Layered soundscape healing method for elderly emotional and sleep intervention

Abstract

The application relates to a layered soundscape healing method for emotion and sleep intervention for the elderly, which comprises the following steps: when the elderly user is in a night sleep state, in a first period, a soundscape intervention audio signal obtained by mixing each target audio signal in an audio signal set according to the gain of the first period is played; in a second period after the first period, a soundscape intervention audio signal obtained by mixing each target audio signal according to the gain of the second period is played; when it is determined that the elderly user is in a sleep state in the second period, in a third period after the second period, a soundscape intervention audio signal obtained by mixing each target audio signal according to the gain of the third period is played; in a fourth period after the third period, a soundscape intervention audio signal obtained by mixing each target audio signal according to the gain of the fourth period is played.

Description

Technical Field

[0001] This application belongs to the fields of audio processing and smart healthcare technology, specifically relating to a layered soundscape therapy method for emotional and sleep intervention in the elderly. Background Technology

[0002] In related technologies, soundscape therapy involves playing pre-set music. Currently, there is a lack of soundscape therapy methods specifically for elderly users. Summary of the Invention

[0003] This application provides a layered soundscape therapy method for emotional and sleep intervention in the elderly, in order to solve the problems in related technologies.

[0004] This application provides a layered soundscape therapy method for emotional and sleep intervention in the elderly, including: When elderly users are asleep at night, during the first period, a soundscape intervention audio signal is played, which is obtained by mixing the gain of each target audio signal in the audio signal set during the first period. The audio signal set includes the following target audio signals: natural ambient sound layer audio signal, rhythm guidance layer audio signal, background noise layer audio signal, environmental masking compensation layer audio signal, and voice soothing layer audio signal. During the second period following the first period, a soundscape intervention audio signal is played, which is obtained by mixing the gain of each target audio signal in the second period. During the second period, the beats per minute (bpm) of the rhythm guidance layer audio signal is reduced at a preset first bpm reduction rate. When it is determined that the elderly user is asleep during the second time period, in the third time period after the second time period, the soundscape intervention audio signal obtained by mixing the gain of each target audio signal in the third time period is played. In the third time period, the gain of the natural ambient sound layer audio signal, the rhythm guidance layer audio signal, the environmental masking compensation layer audio signal, and the background noise layer audio signal fades out in the third time period. During the fourth period following the third period, a soundscape intervention audio signal generated based on the gain of each target audio signal in the fourth period is played, wherein the gain of the background noise layer audio signal fades out in the fourth period.

[0005] This application provides a layered soundscape therapy player for emotional and sleep intervention in the elderly, including a processor and a memory. The memory stores a computer program; when the computer program is loaded by the processor, the processor executes the method provided in this application.

[0006] Beneficial effects: The method provided in this application embodiment can realize soundscape therapy for elderly users.

[0007] The method provided in this application takes into account the characteristics of elderly users, such as needing music assistance to fall asleep at night, being prone to waking up at night, and having high blood pressure. In multiple modes related to the characteristics of elderly users, such as nighttime sleep mode, nighttime wake-up and fall asleep mode, blood pressure relief mode, anxiety relief mode, and noise adaptation mode, corresponding sound scene intervention audio signals can be generated for sound scene therapy.

[0008] The method provided in this application utilizes multiple signals to generate soundscape intervention audio signals, including natural ambient sound layer audio signals for reducing user anxiety in the medical environment, ambient masking compensation layer audio signals for providing targeted masking sounds based on the ambient noise spectrum, rhythm guidance layer audio signals for guiding the user's breathing rate and heart rate to decrease synchronously through regular beats, and voice soothing layer audio signals for cognitive guidance and emotional calming. By comprehensively considering factors affecting the effectiveness of soundscape therapy, better soundscape therapy results can be achieved.

[0009] The method provided in this application embodiment involves reducing the beats per minute (bpm) of the rhythm-guided audio signal at a preset first bpm reduction rate during a second time period when the elderly user is asleep at night. This reduces both the elderly user's respiratory rate and heart rate, thereby achieving a soundscape therapy effect.

[0010] The method provided in this application embodiment involves reducing the bpm of the rhythm-guided layer audio signal at a preset second bpm reduction rate during the fifth time period when an elderly user is awake at night and then falling back asleep. This reduces both the elderly user's respiratory rate and heart rate, thereby achieving a soundscape therapy effect.

[0011] The method provided in this application allows the gain of the audio signal in the environmental masking compensation layer to be dynamic and adaptively adjusted according to the current environmental noise spectrum. Based on spectral analysis results, the masking compensation layer enhances the natural sound in the same frequency band where noise energy is concentrated, thereby improving the noise masking effect.

[0012] The method provided in this application embodiment has a greater gain in the ninth time period than the gain in the ninth time period of the component of the audio signal of the environmental masking compensation layer in the frequency band other than the target energy frequency band of the target environmental noise. This is equivalent to increasing the gain in the ninth time period of the component of the audio signal of the environmental masking compensation layer in the target energy frequency band of the target environmental noise, relative to the gain in the ninth time period of the component of the audio signal of the environmental masking compensation layer in the frequency band other than the target energy frequency band of the target environmental noise. For the frequency band with strong energy of the target environmental noise component, i.e. the target energy frequency band, the volume of the naturally generated noise used to mask the target energy frequency band can be increased, thereby improving the noise masking effect.

[0013] The method provided in this application provides a gentler and safer therapeutic soundscape output. By limiting peak volume (nighttime peak ≤ 45 dB SPL, daytime peak ≤ 55 dB SPL, both below the WHO-recommended nighttime noise exposure limit of 50 dB SPL), constraining the frequency band (high-pass cutoff frequency ≥ 80 Hz to remove low-frequency noise and vibration, low-pass cutoff frequency ≤ 8 kHz to protect the damaged high-frequency hearing of elderly users), smoothing the transition slope (fade-in time 3-5 seconds using a logarithmic curve, fade-out time 5-10 seconds, cross-fade mechanism to avoid abrupt changes in sound image during layer transitions), and compressing the dynamic range (compression ratio ≤ 2:1 to avoid sudden volume changes), the interference of sudden high-pitched sounds, sharp high frequencies, and abrupt sound changes on the emotions and sleep of the elderly is significantly reduced. These parameter constraints are based on research findings in geriatric audiology, rather than simple general audio playback settings.

[0014] The method provided in this application can provide targeted intervention based on different states. It can automatically select different soundscape templates and parameters according to different scenarios such as before falling asleep, waking up at night, calming down in the morning, daytime anxiety, and low mood, thereby improving the targeted nature of its use. For example, in the "Sleep Induction" template, the natural ambient sound layer gain is 60%, the background noise layer is 40%, and the rhythm guidance layer is 50% (the frequency gradually decreases from 12 bpm to 6 bpm). Through the synergistic effect of multiple layers, the system guides the user from a waking state to a sleep state. In the "Night Wake-up and Re-sleep" template, after the system detects an abnormal increase in body movement frequency, it first uses a gradual confirmation mechanism (increasing the background sound gain by 10%) to avoid false triggering. After 30 seconds of no active operation, it automatically switches to the Re-sleep template, where the rhythm guidance frequency gradually decreases from 10 bpm to 5 bpm. The system guides the user into a deep and slow breathing state through a beat slower than the normal breathing rate. In the "Blood Pressure Relief" template, when the blood pressure data received via Bluetooth reaches the warning threshold, the system automatically increases the rhythm guidance layer gain to 70% and fixes it at 6 bpm (corresponding to deep and slow breathing), while reducing the sudden sound components in the voice soothing layer and the natural ambient sound layer. The system uses soundscape intervention to assist the autonomic nervous system in regulation.

[0015] The method provided in this application reduces disturbance to family members and the surrounding environment. Through near-field directional sound generation (effective sound field range 10-30cm, directivity index ≥6dB), pillow-side sound field focusing (using a flat edge structure and near-field speaker layout to concentrate sound in the pillow area), or neck-mounted near-ear output (speakers located 10-15cm in front of the ear, no need to insert into the ear), the sound is primarily concentrated near the user's head and neck, avoiding interference caused by traditional external speakers to roommates. Actual measurements show that when the sound pressure level is 45dB SPL at 30cm from the device, it attenuates to below 30dB SPL at 2m from the device, lower than the general background noise level, and will not significantly affect the normal sleep of roommates.

[0016] The method provided in this application can be extended to access physiological or behavioral state information. Through the Bluetooth BLE (Bluetooth Low Energy) protocol, it can connect to blood pressure monitors, heart rate monitoring devices, motion sensors, sleep monitoring devices, or environmental noise sensors, making sound intervention strategies more intelligent and personalized. For example, when the Bluetooth blood pressure monitor transmits a systolic blood pressure data of 162 mmHg at 3:00 PM, the system determines it to be a high-risk state for hypertension and automatically switches to the "blood pressure diastolic template"; when the sleep monitoring device detects that the user has entered REM sleep, the system automatically reduces the soundscape intensity and prepares for fading out; when the environmental noise sensor detects sudden noise in the 500-2000Hz frequency band (such as construction noise outside the window), the system activates the environmental masking compensation layer, specifically amplifying low-stimulation natural sounds in the same frequency band (such as the sound of a stream), rather than increasing the overall volume. This multi-source data fusion capability distinguishes this invention from all existing fixed audio playback devices.

[0017] The method provided in this application embodiment is highly age-friendly. Through multiple age-friendly designs such as one-button start, large buttons, voice broadcast, memory of frequently used modes (automatically restores the last used mode upon startup), and nighttime accidental touch lock (automatically locks after 30 seconds of entering night mode, retaining only the "emergency stop" and "volume increase" functions), the operation difficulty for elderly users is significantly reduced.

[0018] The method provided in this application has a wide range of applications. It can be used in scenarios such as home bedrooms, living rooms, elderly care facilities, rehabilitation centers, and nap rooms, and is suitable for sleep assistance, mood stabilization, relaxation, and stress relief. In home settings, the elderly can use the device independently to help them fall asleep and fall back asleep after waking up at night. In elderly care facilities, caregivers can provide personalized soundscape services to multiple elderly people through the neck-hanging or backrest attachment structure. In rehabilitation centers, the device can be used simultaneously with rehabilitation training, assisting in breathing training and improving heart rate variability (HRV) through a rhythm-guided layer. Furthermore, since the device does not rely on a mobile app, it can operate independently in environments without network connectivity, making it suitable for elderly care facilities in remote areas or with poor network conditions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings.

[0020] Figure 1 This is a flowchart illustrating a layered soundscape therapy method for emotional and sleep intervention in the elderly, as provided in an embodiment of this application. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. It should be noted that, unless otherwise specified, the implementation methods and features in the implementation methods in this disclosure can be combined, separated, interchanged, and / or rearranged. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] The terminology used herein is for the purpose of describing particular embodiments and is not intended to be limiting. As used herein, unless the context clearly indicates otherwise, the singular forms “a” and “the” are intended to include the plural forms as well. Furthermore, when the terms “comprising” and / or “including” and variations thereof are used in this specification, it indicates the presence of the stated features, integrals, steps, operations, parts, components, and / or groups thereof, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, parts, components, and / or groups thereof. It should also be noted that, as used herein, the terms “substantially,” “about,” and other similar terms are used as approximate terms rather than as terms of degree, thus explaining the inherent biases in measurements, calculated values, and / or provided values ​​that would be recognized by one of ordinary skill in the art.

[0023] In the embodiments of this application, any method can be used to process a given audio signal to generate a playback audio signal to generate a soundscape intervention audio signal based on the gain of the audio signal.

[0024] In this embodiment, each audio track in the audio signal set is a digital audio source used for mixing. Each audio track in the audio signal set can be used as an audio track for mixing. The file of each audio track in the audio signal set can be a PCM / WAV file stored in the storage module.

[0025] In this embodiment, the audio signal for sound scene intervention is a digital signal.

[0026] It should be noted that, in this embodiment, the playback of the soundscape intervention audio signal by the player's speaker can be understood as playing the sound represented by the soundscape intervention audio signal. Any existing audio processing method for playing sound based on a given digital audio signal can be used to generate a sound signal that can be played by the speaker based on the soundscape intervention audio signal.

[0027] In this application embodiment, the audio signal of the natural ambient sound layer, the audio signal of the rhythm guidance layer, the audio signal of the background noise layer, the audio signal of the environmental masking compensation layer, and the audio signal of the voice soothing layer can be referred to as the natural ambient sound layer, the rhythm guidance layer, the background noise layer, the environmental masking compensation layer, and the voice soothing layer, respectively.

