Musical instrument sound sensitivity detection method based on discrete frequency point pure tone burst

Abstract

The application relates to a discrete frequency point pure tone burst-based musical instrument sound sensitivity detection method, which comprises a contact type patch pickup (2), a resonance sound box (3) for playing a discrete pure tone burst excitation sequence and a multi-channel USB audio interface (4); the contact type patch pickup (2) is pasted on the surface area of a measured musical instrument (6) through a patch fixing structure; the resonance sound box (3) is pasted on the surface side of the measured musical instrument (6) through an adhesive; a Raspberry Pi master control module (1) is connected with the multi-channel USB audio interface (4) through a shared data line; the Raspberry Pi master control module (1) is internally provided with a data processing module (5); the discrete frequency point pure tone burst-based musical instrument sound sensitivity detection method is run and detected on the Raspberry Pi master control module (1) by using the data processing module (5) and specifically comprises excitation sequence generator, synchronous playback recorder, time domain envelope detector, starting time extractor, score calculator, graph generator and PDF report generator steps. The application multiplexes existing hardware (Raspberry Pi, resonance sound box, patch pickup, USB audio interface) and can realize rapid, standardized and repeatable sound sensitivity quantitative evaluation in an ordinary room environment.

Description

Technical Field

[0001] This invention relates to the field of acoustic performance testing technology for musical instruments, specifically to a method for detecting the sounding sensitivity of musical instruments based on discrete frequency pure tone bursts. Background Technology

[0002] The sound sensitivity of a musical instrument (i.e., the instrument's transient response to excitation signals) is one of the core indicators for evaluating its performance quality, directly affecting the player's feedback on string contact / bow movement and the clarity of note onset. However, existing testing methods suffer from the following drawbacks: Significant differences in subjective auditory assessment: Traditional methods rely on the experience of the player or luthier to judge "how loud the instrument is" and "how fast the onset is," resulting in scoring deviations often exceeding ±30%. Furthermore, these methods are susceptible to interference from factors such as playing force, bow speed, and string contact point, leading to extremely poor repeatability. Frequency sweep measurement cannot separate transient response: Existing automated solutions (such as frequency sweep excitation + spectrum analysis) can only acquire the instrument's steady-state frequency response, failing to quantify the transient characteristics during the onset phase. The crucial dynamic parameter of onset time is completely ignored. Confusion between continuous excitation and response: When using continuous frequency sweep or long-duration pure-tone excitation, the superposition of vibrations from preceding and following excitations makes it impossible to accurately extract the independent onset envelope of a single excitation, resulting in unreliable measurement results. Lack of a targeted evaluation system: There are no dedicated testing devices or scoring standards on the market specifically for the single indicator of "sound sensitivity," leaving instrument factories and music stores to rely on subjective judgment. Existing research has included methods such as using impact hammers and accelerometers to measure the admittance characteristics of instruments, but these devices are expensive and complex to operate, requiring the attachment of multiple sensors. Others employ frequency sweeping methods to obtain the frequency response function, but the sweeping signal is continuous in time, and the responses of adjacent frequencies overlap, making it impossible to accurately separate the independent onset process of each frequency point. A few studies have proposed using "pure-tone bursts" to measure the transient response of loudspeakers or rooms, but this has never been applied to the sound sensitivity testing of musical instruments, and there is a lack of automated multi-frequency measurement, envelope time-domain extraction algorithms, and dedicated scoring systems. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention proposes an automated testing method that employs discrete frequency pure tone burst excitation, collects the vibration response of the instrument body through a patch pickup, automatically extracts the start-up time of multiple frequency points, and evaluates the instrument's sound sensitivity. This method is suitable for rapid quality assessment of small and medium-sized stringed instruments (such as violins, violas, guitars, etc.) in ordinary room environments and falls within the intersection of international patent classifications G01H (acoustic measurement) and G10H (electroacoustic instrument evaluation).

[0004] This invention uses burst pure tones of multiple standard frequencies as excitation signals and employs an adaptive envelope detection algorithm to extract the attack time at each frequency point. It then calculates the sensitivity score and response flatness coefficient, outputting a professional test report. This invention reuses existing hardware (Raspberry Pi, resonant speaker, patch microphone, USB audio interface) to achieve rapid, standardized, and repeatable quantitative evaluation of vocal sensitivity in ordinary room environments.

