] When calculating the amplitude difference, the amplitude difference between the reference sound field contribution and the actual sound field contribution at two reference listening positions is obtained. When calculating the phase difference, the minimum absolute phase difference between the reference sound field contribution and the actual sound field contribution is obtained after the phase is processed into a continuous curve, and the larger phase difference between the two reference listening positions is taken as the phase difference of the current processing frequency band. When determining the sound field contribution to be compensated, the reference sound field contribution and the actual sound field contribution at the same reference listening position are processed into a vector difference according to amplitude and phase, and the amplitude and phase of the vector difference are taken as the amplitude and phase of the sound field contribution to be compensated at the reference listening position, respectively.

[0040] Subsequently, the stereo audio is divided into multiple processing frequency bands. In one embodiment, within the effective playback range of the speaker, 200 Hz to less than 500 Hz is set as the low-frequency processing band, 500 Hz to less than 4000 Hz as the mid-frequency processing band, and 4000 Hz to 20000 Hz as the high-frequency processing band. The same processing frequency band boundaries are used during the occlusion calibration phase and the normal playback phase. For speakers with an effective playback range different from the example above, the processing frequency bands can be reset according to the measured effective frequency response of the speaker, but the processing frequency band boundaries remain unchanged for the same tablet computer during calibration and playback.

[0041] For each processed frequency band, the amplitude and phase differences between the reference sound field contribution and the actual sound field contribution at two reference listening positions are determined. The sound field contribution to be compensated is not a complete replication of the original audio components of the target loudspeaker, but rather the portion of the sound field contribution that should have been formed under unobstructed conditions but failed to form under the current partially obstructed conditions. Therefore, the effective sound field contribution that the target loudspeaker can still form under partially obstructed conditions is retained and not repeatedly generated by other loudspeakers.

[0042] At least two compensation speakers are selected from the remaining three speakers. Based on the acoustic transfer parameters from the compensation speakers to the two reference listening positions, the contribution of the sound field to be compensated in each processing frequency band is decomposed into compensation components borne by each compensation speaker. The gain adjustment and phase adjustment are determined based on the amplitude and phase of the compensation components. The multi-channel audio engine corrects the multi-channel allocation relationship based on the gain adjustment and phase adjustment, performs gain processing and phase processing on the stereo audio, generates four speaker drive signals, and controls the output of the four speakers.

[0043] In this embodiment, the occlusion calibration data is not merely used to determine whether a speaker is occluded, but rather to convert the impedance changes obtainable during operation into the acoustic transmission relationships that still actually exist in the target speaker under the current partially occluded state. This transforms the subsequent compensation target from the original channel of the occluded speaker to the missing part between the unoccluded sound field and the actual residual sound field. Therefore, under the partially occluded state where the target speaker still has effective output, the sound field contribution already formed by the target speaker can be preserved, reducing the redundant superposition caused by directly copying the channel or increasing the overall gain of other speakers. This reduces the possibility of sound image shift and local frequency enhancement or cancellation from the source of sound field formation.

[0044] Specifically, when a user holds a tablet, moving their finger along the sound outlet may cause the target speaker's current impedance data to gradually change from one set of adjacent partial occlusion states to the next set of adjacent partial occlusion states. For example, when the current interpolation uses a first and a second partial occlusion state, the first partial occlusion state is designated as the first matched occlusion state, and the second partial occlusion state is designated as the second matched occlusion state. As the degree of occlusion increases, the two states used for interpolation switch to the second and a third partial occlusion state. At this point, the second partial occlusion state is a partial occlusion state shared before and after the switch. The acoustic transfer parameters of the second partial occlusion state are retained, and only the other matched occlusion state is switched to the third partial occlusion state. The interpolation ratio is then re-determined based on the current impedance data. This ensures that the current acoustic transfer parameters continuously transition from one interpolation result to the next when crossing two calibration intervals, avoiding abrupt changes caused by directly switching calibration parameters.

[0045] For each processing frequency band, this embodiment further utilizes the phase difference to determine whether the residual output of the target loudspeaker after partial obstruction continues to participate in sound field formation.

[0046] The preset phase threshold is determined during the prototype calibration phase. For each processing frequency band, the degree of obstruction of the target loudspeaker is gradually increased while maintaining the output of the target loudspeaker and the other three loudspeakers according to the multi-channel allocation relationship before adjustment. The synthesized sound field error of the two reference listening positions relative to the unobstructed state is calculated. In this embodiment, 1.5 dB is used as the preset amplitude error threshold and 15 degrees is used as the preset phase error threshold. Both thresholds are used simultaneously for the calibration of the preset phase threshold and the subsequent selection of candidate parameters. When the amplitude error of any reference listening position exceeds the preset amplitude error threshold for the first time, or the phase error exceeds the preset phase error threshold for the first time, the phase difference obtained under the previous obstruction state is used as the preset phase threshold for the current processing frequency band. If no error exceeds the limit in all calibration states, the phase difference under the most severe obstruction state is used as the preset phase threshold.

