Pitch Detection Research Articles

A Robust and Low Computational Cost Pitch Estimation Method.

Pitch estimation is widely used in speech and audio signal processing. However, the current methods of modeling harmonic structure used for pitch estimation cannot always match the harmonic distribution of actual signals. Due to the structure of vocal tract, the acoustic nature of musical equipment, and the spectrum leakage issue, speech and audio signals’ harmonic frequencies often slightly deviate from the integer multiple of the pitch. This paper starts with the summation of residual harmonics (SRH) method and makes two main modifications. First, the spectral peak position constraint of strict integer multiple is modified to allow slight deviation, which benefits capturing harmonics. Second, a main pitch segment extension scheme with low computational cost feature is proposed to utilize the smooth prior of pitch more efficiently. Besides, the pitch segment extension scheme is also integrated into the SRH method’s voiced/unvoiced decision to reduce short-term errors. Accuracy comparison experiments with ten pitch estimation methods show that the proposed method has better overall accuracy and robustness. Time cost experiments show that the time cost of the proposed method reduces to around 1/8 of the state-of-the-art fast NLS method on the experimental computer.

Cortical tracking of voice pitch in the presence of multiple speakers depends on selective attention.

Voice pitch carries linguistic and non-linguistic information. Previous studies have described cortical tracking of voice pitch in clean speech, with responses reflecting both pitch strength and pitch value. However, pitch is also a powerful cue for auditory stream segregation, especially when competing streams have pitch differing in fundamental frequency, as is the case when multiple speakers talk simultaneously. We therefore investigated how cortical speech pitch tracking is affected in the presence of a second, task-irrelevant speaker. We analyzed human magnetoencephalography (MEG) responses to continuous narrative speech, presented either as a single talker in a quiet background or as a two-talker mixture of a male and a female speaker. In clean speech, voice pitch was associated with a right-dominant response, peaking at a latency of around 100 ms, consistent with previous electroencephalography and electrocorticography results. The response tracked both the presence of pitch and the relative value of the speaker’s fundamental frequency. In the two-talker mixture, the pitch of the attended speaker was tracked bilaterally, regardless of whether or not there was simultaneously present pitch in the speech of the irrelevant speaker. Pitch tracking for the irrelevant speaker was reduced: only the right hemisphere still significantly tracked pitch of the unattended speaker, and only during intervals in which no pitch was present in the attended talker’s speech. Taken together, these results suggest that pitch-based segregation of multiple speakers, at least as measured by macroscopic cortical tracking, is not entirely automatic but strongly dependent on selective attention.

Statistical analysis for the pitch of mask-wearing Arabic speech

The study is a comparison between the statistical properties of pitch (F0) for mask-wearing speech and unmasked. The speakers are Arab, of different ages and genders. A robust algorithm for pitch tracking (RAPT) is used for estimating F0. The subjective tests denote that masked speech is attenuated, and noisy-background speech has fewer F0 candidates. Using objective tests, 60% of female and male F0s do not change when wearing masks. The remaining 40% of speech F0s change (the percentage gross error), by an approximately 20% increase and 20% decrease. The percentage classification error is about 10%. The F0 changes in females younger than 12 years old are fewer compared with similarly-aged males. The F0 changes of females older than 12 years old were approximately equal compared with similarly-aged males. An average of F0 (M) is used for each speech to divide its F0 band (50-500) Hz into two bands, lower-band (LB) (50-M) Hz and upper-band (UB) (M-500) Hz. The attributes of the two bands have been statistically analyzed. The F0 classification error (CE) for females is higher than for males, but the gross error (GE) for males is higher than for females. The F0 change values are directly proportional to the probability of F0 change.

Smart-Median: A New Real-Time Algorithm for Smoothing Singing Pitch Contours

Pitch detection is usually one of the fundamental steps in audio signal processing. However, it is common for pitch detectors to estimate a portion of the fundamental frequencies incorrectly, especially in real-time environments and when applied to singing. Therefore, the estimated pitch contour usually has errors. To remove these errors, a contour smoother algorithm should be employed. However, because none of the current contour-smoother algorithms has been explicitly designed to be applied to contours generated from singing, they are often unsuitable for this purpose. Therefore, this article aims to introduce a new smoother algorithm that rectifies this. The proposed smoother algorithm is compared with 15 other smoother algorithms over approximately 2700 pitch contours. Four metrics were used for the comparison. According to all the metrics, the proposed algorithm could smooth the contours more accurately than other algorithms. A distinct conclusion is that smoother algorithms should be designed according to the contour type and the result’s final applications.