[0005] The technical solution of the present invention is: a method for detecting the sounding sensitivity of a musical instrument based on discrete pure tone bursts, including hardware: a musical instrument under test (6), a Raspberry Pi main control module (1), a contact patch pickup for collecting vibrations of the instrument body through solid conduction (2), a resonant speaker for playing discrete pure tone burst excitation sequences (3), and a multi-channel USB audio interface (4). The contact patch pickup (2) is attached to the soundboard area of ​​the instrument under test (6) through a patch fixing structure. The resonance phonograph (3) is attached to the side of the soundboard of the instrument under test (6) with adhesive. The Raspberry Pi main control module (1) is connected to the multi-channel USB audio interface (4) through a shared data cable. The output end of the contact patch pickup (2) is connected to the recording input channel of the multi-channel USB audio interface (4) through a USB interface data cable. The input end of the resonance phonograph (3) is connected to the playback output channel of the multi-channel USB audio interface (4) through a signal line. The Raspberry Pi main control module (1) has a data processing module (5) inside. The instrument sounding sensitivity detection method based on discrete frequency pure tone bursts is run and detected on the Raspberry Pi main control module (1) using the data processing module (5). Specifically, it includes the steps of an excitation sequence generator, a synchronous playback recorder, a time domain envelope detector, an oscillation time extractor, a scoring calculator, a chart generator, and a PDF report generator. The specific steps are as follows: Step 11: Set detection parameters The data processing module (5) presets a timestamp and a preset detection parameter table, and reads the preset detection parameter table. The detection parameter table includes 5 standard frequencies: 196.0Hz G3, 293.66Hz D4, 440.0Hz A4, 587.33Hz D5 and 880.0Hz A5 as a group. Each frequency is read repeatedly 3 times. The burst sound reading time is 0.4 seconds, the fade-in and fade-out sound reading time is 5ms, and the interval between each frequency in the group is 0.7 seconds. When repeatedly reading the frequency in this group, the extra interval between groups is 1.0 second. At the same time, when starting to read, there is a 2.0 second pre-mute. Step 12: Generate discrete pure tone burst excitation sequence The data processing module (5) generates 15 burst sounds in sequence according to the parameters in step 11, splices them in order and inserts silence intervals to form a 48kHz / 16bit PCM excitation array with a total duration of about 21.8 seconds. Step 13: Synchronize playback stimulus and recording response The data processing module (5) is responsible for generating discrete frequency pure tone burst excitation sequences. The data processing module (5) calls the multi-channel USB audio interface (4) and starts the playback and recording thread contained in the data processing module (5). The playback and recording thread outputs the PCM excitation array to the resonant speaker (3). At the same time, the playback and recording thread records the vibration response of the instrument body through the patch pickup (2) to achieve sample-level alignment (delay ≤ 1 sampling point), forming a response segment sequence and frequency timestamp, which together form the total recorded signal. Step 14: Extract response segments at each frequency point Based on the preset timestamp, 15 response segments are extracted from the total recorded signal, and each segment corresponds to the excitation response of a burst tone; Step 15: Zero-phase bandpass filtering For each response segment, a fourth-order Butterworth zero-phase filter is used, with a center frequency equal to the excitation frequency of the burst and a bandwidth of ±40Hz, to filter out crosstalk from other frequencies and environmental noise. Step 16: Extract the energy envelope The absolute value of the filtered signal is taken, and then passed through a low-pass filter with a cutoff frequency of 150Hz to smooth it out, thus obtaining the energy envelope curve. Step 17: Calculate the noise threshold and the starting point t0 The median of the envelope in the first 30ms of the segment is taken as the environmental noise benchmark, and the noise threshold is calculated as noise benchmark × 5. The first position where five consecutive sampling points of the envelope from the start of the segment are greater than the noise threshold is recorded as the starting point t0. Step 18: Calculate the steady-state energy and the 90% point t90. The average value of the envelope over the last 100ms of the segment is taken as the steady-state energy reference value. Starting from t0, the search begins at the point where the envelope first reaches 90% of the steady-state energy reference value, denoted as t90. Step 19: Calculate the start-up time, sensitivity score, and output report. The start-up time AT = t90 - t0 (milliseconds). Outliers are removed from three repeated measurements at each frequency (1.5 times IQR), and the average AT for that frequency is obtained by averaging. Calculate the overall sensitivity score S = 1000 / AT̄ (AT̄ is the mean of the average AT values ​​at 5 frequency points), and the response flatness coefficient CV = (σ_AT / AT̄) × 100%. Determine the sensitivity level based on the S value (S>80: extremely sensitive, 50-80: sensitive, 30-50: moderate, <30: insensitive). Determine the uniformity level based on the CV value (<15%: highly consistent, 15-30%: slightly different, >30%: uneven response across frequency ranges). Generate a frequency-onset time line graph and a typical frequency envelope diagram, and output a PDF test report.