[0047] Different preset phase thresholds can be set for different processing frequency bands. For example, in a prototype tuning scheme, a larger phase change is allowed in the low-frequency processing band, while a smaller phase threshold is used in the mid-to-high frequency processing band. As an example of implementation parameters, the preset phase threshold can be set to 60 degrees for the low-frequency processing band, 45 degrees for the mid-frequency processing band, and 30 degrees for the high-frequency processing band. The above values ​​are used to illustrate the threshold adaptation method; the actual values ​​are determined through calibration based on the specific speaker location, cavity structure, and sound outlet layout.

[0048] If the phase difference in the current processing frequency band does not exceed the preset phase threshold, the current output of the target loudspeaker in the current processing frequency band is maintained, allowing the residual sound field contribution already formed by the target loudspeaker to continue participating in the four-loudspeaker synthesis. The sound field contribution to be compensated only represents the part of the sound field not formed by the target loudspeaker. When the phase difference exceeds the preset phase threshold, it indicates that although the residual output of the target loudspeaker in the current processing frequency band still has a certain amplitude, its phase relationship with the unobstructed state has deviated significantly. Maintaining the original output may aggravate the coherent cancellation or sound image shift between the target loudspeaker and the compensation loudspeaker, thus reducing the output of the target loudspeaker in the current processing frequency band.

[0049] The target loudspeaker employs multiple pre-set attenuation levels. The phase difference of the current processing frequency band is subtracted from a preset phase threshold to obtain the phase excess. When the phase excess is no greater than 10 degrees, the output of the target loudspeaker in the current processing frequency band is reduced by 3 dB. When the phase excess is greater than 10 degrees but no greater than 20 degrees, it is reduced by 6 dB. When the phase excess is greater than 20 degrees, it is reduced by 12 dB. The attenuation is gradually applied between adjacent audio frames until the corresponding attenuation level is reached. After the target loudspeaker output is reduced, the previously calculated actual sound field contribution becomes invalid. Therefore, the actual sound field contribution is re-determined based on the reduced target loudspeaker output and the current acoustic transfer parameters. The amplitude and phase differences between the baseline sound field contribution and the updated actual sound field contribution are then re-determined to update the sound field contribution to be compensated in the current processing frequency band. This avoids insufficient or excessive compensation caused by continuing to use the sound field contribution before the target loudspeaker was reduced.

[0050] For each processing frequency band, the amplitude transfer parameters from each compensated loudspeaker to the left and right reference listening positions are read separately. The smaller amplitude transfer parameter between the two reference listening positions characterizes the effective transfer capability of the compensated loudspeaker, and the initial allocation ratio of the sound field contribution to be compensated is determined according to the ratio between the effective transfer capabilities of each compensated loudspeaker. The compensation component of each compensated loudspeaker is obtained based on the initial allocation ratio. Based on the amplitude of the compensation component and the effective transfer capability of the compensated loudspeaker, the output amplitude required for the compensated loudspeaker to form the compensation component is determined, and the gain change of the required output amplitude relative to the output amplitude under the original multi-channel allocation relationship is used as the initial gain adjustment amount. For phase, based on the phase adjustment values ​​required for the compensated loudspeaker to form the sound field contribution to be compensated at the two reference listening positions, the two phase adjustment values ​​are continuously processed and then averaged to serve as the initial phase adjustment amount for the current compensated loudspeaker.

[0051] Candidate gain adjustments are generated within a 6 dB range above and below the initial gain adjustment, in 0.5 dB increments. Similarly, candidate phase adjustments are generated within a 30-degree range before and after the initial phase adjustment, in 5-degree increments. For each set of candidate parameters, the output of each compensated loudspeaker is determined based on the candidate gain and phase adjustments. Then, combined with the acoustic transfer parameters of each compensated loudspeaker at two reference listening positions, the sound field contribution of each loudspeaker is determined. Finally, these contributions are synthesized according to amplitude and phase to obtain the combined sound field contribution at the two reference listening positions.

