Learned vocalizations in budgerigars (Melopsittacus undulatus): the relationship between contact calls and warble song

The budgerigar (Melopsittacus undulatus) has an extraordinarily complex, learned, vocal repertoire consisting of both the long rambling warble song of males and a number of short calls produced by both sexes. In warble, the most common elements (>30%) bear a strong resemblance to the highly frequency-modulated, learned contact calls that the birds produce as single utterances. However, aside from this apparent similarity, little else is known about the relationship between contact calls and warble call elements. Here, both types of calls were recorded from four male budgerigars. Signal analysis and psychophysical testing procedures showed that the acoustic features of these two vocalizations were acoustically different and perceived as distinctive vocalizations by birds. This suggests that warble call elements are not simple insertions of contact calls but are most likely different acoustic elements, created de novo, and used solely in warble. Results show that, like contact calls, warble call elements contain information about signaler identity. The fact that contact calls and warble call elements are acoustically and perceptually distinct suggests that they probably represent two phonological systems in the budgerigar vocal repertoire, both of which arise by production learning.     

Acoustic features of male baboon loud calls: influences of context, age, and individuality

The acoustic structure of loud calls ("wahoos") recorded from free-ranging male baboons (Papio cynocephalus ursinus) in the Moremi Game Reserve, Botswana, was examined for differences between and within contexts, using calls given in response to predators (alarm wahoos), during male contests (contest wahoos), and when a male had become separated from the group (contact wahoos). Calls were recorded from adolescent, subadult, and adult males. In addition, male alarm calls were compared with those recorded from females. Despite their superficial acoustic similarity, the analysis revealed a number of significant differences between alarm, contest, and contact wahoos. Contest wahoos are given at a much higher rate, exhibit lower frequency characteristics, have a longer "hoo" duration, and a relatively louder "hoo" portion than alarm wahoos. Contact wahoos are acoustically similar to contest wahoos, but are given at a much lower rate. Both alarm and contest wahoos also exhibit significant differences among individuals. Some of the acoustic features that vary in relation to age and sex presumably reflect differences in body size, whereas others are possibly related to male stamina and endurance. The finding that calls serving markedly different functions constitute variants of the same general call type suggests that the vocal production in nonhuman primates is evolutionarily constrained.     

Development of a contact call in black-capped chickadees (Poecile atricapillus) hand-reared in different acoustic environments

The tseet contact call, common to black-capped (Poecile atricapillus) and mountain chickadees (P. gambeli), is the most frequently produced vocalization of each species. Previous work has characterized the tseet call of black-capped and mountain chickadees from different geographic locations in terms of nine acoustic features. In the current study, using similar methods, the tseet call of black-capped chickadees that were hand reared with either conspecifics, heterospecifics (mountain chickadees), or in isolation from adult chickadees are described. Analysis of call features examined which acoustic features were most affected by rearing environment, and revealed that starting frequency and the slope of the descending portion of the tseet call differed between black-capped chickadees reared with either conspecific or heterospecific adults. Birds reared in isolation from adults differed from the other hand-reared groups on almost every acoustic feature. Chickadee tseet calls are more individualized when they are reared with adult conspecifics or heterospecifics compared to chickadees that are reared in isolation from adults. The current results suggest a role of learning in this commonly used contact call.     

Quantifying complex patterns of bioacoustic variation: use of a neural network to compare killer whale (Orcinus orca) dialects

A quantitative measure of acoustic similarity is crucial to any study comparing vocalizations of different species, social groups, or individuals. The goal of this study was to develop a method of extracting frequency contours from recordings of pulsed vocalizations and to test a nonlinear index of acoustic similarity based on the error of an artificial neural network at classifying them. Since the performance of neural networks depends on the amount of consistent variation in the training data, this technique can be used to assess such variation from samples of acoustic signals. The frequency contour extraction and the neural network index were tested on samples of one call type shared by nine social groups of killer whales. For comparison, call similarity was judged by three human subjects in pairwise classification tasks. The results showed a significant correlation between the neural network index and the similarity ratings by the subjects. Both measures of acoustic similarity were significantly correlated with the groups' association patterns, indicating that both methods of quantifying acoustic similarity are biologically meaningful. An index based on neural network analysis therefore represents an objective and repeatable means of measuring acoustic similarity, and allows comparison of results across studies, species and time.     