[0006] Preferably, the instrument under test (6) is a violin, and the resonant sound (3) is attached to the right side of the sound hole on the panel of the instrument under test (6) about 40mm away by adhesive.

[0007] Preferably, the patch pickup (2) is a piezoelectric contact pickup, which is directly attached to the instrument under test (6) directly below the bridge or the nut area of ​​the guitar using traceless acrylic double-sided adhesive.

[0008] Preferably, the instrument under test (6) is a guitar, and the patch pickup (2) is a piezoelectric contact pickup, which is directly attached to the nut area of ​​the instrument under test (6) using traceless acrylic double-sided tape.

[0009] Preferably, the multi-channel USB audio interface (4) adopts a 16bit / 48kHz USB sound card that supports full-duplex synchronous sampling.

[0010] Preferably, the control and data processing module (5) is provided with an excitation sequence generator. The excitation sequence generated by the excitation sequence generator consists of multiple pure tone bursts of standard frequency. Each burst tone has a duration of 0.2–0.6 seconds. Cosine fade-in and fade-out windows are provided before and after each burst tone. A silence interval is provided between burst tones. A pre-silent segment is provided at the beginning of the sequence.

[0011] Preferably, the control and data processing module (5) is provided with a detection parameter table, which includes three or more frequencies of 196.0Hz, 293.66Hz, 440.0Hz, 587.33Hz, and 880.0Hz, with each frequency repeated 2–4 times.

[0012] The present invention has the following quantitative beneficial effects: Achieving specialized testing of musical instrument sound sensitivity: Filling the gap in existing frequency sweep methods that can only measure steady-state frequency response and cannot measure transient oscillation characteristics, quantifying subjective experience into objective oscillation time (millisecond level), with a measurement accuracy of ±0.5ms.

[0013] High measurement repeatability: Each frequency point is excited 3 times. With the anomaly elimination algorithm, the same instrument is measured 3 times in a row. The standard deviation of AT at each frequency point is less than 10% of the mean, which is a significant improvement compared to the ±30% dispersion of human subjective judgment.

[0014] Strong environmental adaptability: It uses solid-conductive sound pickup to isolate most air noise, and the time-domain envelope detection method is not sensitive to stationary noise. Reliable results can be obtained in a normal office environment (ground noise ≤45dBA) without the need for an anechoic chamber.

[0015] High testing efficiency: The total time for a single test (stimulation, recording, analysis, and reporting) is ≤35 seconds, making it suitable for batch sampling inspections on factory production lines.

[0016] The evaluation system is intuitive: it outputs a comprehensive sensitivity score S and a flatness coefficient CV, which are highly correlated with the performer's subjective feelings (blind test correlation coefficient ≥ 0.85), and provides a two-level classification to facilitate understanding by non-professionals. Attached Figure Description

[0017] Figure 1 is a schematic diagram of the overall structure of the present invention; Figure 2 is a magnified view of the attachment position of the patch pickup and resonant speaker on the musical instrument; Figure 3 is a flowchart of the detection method of the present invention; The steps are numbered 10 to 18, totaling 9 steps. Each step is represented by a simplified graphic symbol, and the steps are connected by a right-pointing arrow. The corresponding operations for each step are as follows: 10. Set detection parameters; 11. Generate discrete pure tone burst excitation sequence; 12. Synchronize playback of excitation and recording of response; 13. Extract response segments at each frequency point; 14. Bandpass filtering (zero phase); 15. Extract energy envelope; 16. Calculate noise threshold and oscillation point t0; 17. Calculate steady-state energy and 90% point t90; 18. Calculate oscillation time AT, sensitivity score S, and output report.