[0028] In the embodiments of this application, the nighttime sleep mode, the nighttime wake-up and fall back asleep mode, the blood pressure relief mode, the anxiety relief mode, the noise adaptation mode, the daytime soothing mode, and the emotion calming mode are all soundscape intervention modes.

[0029] The player used to execute the embodiments of this application is referred to as the player.

[0030] The player can operate in the following modes: low power standby mode, nighttime sleep mode, nighttime wake-up and then sleep mode, blood pressure relief mode, etc.

[0031] In this embodiment of the application, when an elderly user is asleep at night, the player's operating mode can be switched to the nighttime sleep mode.

[0032] In this embodiment of the application, when an elderly user is in a state of waking up at night and then falling asleep again, the mode in which the player works can be switched to the mode of waking up at night and then falling asleep again.

[0033] In this embodiment of the application, when it is detected that the target blood pressure level of an elderly user belongs to the hypertension risk level, the mode in which the player works can be switched to the blood pressure relief mode.

[0034] In this embodiment of the application, when the target blood pressure of an elderly user is slightly elevated during a retest, the player can be switched to the anxiety relief mode.

[0035] In one possible implementation, when an elderly user presses the button or knob on the player corresponding to the nighttime sleep mode, it can be determined that the elderly user is in nighttime sleep mode. The player's operating mode can then be switched to nighttime sleep mode.

[0036] In one possible implementation, when an elderly user presses the button or knob on the player corresponding to the "wake up and fall back asleep" mode, it can be determined that the elderly user is in the "wake up and fall back asleep" mode. The player's operating mode can then be switched to this mode.

[0037] The natural ambient sound layer provides the acoustic background of the natural environment, including sounds of rain, ocean waves, forests, streams, and campfires. These sounds are obtained through high-fidelity recording or parametric generation. The main function of this layer is to create a comfortable acoustic environment and reduce user anxiety in medical settings. The advantage of parametric generation is that it allows for real-time adjustment of parameters such as rainfall intensity, wave strength, and forest density without storing large amounts of audio files. In terms of spectral characteristics, the natural ambient sound layer is dominated by low to mid-frequency components (80Hz-4kHz), avoiding sudden high-frequency components (such as thunder and sharp bird calls).

[0038] Background noise layer audio signal: Provides a smooth acoustic masking, including white noise, pink noise, and brown noise. The main function of this layer is to mask sudden, unpredictable noises in the environment (such as door slams or footsteps) by providing a continuous, smooth acoustic background, thus reducing the perceived salience of environmental noise. In noise-adaptive mode, the spectral characteristics of this layer can be dynamically adjusted based on the spectral analysis results of the environmental noise: if the environmental noise is predominantly low-frequency (such as traffic noise), the brown noise component is enhanced; if the environmental noise is predominantly mid-to-high frequency (such as human conversation), the pink noise component is enhanced.

[0039] The rhythm-guided layer audio signal provides low-frequency beats ranging from 4-12 bpm, corresponding to the respiratory rate range. The sound format of this layer is a sine wave or triangular wave (fundamental frequency 50-200 Hz, each beat lasting 0.3-0.8 seconds, with a slowly rising and falling envelope to avoid transient impacts), guiding the user's respiratory rate and heart rate to decrease synchronously through regular beats. In the sleep induction template, the rhythm frequency gradually decreases from 12 bpm (corresponding to normal resting breathing) to 6 bpm (corresponding to deep, slow breathing); in the blood pressure relief template, the rhythm frequency is fixed at 6 bpm (4 seconds of inhalation, 6 seconds of exhalation, consistent with the recommended breathing training frequency for hypertensive patients). The rhythm-guided layer is the core acoustic means for achieving non-pharmacological autonomic nervous system regulation in this application, and its effect is based on the respiratory synchrony phenomenon of heart rate variability (HRV) (Respiratory Sinus Arrhythmia, RSA).

[0040] The voice soothing layer audio signal provides pre-recorded guidance or text-to-speech synthesized voice content for cognitive guidance and emotional soothing in daytime mode. Voice content includes: breathing instructions ("Please inhale slowly, feel your abdomen rise...exhale slowly, feel your body relax..."), mindfulness instructions ("Pay attention to the sounds around you, don't judge, just listen..."), and positive cues ("You are safe, relaxed, everything is fine..."). The voice soothing layer is only enabled in daytime mode or the "emotional soothing template" (gain is forcibly set to 0% in nighttime mode to avoid cognitive activation caused by language comprehension). The voice broadcast uses a warm, slow, low-pitched female voice (fundamental frequency 150-200Hz, speech rate 120-140 words / minute, lower than normal conversation speed).

[0041] The ambient masking compensation layer audio signal analyzes the ambient noise spectrum in real time to generate targeted masking sounds. This layer differs from the background noise layer in that the background noise layer provides a general, continuous background sound, while the ambient masking compensation layer is dynamic and adaptively adjusts to the current ambient noise spectrum. Based on the spectrum analysis results, the masking compensation layer enhances the natural sound in the same frequency band where noise energy is concentrated, rather than simply increasing the overall volume. The ambient masking compensation layer audio signal is used to mask noise. For frequency bands with high noise energy, the gain of the components of the ambient masking compensation layer audio signal in these frequency bands is increased compared to the gain of its components in other frequency bands. Therefore, the higher volume of natural sound in the high-noise-energy frequency bands can be used to mask the noise in those bands, improving the masking effect.

[0042] The five audio layers are mixed in the digital domain using a mixer. The mixing algorithm is based on a gain matrix G = [g1, g2, g3, g4, g5], where g1-g5 represent the gain coefficients (range 0-100%, in dB, 0% corresponds to -∞dB and 100% corresponds to 0dB) of the natural ambient sound layer, background noise layer, rhythm guidance layer, voice soothing layer, and environmental masking compensation layer, respectively.

[0043] Mixing Process: Each audio source layer first passes through an independent digital equalizer (EQ, each EQ is an 8-band parametric equalizer with frequency bands of 80Hz, 160Hz, 315Hz, 630Hz, 1.25kHz, 2.5kHz, 5kHz, and 8kHz, with a gain range of -12dB to +12dB for each band) and an independent dynamic range compressor (DRC, compression ratio ≤ 2:1, threshold -20dBFS, attack time 5ms, release time 50ms). Then, each layer is multiplied by its respective gain coefficient, and finally, additive mixing is performed: Y_mix = g1×Y1_EQ+DRC + g2×Y2_EQ+DRC + g3×Y3_EQ+DRC + g4×Y4_EQ+DRC + g5×Y5_EQ+DRC. Where Y1-Y5 are the original audio sample values ​​of each layer (16-bit or 24-bit signed integers, ranging from -32768 to +32767).

[0044] Bus post-processing: The mixed audio stream passes through a bus compressor and a peak limiter to ensure that the output signal is not overloaded and the dynamic range is controlled.

[0045] Dynamic Range Control: To prevent sudden volume changes from disturbing elderly users, the dynamic range of each layer and the bus is limited. The compression ratio within a single layer is ≤2:1, the bus compression ratio is ≤1.5:1, and the overall system dynamic range is limited to ≤10dB (night mode) or ≤15dB (day mode). This means that even if there are sudden volume changes in the original audio file (such as thunder in natural sounds or emotional outbursts in speech), the sound pressure level output to the user will not change by more than 10-15dB.

[0046] Fade-in / Fade-out Rules: The main control module maintains an independent gain envelope state machine for each layer. When a layer needs to be enabled, its gain fades in from 0% to the target gain using a 3-second logarithmic curve; when a layer needs to be disabled or reduced, its gain fades out using a 5-second logarithmic curve. During template switching, the layers of the old template fade out first, and the layers of the new template fade in synchronously, with a 2-second overlap in cross-fade time. The main control module updates the gain envelope value every 1ms (linearly interpolated to the logarithmic gain scale) to ensure the smoothness of the fade-in / fade-out process.

[0047] Audio signal chain: Digital audio source (PCM / WAV file stored on SD card or Flash) → Independent EQ processing for each layer (Natural ambient sound layer: 80Hz-4kHz bandpass; Background noise layer: 100Hz-5kHz bandpass; Rhythm guiding layer: 50-200Hz low pass; Masking layer: adaptive according to ambient noise spectrum) → Independent compression for each layer (compression ratio 2:1, threshold -20dBFS) → Gain matrix weighting → Mixer mixing → Bus compression (compression ratio 1.5:1) → Peak limiting (Limiter, limiting output peak ≤ -3dBFS, corresponding to 45dB SPL) → DAC (24bit / 48kHz) → Power amplifier (Class-D, efficiency ≥85%) → Near-field directional loudspeaker (full-range unit, frequency response 200Hz-8kHz).

[0048] refer to Figure 1 It shows a flowchart of a layered soundscape therapy method for emotional and sleep intervention in the elderly provided in an embodiment of this application.

[0049] The layered soundscape therapy method for emotional and sleep intervention in the elderly provided in this application includes steps S101-S104.

[0050] In step S101, when the elderly user is asleep at night, during the first time period, a soundscape intervention audio signal is played, which is obtained by mixing the gain of each target audio signal in the audio signal set during the first time period. The audio signal set includes the following target audio signals: natural ambient sound layer audio signal, rhythm guidance layer audio signal, background noise layer audio signal, environmental masking compensation layer audio signal, and voice soothing layer audio signal. In step S102, during the second time period after the first time period, a soundscape intervention audio signal obtained by mixing the gain of each target audio signal in the second time period is played, wherein, during the second time period, the beats per minute (bpm) of the rhythm guidance layer audio signal is reduced at a preset first bpm reduction rate. In step S103, when it is determined that the elderly user is asleep during the second time period, in the third time period after the second time period, the soundscape intervention audio signal obtained by mixing the gain of each target audio signal in the third time period is played. In the third time period, the gain of the natural ambient sound layer audio signal, the rhythm guidance layer audio signal, the environmental masking compensation layer audio signal, and the background noise layer audio signal fades out in the third time period. In step S104, during the fourth time period following the third time period, a soundscape intervention audio signal generated according to the gain of each target audio signal in the fourth time period is played, wherein the gain of the background noise layer audio signal fades out in the fourth time period.

[0051] The start time of the first time period can be determined when the elderly user is in a state of sleep at night. The duration of the first time period can be preset.

[0052] The start time of the second period can be the end time of the first period. The end time of the second period can be the time when the elderly user is determined to be asleep.

[0053] The start time of the third period can be the end time of the second period.

[0054] In one possible implementation, the duration of the third time period is preset.

[0055] In one possible implementation, the end time of the third time period can be the latest of the following: the time when the gain fade-out of the natural ambient sound layer audio signal in the third time period is completed; the time when the gain fade-out of the rhythm guidance layer audio signal in the third time period is completed; the time when the gain fade-out of the environmental masking compensation layer audio signal in the third time period is completed; and the time when the gain fade-out of the background noise layer audio signal in the third time period is completed.

[0056] The start time of the fourth period can be the end time of the third period.

[0057] In one possible implementation, the duration of the fourth time slot can be preset.

[0058] In this embodiment of the application, for a target audio signal, the gain of the target audio signal at the beginning of the first time period is: the initial gain of the target audio signal for the nighttime sleep state.

[0059] In one possible implementation, the initial gain of a target audio signal for nighttime sleep can be preset.

[0060] In this embodiment of the application, the gain of the audio signal of the environmental masking compensation layer in the third time period can be a low gain less than the gain threshold.

[0061] Given a target audio signal and a time period, if the target audio signal has a gain during that time period, then at the corresponding moment within that time period, the gain of the target audio signal is the gain of the target audio signal during that time period.

[0062] It should be noted that for a target audio signal and a time period, if the target audio signal has a gain during that time period, then the gain of the target audio signal during that time period may be fixed, or the gain of the target audio signal during that time period may vary.

[0063] The natural ambient sound layer audio signal can be a rain sound audio signal, and the background noise layer audio signal can be a powder noise audio signal.

[0064] In one possible implementation, the bpm of the rhythm-guided layer audio signal at the beginning of the first time period is 12 bpm, corresponding to the normal resting respiratory rate.

[0065] It should be noted that the audio signal of the voice soothing layer is not used to generate the soundscape intervention audio signal when the elderly user is asleep at night. The gain of the audio signal of the voice soothing layer in the first time period, the gain of the audio signal of the voice soothing layer in the second time period, the gain of the audio signal of the voice soothing layer in the third time period, and the gain of the audio signal of the voice soothing layer in the fourth time period are 0%.

[0066] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the first time period, so as to obtain the sound scene intervention audio signal obtained by mixing the gain of each target audio signal in the first time period.