人工神经网络分类鉴别苦丁茶红外光谱

为了分类鉴别苦丁茶,采用竞争神经网络(CNN)和反向传播人工神经网络(BP网络)两种模式的人工神经网络(ANN)分别分析了各种苦丁茶的红外谱图。作者采用25个样本作训练集,11个样本作检验集,用两种网络进行了训练。结果表明,CNN网络和BP网络均能够有效地实现苦丁茶产地的鉴别,但CNN网络能够进一步地区分苦丁茶的级别。实验表明,CNN速度快,预测结果准确,可望用竞争神经网络(CNN)和红外光谱法结合分类鉴别苦丁茶。     

基于深度学习的气溶胶荧光光谱识别应用研究

空气中的高危病原微生物对人类社会存在着极大威胁,而传统的监测方法无法对空气中的微生物实现准确的识别与分类。因此采用激光诱导荧光技术原理,以单光子探测器为核心器件,设计并搭建了一种高效的荧光光谱仪用于空气中高危病原微生物的识别与分类,并且该光谱仪可以实现对微生物浓度的预测,其对于环境安全具有重要意义。对于该光谱仪采集的数据,探索了以一维向量和二维矩阵2种输入形式来实现荧光光谱的识别与分类,并研究对比了主成分分析网络、卷积神经网络和全卷积网络等深度学习网络的识别与分类效果。实验结果表明以矩阵形式输入的卷积神经网络模型在测试集中识别分类准确率达到98.05%。采用矩阵形式输入的全卷积网络模型在测试集中微生物浓度预测准确率达到98.97%。     

Vocal learning in Budgerigars (Melopsittacus undulatus): effects of an acoustic reference on vocal matching

Budgerigars were trained to produce specific vocalizations (calls) using operant conditioning and food reinforcement. The bird's call was compared to a digital representation of the call stored in a computer to determine a match. Once birds were responding at a high level of precision, we measured the effect of several manipulations upon the accuracy and the intensity of call production. Also, by differentially reinforcing other aspects of vocal behavior, budgerigars were trained to produce a call that matched another bird's contact call and to alter the latency of their vocal response. Both the accuracy of vocal matching and the intensity level of vocal production increased significantly when the bird could hear the template immediately before each trial. Moreover, manipulating the delay between the presentation of an acoustic reference and the onset of vocal production did not significantly affect either vocal intensity or matching accuracy. Interestingly, the vocalizations learned and reinforced in these operant experiments were only occasionally used in more natural communicative situations, such as when birds called back and forth to one another in their home cages.     

Effects of deafening on the calls and warble song of adult budgerigars (Melopsittacus undulatus)

Budgerigars are small Australian parrots that learn new vocalizations throughout adulthood. Earlier work has shown that an external acoustic model and auditory feedback are necessary for the development of normal contact calls in this species. Here, the role of auditory feedback in the maintenance of species-typical contact calls and warble song in adult budgerigars is documented. Deafened adult birds (five male, one female) vocalized less frequently and showed both suprasegmental and segmental changes in their contact calls and warble song. Contact calls of all adult-deafened budgerigars showed abnormalities in acoustic structure within days to a few weeks following surgery. Within 6 months of surgery, nearly all contact calls produced by deafened birds were strikingly abnormal, showing highly variable patterns of frequency modulation and duration. The warble song of deafened male budgerigars also differed significantly from that of normal budgerigars on several acoustic measures. These results show that auditory feedback is necessary for the maintenance of a normal, species-typical vocal repertoire in budgerigars.     

Low frequency vocalizations attributed to sei whales (Balaenoptera borealis)

Low frequency (<100 Hz) downsweep vocalizations were repeatedly recorded from ocean gliders east of Cape Cod, MA in May 2005. To identify the species responsible for this call, arrays of acoustic recorders were deployed in this same area during 2006 and 2007. 70 h of collocated visual observations at the center of each array were used to compare the localized occurrence of this call to the occurrence of three baleen whale species: right, humpback, and sei whales. The low frequency call was significantly associated only with the occurrence of sei whales. On average, the call swept from 82 to 34 Hz over 1.4 s and was most often produced as a single call, although pairs and (more rarely) triplets were occasionally detected. Individual calls comprising the pairs were localized to within tens of meters of one another and were more similar to one another than to contemporaneous calls by other whales, suggesting that paired calls may be produced by the same animal. A synthetic kernel was developed to facilitate automatic detection of this call using spectrogram-correlation methods. The optimal kernel missed 14% of calls, and of all the calls that were automatically detected, 15% were false positives.     

Knowledge-Based Artificial Neural Network Models for Microstrip Radial Stubs

Radial stubs are a superior choice over low characteristic impedance rectangular stubs in terms of providing an accurate localized zero-impedance reference point and maintaining a low input impedance value over a wide frequency range. In this paper, knowledge-based artificial neural networks are used to model the microstrip radial stubs. Using space-mapping technology and Huber optimization make the neural network models for radial stubs decrease the number of training data, improve generalization ability, and reduce the complexity of the neural network topology with respect to the classical neuromodeling approach. The neural networks are developed for design and optimization of radial stubs, which are robust both from the angle of time of computation and accuracy.     