[0052] Candidate parameters are selected using the aforementioned preset amplitude error threshold and preset phase error threshold. If multiple sets of candidate parameters meet the thresholds, the set with the smallest sum of absolute gain changes of each compensated speaker relative to the original multi-channel distribution is selected first. If multiple sets still exist, the set with the smallest sum of absolute phase changes of each compensated speaker is selected.

[0053] If no parameter combination in the current candidate range satisfies all error thresholds, the search step size remains unchanged at 0.5 dB and 5 degrees. Each time, the upper and lower limits of the gain candidate range are increased by 2 dB, and the front and rear boundaries of the phase candidate range are increased by 10 degrees, until the speaker output range allowed by the audio amplifier and the phase adjustment range allowed by the multi-channel audio engine are reached. If no parameter combination still satisfies the conditions after reaching the allowed range, the amplitude error at each of the two reference listening positions is divided by a preset amplitude error threshold, and the phase error is divided by a preset phase error threshold. The four normalized errors are summed, and the candidate parameter set with the smallest sum of normalized errors is selected.

[0054] This embodiment does not simply aim for smooth parameter changes when the degree of occlusion changes. Instead, it maintains a common calibration state as a continuous basis during interpolation state switching, and determines whether to continue using the existing output of the target loudspeaker based on the phase deviation of the residual sound field. This avoids discrete jumps in compensation parameters when the degree of occlusion crosses the calibration state, and also prevents residual sound with significant phase deviation from continuing to superimpose with the compensation sound. This embodiment addresses two practical processes: continuous changes in occlusion and the presence of sound pressure levels that are no longer suitable for continued use in terms of phase, reducing the risk of sudden sound image jumps and local coherent cancellation.

[0055] In one example embodiment, a tablet computer has a speaker positioned near each of its four corners, with a normal viewing distance of approximately 400 mm. A left and right reference listening position are located in front of the center of the screen, approximately 180 mm apart. Four calibration states are established for each speaker: unobstructed, 25% obstructed, 50% obstructed, and 75% obstructed. The obstruction ratio is determined by the proportion of the projected area of ​​the speaker's sound-emitting area covered by the obstruction fixture to the total area of ​​the speaker's sound-emitting area. The obstruction fixture gradually increases the coverage area along the same direction. The same broadband calibration audio is played in each state, and the amplitude and phase of each processed frequency band are obtained at the two reference listening positions, along with speaker impedance data. The effective audio frequency range is divided into three processed frequency bands: low frequency, mid frequency, and high frequency. The specific frequency boundaries are determined based on the actual effective frequency response of the speakers, without requiring the four speakers to use different frequency band boundaries.

[0056] During a lateral grip test, the current impedance data of the upper left speaker exceeded the normal fluctuation range of the unobstructed state. Comparing the current impedance data with the impedance data of the upper left speaker under various obstructed states, the impedance difference was smallest in the 50% obstruction state, which was then identified as the first matched obstruction state. Among the adjacent 25% and 75% obstruction states, the impedance difference between the 75% obstruction state and the current impedance data was smaller, thus the 75% obstruction state was identified as the second matched obstruction state. Assuming the interpolation ratios obtained from the differences between the two states and the current impedance data are 0.7 and 0.3 respectively, the current acoustic transfer parameters are obtained using the acoustic transfer parameters of the 50% obstruction state (70%) and the 75% obstruction state (30%). The phase parameters are converted to a continuous phase change form before interpolation to avoid interpolation direction errors caused by the ±180 degree boundary.

[0057] Subsequently, the stereo audio components currently handled by the upper left speaker are combined with the reference acoustic transfer parameters in the unobstructed state and the interpolated current acoustic transfer parameters to calculate the reference sound field contribution and the actual sound field contribution at the two reference listening positions. Taking the mid-frequency processing band as an example, if the phase difference between the actual sound field contribution and the reference sound field contribution is 28 degrees, and the preset phase threshold of the mid-frequency processing band is 45 degrees, then the current output of the upper left speaker in the mid-frequency processing band is maintained, and only the part missing between the reference sound field contribution and the actual sound field contribution is taken as the sound field contribution to be compensated. If the phase difference in the high-frequency processing band reaches 42 degrees, which is higher than the preset phase threshold of 30 degrees for the high-frequency processing band, then the output of the upper left speaker in the high-frequency processing band is reduced, for example, by 6 dB, and then the actual sound field contribution and the sound field contribution to be compensated in the high-frequency processing band are re-determined based on the reduced output.