[0067] In one possible implementation, the total volume is increased from 0 to the target volume of 35 dB SPL during the first time period.

[0068] It should be noted that the total volume refers to the volume used when playing the audio signal for sound scene intervention.

[0069] In this embodiment, each time period corresponds to audio signal parameter information. The first time period, second time period, third time period, fourth time period, fifth time period, sixth time period, seventh time period, eighth time period, ninth time period, and tenth time period each correspond to audio signal parameter information.

[0070] For a given time period, the corresponding audio signal parameter information includes: sub-audio signal parameter information for each target audio signal within that time period. The sub-audio signal parameter information for a target audio signal within that time period indicates at least some of the following: the gain of the target audio signal at the beginning of the time period, the gain of the target audio signal at the end of the time period, whether the gain of the target audio signal decreases during the time period, how the gain of the target audio signal decreases during the time period, and to what extent the gain of the target audio signal ultimately decreases during the time period. If the target audio signal is a rhythmic guidance layer audio signal, the sub-audio signal parameter information for the target audio signal within that time period also indicates at least some of the following: whether the bpm of the target audio signal decreases during the time period, how the bpm of the target audio signal decreases during the time period, and to what extent the bpm of the target audio signal ultimately decreases during the time period.

[0071] It should be noted that, in the embodiments of this application, fading out can mean: gradually decreasing.

[0072] Fade-out can be achieved using any existing algorithm that smooths the reduction of audio parameters such as gain.

[0073] In one possible implementation, the total volume is linearly reduced from 35 dB SPL to 15 dB SPL during the second time period.

[0074] In one possible implementation, the initial bpm reduction rate is preset to be 1 bpm every 5 minutes.

[0075] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the second time period, so as to obtain the sound scene intervention audio signal obtained by mixing the gain of each target audio signal in the second time period.

[0076] In step S102, during the second time period, the beats per minute (bpm) of the rhythm guidance layer audio signal decreases at a preset first bpm decreasing rate, which can guide the elderly user to slow down their breathing, thereby reducing the elderly user's respiratory rate and heart rate.

[0077] In one possible implementation, in step S102, during the second time period, the beats per minute (bpm) of the rhythm-guided layer audio signal gradually decreases from 12 bpm, corresponding to a normal resting respiratory rate, to 6 bpm, corresponding to a deep and slow respiratory rate.

[0078] In one possible implementation, if the elderly user's body movement frequency is detected to be less than 1 time / minute during the second time period and there is no significant body movement for 5 minutes, it is determined that the elderly user is in a sleep state.

[0079] In one possible implementation, the elderly user is determined to be asleep when their body movement frequency is less than 2 beats / minute and their heart rate shows a downward trend during the second time period. The downward trend can be a slope of less than -0.5 bpm / minute.

[0080] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the third time period, so as to obtain the sound scene intervention audio signal obtained by mixing the gain of each target audio signal in the third time period.

[0081] In this embodiment of the application, during the third time period, the gain of the natural ambient sound layer audio signal and the rhythm guidance layer audio signal can be faded to 0% during the third time period.

[0082] In this embodiment of the application, during the third time period, the gain of the background noise layer audio signal fades out to a preset background noise retention gain.

[0083] In one possible implementation, during the third time period, the gain of the ambient masking compensation layer audio signal fades to 0%.

[0084] In one possible implementation, during a third time period, the gain of the ambient masking compensation layer audio signal fades out to a preset ambient masking compensation layer retain gain.

[0085] The preset background noise retention gain is 10%. The preset environment masking compensation layer retention gain is 10%.

[0086] It should be noted that in the embodiments of this application, gain is expressed in percentage or decibels (dB). The conversion relationship between percentage and decibel of gain is existing technology and will not be described in detail in this application.

[0087] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the fourth time period, so as to obtain the sound scene intervention audio signal obtained by mixing according to the gain of each target audio signal in the fourth time period.

[0088] During the fourth time period, the gain of the background noise layer audio signal fades to 0%.

[0089] It should be noted that the gain of the natural ambient sound layer audio signal and the rhythm guidance layer audio signal is 0% in the fourth time period.

[0090] It should be noted that at the beginning of the fourth time period, the gain of the background noise layer audio signal is the preset background noise preservation gain. During the fourth time period, the gain of the background noise layer audio signal fades from the preset background noise preservation gain to 0%.

[0091] It should be noted that if the gain of the ambient masking compensation layer audio signal fades to 0 during the third time period, then the gain of the ambient masking compensation layer audio signal in the fourth time period will be 0%. If the ambient masking compensation layer audio signal fades to the preset ambient masking compensation layer retainable gain during the third time period, then at the beginning of the fourth time period, the gain of the ambient masking compensation layer audio signal will be the preset ambient masking compensation layer retainable gain, and during the fourth time period, the gain of the ambient masking compensation layer audio signal will fade to 0 from the preset ambient masking compensation layer retainable gain.

[0092] It should be noted that, in the embodiments of this application, gain may specifically refer to EQ gain.

[0093] For the target audio signal, multiple frequency bands are preset. The target audio signal has gain in the corresponding preset frequency bands.

[0094] As an example, eight frequency bands are pre-set, with the right endpoints of the eight bands being 80Hz, 160Hz, 315Hz, 630Hz, 1.25kHz, 2.5kHz, 5kHz, and 8kHz, respectively. The gain range of each band is -12dB to +12dB. The left endpoint of the band with the smallest right endpoint is 0.

[0095] As an example, user A, a 75-year-old woman, suffers from mild hypertension (systolic blood pressure 135-150 mmHg) and difficulty falling asleep (sleep latency >30 minutes). She goes to bed at 10:00 PM every night to prepare for sleep. User A lives in the family bedroom, sleeps with her spouse, is sensitive to sudden changes in sound, and does not wish to disturb her spouse.

[0096] Night Mode Strategy: In Night Mode (default 22:00-7:00), the system implements the following parameter constraints: the high-frequency cutoff frequency is further reduced from 8kHz to 6kHz to reduce the inhibition of melatonin secretion by high-frequency stimulation; the dynamic range limit is tightened from 15dB to 10dB to avoid any sudden volume changes; the voice soothing layer is disabled (forced gain to 0%) to avoid cognitive activation caused by language understanding; and stable and continuous natural sounds (such as continuous rain or calm ocean waves) are prioritized over soundscapes containing sudden events (such as intermittent thunder or sudden bird calls) to reduce the unpredictability of sound events.

[0097] In this example, at 22:00: After the elderly person A went to bed, they pressed the large "Sleep Mode" button on the front of the device. The device confirmed the operation through vibration feedback and announced "Sleep Mode has been activated" at a low volume through the neckband speaker.

[0098] In this example, parameter initialization is performed as follows: The main control module reads the time information (22:00, indicating night mode) and loads the initial parameters of the "sleep induction template". The initial gain matrix G0 = [g1 = 60%, g2 = 40%, g3 = 50%, g4 = 0%, g5 = 20%], where the natural ambient sound layer selects "rain sound" (pre-recorded high-fidelity rain sound, with a spectrum mainly in the mid-low frequency range and no sudden thunder), the background noise layer selects "pink noise", the initial frequency of the rhythm guidance layer is set to 12 bpm (corresponding to the normal resting breathing frequency), the voice soothing layer gain is 0% (the voice layer is disabled in night mode to avoid cognitive activation), and the environmental masking compensation layer gain is 20% (basic masking is performed based on the current bedroom ambient noise level).

[0099] In this example, a smooth start is achieved by executing a smooth transition algorithm, which uses a 3-second logarithmic fade-in to increase the total volume from 0 to the target volume of 35 dB SPL, with each sound layer fading in synchronously according to its own gain ratio.

[0100] In this example, dynamic adjustments were made during playback. From 22:05 to 22:30: During playback, the motion sensor (three-axis accelerometer, sampling rate 50Hz) continuously monitored the elderly person A's motion frequency. The initial motion frequency was 4 times / minute (turning over, adjusting sleeping position), and the motion frequency gradually decreased as the soundscape intervention progressed. The rhythm guidance layer frequency gradually decreased from 12 bpm to 6 bpm (corresponding to a deep and slow breathing frequency) according to a preset curve, decreasing by 1 bpm every 5 minutes, guiding the user's breathing to slow down synchronously.

[0101] In this example (Auto-fade-out, 22:30): After 30 minutes of playback, the motion sensor detected that the motion frequency had dropped below 1 movement per minute, and there was no significant motion for 5 minutes. The main control module determined that the user had entered the early stage of stable sleep. The system initiated Auto-fade-out: the total volume was linearly reduced from 35dB SPL to 15dB SPL over 5 minutes, while maintaining a low-intensity background sound layer (pink noise, gain reduced to 10%) as a continuous environmental masking. The natural ambient sound layer and rhythm guidance layer faded out first during the initial fade-out, and the background noise layer faded out last.

[0102] In this example, a multi-parameter comprehensive judgment method is used to determine the sleep stage: when the body movement frequency is <2 beats / minute and the heart rate shows a downward trend (slope < -0.5 bpm / minute), and the proximity sensor detects the user's presence, the main control module determines that the user has entered a stable sleep stage. After the judgment lasts for 5 minutes, automatic fading is initiated: the total volume is reduced from the current value to 0dB within 5-10 minutes.

[0103] In this example, the rhythm guidance layer begins to fade out after 5 minutes to avoid adaptive saturation of the autonomic nervous system due to prolonged rhythm guidance; the natural ambient sound layer begins to fade out 10 minutes after the rhythm guidance layer fades out; and the background noise layer fades out last.

[0104] In this example, the layered fading strategy ensures a smooth and natural transition for the user from soundscape intervention to natural sleep, avoiding the "sound withdrawal" effect caused by sudden cessation.

[0105] In this example, after the fade-out is complete, the device switches to a low-power standby state (standby current ≤2mA, only RTC, motion sensor low-power listening and BLE reception are retained).

[0106] In this example, at 23:30, a complete stop is implemented: a low-intensity background noise layer (pink or brown noise, gain 10-15%, corresponding to a sound pressure level of 15-20 dB SPL) is retained as a continuous environmental masking layer for 60 minutes before being completely turned off. After the low-intensity background noise plays continuously for 60 minutes, the main control module executes a complete stop: the background noise layer fades to zero with a 10-second logarithmic curve, the device switches to a low-power standby state (standby power consumption ≤ 0.5W), and only the low-power monitoring of the motion sensor and Bluetooth receiver module is retained (waking up every 30 seconds to detect motion).

[0107] Table 1 shows examples of how the gain of each sound layer changes over time when elderly users are asleep at night.

[0108] Table 1 Table 2 shows an example of filtering each acoustic layer.

[0109] Table 2 In one possible implementation, step S105 is also included.

[0110] In step S105, when the elderly user is in a state of waking up at night and then falling back asleep, during the fifth time period, a soundscape intervention audio signal generated according to the gain of each target audio signal in the fifth time period is played. During the fifth time period, the bpm of the rhythm guidance layer audio signal decreases at a preset second bpm decrease rate. When the elderly user's heart rate is detected to meet the heart rate decrease condition during the fifth time period, during the sixth time period after the fifth time period, a soundscape intervention audio signal generated according to the gain of each target audio signal in the sixth time period is played. During the sixth time period, the gain of the rhythm guidance layer audio signal, the natural ambient sound layer audio signal, and the background noise layer audio signal fades out in the sixth time period.

[0111] It should be noted that the state of waking up at night and then falling back asleep can be understood as: the state of an elderly user waking up at night and then falling back asleep.

[0112] The start time of the fifth time period can be the moment when the elderly user is confirmed to be in a state of waking up at night and then falling back asleep. The end time of the fifth time period can be the moment when the elderly user's heart rate is detected to meet the condition of heart rate decrease. The start time of the sixth time period can be the end time of the fifth time period. The duration of the sixth time period can be preset.

[0113] For a target audio signal, at the beginning of the fifth time period, the gain of the target audio signal can be: the initial gain of the target audio signal for the nighttime wake-up and fall back asleep mode.

[0114] It should be noted that the voice-soothing layer audio signal is not used to generate soundscape intervention audio signals when elderly users are in a state of waking up at night and then falling back asleep. The gain of the voice-soothing layer audio signal in the fifth time period and the gain of the voice-soothing layer audio signal in the sixth time period are both 0%. The initial gain of the voice-soothing layer audio signal for the waking up at night and then falling back asleep mode is 0%.

[0115] In this embodiment of the application, during the sixth time period, the gain of the audio signal of the environmental masking compensation layer can fade out during the sixth time period.