[0018] Figure 4 This is a test diagram of the present invention; Figure 5 This is a bar chart showing the start-up time at various frequencies of the present invention; Figure 6 These are 15 original measurement data graphs from this invention.

[0019] In the diagram, the labels are: 1 - Raspberry Pi main control module, 2 - Contact patch microphone, 3 - Resonant speaker, 4 - USB audio interface, 5 - Control and data processing software, 6 - Instrument under test (violin illustration). The arrows indicate the signal flow.

[0020] Markings: 7 - Piezoelectric ceramic sheet, 8 - Non-marking adhesive layer, 9 - Resonant acoustic oscillator with flexible silicone pad. Detailed Implementation

[0021] See Figures 1 to 6 As shown, the present invention includes a musical instrument under test (6), a Raspberry Pi main control module (1), a contact patch pickup (2) for collecting vibrations of the instrument body through solid conduction, a resonant speaker (3) for playing discrete pure tone burst excitation sequences, and a multi-channel USB audio interface (4). The contact patch pickup (2) is attached to the surface area of ​​the musical instrument under test (6) through a patch fixing structure. The resonant speaker (3) is attached to the side of the surface of the musical instrument under test (6) with adhesive. The Raspberry Pi main control module (1) is connected to the multi-channel USB audio interface (4) through a shared data cable. The output end of the contact patch pickup (2) is connected to the recording input channel of the multi-channel USB audio interface (4) through a USB interface data cable. The input end of the resonant speaker (3) is connected to the playback output channel of the multi-channel USB audio interface (4) through a signal line. The Raspberry Pi main control module (1) has a data processing module (5) inside. The instrument under test (6) is a violin, and the resonant sound (3) is attached to the right side of the sound hole on the soundboard of the instrument under test (6) with adhesive. The patch pickup (2) is a piezoelectric contact pickup, which is directly attached to the instrument under test (6) directly below the bridge or the nut area of ​​the guitar with traceless acrylic double-sided tape. The instrument under test (6) can also be a guitar, and the patch pickup (2) is a piezoelectric contact pickup, which is directly attached to the nut area of ​​the instrument under test (6) with traceless acrylic double-sided tape.

[0022] The multi-channel USB audio interface (4) adopts a 16bit / 48kHz USB sound card that supports full-duplex synchronous sampling. The control and data processing module (5) is equipped with an excitation sequence generator. The excitation sequence generated by the excitation sequence generator consists of multiple pure tone bursts of standard frequencies. Each burst lasts for 0.2–0.6 seconds and is preceded and followed by cosine fade-in and fade-out windows. There is a silence interval between bursts and a pre-silence segment at the beginning of the sequence.

[0023] The control and data processing module (5) is provided with a detection parameter table, which includes at least three frequencies: 196.0Hz, 293.66Hz, 440.0Hz, 587.33Hz, and 880.0Hz, with each frequency repeated 2–4 times.

[0024] This invention is divided into 5 core components according to signal flow and control logic, and all mechanical structures are labeled in the accompanying drawings.

[0025] The core control module (1) is a Raspberry Pi 4 or Raspberry Pi 5 single-board computer, pre-installed with a 64-bit Raspberry Pi OS and Python environment. This module is responsible for generating discrete frequency pure tone burst excitation sequences, outputting them to the resonant speaker (3) through the playback output of the USB audio interface (4); simultaneously controlling the recording input of the USB audio interface (4) to collect the vibration response signal of the instrument body from the patch pickup (2); running a dedicated time-domain processing algorithm to extract the start-up time of each frequency point; calculating the sensitivity score and flatness coefficient; and generating a PDF test report.

[0026] The patch pickup (2) is a piezoelectric contact pickup, which is directly attached to the underside of the violin bridge or the nut area of ​​the guitar using non-marking acrylic double-sided tape. It collects the time-domain waveform of the forced vibration of the violin body through solid conduction. The output is connected to the recording input channel of the USB audio interface (4) via a shielded cable.

[0027] The resonant sound (3) is a miniature flat plate resonant sound, which is attached to the side of the sound hole on the instrument panel or near the bridge by a flexible silicone pad. Its input is connected to the playback output channel of the USB audio interface (4) to convert pure tone burst excitation signals into mechanical vibrations and couple them to the instrument body.