A Novel Pitch Detection Algorithm Based on Instantaneous Frequency for Clean and Noisy Speech

A Novel Pitch Detection Algorithm Based on Instantaneous Frequency for Clean and Noisy Speech

Modern Technological Development In Music Information Retrieval

In present era of technology, everything is in fingertips likewise ,any unstructured data such as video ,image is captured and analyzed, but audio data such as music, sounds are very complicatedly classified and analyzed , Music information retrieval is abbreviated as ‘ MIR ‘therefore it is a interdisciplinary science of retrieval of data from music by data sourcing ,transcribing,feature extraction, machine learning algorithms, genre categorizing, pitch tracking, chord recognition and MIR applications A simple MIR system retrieves data according to a user-introduced text query. B. "David Bowie Heroes". In such cases, the system is basically the same as a text- based search engine (Google, Yahoo, etc.) because the text is compared to the text data associated with the album or track. However, given the characteristics of the retrieved content, you need a system that can accept "musical" queries such as sheet music, sung melodies (queries by ham), and recorded audio segments (queries by example). This proposal concerns the latter case. The purpose of the sample query is to retrieve music from a large collection of digital music content based on its similarity to the sample audio document. The ability to query using examples is an important requirement of the MIR system. There are many challenges, such as computational and complexity issues, proper testbed design, and choosing the right representation of audio for queries and music collections...

A Music Cognition–Guided Framework for Multi-pitch Estimation

As one of the most important subtasks of automatic music transcription (AMT), multi-pitch estimation (MPE) has been studied extensively for predicting the fundamental frequencies in the frames of audio recordings during the past decade. However, how to use music perception and cognition for MPE has not yet been thoroughly investigated. Motivated by this, this demonstrates how to effectively detect the fundamental frequency and the harmonic structure of polyphonic music using a cognitive framework. Inspired by cognitive neuroscience, an integration of the constant Q transform and a state-of-the-art matrix factorization method called shift-invariant probabilistic latent component analysis (SI-PLCA) are proposed to resolve the polyphonic short-time magnitude log-spectra for multiple pitch estimation and source-specific feature extraction. The cognitions of rhythm, harmonic periodicity and instrument timbre are used to guide the analysis of characterizing contiguous notes and the relationship between fundamental frequency and harmonic frequencies for detecting the pitches from the outcomes of SI-PLCA. In the experiment, we compare the performance of proposed MPE system to a number of existing state-of-the-art approaches (seven weak learning methods and four deep learning methods) on three widely used datasets (i.e. MAPS, BACH10 and TRIOS) in terms of F-measure ({F}_{1}) values. The experimental results show that the proposed MPE method provides the best overall performance against other existing methods.

Analysis of audio signal using various transforms for enhanced audio processing

Audio Signals are the portrayal of sounds. It changes with respect to frequencies rather than time, and it shows more information in the frequency domain. So it is much appropriate to evaluate in the frequency domain rather than the time domain. By using different transforms like DFT, DST, DCT, MDCT, Integer MDCT, the time domain audio signal can be converted into a frequency domain signal. The signal is reconstructed to analyze the features like mean square error, Signal to noise ratio, Peak signal to noise ratio between the original and reconstructed signal. Other features like energy, entropy, zero crossing rates (ZCR) were also considered for the evaluation. In this paper, different audio file formats were taken for interpretation. It includes wave file, mp3 file, m4a file, aac file, where wave file is in uncompressed format and mp3, m4a, aac are in compressed format. These compressed files come under lossy compression. The above-mentioned features are used for applications like music information retrieval (MIR). MIR includes onset detection, pitch detection and to measure the noise and loudness of the music.