[0116] In this embodiment, during the sixth time period, the gain of the rhythm guidance layer audio signal, the natural ambient sound layer audio signal, and the background noise layer audio signal can each be faded to 0% during the sixth time period. During the sixth time period, the gain of the ambient masking compensation layer audio signal can also be faded to 0% during the sixth time period.

[0117] In this embodiment, when an elderly user presses the button or knob on the player corresponding to the "wake up and fall back asleep" mode, it can be determined that the elderly user is in the "wake up and fall back asleep" mode. The player's operating mode can then be switched to the "wake up and fall back asleep" mode.

[0118] In this embodiment of the application, for a target audio signal, the gain of the target audio signal at the beginning of the fifth time period is: the initial gain of the target audio signal for the state of waking up at night and then falling back asleep.

[0119] In one possible implementation, the initial gain of a target audio signal for the state of waking up and falling back asleep can be preset.

[0120] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the fifth time period, so as to obtain the sound scene intervention audio signal obtained by mixing according to the gain of each target audio signal in the fifth time period.

[0121] It should be noted that if the heart rate of an elderly user meets the condition of decreasing heart rate, it can be determined that the elderly user is relaxed and tends to fall asleep again.

[0122] In one possible implementation, the condition for a decrease in heart rate is: the elderly user's heart rate is less than a heart rate threshold for a preset number of minutes, such as 2 minutes.

[0123] In one possible implementation, the total volume is 38 dB SPL during the fifth time slot.

[0124] In one possible implementation, the second bpm reduction rate is preset to decrease by 1 bpm every 3 minutes.

[0125] In one possible implementation, the bpm corresponding to a deep, slow breathing rate is 5 bpm.

[0126] In one possible implementation, the total volume is 38 dB SPL during the fifth time slot.

[0127] In one possible implementation, during the sixth period, the total volume is gradually reduced from 38dB SPL to 10dB SPL over an 8-minute period.

[0128] In one possible implementation, during the fifth time period, the bpm of the rhythm-guided layer audio signal starts at 10 bpm and gradually decreases at a rate of 1 bpm every 3 minutes to 5 bpm, corresponding to a deep, slow breathing frequency.

[0129] In one possible implementation, when the elderly user is detected to meet the preset condition of increased body movement frequency, it is determined that the elderly user is awake at night; when the elderly user does not perform any operation corresponding to the state of falling back asleep after waking up at night within a preset time period from the time the elderly user is determined to be awake at night, and the elderly user's current body movement frequency is greater than the upper limit threshold of body movement frequency, it is determined that the elderly user is in the state of falling back asleep after waking up at night.

[0130] The operation corresponding to the sleep-back mode after waking up at night is for elderly users to press the corresponding button or knob on the player.

[0131] The preset condition for increasing body movement frequency can be: the difference between the most recently determined body movement frequency of elderly users and the body movement frequency of the most recently determined elderly users is greater than a threshold.

[0132] The difference between the most recently determined physical activity frequency of elderly users and the most recently determined physical activity frequency of elderly users is: the most recently determined physical activity frequency of elderly users minus the most recently determined physical activity frequency of elderly users.

[0133] The system can determine the frequency of physical activity of elderly users within a one-minute time period at one-minute intervals, and check whether the elderly users meet the preset conditions for increasing the frequency of physical activity at one-minute intervals.

[0134] As an example, at 2:30, night wake-up detection: The body movement sensor detected that the elderly B's body movement frequency suddenly increased from 0.5 times / minute to 8 times / minute (turning over, adjusting the pillow, moving the body). The main control module recorded this abnormal body movement event, but did not immediately switch modes (to avoid false triggering, such as the elderly only briefly adjusting their sleeping position).

[0135] In this example, at 2:30:30, progressive confirmation: when the detected body movement frequency suddenly increases from the baseline (<1 time / minute) to >6 times / minute (lasting 30 seconds), the progressive confirmation mechanism is activated: first, the background noise layer gain is increased by 10% (from 10% to 20%, and the total volume is increased from 15dB SPL to 20dB SPL). This gentle sound change serves as a notification to the user that the device has detected their awake state. If active user interaction (button / knob) is detected within 30 seconds, it is determined that the user is actively seeking intervention, and the system immediately switches to the "Night Wake-up and Resleep Template"; if there is no active interaction within 30 seconds but body movement remains frequent, Automatically switch to the "Wake Up and Fall Back to Sleep Template". If body movement returns to baseline within 30 seconds, it is considered a brief adjustment of sleeping position, and the current state is maintained. The purpose of the gradual confirmation mechanism is to avoid accidental triggering (such as the user only briefly turning over or adjusting the pillow) and reduce unnecessary intervention.

[0136] In this example, the output mode switching command (switch to the "wake up and fall asleep template") or the command to maintain the current state is given.

[0137] In this example, the elderly person B did not make any active movements within 30 seconds, but the frequency of body movements remained above 6 times per minute. The main control module determined that "it is difficult to fall back asleep after waking up at night" and started the automatic switching process.

[0138] In this example, at 2:31, template switching occurs: the main control module switches its operating mode from "low-power standby" to "wake up and fall asleep again" mode, loading the "wake up and fall asleep again template". The initial gain matrix G0 = [g1 = 50%, g2 = 50%, g3 = 60%, g4 = 0%, g5 = 30%]. The natural ambient sound layer is selected as "gentle ocean waves" (no sudden wave crashing sound, continuous and stable spectrum), the background noise layer is selected as "mixed pink noise + brown noise" (Brown Noise, stronger low-frequency energy, sounds more muffled and soothing), the initial frequency of the rhythm guidance layer is set to 10 bpm (slightly lower than the normal breathing rate, slowing down the guidance at the beginning), the voice soothing layer gain is 0% (disabled at night), and the masking compensation layer gain is 30% (to enhance and mask sudden noises outside the window).

[0139] In this example, the type of the natural ambient sound layer audio signal is "gentle ocean waves" (no sudden splashing sound, continuous and stable spectrum), the background noise layer audio signal is "mixture of pink noise and brown noise" (with stronger low-frequency energy, sounding more muffled and soothing), and the initial frequency of the rhythm guidance layer audio signal is set to 10 bpm (slightly lower than the normal breathing rate, to begin guiding the slowdown).

[0140] In this example (2:31-2:46, rhythmic breathing guidance): the rhythmic guidance layer starts playing a low-frequency beat (sine wave, 80Hz fundamental frequency, once every 6 seconds, lasting 0.5 seconds, with a slowly rising and falling envelope) at 10 bpm, and gradually decreases to 5 bpm (corresponding to a deep and slow breathing rate) at a rate of 1 bpm every 3 minutes. Simultaneously, the main control module monitors the elderly person B's heart rate changes in real time via a heart rate monitor connected via Bluetooth BLE. The initial heart rate is 78 bpm (sympathetic nerve excitation after awakening), and gradually decreases as the rhythmic breathing guidance progresses.

[0141] In this example, at 2:46, heart rate drop detection: after 15 minutes of playback, the heart rate monitor detected that the heart rate had dropped from 78 bpm to 62 bpm and remained below 65 bpm for 2 consecutive minutes. The main control module determined that the user had entered a relaxed state and was tending to fall asleep again. The system then initiated automatic fading: over 8 minutes, the total volume gradually decreased from 38 dB SPL to 10 dB SPL. The rhythm guide layer faded out first, while the natural ambient sound layer and background noise layer remained until the end.

[0142] In this example, at 2:54, complete stop: after fading out, the device switches back to low-power standby mode. The entire nighttime wake-up and sleep-back intervention process lasted approximately 24 minutes, during which the sleeping spouse was not disturbed (near-field directional output, sound pressure level at 2m <30dBSPL), and the elderly person B did not perform any manual operation.

[0143] Table 3 shows an example of how the parameters change when an elderly user wakes up at night and then falls back asleep.

[0144] Table 3 In one possible implementation, step S106 is also included.

[0145] In step S106, when it is detected that the target blood pressure level of the elderly user belongs to the hypertension risk level, during the seventh time period, a soundscape intervention audio signal generated according to the gain of each target audio signal in the seventh time period is played, wherein the rhythm guidance layer audio signal is the target audio signal with the highest gain in the seventh time period.

[0146] The target blood pressure is either systolic or diastolic.

[0147] The start time of the seventh time period can be the moment when the elderly user's systolic blood pressure level is determined to be within the hypertension risk category. The duration of the seventh time period can be preset, or the end time of the seventh time period can be the moment when the elderly user's heart rate is detected to drop to a heart rate level not exceeding the heart rate threshold.

[0148] In this embodiment of the application, for a target audio signal, the gain of the target audio signal at the beginning of the seventh time period is: the initial gain of the target audio signal for the blood pressure relief mode.

[0149] In one possible implementation, the initial gain of the target audio signal for the blood pressure relief mode can be preset.

[0150] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the seventh time period, so as to obtain the soundscape intervention audio signal obtained by mixing the gain of each target audio signal in the seventh time period.

[0151] In one possible implementation, the total volume is 45 dB SPL during the seventh time slot.

[0152] During the seventh time period, the rhythm-guided audio signal is the target audio signal with the highest gain during this period. Therefore, a strong respiratory rhythm can guide the activation of the parasympathetic nervous system of the assisted autonomic nervous system, quickly reducing the respiratory rate and heart rate of elderly users.

[0153] During the seventh time period, the bpm of the rhythm-guided layer audio signal is fixed at the bpm corresponding to the deep and slow breathing frequency.

[0154] In one possible implementation, the gain of the rhythmic guidance layer audio signal in the seventh time slot is: 70% of the time, during the seventh time period, the bpm of the rhythm-guided layer audio signal was fixed at 6 bpm. The deep, slow breathing rate was 4 seconds for each inhalation and 6 seconds for each exhalation.

[0155] In one possible implementation, step S107 is also included.

[0156] In step S107, when the heart rate of an elderly user is detected to drop to a level not exceeding the heart rate threshold during the seventh time period, the elderly user is prompted to retest their blood pressure. When the target blood pressure level of the elderly user's retest is slightly elevated, a soundscape intervention audio signal generated based on the gain of each target audio signal in the eighth time period is played during the eighth time period following the seventh time period.

[0157] Blood pressure retesting refers to measuring blood pressure again. The target blood pressure for the retest is obtained through blood pressure retesting.

[0158] The start time of the eighth time slot can be used to determine whether the target blood pressure level for elderly users is slightly elevated.

[0159] When an elderly user's target blood pressure is found to be slightly elevated during a retest, the player can be switched to the anxiety relief mode.

[0160] In this embodiment of the application, for a target audio signal, the gain of the target audio signal at the beginning of the eighth time period is: the initial gain of the target audio signal for the anxiety relief mode.

[0161] In one possible implementation, the initial gain of the target audio signal for the anxiety relief mode can be preset.

[0162] In the embodiments of this application, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the eighth time period, so as to obtain the sound scene intervention audio signal obtained by mixing the gain of each target audio signal in the eighth time period.

[0163] In one possible implementation, during the eighth segment, the gain of the rhythm guidance layer is 30%, the gain of the voice soothing layer is 20%, and relaxing natural music is played, with the total volume gradually decreasing from 45dB SPL to 35dB SPL. The duration of the eighth segment is 15 minutes.

[0164] As an example, Ms. C, a 78-year-old woman with stage 2 hypertension (regularly taking amlodipine), used a backrest-mounted player for daytime relaxation in her family living room at 3:00 PM. That day, due to family troubles, she felt anxious and tense.

[0165] In this example (15:00, blood pressure data is automatically received): Elderly person C uses a Bluetooth electronic blood pressure monitor (already paired with the player via BLE protocol) to measure blood pressure. The measurement results are automatically transmitted to the player via Bluetooth: systolic blood pressure 162 mmHg, diastolic blood pressure 98 mmHg, heart rate 85 bpm. After receiving the data, the main control module compares it with the preset blood pressure linkage threshold table.

[0166] Table 4 shows an example of blood pressure levels and responses based on blood pressure levels.

[0167] Table 4 In this example (15:00:10, automatic template switching): According to the threshold table, a systolic blood pressure of 162 mmHg reaches the "hypertension risk" level. The main control module automatically switches to the "blood pressure relief template" and prompts "Blood pressure is high, relief mode has been activated" via voice broadcast (volume 45dB SPL). Elderly person C does not need to perform any operation.