[0058] Of the remaining three speakers, the upper right speaker and the lower left speaker can be selected as compensation speakers. Based on the acoustic transfer parameters from the two compensation speakers to the left and right reference listening positions, the contribution of the sound field to be compensated is decomposed into two compensation components, which are then used to form candidate gain adjustment and candidate phase adjustment. The synthetic sound field contribution of the left and right reference listening positions is calculated for each set of candidate parameters. If a set of candidate parameters results in an amplitude error of 1.0 dB and a phase error of 9 degrees for the left reference listening position, and an amplitude error of 1.2 dB and a phase error of 11 degrees for the right reference listening position, and all four errors are within the allowable range when the preset amplitude error threshold is 1.5 dB and the preset phase error threshold is 15 degrees, then this set of candidate gain adjustment and candidate phase adjustment is used to correct the current multi-channel allocation relationship. The multi-channel audio engine processes the stereo audio according to the corrected multi-channel allocation relationship and generates four speaker drive signals.

[0059] If the user continues to move their finger, gradually shifting the current impedance data of the upper left speaker from the 50% to 75% obstruction range to around 75%, the adjacent calibration states and interpolation ratios are redefined based on the new impedance data. During interpolation state switching, the shared calibration states before and after the switch are retained, ensuring continuous updates to the current acoustic transfer parameters, while gradually updating the gain and phase adjustments. Therefore, the entire process does not require real-time identification of finger shape, palm position, or calculation of the actual obstruction area; it only uses the speaker's own electroacoustic state changes and pre-established acoustic calibration relationships to complete the sound field adjustment.

[0060] This embodiment can cover a large number of continuously changing grip states in actual use with a limited set of partially occluded calibration states. It does not require establishing separate rules for each gesture, nor does it require additional hand visual recognition or complex environmental perception processes. Furthermore, the compensation result is jointly filtered by the actual amplitude and phase constraints of two reference listening positions, rather than compensating solely based on the volume change of a single speaker. Therefore, while maintaining the processing chain suitable for real-time execution on existing tablet audio hardware, it can more directly maintain the sound field relationship between the left and right listening positions, making the solution feasible, calibrable, and with a low computational burden.

[0061] Based on another preferred embodiment described above, see [link to preferred embodiment]. Figure 5 As shown, this embodiment provides a four-speaker stereo audio optimization system for a tablet computer, used to execute the aforementioned four-speaker stereo audio optimization method for a tablet computer, including: The data acquisition module is used to acquire stereo audio and collect the current impedance data of the four speakers; The occlusion analysis module is used to determine the target loudspeaker based on the current impedance data and occlusion calibration data, and to obtain the current acoustic transmission parameters of the target loudspeaker. The sound field contribution processing module is used to determine the reference sound field contribution, the actual sound field contribution, and the sound field contribution to be compensated based on the stereo audio, the reference acoustic transfer parameters, and the current acoustic transfer parameters. The multi-channel adjustment module is used to adjust the multi-channel allocation relationship based on the contribution of the sound field to be compensated and the acoustic transmission parameters of the compensated loudspeakers; The audio output module includes a multi-channel sound engine, which performs gain and phase processing on stereo audio according to the adjusted multi-channel allocation relationship, generates four-speaker drive signals, and controls the output of the four speakers.

[0062] When the above system executes the aforementioned method, it can achieve the same technical effect as the aforementioned method.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for optimizing stereo audio in a tablet computer using four speakers, characterized in that, include: Acquire stereo audio and collect the current impedance data of the four speakers; The current impedance data is matched with the occlusion calibration data to determine the target loudspeaker that is partially occluded, and the current acoustic transmission parameters of the target loudspeaker are obtained. The occlusion calibration data includes the correlation between impedance data and acoustic transmission parameters under different occlusion states. The reference sound field contribution is determined based on the stereo audio and the reference acoustic transmission parameters of the target loudspeaker in an unobstructed state. The actual sound field contribution is determined based on the stereo audio and the current acoustic transmission parameters. The amplitude difference and phase difference between the reference sound field contribution and the actual sound field contribution are determined. The sound field contribution to be compensated is obtained based on the amplitude difference and phase difference. Select at least two compensation speakers from the three speakers other than the target speaker, and adjust the multi-channel allocation relationship according to the contribution of the sound field to be compensated and the acoustic transfer parameters of the compensation speakers; The stereo audio is processed by a multi-channel sound engine according to the adjusted multi-channel allocation relationship, and the gain and phase are processed to generate four-way speaker drive signals to control the output of the four speakers.

2. The method according to claim 1, characterized in that, When establishing the occlusion calibration data, the following steps are included: Each speaker is sequentially placed in an unobstructed state and multiple partially obstructed states arranged according to the degree of obstruction; impedance data is collected in the unobstructed state and each partially obstructed state, the same calibration audio is played, and amplitude and phase are collected at two reference listening positions to form acoustic transfer parameters; the correlation between impedance data and acoustic transfer parameters in each state is established.