Generalized focal plane array modulation transfer function estimation approach for nonstationary laser speckle random processes

The infrared (IR) imaging community has a need for direct IR detector evaluation due to the continued demand for small pixel pitch detectors, the emergence of strained-layer-superlattice devices, and the associated lateral carrier diffusion issues. Conventional laser speckle-based modulation transfer function estimation is dependent on Fresnel propagation and limited to paraxial geometries, preventing the utilization of this approach on small pixel pitch IR devices. This paper presents a generalized approach for determining a focal plane array (FPA) system input power spectral density, utilizing numerical evaluation of Rayleigh–Sommerfeld speckle irradiance autocorrelation functions, speckle irradiance spectral analysis using the Wigner distribution function, and experimental error quantification incurred from making wide-sense-stationary assumptions regarding the associated laser speckle random process. The effort’s final result is an experimental demonstration of an FPA resolution estimation technique valid in nonparaxial test scenarios.

0349 Robust inter-beat interval estimation algorithm using clustering method

Abstract Introduction The heart rate is a basic indicator of the state of the human body and therefore it is important to measure the heart rate continuously, reliably, and accurately. However, it is difficult to measure the heart rate in various environments due to the limitations of Contact sensors, such as ECG. Therefore, heart rate measurement using non-contact sensors such as PVDF, Radar, and cameras has been widely studied. Among these non-contact sensors to measure the heart rate, the PVDF sensor, installed under a bed or chair, can measure the heart rate from the J-peak of Ballistocardiograph (BCG). However, it is challenging to accurately measure the heart rate due to noise-induced by the external environment. To overcome these challenges in this study, we have proposed a new algorithm for accurate heart rate measurement using PVDF sensor. The proposed algorithm includes estimators and clustering techniques to estimate the beat-to-beat of BCG signals. Methods In order to continuously estimate inter-beat interval, basic, auto-correlation, average magnitude difference function, maximum amplitude pairs, and a Bayesian approach are used to obtain basic information to calculate inter-beat interval. Auto-correlation is an estimator that estimates heart rate intervals by calculating for all discrete lags. The average magnitude difference function (AMDF), which is often used for pitch tracking, also calculates for discrete lags like auto-correlation. If the waveforms of BCG are similar, the calculated value with AMDF is small. Therefore, the reciprocal of the ADMF output is used to take the larger value for the most likely interval. This is a complementary relationship from auto-correlation because the noise characteristics are different. Maximum amplitude pairs is an estimator used for indirect peak detectors with the amplitude information of the signal. The estimated IBI was determined by combining the estimation results of three estimators with a Bayesian approach. After the estimated IBI was obtained, continuously estimated inter-beat intervals were clustered into locally similar values. The inter-beat interval values were determined by voting for the most estimated interval value in the clustered sets.All inter-beat intervals estimated by the Bayesian method have their own probability density function and the confidence value was calculated using the difference between the largest peak and the second largest peak of the probability density function. By excluding inter-beat intervals with low confidence values, the accuracy of the estimation was improved. After, the final continuous inter-beat intervals were calculated by performing the ectopic beats removal algorithm and interpolation. Results To evaluate the performance, BCG data from PVDF sensor was simultaneously measured with PSG. The PSG datasets consisted of three male patients with moderate sleep apnea, a patient with snoring problems, and a normal person. Because of the difference in timing between the peak signal of ECG and that of BCG, less than 1 HR count was allowed when the inter-beat intervals were converted to HR. When the 21-hour inter-beat intervals were compared, the difference in HR was less than 1 in 91.4%. The rest 8.6% had a difference of 1 HR or more, and it was confirmed that 6.9% of them were caused by movement, and the remaining 1.7% were caused by a temporary failure of the sensor. Conclusion The proposed algorithm showed better coverage by 18.71% compared to the coverage of 72.69% from the previous study. It was confirmed that the proposed algorithm provides a more accurate heart rate than previous studies, and the measured inter-beat intervals will be used to estimate, not only the heart rate but also the sleep stage in the future. Support (If Any)