[0168] In this example, (soundscape parameter initialization): initial gain matrix G0=[g1=60%, g2=50%, g3=70%, g4=10%, g5=20%]. The rhythm guidance layer has the highest gain (70%), with a fixed frequency of 6 bpm (corresponding to a deep, slow breathing frequency, 4 seconds of inhalation and 6 seconds of exhalation), using a strong breathing rhythm to guide and assist the activation of the parasympathetic nervous system of the autonomic nervous system. The voice soothing layer has a gain of 10% (can be enabled during the day), playing a pre-recorded short guiding phrase: "Please follow the beat, inhale slowly… exhale slowly…". The natural environment sound layer is set to "Forest Stream" (removing sudden bird calls and enhancing the continuous flowing water sound to over 90%).

[0169] If blood pressure data is received and the systolic blood pressure remains high (≥160 mmHg, for two consecutive measurements), the gain of the rhythm guidance layer is increased (by 10% from the current value, with an upper limit of 70%), while the gain of the voice soothing layer is reduced (to avoid cognitive load), and the sudden sound components in the natural ambient sound layer are reduced (by dynamically detecting transient events in the natural sound and temporarily reducing the gain of that layer).

[0170] Table 5 shows an example of parameters for blood pressure relief mode.

[0171] Table 5 In this example (15:00-15:30, playback intervention): During the 30-minute playback, the main control module continuously monitored heart rate changes (via a Bluetooth heart rate monitor), with the heart rate gradually decreasing from an initial 85 bpm to 68 bpm. The rhythm guidance layer maintained a fixed frequency of 6 bpm, without further decreasing due to the drop in heart rate (to avoid the risk of respiratory alkalosis caused by excessive guidance).

[0172] In this example (15:30, Blood Pressure Retest and Data Recording): After 30 minutes of playback, the system automatically prompts Elderly C to retest their blood pressure. The retest results are: Systolic blood pressure 148 mmHg, Diastolic blood pressure 92 mmHg. The main control module records the blood pressure data before and after the intervention, the heart rate change curve, the template parameters used, and the intervention duration to local storage (which can be exported to a mobile app or the cloud for family doctors to reference).

[0173] In this example, at 15:31, mode switching: Since the retested blood pressure has dropped from "high blood pressure risk" to "mildly high", the main control module automatically switches the soundscape from "blood pressure relief template" to "anxiety relief template" (gain matrix G=[50%, 40%, 30%, 20%, 10%], rhythm guidance layer reduced to 30%, voice soothing layer increased to 20%, playing relaxing natural music), the total volume gradually decreases from 45dB SPL to 35dB SPL, and stops completely after playing for 15 minutes.

[0174] In anxiety relief mode, the rhythm guidance layer is prioritized for enhancement (gain ≥40%) to guide a decrease in breathing rate, while sudden sound components in the natural ambient sound layer are reduced (using a real-time transient detection algorithm, when a transient event is detected in the natural sound, the gain of this layer is temporarily reduced by 6dB, lasting for 2 seconds before being restored). Slow, fluctuating soundscape changes (volume change rate ≤2dB / min, EQ change rate ≤1dB / min) avoid any sudden changes that might trigger alertness.

[0175] In one possible implementation, step S108 is also included.

[0176] In step S108, when target environmental noise is detected, during the ninth time period, a soundscape intervention audio signal generated based on the gain of the environmental masking compensation layer audio signal during the ninth time period is played. The gain of the environmental masking compensation layer audio signal during the ninth time period is determined based on the gain of the component of the environmental masking compensation layer audio signal in the target energy frequency band of the target environmental noise during the ninth time period and the gain of the component of the environmental masking compensation layer audio signal in frequency bands other than the target energy frequency band during the ninth time period. The energy of the component of the target environmental noise in the target energy frequency band is greater than the energy of the component of the target environmental noise in frequency bands other than the target energy frequency band, and the gain of the component of the environmental masking compensation layer audio signal in the target energy frequency band during the ninth time period is greater than the gain of the component of the environmental masking compensation layer audio signal in frequency bands other than the target energy frequency band during the ninth time period.

[0177] It should be noted that the target environmental noise can be any type of environmental noise.

[0178] The start time of the ninth time period can be the moment when the target environmental noise is detected.

[0179] When ambient noise is detected, the player can be switched to noise adaptive mode.

[0180] In this embodiment of the application, for a target audio signal, the gain of the target audio signal at the beginning of the ninth time period is: the initial gain of the target audio signal for the noise adaptive mode.

[0181] In one possible implementation, the initial gain of the target audio signal for the noise-adaptive mode can be preset.

[0182] It should be noted that the gain of each of the following audio signals in the ninth time period is 0%: natural ambient sound layer audio signal, rhythm guidance layer audio signal, background noise layer audio signal, and voice soothing layer audio signal. The initial gain of each of these audio signals for the noise adaptive mode is 0%.

[0183] In this embodiment, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the ninth time period, so as to obtain the sound scene intervention audio signal generated according to the gain of the ambient masking compensation layer audio signal in the ninth time period.

[0184] For a target audio signal, the initial gain of the target audio signal can be determined by the initial gains of the components of the target audio signal in multiple frequency bands.

[0185] The target energy frequency band of the target environmental noise can be understood as the frequency band with high energy of the components of the target environmental noise.

[0186] The energy of the target environmental noise component in the target energy frequency band of the target environmental noise is greater than the energy of the target environmental noise component in frequency bands other than the target energy frequency band of the target environmental noise.

[0187] The gain of the component of the audio signal in the target energy frequency band of the target ambient noise in the ninth time period is greater than the gain of the component of the audio signal in the target ambient noise other than the target energy frequency band in the ninth time period. This is equivalent to increasing the gain of the component of the audio signal in the target ambient noise in the ninth time period relative to the gain of the component of the audio signal in the target ambient noise other than the target energy frequency band in the ninth time period. For the frequency band with strong energy of the target ambient noise component, i.e. the target energy frequency band, the volume of the naturally generated noise used to mask the target energy frequency band can be increased, thus improving the effect of noise masking.

[0188] As an example, the audio signal for the environmental masking compensation layer can be a stream sound audio signal. The target energy frequency band of the target environmental noise is 800-1500Hz, and the stream sound audio signal has continuous and stable spectral characteristics in this target energy frequency band, and the subjective listening experience is soothing.

[0189] During the ninth time period, the gain of the audio signal of the environmental masking compensation layer can be fixed.

[0190] In one possible implementation, the gain of the ambient masking compensation layer audio signal is 40% in the ninth time segment.

[0191] In one possible implementation, during the ninth time period, the total volume is strictly limited to ≤45dB SPL (nighttime) or ≤55dB SPL (daytime) by a peak limiter, so that even if the masking layer is enhanced, it will not exceed the safety threshold.

[0192] As an example, elderly person D lives in a double room in a nursing home. Suddenly, construction noise (electric drill sound, mainly concentrated in the 500-2000Hz frequency band, with a peak sound pressure level of about 65dB SPL) comes from outside the window. Elderly person D is using a pillow-side player to relax during their afternoon nap.

[0193] In this example, the noise masking strategy is as follows: when ambient noise increases, the system does not adopt a simple "overall increase in volume" approach. Instead, it activates a targeted enhancement of the ambient masking compensation layer. Specifically, the system analyzes the main energy frequency bands of the ambient noise (obtaining the top 3 bands via FFT), and then enhances natural sounds within the same frequency band (such as stream sounds, which have rich spectral content in the 500-2000Hz range, suitable for masking human voices and construction noise) within that range. The gain increase of the masking compensation layer is simultaneously constrained by a total volume limiter to ensure that the total output does not exceed a safe threshold. When the ambient noise disappears, the masking compensation layer smoothly returns to its base gain within 30 seconds.

[0194] In this example, ambient noise adapts during playback: the FFT spectrum analysis is updated every 10 seconds; if the ambient noise LAeq changes by more than ±5dB, the gain and EQ of the masking compensation layer are adjusted accordingly. If the ambient noise increases, the masking layer gain is increased (not exceeding 50%), with a boost of 3-6dB in the main noise frequency bands; if the ambient noise decreases, the masking layer gain is reduced accordingly. The total volume is always limited by the peak limiter and does not exceed the safety threshold due to masking enhancement.

[0195] In this example, noise detection: An ambient noise sensor (MEMS microphone, 48kHz sampling, built-in FFT accelerator) performs a Fast Fourier Transform (FFT, 2048 points, Hanning window) every 10 seconds to analyze the ambient noise spectrum. At 12:15, a sudden energy peak was detected in the 500-2000Hz frequency band, which increased by about 25dB compared to the baseline ambient noise, and was determined to be sudden interference noise.

[0196] In this example, spectrum analysis: the main control module reads the FFT results, identifies the main noise frequency band as 800-1500Hz (the characteristic frequency band of electric drill sound), and calculates the masking threshold for this frequency band.

[0197] In this example, masking compensation is activated: the main control module activates the environmental masking compensation layer (originally with a gain of 15%) and makes targeted adjustments based on the noise spectrum analysis results. In this example, the system does not increase the overall output volume (maintaining the current 38dB SPL), but instead enhances the natural sound energy in the 800-1500Hz frequency band of the masking compensation layer (the "stream sound" is chosen as the masking sound source because the stream sound has a continuous and stable spectral characteristics in this frequency band and is subjectively soothing).

[0198] In this example, in-band enhancement: the EQ of the masking compensation layer is set to boost by 6dB in the 800-1500Hz band, while other bands remain unchanged. The masking layer gain is increased from 15% to 40%.

[0199] In this example, a safety threshold is maintained: the total output volume is strictly limited to ≤45dB SPL (nighttime) or ≤55dB SPL (daytime) by a peak limiter, and the safety threshold will not be exceeded even if the masking layer is enhanced.

[0200] In this example, continuous monitoring is performed: the FFT analysis is updated every 10 seconds to dynamically track noise changes. If the noise persists, the current masking parameters are maintained; if the noise spectrum changes (e.g., the electric drill stops and is replaced by a hammer sound, with the main energy shifting to 200-500Hz), the EQ settings of the masking layer are adjusted accordingly (shifting the boost band to 200-500Hz).

[0201] In this example, the noise disappeared and then recovered: at 12:35, the construction noise stopped, and the ambient noise returned to the baseline level (30dB SPL). The main control module detected the disappearance of the energy peak in the 500-2000Hz frequency band, and within 30 seconds, it smoothly and gradually reduced the masking compensation layer gain from 40% to 15%, the EQ returned to a flat state, and the total volume remained unchanged.

[0202] In this example, through a targeted masking strategy, elderly person D was not significantly disturbed by noise throughout the construction period, nor did they experience hearing discomfort or disturb other elderly people in the same room due to the overall increase in volume.

[0203] In one possible implementation, step S109 is also included.

[0204] In step S109, when the target environmental noise is detected to disappear, during the tenth time period after the ninth time period, a soundscape intervention audio signal generated based on the gain of the environmental masking compensation layer audio signal during the tenth time period is played, wherein the gain of the environmental masking compensation layer audio signal gradually decreases within a preset duration during the tenth time period.

[0205] The start time of the tenth time period can be the moment when the target environmental noise is detected to have disappeared.

[0206] It should be noted that the gain of the audio signals of the natural ambient sound layer, rhythm guidance layer, background noise layer, and voice soothing layer is 0% in the tenth time period.

[0207] In this embodiment, any existing mixing algorithm can be used to mix the audio signal set according to the gain of each target audio signal in the tenth time period, so as to obtain the sound scene intervention audio signal generated according to the gain of the ambient masking compensation layer audio signal in the tenth time period.

[0208] In this embodiment of the application, for a target audio signal, the gain of the target audio signal at the beginning of the tenth time period is: the initial gain of the target audio signal with respect to the disappearance of noise, which can be preset.

[0209] The gradual decrease in the gain of the audio signal of the environmental masking compensation layer within the tenth time period and within the preset duration can be interpreted as: within the tenth time period, the gain of the audio signal of the environmental masking compensation layer gradually decreases within the time period that begins at the start of the tenth time period and lasts for the preset duration.

[0210] In one possible implementation, the preset duration is 30 seconds.

[0211] In one possible implementation, during the tenth time period, the gain of the ambient masking compensation layer audio signal is smoothly reduced from 40% to 15%.

[0212] In one possible implementation, the total volume remains constant during the tenth time period.

[0213] In one possible implementation, step S110 is also included.

[0214] In step S110, it is determined whether there is a frequency band of hearing loss for elderly users; if so, the initial gain of the component of the target audio signal in the frequency band of hearing loss for elderly users is set to be higher than the initial gain of the component of the target audio signal in frequency bands other than the frequency band of hearing loss for elderly users.

[0215] In step S110, it can be determined whether there is a frequency band of hearing loss for elderly users based on the hearing compensation curve of the elderly users.