[0028] The USB audio interface (4) is a multi-channel (≥2 inputs, ≥2 outputs) external sound card that supports 24bit / 48kHz synchronous full-duplex sampling. The playback output of this module is electrically connected to the resonant speaker (3), the recording input is electrically connected to the patch microphone (2), and it is electrically connected to the Raspberry Pi (1) via the USB interface to achieve sample-level synchronization of excitation signal playback and response recording.

[0029] The control and data processing module (5) is a dedicated detection software (written in Python) running on the Raspberry Pi (1), which includes the following sub-functions: excitation sequence generator, synchronous playback recorder, temporal envelope detector, start-up time extractor, scoring calculator, chart generator and PDF report generator.

[0030] The workflow of this invention includes the following steps: Step 11: Set detection parameters The data processing module (5) presets a timestamp and a preset detection parameter table, and reads the preset detection parameter table. The detection parameter table includes 5 standard frequencies: 196.0Hz G3, 293.66Hz D4, 440.0Hz A4, 587.33Hz D5 and 880.0Hz A5 as a group. Each frequency is read repeatedly 3 times. The burst sound reading time is 0.4 seconds, the fade-in and fade-out sound reading time is 5ms, and the interval between each frequency in the group is 0.7 seconds. When repeatedly reading the frequency in this group, the extra interval between groups is 1.0 second. At the same time, when starting to read, there is a 2.0 second pre-mute. Step 12: Generate discrete pure tone burst excitation sequence The data processing module (5) generates 15 burst sounds in sequence according to the parameters in step 11, splices them in order and inserts silence intervals to form a 48kHz / 16bit PCM excitation array with a total duration of about 21.8 seconds. Step 13: Synchronize playback stimulus and recording response The data processing module (5) is responsible for generating discrete frequency pure tone burst excitation sequences. The data processing module (5) calls the multi-channel USB audio interface (4) and starts the playback and recording thread contained in the data processing module (5). The playback and recording thread outputs the PCM excitation array to the resonant speaker (3). At the same time, the playback and recording thread records the vibration response of the instrument body through the patch pickup (2) to achieve sample-level alignment (delay ≤ 1 sampling point), forming a response segment sequence and frequency timestamp, which together form the total recorded signal. Step 14: Extract response segments at each frequency point Based on the preset timestamp, 15 response segments are extracted from the total recorded signal, and each segment corresponds to the excitation response of a burst tone; Step 15: Zero-phase bandpass filtering For each response segment, a fourth-order Butterworth zero-phase filter is used, with a center frequency equal to the excitation frequency of the burst and a bandwidth of ±40Hz, to filter out crosstalk from other frequencies and environmental noise. Step 16: Extract the energy envelope The absolute value of the filtered signal is taken, and then passed through a low-pass filter with a cutoff frequency of 150Hz to smooth it out, thus obtaining the energy envelope curve. Step 17: Calculate the noise threshold and the starting point t0 The median of the envelope in the first 30ms of the segment is taken as the environmental noise benchmark, and the noise threshold is calculated as noise benchmark × 5. The first position where five consecutive sampling points of the envelope from the start of the segment are greater than the noise threshold is recorded as the starting point t0. Step 18: Calculate the steady-state energy and the 90% point t90. The average value of the envelope over the last 100ms of the segment is taken as the steady-state energy reference value. Starting from t0, the search begins at the point where the envelope first reaches 90% of the steady-state energy reference value, denoted as t90. Step 19: Calculate the start-up time, sensitivity score, and output report. The start-up time AT = t90 - t0 (milliseconds). Outliers are removed from three repeated measurements at each frequency (1.5 times IQR), and the average AT for that frequency is obtained by averaging. Calculate the overall sensitivity score S = 1000 / AT̄ (AT̄ is the mean of the average AT values ​​at 5 frequency points), and the response flatness coefficient CV = (σ_AT / AT̄) × 100%. Determine the sensitivity level based on the S value (S>80: extremely sensitive, 50-80: sensitive, 30-50: moderate, <30: insensitive). Determine the uniformity level based on the CV value (<15%: highly consistent, 15-30%: slightly different, >30%: uneven response across frequency ranges). Generate a frequency-onset time line graph and a typical frequency envelope diagram, and output a PDF test report.