3. The method according to claim 2, characterized in that, Obtaining the current acoustic transfer parameters includes: The current impedance data of the target loudspeaker is compared with the impedance data of each part of the occlusion state to determine the first matching occlusion state with the smallest impedance difference. Determine the second matching blockage state with the smallest impedance difference from the current impedance data from the adjacent partial blockage states of the first matching blockage state; Based on the impedance difference between the current impedance data and the first and second matched blocking states, the interpolation ratio is determined, and the acoustic transmission parameters of the first and second matched blocking states are interpolated to obtain the current acoustic transmission parameters.

4. The method according to claim 3, characterized in that, When obtaining the current acoustic transfer parameters, the method further includes: When the first and second matching occlusion states used for interpolation switch as the current impedance data changes, the acoustic transfer parameters of the common occlusion states before and after the switch are retained; the interpolation ratio is re-determined based on the first and second matching occlusion states after the switch to obtain the continuously changing current acoustic transfer parameters.

5. The method according to claim 1, characterized in that, When obtaining the sound field contribution to be compensated, it includes: The amplitude and phase of the reference acoustic transfer parameters and the current acoustic transfer parameters are obtained according to the same processing frequency band; the amplitude difference and phase difference between the reference sound field contribution and the actual sound field contribution in each processing frequency band are determined respectively, and the sound field contribution to be compensated in the current processing frequency band is determined according to the amplitude difference and phase difference.

6. The method according to claim 5, characterized in that, When obtaining the contribution of the sound field to be compensated, it also includes: when the phase difference does not exceed the preset phase threshold, maintaining the current output of the target loudspeaker in the current processing frequency band; When the phase difference exceeds the preset phase threshold, reduce the current output of the target loudspeaker in the current processing frequency band and update the sound field contribution to be compensated in the current processing frequency band.

7. The method according to claim 6, characterized in that, When updating the sound field contribution to be compensated, the following should be included: After the target loudspeaker reduces its output in the current processing frequency band, the actual sound field contribution is updated based on the reduced target loudspeaker output and the current acoustic transfer parameters. The amplitude and phase differences between the reference sound field contribution and the updated actual sound field contribution are redefined, and the sound field contribution to be compensated in the current processing frequency band is updated based on the redefined amplitude and phase differences.

8. The method according to claim 1, characterized in that, Adjusting the multi-channel allocation relationship includes: decomposing the contribution of the sound field to be compensated into a compensation component undertaken by each compensation loudspeaker based on the acoustic transfer parameters of each compensation loudspeaker at two reference listening positions; determining the gain adjustment amount and phase adjustment amount of each compensation loudspeaker based on the amplitude and phase of each compensation component; and correcting the multi-channel allocation relationship using the gain adjustment amount and phase adjustment amount.

9. The method according to claim 8, characterized in that, Determining the gain adjustment and phase adjustment includes: Using multiple sets of candidate gain adjustment and candidate phase adjustment, the contribution of the synthetic sound field formed by the compensation loudspeaker at two reference listening positions is calculated, and the amplitude error and phase error of the synthetic sound field contribution relative to the contribution of the sound field to be compensated are determined. A set of candidate gain adjustment and candidate phase adjustment values ​​is selected from those whose amplitude error does not exceed a preset amplitude error threshold and whose phase error does not exceed a preset phase error threshold at both reference listening positions, and these are used as the gain adjustment and phase adjustment values.

10. A four-speaker stereo audio optimization system for a tablet computer, used to perform the four-speaker stereo audio optimization method for a tablet computer as described in any one of claims 1 to 9, characterized in that, include: The data acquisition module is used to acquire stereo audio and collect the current impedance data of the four speakers; The occlusion analysis module is used to determine the target loudspeaker based on the current impedance data and occlusion calibration data, and to obtain the current acoustic transmission parameters of the target loudspeaker. The sound field contribution processing module is used to determine the reference sound field contribution, the actual sound field contribution, and the sound field contribution to be compensated based on the stereo audio, the reference acoustic transfer parameters, and the current acoustic transfer parameters. A multi-channel adjustment module is used to adjust the multi-channel allocation relationship according to the contribution of the sound field to be compensated and the acoustic transmission parameters of the compensation loudspeaker; The audio output module includes a multi-channel sound engine, which performs gain and phase processing on the stereo audio according to the adjusted multi-channel allocation relationship, generates four-speaker drive signals, and controls the output of the four speakers.

Images