[0216] In step S110, if there is a frequency band of hearing loss for elderly users, for a target audio signal, the initial gain of the component of the target audio signal in the frequency band of hearing loss for elderly users is set to be higher than the initial gain of the component of the target audio signal in frequency bands other than the frequency band of hearing loss for elderly users.

[0217] The frequency band indicating hearing loss for elderly users can indicate that the elderly user cannot hear the audio signal representing the frequency band of their hearing loss.

[0218] Setting the initial gain of the target audio signal in the frequency band of hearing loss for elderly users to be higher than the initial gain of the target audio signal in frequency bands other than the frequency band of hearing loss for elderly users is equivalent to increasing the gain of the target audio signal in the frequency band of hearing loss for elderly users, thus preventing elderly users from not being able to perceive the sound represented by the target audio signal in the frequency band of hearing loss for elderly users.

[0219] In one possible implementation, step S111 is also included.

[0220] In step S111, when the volume of the ambient noise is greater than the ambient noise threshold, the initial gain of the ambient masking compensation layer audio signal for the target sound scene intervention mode is calculated based on the base gain of the ambient masking compensation layer audio signal for the target sound scene intervention mode and the volume of the ambient noise.

[0221] The target soundscape intervention mode can be any soundscape intervention mode. The base gain of the ambient masking compensation layer audio signal for the target soundscape intervention mode is preset.

[0222] When calculating the initial gain of the audio signal of the environmental masking compensation layer, if the volume of the environmental noise is greater than the environmental noise threshold, the initial gain of the audio signal of the environmental masking compensation layer for the target sound scene intervention mode is calculated based on the base gain of the audio signal of the environmental masking compensation layer for the target sound scene intervention mode and the volume of the environmental noise.

[0223] When the ambient noise volume is greater than the ambient noise threshold, the initial gain of the ambient noise masking compensation layer audio signal for the target sound scene intervention mode is calculated based on the base gain of the ambient noise masking compensation layer audio signal for the target sound scene intervention mode and the ambient noise volume. This is equivalent to increasing the initial gain of the ambient noise masking compensation layer audio signal for the target sound scene intervention mode on the basis of the base gain of the ambient noise masking compensation layer audio signal for the target sound scene intervention mode when the ambient noise volume is greater than the ambient noise threshold. This improves the ambient noise masking effect when the ambient noise volume is greater than the ambient noise threshold.

[0224] The initial gain of the ambient masking compensation layer audio signal for the target soundscape intervention mode can be calculated using the following formula, based on the base gain of the ambient masking compensation layer audio signal for the target soundscape intervention mode and the volume of the ambient noise: The initial gain of the ambient masking compensation layer audio signal for the target sound scene intervention mode = the base gain of the ambient masking compensation layer audio signal for the target sound scene intervention mode + (the volume of ambient noise - the ambient noise threshold) × 1.5%.

[0225] In this embodiment, the initial gain of the audio signal of the environmental masking compensation layer in the frequency band to which the noise energy of the highest component of the environmental noise belongs is higher than the initial gain of the components of the audio signal of the environmental masking compensation layer in frequency bands other than the frequency band to which the noise energy of the highest component of the environmental noise belongs.

[0226] In one possible implementation, the main control module can initialize the soundscape parameters. Inputs include: template ID, ambient noise spectrum (env_noise_spectrum), and user hearing compensation parameters (read from the storage module, such as high-frequency hearing loss compensation curves). The main control module retrieves the basic parameters of the corresponding template (initial gain matrix G0, EQ configurations for each layer, and special parameters such as the initial value of rhythm frequencies) from the template library in the storage module. Then, the following corrections are performed sequentially: Ambient noise correction: The initial gain of the masking compensation layer is calculated based on env_noise_spectrum. If the ambient noise LAeq > 40dB, then the masking layer gain = basic gain + (LAeq - 40) × 1.5%, with an upper limit of 50%. Simultaneously, the EQ configuration of the masking layer is adjusted based on the spectrum analysis results (increasing the noise energy by 3-6dB in the highest noise frequency band).

[0227] User hearing compensation correction: Adjust the EQ of each layer according to the user's hearing compensation curve (obtained through pure tone audiometry and stored in the device). For example, if the user has a hearing loss of 20dB at 4kHz, the EQ gain of all sound layers in the 4kHz band will not exceed 10dB (to avoid overcompensation leading to distortion), while the low-pass filter attenuation slope will be increased at 8kHz (to protect residual high-frequency hearing).

[0228] Time-corrected volume: If the current mode is night mode, the gain of all sound layers is multiplied by a night factor of 0.8, and the total volume limit is reduced to 45dB SPL; if the current mode is day mode, the gain is multiplied by 1.0, and the total volume limit is 55dB SPL. Output: Initial gain matrix G0 = [g1, g2, g3, g4, g5] (values ​​range from 0-100%), initial EQ parameters (8-band EQ gain values ​​for each layer), initial rhythm frequency f0 (unit: bpm).

[0229] In one possible implementation, the user presses and holds the "Emergency Stop" button (≥2 seconds) or abnormal physiological data is detected (e.g., heart rate >120 bpm or <40 bpm). The main control module immediately performs a rapid fade-out of all sound layers (reducing to 0 within 1 second), shuts down the power amplifier, saves the current state (playback duration, used template, physiological data records), and announces "Stopped, please take care of your health" (only in daytime mode). If abnormal physiological data is detected, an alarm notification is simultaneously sent to the paired mobile app via Bluetooth. Output: Stop command and alarm notification (if applicable).

[0230] In one possible implementation, the user presses a mode switching button or knob (short press <1 second). The sequence is: Sleep Induction → Nighttime Awakening and Re-sleep → Anxiety Relief → Daytime Soothing → Emotional Calming → Blood Pressure Reduction → Return to Sleep Induction. The current template name is announced via voice. The new template ID and corresponding audio parameters are also announced.

[0231] In one possible implementation, the user double-clicks the "Like" button or the "Dislike" button (optional). If the user marks it as "Like," the current template and parameters are recorded as the user's preference, increasing the template's weight in automatic matching; if the user marks it as "Dislike," the template's weight is decreased, and the user is prompted to switch to another template. This feedback is used for long-term personalized optimization. Output: Updated user preference parameters.

[0232] In one possible implementation, the system records the time, duration, template, physiological data, environmental noise data, and user actions for each use. The main control module records this data in a structured format (JSON or CSV) to the storage module, automatically saving it after each use. The data can be used for: user self-management (exporting to a mobile app to view usage trends and effects), remote monitoring by doctors / nurses (uploading to the cloud via Bluetooth or WiFi), and algorithm optimization (optimizing template matching weights and parameters based on historical data). Data recording adheres to privacy protection principles, with local storage encrypted (AES-128) and cloud transmission using TLS 1.2 encryption. Output: local data file or cloud upload confirmation.

[0233] The method provided in this application embodiment is executed by the player provided in this application embodiment. The player provided in this application embodiment can be called a layered soundscape therapy player for emotional and sleep intervention in the elderly.

[0234] The soundscape therapy player comprises: a housing assembly, an audio output assembly, a main control module, a storage module, a soundscape layering generation module, a status input module, an age-friendly interactive module, and a power supply module. These modules are interconnected via standardized communication and audio interfaces, forming a complete audio signal chain and control signal chain.

[0235] The shell structure of the soundscape therapy player can be in one of the following forms: bedside, pillowside, neck-hanging, or backrest-attached.

[0236] Neck-mounted and backrest-attached types have mounting positions for vibration transmission units; pillow-side and bedside types rely more on the sound field focusing of near-field directional speakers; neck-mounted types emphasize lightweight (total weight ≤150g) and flexible support.

[0237] The housing assembly is used to house and protect the various functional modules, and offers four switchable structural forms to suit different application scenarios and user preferences.

[0238] Neckband style: Employs a flexible or semi-flexible neck support structure made of medical-grade silicone or shape memory alloy (such as NiTiNol) to conform to different neck curves. A near-field directional speaker is located at each end, with the speaker unit's angle to the horizontal plane adjustable within ±30° (via ball joints or damping hinges) to accommodate different head positions and sleeping postures. The total weight is kept under 150g (including battery), with a symmetrical weight distribution design to avoid unilateral pressure. The speakers are positioned 10-15cm from the ears to create a near-field sound field. A vibration transmission unit (eccentric rotary motor or linear resonant motor, frequency range 50-200Hz) can be integrated into the back of the housing for tactile feedback and low-frequency rhythm transmission.

[0239] Pillow-side type: Featuring a flattened edge structure and an overall thickness of ≤25mm, it can be placed on the side of the pillow or the edge of the mattress, without taking up bedside table space. The speaker uses a narrow directional design, focusing the sound field onto the area above the pillow through a sound duct or acoustic lens (sound field focusing range: a hemispherical area with a radius of 20cm centered on the speaker). The bottom of the shell has an anti-slip silicone pad or Velcro closure to prevent displacement due to body movement during use. The pillow-side type is particularly suitable for side sleeping positions, providing balanced sound pressure levels (difference ≤3dB) to both ears.

[0240] Bedside type: The housing features three mounting options: a support base (for placement on a bedside table), a clamp (for clamping onto the bed frame or headboard, clamping range 20-40mm), or a magnetic fastener (for strong magnetic attachment to a metal bed frame). The speaker is mounted on an adjustable arm, which uses a multi-section damping joint to freely adjust the height (adjustment range 150-300mm) and angle (tilt ±45°, horizontal rotation ±90°), ensuring precise sound field coverage of the user's head and neck area. The bedside type is suitable for use in a sitting or semi-reclining position (such as aiding relaxation while reading before bed).

[0241] Backrest-attached type: The back of the housing features a flexible strap (≥30mm wide, made of elastic fabric, adjustable in length 300-500mm) or a buckle structure (compatible with the fixing rods of wheelchairs, recliners, and sofa backrests) to securely attach the player to the backrest surface. The speaker faces the user's head and back, at a distance of approximately 20-30cm. A vibration transmission unit mounting position (vibration motor with a diameter ≥20mm) is provided in the middle or lower part of the housing, allowing vibration to be directly transmitted to the user's back through the backrest, creating a combined "sound + vibration" sensory effect. The backrest-attached type is particularly suitable for relaxing use in living rooms and public areas of elderly care facilities during the day.

[0242] The soundscape healing player supports a five-layer soundscape architecture and adaptive control method; all are limited by elderly-safe acoustic parameters such as peak volume ≤55dB SPL (daytime) / ≤45dB SPL (nighttime), high frequency cutoff ≤8kHz, and low frequency cutoff ≥80Hz; it is equipped with an age-friendly interactive module (large buttons, knobs, voice broadcast).

[0243] The audio output component includes at least one speaker unit, preferably two near-field directional speakers (left and right), for outputting therapeutic soundscapes to the user's head or periauricular region. In some embodiments, it may also include a vibration conduction unit or a bone conduction output unit to create a composite non-in-ear playback effect.

[0244] The main control module is used to perform audio scheduling, mode switching, status recognition, parameter adjustment, and interactive control.

[0245] MCU Selection Recommendations: We recommend using ARM Cortex-M4 or Cortex-M7 series microcontrollers (such as STM32F4xx, NXP i.MX RT series), with a main frequency ≥80MHz, and an integrated single-precision floating-point unit (FPU) and DSP instruction set (such as SIMD instructions) to support real-time audio processing (EQ, compression, mixing) and sensor data fusion. Memory Configuration: Flash ≥512KB (for storing firmware and configuration data), SRAM ≥128KB (for runtime data and audio buffers). For more complex audio processing (such as multi-channel FFT, advanced dynamic range control), a dedicated low-power DSP (such as TI C55x series or Analog Devices SHARC series) can be used as an audio coprocessor, with the main MCU handling task scheduling and communication.

[0246] Communication interfaces with each module: I2S audio interface (at least 1 set, supporting master mode, clock frequency 1.536MHz or 3.072MHz, corresponding to 48kHz sampling rate × 32bit × 1 channel or 2 channels); I2C sensor interface (at least 2 sets, standard mode 100kHz or fast mode 400kHz, connecting to accelerometer, MEMS microphone, proximity sensor); UART Bluetooth interface (1 set, baud rate 115200bps, connecting to BLE module, supporting AT commands or custom protocols); GPIO button interface (at least 8 GPIOs, supporting interrupt triggering and matrix keyboard scanning, connecting to physical buttons, rotary encoder, LED, vibration motor); SPI / SDIO storage interface (1 set, clock frequency ≥ 25MHz, connecting to SD card or Flash chip).