[0031] Example 1: Violin sound sensitivity detection device based on Raspberry Pi 4 and resonant acoustics (optimal implementation)

[0032] This embodiment represents the optimal implementation method for violin factory quality inspection, violin-making workshops, and teaching laboratories.

[0033] Instrument Number: TEST-001 Test environment: Temperature 25°C / Humidity 60% Hardware platform: Raspberry Pi 4 + Resonant speaker + Yamaha MG-XU mixer (USB) + Piezoelectric pickup Test signal: 5 frequencies (196 / 294 / 440 / 587 / 880Hz), burst signal, repeated 3 times per frequency. Evaluation metrics: Audio attack time, total sensitivity score, frequency response flatness Test process: 1. Preparation: Place the resonant speaker near the bridge and ensure a quiet environment.

[0034] 2. Generation: The PC sends a trigger signal, and the Raspberry Pi generates a burst sequence.

[0035] 3. Recording: Yamaha mixer USB, 48kHz / 24bit simultaneous recording.

[0036] 4. Processing: Bandpass filter (center frequency ±40Hz) → Envelope extraction → Dynamic threshold → Start-up time.

[0037] 5. Average: Median of three occurrences at the same frequency (IQR outliers removed) → Start-up time at each frequency + score.

[0038] 6. Scoring: Shorter start-up time = higher sensitivity; flatter frequency response = better consistency.

[0039] Core main control module (1): It adopts Raspberry Pi 4B (4GB memory), pre-installed with 64-bit Raspberry Pi OSLite, and installs Python 3.9 and above and dependent libraries: numpy, scipy, matplotlib, sounddevice, and reportlab.

[0040] Surface mount pickup (2): A PZT-5H piezoelectric ceramic plate, measuring 10mm×8mm×0.2mm, is attached directly below the violin bridge and above the bass bar. The output is connected to the input channel 1 of the USB audio interface (4) via a 1.5m shielded cable.

[0041] Resonance speaker (3): A miniature flat-panel resonant speaker (30mm in diameter, 3W in power) is attached to the violin soundboard about 40mm to the right of the sound hole using 2mm thick silicone double-sided adhesive. The input is connected to the left channel output of the USB audio interface (4).

[0042] Multi-channel USB audio interface (4): Uses a 16bit / 48kHz USB sound card that supports full-duplex synchronous sampling.

[0043] Control and data processing software (5): The Python script pi_sensitivity_test.py running on the Raspberry Pi realizes excitation sequence generation, synchronous playback recording, bandpass filtering, envelope extraction, start-up time calculation, scoring and PDF report output.

[0044] Measurement Procedure: Attach the pickup and resonator to their standard positions, run the command `python3 pi_sensitivity_test.py --violin-id VN-001`, and the software will automatically complete steps ⑪ to ⑲ and output a PDF report. The test results for a fully handmade violin were: average AT = 16.5ms, S = 60.6 (sensitivity), CV = 57.2% (uneven range response), consistent with the subjective evaluations of the three luthiers.

[0045] Test results are available Figure 4 , Figure 5 and Figure 6 .

[0046] Example 2: Guitar Sound Sensitivity Detection Based on the Same Device

[0047] This embodiment applies the device to a guitar. Simply move the patch pickup below the nut and attach the resonant pickup to the side of the soundhole. The frequency list can be changed to the guitar's open string fundamental frequency or retain the original five frequencies. The measurement procedure is the same as in Embodiment 1, demonstrating the portability of this invention.