Vehicle Positioning Algorithm Based on NHC/Virtual-MINS/OD

Aiming at the problem of large navigation errors for unmanned vehicle based on MEMS inertial navigation systems (MEMS-INS, MINS) due to low output accuracy of MEMS inertial sensors, a redundant MINS is designed and a high precision positioning algorithm is proposed. In redundant MINS, the output of the redundant gyroscopes is estimated by the improved virtual gyro algorithm based on improved current statistical model and interacting multiple model Kalman filter (ICS-IMMKF). The weighted fusion algorithm is used to fuse the redundant accelerometer information, thus the virtual micro inertial navigation system (Virtual-MINS) is obtained. In the process of vehicle positioning based on virtual MINS, the centripetal acceleration is solved based on the kinematic constraints. Combined with the navigation information of Virtual-MINS and odometer (OD), the difference of centripetal acceleration and velocity are used as observation, and a Virtual-MINS/OD positioning algorithm based on kinematic nonholonomic constraints (NHC/Virtual-MINS/OD) is proposed. The simulation results show that the ICS-IMMKF algorithm can achieve high precision fusion of redundant gyro. In the test, azimuth maneuver is not constrained by the ground, therefore the ICS-IMMKF algorithm can obtain the azimuth rate with high precision. The experimental results show that the root mean square (RMS) of the East and North positioning errors, as well as the RMS of pitch, roll and azimuth estimation error based on the NHC/Virtual-MINS/ OD positioning algorithm are reduced to 34.15%, 22.65%, 55.11%, 64.93% and 41.37% of the NHC/#1-MINS/OD positioning algorithm, respectively. The method will provide a solution for high-precision positioning when GNSS fails.

Age-related changes to vestibular heave and pitch perception and associations with postural control

Falls are a common cause of injury in older adults (OAs), and age-related declines across the sensory systems are associated with increased falls risk. The vestibular system is particularly important for maintaining balance and supporting safe mobility, and aging has been associated with declines in vestibular end-organ functioning. However, few studies have examined potential age-related differences in vestibular perceptual sensitivities or their association with postural stability. Here we used an adaptive-staircase procedure to measure detection and discrimination thresholds in 19 healthy OAs and 18 healthy younger adults (YAs), by presenting participants with passive heave (linear up-and-down translations) and pitch (forward–backward tilt rotations) movements on a motion-platform in the dark. We also examined participants’ postural stability under various standing-balance conditions. Associations among these postural measures and vestibular perceptual thresholds were further examined. Ultimately, OAs showed larger heave and pitch detection thresholds compared to YAs, and larger perceptual thresholds were associated with greater postural sway, but only in OAs. Overall, these results suggest that vestibular perceptual sensitivity declines with older age and that such declines are associated with poorer postural stability. Future studies could consider the potential applicability of these results in the development of screening tools for falls prevention in OAs.

Aging Affects Subcortical Pitch Information Encoding Differently in Humans With Different Language Backgrounds.

Aging and language background have been shown to affect pitch information encoding at the subcortical level. To study the individual and compounded effects on subcortical pitch information encoding, Frequency Following Responses were recorded from subjects across various ages and language backgrounds. Differences were found in pitch information encoding strength and accuracy among the groups, indicating that language experience and aging affect accuracy and magnitude of pitch information encoding ability at the subcortical level. Moreover, stronger effects of aging were seen in the magnitude of phase-locking in the native language speaker groups, while language background appears to have more impact on the accuracy of pitch tracking in older adult groups.

Periodicity Pitch Perception Part III: Sensibility and Pachinko Volatility.

Neuromorphic computer models are used to explain sensory perceptions. Auditory models generate cochleagrams, which resemble the spike distributions in the auditory nerve. Neuron ensembles along the auditory pathway transform sensory inputs step by step and at the end pitch is represented in auditory categorical spaces. In two previous articles in the series on periodicity pitch perception an extended auditory model had been successfully used for explaining periodicity pitch proved for various musical instrument generated tones and sung vowels. In this third part in the series the focus is on octopus cells as they are central sensitivity elements in auditory cognition processes. A powerful numerical model had been devised, in which auditory nerve fibers (ANFs) spike events are the inputs, triggering the impulse responses of the octopus cells. Efficient algorithms are developed and demonstrated to explain the behavior of octopus cells with a focus on a simple event-based hardware implementation of a layer of octopus neurons. The main finding is, that an octopus' cell model in a local receptive field fine-tunes to a specific trajectory by a spike-timing-dependent plasticity (STDP) learning rule with synaptic pre-activation and the dendritic back-propagating signal as post condition. Successful learning explains away the teacher and there is thus no need for a temporally precise control of plasticity that distinguishes between learning and retrieval phases. Pitch learning is cascaded: At first octopus cells respond individually by self-adjustment to specific trajectories in their local receptive fields, then unions of octopus cells are collectively learned for pitch discrimination. Pitch estimation by inter-spike intervals is shown exemplary using two input scenarios: a simple sinus tone and a sung vowel. The model evaluation indicates an improvement in pitch estimation on a fixed time-scale.