[0247] Real-time operating system or interrupt-driven task scheduling: The main control module can use a real-time operating system (RTOS, such as FreeRTOS or RT-Thread) for multi-task scheduling, or a pure interrupt-driven polling architecture. The main tasks include: audio processing (highest priority, executed every 1ms, handling audio buffer filling and DSP operations); sensor sampling (executed every 10ms, reading accelerometer and microphone data); Bluetooth communication (executed every 50ms, handling BLE data reception and transmission); user interaction (executed every 20ms, scanning button and knob states); and state machine (executed every 100ms, performing template matching, state determination, and parameter adjustment). The audio processing task employs a double buffering mechanism to ensure uninterrupted audio playback.

[0248] The main control module (ARM Cortex-M4 or equivalent low-power MCU, clock speed ≥80MHz, built-in DSP instruction set) is located at the center of the system. It connects to the DAC and power amplifier of the audio output component via the I2S (Inter-IC Sound) bus to transmit 24bit / 48kHz digital audio stream; it connects to the motion sensor (three-axis accelerometer), environmental noise sensor (MEMS microphone), and proximity sensor (infrared or capacitive) via the I2C (Inter-Integrated Circuit) bus to read sensor data; it connects to the Bluetooth BLE module via UART (Universal Asynchronous Receiver-Transmitter) to receive data from external health devices (blood pressure monitor, heart rate monitor); it connects to the storage module (SD card or eMMC Flash) via SPI (Serial Peripheral Interface) or SDIO to read soundscape audio files and template parameters; and it connects to physical buttons, rotary encoders, LED displays, and vibration motors via GPIO (General Purpose Input / Output) to process user interaction input and output feedback. The power module is powered by a lithium battery (3.7V, capacity ≥2000mAh) or an external USB power source, providing the required voltage to each module through an LDO and a DC-DC converter.

[0249] The storage module stores multiple soundscape healing templates and their corresponding parameters, as well as digital audio source files. Storage media includes MicroSD cards (capacity ≥ 16GB, Class 10 or higher) or onboard eMMC Flash (capacity ≥ 8GB). Audio file formats are lossless PCM (WAV) or high-bitrate lossy compression (AAC, 256kbps or higher) to ensure sound quality meets healing requirements.

[0250] Table 6 shows an example of a soundscape healing template.

[0251] Each template defines the initial gain ratio, frequency band characteristics, and special parameters for five layers of audio signals: Table 6 Template parameter data structure: Each template is stored in structured data format (such as JSON or binary protocol), including template ID (1 byte), template name (UTF-8 string, maximum 16 Chinese characters), five initial gains (1 byte each, range 0-100), EQ configuration for each layer (gain values ​​for 8 frequency bands per layer, 1 byte each, range -12dB to +12dB), and special parameters (such as rhythm frequency start value, end value, and rate of change, 2 bytes each). The main control module reads the corresponding template from the storage module during mode switching and parses it into runtime parameters.

[0252] Audio file organization: Each template corresponds to a set of audio file resources, including natural ambient sound audio (such as "Rain_48k.wav", "Ocean Waves_48k.wav", "Forest_48k.wav", etc., each with a duration of ≥10 minutes and can be looped), background noise audio (pre-generated files of white noise / pink noise / brown noise, with a duration of ≥10 minutes), rhythm guidance audio (low-frequency beat sequences, pre-generated according to different bpm, such as "beat_6bpm.wav", "beat_8bpm.wav", etc.), voice soothing audio (pre-recorded guidance or TTS synthesized speech, such as "Deep Breath.wav", "Relax Shoulders.wav", etc.), and masking compensation audio (sharing the resource library with the natural ambient sound layer, but with different EQ parameters). All audio files are uniformly in 48kHz / 16bit stereo format to ensure consistent sampling rate and bit depth during mixing.

[0253] The soundscape layering generation module is responsible for organizing and mixing audio content into layers for output. This module adopts an architecture of "five-layer parallel processing + independent gain control + bus post-processing", which allows each layer of audio components to be adjusted independently according to the user's state, thereby achieving fine-grained soundscape intervention.

[0254] Complete description of the audio signal chain: Digital audio source (PCM / WAV files in the storage module, sampling rate 48kHz, bit depth 16bit or 24bit) → Soundscape layering generation module (assigns the digital audio source to 5 independent audio layers) → Independent DSP processing for each layer (digital equalizer EQ, dynamic range compression DRC, gain envelope control) → Mixer (digital domain mixing, floating-point arithmetic) → Bus processing (bus compression, peak limiter) → DAC (digital-to-analog conversion, 24bit / 48kHz) → Power amplifier (Class-D, efficiency ≥85%, output power ≤1W) → Speaker (near-field directional speaker, frequency response 200Hz-8kHz) or vibration conduction unit (bone conduction or vibration motor, frequency range 50-500Hz) → User's ear / head and neck. The entire signal chain, from the digital domain to the analog domain, is controlled by the real-time parameter adjustment of the main control module.

[0255] The soundscape layer generation module performs audio layer generation and mixing. The inputs to the soundscape layer generation module are: G0 and the audio sources for each layer (PCM audio data read from the storage module). The module performs independent digital signal processing on each audio layer: first, digital equalization (EQ, 8-band parametric equalization, adjusting the gain of each band according to the EQ parameters output by S3); then, dynamic range compression (DRC, compression ratio ≤ 2:1, threshold -20dBFS, attack time 5ms, release time 50ms); followed by gain envelope control (setting the gain of each layer according to G0, with the gain change rate limited to 5dB / s). The processed audio streams are then mixed using floating-point addition in the mixer: Y_mix = Σ(gi × Yi_EQ + DRC). The mixed audio stream undergoes bus compression (compression ratio 1.5:1) and peak limiting (Limiter, threshold -3dBFS). Output: Mixed digital audio stream (24bit / 48kHz, I2S format).

[0256] The soundscape layering generation module performs smooth transitions and outputs. Inputs to the soundscape layering generation module: target audio stream Y_mix, current playback status (current volume, current play / stop status). The soundscape layering generation module performs a fade-in: increasing the volume from 0 to the target volume within 3-5 seconds using a logarithmic curve (volume(t) = V_target × log10(1 + 9×t / T_fadein) / log10(10), where t is the current time and T_fadein is the total fade-in time). If playback is currently in progress and the target volume changes, a smooth transition is performed: the volume change rate is limited to 5dB / second. If template switching is in progress, crossfade is performed: the old template fades out within 5 seconds, and the new template fades in within 3 seconds, with a 2-second overlap. The audio stream is converted to an analog signal via a DAC (24bit / 48kHz), and then driven by a power amplifier (Class-D, output power ≤1W) to power speakers or vibration conduction units. Output: analog audio signal (driving speakers) or vibration signal (driving vibration motors).

[0257] The status input module is used to acquire user or environmental status information, providing a data foundation for adaptive control methods. The input sources for the status input module are divided into: Ambient noise sensor: A MEMS microphone (such as Knowles SPH0645 or TDK InvenSense ICS-43434) is used, with a sampling rate of 48kHz, a bit depth of 24bit, and a signal-to-noise ratio ≥65dB. The microphone is mounted on the housing surface, facing the environment (opposite to the user's direction) to minimize crosstalk from the player's own output sound (further suppressed by the echo cancellation algorithm AEC). The main control module performs FFT spectral analysis (2048 points, once every 10 seconds) on the signal collected by the microphone to calculate the main energy bands (Top 3 Energy Bands) and A-weighted equivalent sound pressure level (LAeq) of the ambient noise. The FFT results are used to dynamically adjust the parameters of the environmental masking compensation layer.

[0258] Body movement sensor: Employs a triaxial accelerometer (such as ST LIS3DH or Bosch BMA400), with selectable ranges of ±2g / ±4g / ±8g, a sampling rate of 50Hz, and a resolution of 12bit. The accelerometer is installed inside the housing and comes into direct or indirect contact with the user's body (direct contact for neck-mounted and backrest-attached types, and indirect transmission of body movement signals through the mattress for pillow-side and bedside types). The main control module processes the acceleration data in real time: first, it removes the DC component of gravity through a high-pass filter (cutoff frequency 0.5Hz); then, it calculates the short-time energy of the triaxial acceleration vector sum (window length 2 seconds, step size 0.5 seconds); finally, it identifies body movement events through threshold judgment (energy threshold 0.05g², duration threshold 1 second). Body movement frequency (number of body movement events per minute) is used to determine sleep stage: a body movement frequency >6 times / minute indicates wakefulness or light sleep, 3-6 times / minute indicates light to moderate sleep, and <3 times / minute indicates deep sleep.

[0259] Proximity sensor: An infrared proximity sensor (such as the Vishay VCNL4040, detection distance 0-20cm) or a capacitive proximity sensor (such as the Azoteq IQS266) is used, installed near the speaker. When the system detects a user's head or body entering the effective listening range (10-30cm), it confirms the user's presence and allows normal playback. When the system detects that the user has left the effective range for more than 5 minutes, it automatically lowers the volume or pauses playback to save power and avoid disturbing others.

[0260] Time information module: Employs an RTC real-time clock chip (such as NXP PCF8563 or Seiko Epson RX-8025), powered by an independent button cell battery (CR2032), ensuring no time information is lost when the main power is turned off. The RTC provides year, month, day, hour, minute, and second information for automatically switching between day / night modes (default settings: night mode 21:00-7:00, day mode 7:00-21:00; users can customize this via the interactive module), recording intervention timestamps, and implementing timed playback functionality.

[0261] External Health Device Communication Interface: Wireless communication with external health devices is achieved via a Bluetooth BLE 4.2 / 5.0 module (such as Nordic nRF52832 or TICC2640R2F). Supported device types include: electronic blood pressure monitors (compliant with Bluetooth SIG Blood Pressure Monitor Profile, UUID 0x1810), transmitting systolic blood pressure, diastolic blood pressure, and heart rate data; heart rate monitoring devices (compliant with Bluetooth SIG Heart Rate Profile, UUID 0x180D), transmitting real-time heart rate (1Hz update) and heart rate variability (HRV) data; and sleep monitoring devices (via a custom GATT service), transmitting sleep stage (awake / light sleep / deep sleep / REM) and body movement data. The communication protocol adopts BLE GATT (Generic Attribute Profile), with the main control module acting as the central device and the external health device as the peripheral device. The connection interval is 50-100ms to ensure data real-time performance.

[0262] The age-friendly interactive module includes physical buttons, knobs, tactile feedback structures, a voice broadcast module, and a simplified display module. Its design follows four principles: "one-click access, tactile confirmation, voice assistance, and accidental touch protection."

[0263] Physical button design: The device features at least three large physical buttons on the front, with raised textures on the surface (e.g., a moon for sleep mode, a sun for daytime mode, and a heart for mood-soothing mode) to facilitate tactile identification for elderly users in low-light conditions. The buttons provide clear tactile feedback. The button layout uses a linear or triangular arrangement to reduce the probability of accidental presses.

[0264] Knob Function: A rotary encoder knob is provided for volume adjustment. The knob uses a discrete level design (5 discrete volume levels: mute / low / medium / high / high, corresponding to output sound pressure levels of 0dB / 25dB / 35dB / 45dB / 55dB SPL), rather than continuous sliding adjustment. Each rotation confirms the operation with a "click" mechanical feedback and vibration (50ms short vibration), and announces the current volume level ("Volume: Medium") via voice. This discrete level design prevents incorrect volume settings due to hand tremors or inaccurate operation by elderly users.

[0265] Tactile feedback: The device has a built-in vibration motor for tactile feedback in the following scenarios: button press confirmation (20ms short vibration), knob speed switch confirmation (30ms short vibration), mode switch completion (50ms double short vibration), nighttime accidental touch lock activation (100ms long vibration), and low battery reminder (20ms vibration every 30 seconds). The vibration frequency is 100-200Hz, with moderate amplitude; the vibration intensity is reduced by 50% in nighttime mode.

[0266] Voice prompts: The device has a built-in voice synthesis chip for operation confirmation announcements ("Sleep mode activated"), current mode announcements ("Current mode: Nighttime Soothing"), error prompts ("Please try again"), and status prompts ("High blood pressure, stress relief mode activated"). The voice prompt volume is independent of the music volume and can be set separately in the interaction module (range 20-50dB SPL). In night mode, "Nighttime Silent Prompt" is enabled: voice prompts are output only through headphones / speakers at a very low volume (≤25dB SPL), or voice prompts are completely disabled (only vibration feedback is retained) to avoid waking the sleeping partner.