Claims

1. A method for detecting the sensitivity of musical instrument sound production based on discrete pure tone bursts, comprising hardware: a musical instrument under test (6), a Raspberry Pi main control module (1), a contact patch pickup (2) for collecting vibrations of the instrument body through solid conduction, a resonant speaker (3) for playing discrete pure tone burst excitation sequences, and a multi-channel USB audio interface (4); the contact patch pickup (2) is attached to the surface area of ​​the musical instrument under test (6) through a patch fixing structure, the resonant speaker (3) is attached to the side of the surface of the musical instrument under test (6) with adhesive, the Raspberry Pi main control module (1) is connected to the multi-channel USB audio interface (4) through a shared data cable, the output end of the contact patch pickup (2) is connected to the recording input channel of the multi-channel USB audio interface (4) through a USB interface data cable; the input end of the resonant speaker (3) is connected to the playback output channel of the multi-channel USB audio interface (4) through a signal line, and the Raspberry Pi main control module (1) is internally equipped with a data processing module (5); its characteristic is: The instrument sounding sensitivity detection method based on discrete frequency pure tone bursts is run and tested on the Raspberry Pi main control module (1) using the data processing module (5). Specifically, it includes the steps of an excitation sequence generator, a synchronized playback recorder, a time-domain envelope detector, a start-up time extractor, a scoring calculator, a chart generator, and a PDF report generator. The specific steps are as follows: Step 11: Set detection parameters The data processing module (5) presets a timestamp and a preset detection parameter table, and reads the preset detection parameter table. The detection parameter table includes 5 standard frequencies: 196.0Hz G3, 293.66Hz D4, 440.0Hz A4, 587.33Hz D5 and 880.0Hz A5 as a group. Each frequency is read repeatedly 3 times. The burst sound reading time is 0.4 seconds, the fade-in and fade-out sound reading time is 5ms, and the interval between each frequency in the group is 0.7 seconds. When repeatedly reading the frequency in this group, the extra interval between groups is 1.0 second. At the same time, when starting to read, there is a 2.0 second pre-mute. Step 12: Generate discrete pure tone burst excitation sequence The data processing module (5) generates 15 burst sounds in sequence according to the parameters in step 11, splices them in order and inserts silence intervals to form a 48kHz / 16bit PCM excitation array with a total duration of about 21.8 seconds. Step 13: Synchronize playback stimulus and recording response The data processing module (5) is responsible for generating discrete frequency pure tone burst excitation sequences. The data processing module (5) calls the multi-channel USB audio interface (4) and starts the playback and recording thread contained in the data processing module (5). The playback and recording thread outputs the PCM excitation array to the resonant speaker (3). At the same time, the playback and recording thread records the vibration response of the instrument body through the patch pickup (2) to achieve sample-level alignment (delay ≤ 1 sampling point), forming a response segment sequence and frequency timestamp, which together form the total recorded signal. Step 14: Extract response segments at each frequency point Based on the preset timestamp, 15 response segments are extracted from the total recorded signal, and each segment corresponds to the excitation response of a burst tone; Step 15: Zero-phase bandpass filtering For each response segment, a fourth-order Butterworth zero-phase filter is used, with a center frequency equal to the excitation frequency of the burst and a bandwidth of ±40Hz, to filter out crosstalk from other frequencies and environmental noise. Step 16: Extract the energy envelope The absolute value of the filtered signal is taken, and then passed through a low-pass filter with a cutoff frequency of 150Hz to smooth it out, thus obtaining the energy envelope curve. Step 17: Calculate the noise threshold and the starting point t0 The median of the envelope in the first 30ms of the segment is taken as the environmental noise benchmark, and the noise threshold is calculated as noise benchmark × 5. The first position where five consecutive sampling points of the envelope from the start of the segment are greater than the noise threshold is recorded as the starting point t0. Step 18: Calculate the steady-state energy and the 90% point t90. The average value of the envelope over the last 100ms of the segment is taken as the steady-state energy reference value. Starting from t0, the search begins at the point where the envelope first reaches 90% of the steady-state energy reference value, denoted as t90. Step 19: Calculate the start-up time, sensitivity score, and output report. The start-up time AT = t90 - t0 (milliseconds). Outliers are removed from three repeated measurements at each frequency (1.5 times IQR), and the average AT for that frequency is obtained by averaging. Calculate the overall sensitivity score S = 1000 / AT̄ (AT̄ is the mean of the average AT values ​​at 5 frequency points), and the response flatness coefficient CV = (σ_AT / AT̄) × 100%. Determine the sensitivity level based on the S value (S>80: extremely sensitive, 50-80: sensitive, 30-50: moderate, <30: insensitive). Determine the uniformity level based on the CV value (<15%: highly consistent, 15-30%: slightly different, >30%: uneven response across frequency ranges). Generate a frequency-onset time line graph and a typical frequency envelope diagram, and output a PDF test report.

Images