Study on radar scattering characteristics of double S-shaped nozzles with different aspect ratio

To study the radar scattering characteristics of dual S-bend nozzles with different exit aspect ratios, the iterative physical optic method (IPO) method and equivalent edge currents method (EEC) method were combined to build a code that can be used to solve the electromagnetic scattering field of the nozzle. The reliability of the method has been verified by comparing the reference data. Then, the forward and back-forward IPO and ray tracing methods are used to reduce the calculation time. CFD method is used to calculate and analyse the influence of aspect ratio double S-shaped nozzle on aerodynamic performance. The thrust coefficient and the total pressure recovery coefficient of the double S-shaped nozzle will gradually decrease with the increase of the aspect ratio. the influence of the aspect ratio on the radar scattering characteristics of the dual S-shaped nozzle under the C-band was simulated by the IPO and the results were discussed.it is obtained that the RCS average value in the range of −40°~40° gradually decreases with the increase of the aspect ratio. Under the vertical polarization mode of the pitch detection plane, the average RCS of the AR6 model is 52.8% lower than that of the AR3 model, while it is reduced by 11.0% in the horizontal polarization mode, which has better stealth performance.

Dynamic Performance Enhancement of a Renewable Energy System for Grid Connection and Stand-Alone Operation with Battery Storage

This paper introduces a new formulated control scheme for enhancing the dynamic performance of a wind driven surface permanent magnet synchronous generator. The designed control scheme is based on predictive control theory, in which the shortcomings of previous predictive controllers are avoided. To visualize the effectiveness of the proposed control scheme, the performance of the generator was dynamically evaluated under two different operating regimes: grid connection and standalone operation in which a battery storage system was used to enhance the power delivery to the isolated loads. In addition, a detailed performance comparison between the proposed controller and traditional predictive controllers was carried out. The traditional control topologies used for comparison were the model predictive direct power control, model predictive direct torque control, and model predictive current control. A detailed description of each control scheme is introduced illustrating how it is configured to manage the generator operation. Furthermore, to achieve the optimal exploitation of the wind energy and limit the power in case of exceeding the nominal wind speed, maximum power point tracking and blade pitch angle controls were adopted. A detailed performance comparison effectively outlined the features of each controller, confirming the superiority of the proposed control scheme over other predictive controllers. This fact is illustrated through its simple structure, low ripples, low computation burdens and low current harmonics obtained with the proposed control scheme.

Music to My Ears: Neural modularity and flexibility differ in response to real-world music stimuli.

Music listening involves many simultaneous neural operations, including auditory processing, working memory, temporal sequencing, pitch tracking, anticipation, reward, and emotion, and thus, a full investigation of music cognition would benefit from whole-brain analyses. Here, we quantify whole-brain activity while participants listen to a variety of music and speech auditory pieces using two network measures that are grounded in complex systems theory: modularity, which measures the degree to which brain regions are interacting in communities, and flexibility, which measures the rate that brain regions switch the communities to which they belong. In a music and brain connectivity study that is part of a larger clinical investigation into music listening and stroke recovery at Houston Methodist Hospital’s Center for Performing Arts Medicine, functional magnetic resonance imaging (fMRI) was performed on healthy participants while they listened to self-selected music to which they felt a positive emotional attachment, as well as culturally familiar music (J.S. Bach), culturally unfamiliar music (Gagaku court music of medieval Japan), and several excerpts of speech. There was a marked contrast among the whole-brain networks during the different types of auditory pieces, in particular for the unfamiliar music. During the self-selected and Bach tracks, participants’ whole-brain networks exhibited modular organization that was significantly coordinated with the network flexibility. Meanwhile, when the Gagaku music was played, this relationship between brain network modularity and flexibility largely disappeared. In addition, while the auditory cortex’s flexibility during the self-selected piece was equivalent to that during Bach, it was more flexible during Gagaku. The results suggest that the modularity and flexibility measures of whole-brain activity have the potential to lead to new insights into the complex neural function that occurs during music perception of real-world songs.