[0267] Simplified Display: The front of the device features a small LED dot matrix screen that displays only the most essential information: the current mode icon (represented by LED patterns such as a moon, sun, or heart) and volume level (levels 1-5, indicated by the corresponding number of lit LEDs). It does not display the time, menu text, or complex icons, minimizing cognitive load. Automatic Brightness Adjustment: Daytime mode is 100% brightness, while nighttime mode reduces brightness to 20% and switches to warm red / orange light (wavelength >590nm, reducing the suppression of melatonin secretion by blue light).

[0268] Accidental Touch Lock Mechanism: To prevent accidental operation during half-awake periods at night, the device features a dual locking mechanism. Automatic Lock: After entering night mode (or manually selecting a night-related template), the system automatically locks after 30 seconds. Manual Lock: Press and hold the "Mode Switch" button for 3 seconds; the device will vibrate to confirm and then lock. Valid Operations During Locking: Only the "Emergency Stop" button (press and hold for 2 seconds to immediately stop all playback and turn off the device) and the "Increase Volume" knob (increases by 3dB per level, up to a maximum of 45dB SPL) are valid. All other button and knob operations are ignored, and a short vibration indicates "Locked." Unlock Method: Double-click any button (with an interval of <500ms between presses) or rotate the knob two levels; the device will vibrate and announce "Unlocked," restoring all operational functions.

[0269] Table 7 shows an example of the input and output signals for each module.

[0270] Table 7 Real-time data from environmental noise sensors (48kHz sampling, FFT spectrum results), real-time data from body motion sensors (50Hz acceleration sampling), proximity sensor status (present / absent), time information module (RTC clock, accurate to the second), and external health device communication interface data (blood pressure monitor BLE data, heart rate belt BLE data, optional).

[0271] Upon powering on or waking from standby, the main control module first performs system initialization: initializing peripheral interfaces such as I2S / I2C / UART / GPIO, reading system configuration parameters (such as user hearing compensation parameters, volume preferences, and night mode time periods) from the storage module, and calibrating sensor baselines (ambient noise baseline and body movement / resting baseline). It then establishes the current state vector S = [time, location, activity, env_noise_spectrum, env_noise_Leq, body_movement_freq, proximity_state, systolic_bp, diastolic_bp, heart_rate, sleep_stage, user_preference]. Wherein, time is the 24-hour time (0-23); location is the location information (bedside / pillowside / neckrest / backrest, manually set by the user or determined by proximity sensors); activity is the user's activity state (awake / light sleep / deep sleep / night awakening, determined by a combination of body movement frequency and heart rate); env_noise_spectrum is the ambient noise spectrum vector (energy values ​​of the first 8 FFT frequency bands, in dB); env_noise_Leq is the ambient noise A-weighted equivalent sound pressure level (in dB); body_movement_freq is the body movement frequency. Activity frequency (beats / minute); proximity_state is the proximity state (0 = away, 1 = present); systolic_bp and diastolic_bp are systolic and diastolic blood pressure (unit: mmHg, set to -1 if no external device is available); heart_rate is the heart rate (beats / minute, set to -1 if no external device is available); sleep_stage is the sleep stage (0 = awake, 1 = light sleep, 2 = deep sleep, 3 = REM); user_preference is the user's historical preferences (the template ID used most frequently in the last 7 days).

[0272] The player can employ the following four playback strategy optimization mechanisms: Night Mode Strategy: In Night Mode (default 22:00-7:00), the system implements the following parameter constraints: the high-frequency cutoff frequency is further reduced from 8kHz to 6kHz to reduce the inhibition of melatonin secretion by high-frequency stimulation; the dynamic range limit is tightened from 15dB to 10dB to avoid any sudden volume changes; the voice soothing layer is disabled (forced gain to 0%) to avoid cognitive activation caused by language understanding; and stable and continuous natural sounds (such as continuous rain or calm ocean waves) are prioritized over soundscapes containing sudden events (such as intermittent thunder or sudden bird calls) to reduce the unpredictability of sound events.

[0273] Anxiety calming strategy: In anxiety calming mode, the rhythm guidance layer is enhanced first (gain ≥ 40%) to guide a decrease in breathing rate, while sudden sound components in the natural ambient sound layer are reduced (using a real-time transient detection algorithm, when a transient event is detected in the natural sound, the gain of this layer is temporarily reduced by 6dB, lasting for 2 seconds before being restored). Slow, fluctuating soundscape changes (volume change rate ≤ 2dB / min, EQ change rate ≤ 1dB / min) are used to avoid any sudden changes that might trigger alertness.

[0274] Noise Masking Strategy: When ambient noise increases, the system does not employ a simple "overall increase in volume" strategy. Instead, it activates a targeted enhancement of the ambient masking compensation layer. Specifically, the system analyzes the main energy frequency bands of the ambient noise (obtaining the Top 3 bands via FFT), and then enhances natural sounds within the same frequency band (such as stream sounds, which have rich spectral content in the 500-2000Hz range, suitable for masking human voices and construction noise). The gain increase of the masking compensation layer is simultaneously constrained by a total volume limiter to ensure that the total output does not exceed a safe threshold. When the ambient noise disappears, the masking compensation layer smoothly returns to its base gain within 30 seconds.

[0275] Stable Rest Detection Strategy: Once the system detects that the user has entered stable sleep (body movement frequency < 2 times / minute, heart rate decrease), the following strategy is implemented: The voice soothing layer fades out first (if enabled); the rhythm guidance layer begins to fade out after 5 minutes to avoid adaptive saturation of the autonomic nervous system due to prolonged rhythm guidance; the natural ambient sound layer begins to fade out 10 minutes after the rhythm guidance layer fades out; finally, a low-intensity background noise layer (pink or brown noise, gain 10-15%, corresponding to a sound pressure level of 15-20 dB SPL) is retained as a continuous environmental masking, and completely turned off after 60 minutes. This layered fading strategy ensures a smooth and natural transition from soundscape intervention to natural sleep, avoiding the "sound withdrawal" effect caused by sudden cessation.

[0276] In one possible implementation, the main control module is used for mode selection or automatic determination. Input: State vector S, user's historical preferences (user_preference). Processing: The main control module first determines whether the user manually selected a mode via a physical button. If manual mode selection is detected (e.g., pressing the "Sleep Mode" button), it directly proceeds to S3 and loads the template selected by the user. If no manual selection is detected, the automatic template matching algorithm is initiated. The template matching algorithm uses a weighted scoring method to calculate the matching score for each template: score_i = w1 × time_match_i + w2 × state_match_i + w3 × preference_match_i + w4 × emergency_match_i. Where: time_match_i is the time matching score (1 if the current time falls within the template's recommended time period, 0 otherwise. For example, the sleep induction template recommends 22:00-24:00, and the daytime relaxation template recommends 8:00-20:00); state_match_i is the state matching score (calculated based on the degree of matching between activity, bp, hr, and other state parameters and the applicable state of the template, ranging from 0 to 1); preference_match_i is the preference matching score (1 if the template ID equals user_preference, 0 otherwise); emergency_match_i is the emergency matching score (1 if blood pressure reaches the emergency threshold, 1 otherwise). The weight coefficients are set to w1=0.3, w2=0.4, w3=0.2, w4=1.0 by default (the emergency matching score has the highest priority; once triggered, the corresponding template is selected directly).

[0277] The main control module calculates the score_i of all templates and selects the template with the highest score as the current target template. If the highest score is lower than the threshold (e.g., 0.5), the "daytime soothing template" (the most common template) is selected by default. Output: The ID of the selected soundscape template (an integer from 1 to 6).

[0278] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0279] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented in hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0280] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A layered soundscape therapy method for emotional and sleep intervention in the elderly, characterized by: The method includes: When elderly users are asleep at night, during the first period, a soundscape intervention audio signal is played, which is obtained by mixing the gain of each target audio signal in the audio signal set during the first period. The audio signal set includes the following target audio signals: natural ambient sound layer audio signal, rhythm guidance layer audio signal, background noise layer audio signal, environmental masking compensation layer audio signal, and voice soothing layer audio signal. During the second period following the first period, a soundscape intervention audio signal is played, which is obtained by mixing the gain of each target audio signal in the second period. During the second period, the beats per minute (bpm) of the rhythm guidance layer audio signal is reduced at a preset first bpm reduction rate. When it is determined that the elderly user is asleep during the second time period, in the third time period after the second time period, the soundscape intervention audio signal obtained by mixing the gain of each target audio signal in the third time period is played. In the third time period, the gain of the natural ambient sound layer audio signal, the rhythm guidance layer audio signal, the environmental masking compensation layer audio signal, and the background noise layer audio signal fades out in the third time period. During the fourth period following the third period, a soundscape intervention audio signal generated based on the gain of each target audio signal in the fourth period is played, wherein the gain of the background noise layer audio signal fades out in the fourth period.

2. The method according to claim 1, characterized in that: The method further includes: When elderly users are in a state of waking up at night and then falling back asleep, during the fifth time period, a soundscape intervention audio signal generated according to the gain of each target audio signal in the fifth time period is played. During the fifth time period, the bpm of the rhythm guidance layer audio signal is reduced at a preset second bpm reduction speed. When the heart rate of an elderly user meets the condition of heart rate decline during the fifth time period, a soundscape intervention audio signal generated according to the gain of each target audio signal in the sixth time period is played during the sixth time period after the fifth time period. In the sixth time period, the gain of the rhythm guidance layer audio signal, the natural ambient sound layer audio signal, and the background noise layer audio signal fades out in the sixth time period.

3. The method according to claim 2, characterized in that: The method further includes: When an elderly user is detected to meet the preset condition of increased body movement frequency, it is determined that the elderly user is waking up at night. If, within a preset time period from the moment an elderly user is identified as waking up at night, the elderly user does not perform any actions corresponding to the state of falling back asleep after waking up at night, and the elderly user's current body movement frequency is greater than the upper limit threshold of body movement frequency, the elderly user is determined to be in the state of falling back asleep after waking up at night.

4. The method according to claim 1, characterized in that: The method further includes: When the target blood pressure level of an elderly user is detected to be at the hypertension risk level, a soundscape intervention audio signal generated based on the gain of each target audio signal in the seventh time period is played. The rhythm guidance layer audio signal is the target audio signal with the highest gain in the seventh time period.

5. The method according to claim 4, characterized in that: The method further includes: If the heart rate of an elderly user drops to a level not exceeding the heart rate threshold during the seventh time period, the elderly user is prompted to have their blood pressure retested. When the target blood pressure of an elderly user is retested and is classified as slightly elevated, a soundscape intervention audio signal, generated based on the gain of each target audio signal in the eighth time slot after the seventh time slot, is played.

6. The method according to claim 1, characterized in that: The method further includes: When target ambient noise is detected, during the ninth time period, a soundscape intervention audio signal generated based on the gain of the ambient masking compensation layer audio signal during the ninth time period is played. The gain of the ambient masking compensation layer audio signal during the ninth time period is determined based on the gain of the component of the ambient masking compensation layer audio signal in the target energy frequency band of the target ambient noise during the ninth time period, and the gain of the component of the ambient masking compensation layer audio signal in frequency bands other than the target energy frequency band during the ninth time period. The energy of the component of the target ambient noise in the target energy frequency band is greater than the energy of the component of the target ambient noise in frequency bands other than the target energy frequency band, and the gain of the component of the ambient masking compensation layer audio signal in the target energy frequency band during the ninth time period is greater than the gain of the component of the ambient masking compensation layer audio signal in frequency bands other than the target energy frequency band during the ninth time period.

7. The method according to claim 6, characterized in that: The method further includes: When the target ambient noise is detected to have disappeared, during the tenth period following the ninth period, a soundscape intervention audio signal generated based on the gain of the ambient masking compensation layer audio signal during the tenth period is played. During the tenth period, the gain of the ambient masking compensation layer audio signal gradually decreases over a preset duration.

8. The method according to claim 1, characterized in that: The method further includes: Determine if there are frequency bands in which elderly users experience hearing loss; If so, set the initial gain of the target audio signal in the frequency band of hearing loss for elderly users to be higher than the initial gain of the target audio signal in frequency bands other than the frequency band of hearing loss for elderly users.

9. The method according to claim 1, characterized in that: The method further includes: When the ambient noise volume is greater than the ambient noise threshold, the initial gain of the ambient noise compensation layer audio signal for the target sound scene intervention mode is calculated based on the base gain of the ambient noise compensation layer audio signal for the target sound scene intervention mode and the ambient noise volume.

10. A layered soundscape therapy player for emotional and sleep intervention in the elderly, characterized in that: The layered soundscape therapy player for emotional and sleep intervention in the elderly includes: a processor and a memory, the memory being used to store a computer program; when the computer program is loaded by the processor, it causes the processor to execute the method of any one of claims 1-9.

Images