A Decoupled Orientation Estimation Approach for Robust Roll and Pitch Measurements in Magnetically Disturbed Environment

Orientation estimation using inertial and magnetic sensors has permeated into various applications. However, robust roll and pitch estimates are still challenging for the orientation estimators when the sensors are exposed to magnetic disturbances. In this article, we proposed a decoupled orientation estimation approach (DOEA) to separate and achieve accurate roll and pitch estimates from the magnetic measurements. In the proposed DOEA, we formulate a reference vector, which is perpendicular to the gravitational vector and keeps constant size and direction, and derive its covariances to replace the magnetic field vector as the observation references. The sensor measurements of the two orthogonal homogeneous fields are further fused in the extended Kalman filter frame to correct the predicted orientation of the gyroscope integration. To validate the performance of the proposed DOEA, we compare it with three state-of-the-art approaches in four different experiments. In the magnetically disturbed environment without reliability validation of the sensors, experiment results show that the DOEA could provide robust and accurate roll and pitch measurements with 0.001- and 0.007-rad root-mean-square error (RMSE). The influence of the magnetic disturbances on the DOEA is controlled only on the yaw direction. In the wide-range static and dynamic mixed motion tracking of the robot, experiment results show that the proposed DOEA achieves 0.035-, 0.024-, and 0.018-rad RMSE of roll, pitch, and yaw estimates without singularities. Moreover, the proposed DOEA converges faster than the popular optimization filters and achieves computational efficiency 56% faster than the two-step orientation estimator, which is beneficial for realization on the embedded system.

Cross-Domain Multitask Model for Head Detection and Facial Attribute Estimation

Extracting specific attributes of a face within an image, such as emotion, age, or head pose has numerous applications. As one of the most widely used vision-based attribute extraction models, HPE (Head Pose Estimation) models have been extensively explored. In spite of the success of these models, the pre-processing step of cropping the region of interest from the image, before it is fed into the network, is still a challenge. Moreover, a significant portion of the existing models are problem-specific models developed specifically for HPE. In response to the wide application of HPE models and the limitations of existing techniques, we developed a multi-purpose, multi-task model to parallelize face detection and pose estimation (i.e., along both axes of yaw and pitch). This model is based on the Mask-RCNN object detection model, which computes a collection of mid-level shared features in conjunction with some independent neural networks, for the detection of faces and the estimation of poses. We evaluated the proposed model using two publicly available datasets, <i>Prima</i> and <i>BIWI</i>, and obtained MAEs (Mean Absolute Errors) of 8.0 &#x00B1; 8.6, and 8.2 &#x00B1; 8.1 for yaw and pitch detection on <i>Prima</i>, and 6.2 &#x00B1; 4.7, and 6.6 &#x00B1; 4.9 on <i>BIWI</i> dataset. The generalization capability of the model and its cross-domain effectiveness was assessed on the publicly available dataset of <i>UTKFace</i> for face detection and age estimation, resulting a MAE of 5.3 &#x00B1; 3.2. A comparison of the proposed model&#x2019;s performance on the domains it was tested on reveals that it compares favorably with the state-of-the-art models, as demonstrated by their published results. We provide the source code of our model for public use at: <uri>https://github.com/kahroba2000/MTL_MRCNN</uri>.

Design of Music Melody Contour Extraction Algorithm Based on Fuzzy Clustering Algorithm

Now is the era of big data, and a large amount of data will be generated in any industry. How to mine useful information from this huge amount of data is a huge challenge we face. Efficient clustering algorithm is an important means to mine useful information, and it is also a hot topic of current research. Among many drama algorithms, fuzzy and other methods that use fuzzy mathematics to describe and classify things, In recent years, it has attracted the attention of scholars at home and abroad. With the development of Internet technology, clustering plays an important role in various fields. It can actively group data points, so that data points belonging to the same cluster have high similarity, and data points belonging to different clusters have great differences. The algorithm extracts the song pitch sequence from the humming segment, and converts it into the melody contour sequence after regularization, merging and segmentation, and then converts this sequence into the melody contour feature using the standard tone difference table generated by the standard tone. The results show that the system has good robustness to environmental noise; In the search space containing 405 songs, the success rate of retrieving the top five is more than 90%.The improved algorithm is simulated on rechset music data set, and the results show that the pitch estimation accuracy of 62.04% and the total accuracy of 55.08% are achieved.