The acoustic calls of blue whales off California with gender data

The acoustic calls of blue whales off California are described with visual observations of behavior and with acoustic tracking. Acoustic call data with corresponding position tracks are analyzed for five calling blue whales during one 100-min time period. Three of the five animals produced type A-B calls while two produced another call type which we refer to as type D. One of the animals producing the A-B call type was identified as male. Pauses in call production corresponded to visually observed breathing intervals. There was no apparent coordination between the calling whales. The average call source level was calculated to be 186 dB re: 1 muPa at 1 m over the 10-110-Hz band for the type B calls. On two separate days, female blue whales were observed to be silent during respective monitoring periods of 20 min and 1 h.     

Evidence for an analytic perception of multiharmonic sounds in the bat, Megaderma lyra, and its possible role for echo spectral analysis

For echolocation, the gleaning bat Megaderma lyra relies on short and broadband calls consisting of multiple harmonic components, each of which is downward frequency modulated. The harmonic components in M. lyra's calls have a relatively small frequency excursion and do not overlap spectrally. Broadband calls of other bat species, on the other hand, often consist of only a few harmonics which are modulated over broad and sometimes overlapping frequency ranges. A call consisting of narrow and nonoverlapping harmonic components may provide a less complete representation of target structure than a call which consists of broadly modulated components. However, a multiharmonic call may help the bats to perceive local spectral changes in the echo from shifts in the peak frequencies of single harmonics, and thereby to extract additional information about the target. To assess this hypothesis, the accuracy with which M. lyra can analyze frequency shifts of single partials in multiharmonic complex tones was investigated. A two-alternative, forced-choice behavioral task was used to measure M. lyra's frequency discrimination threshold for the third partial in complex tones whose spectral composition resembled that of the bat's sonar calls. The discrimination threshold for the third partial in a 21.5-kHz harmonic tone amounted to about 2% and was similar to the bat's pure-tone discrimination threshold at 64.5 kHz. Discrimination performance was essentially unaffected by random frequency changes of the other partials and by reducing stimulus duration from 50.5 to 1.5 ms. Both findings are in accordance with predictions made on the basis of the shape of M. Ivra's cochlear filters. The comparison between the observed frequency discrimination performance and a computational estimate of the expected frequency shift in the third harmonic of an echo reflected by a simple, two-front target showed that M. lyra's frequency resolution is sufficient for analyzing the target-specific information conveyed by shifts in the peak frequency of single echo components.     

The acoustic features of vervet monkey alarm calls

Vervet monkeys routinely produce semantic alarm calls upon detection of various predators encountered in their natural environment. Two of these calls, snake and eagle alarms, were analyzed using digital signal processing techniques in order to identify potentially distinctive acoustic cues. Distinctive cues were sought in the periodicity of the source waveform associated with each call type, the probable vocal tract filtering functions, and in temporal patterning. Results were equivocal with respect to source periodicity, but a variety of distinguishing features were found in both supralaryngeal filtering and timing. These data provide a basis for psychoacoustic perceptual testing with vervets as subjects.     

基于可见/近红外光谱和深度学习的早期鸭胚雌雄信息无损检测

胚蛋雌雄识别一直是家禽业发展的瓶颈问题,在禽肉生产过程中倾向于养殖雄性个体,而禽蛋生产产业倾向于养殖雌性家禽。若能在孵化过程中较早鉴别出种蛋的雌雄,不仅能够降低家禽孵化产业的成本,还能够提高禽蛋和禽肉生产行业的经济效益。该文以种鸭蛋为研究对象,为了在种鸭蛋孵化早期实现对种蛋的雌雄识别,构建了可见/近红外透射光谱信息采集系统,在200～1 100 nm的波长范围内采集了345枚孵化了0~8 d的种鸭蛋光谱数据。搭建了适用于种鸭蛋光谱信息的6层卷积神经网络（convolutional neural network, CNN）,其中包括输入层、3个卷积层、全连接层与输出分类层。卷积层可以提取光谱中的有效信息,全连接层通过对卷积层提取的局部特征进行整合供输出层分类决策。另外在卷积神经网络中引入局部响应归一化和dropout操作能够加快网络的收敛速度。利用该卷积神经网络构建鸭胚雌雄信息识别网络,通过对比与分析不同孵化天数的识别效果,发现孵化7d的识别效果最佳。随后将孵化7 d的种鸭蛋原始光谱数据进行噪声去除,选取500~900 nm波段用于后续的特征波长选取和建模。分别运用了竞争性自适应重加权算法(CARS)、连续投影算法( SPA)与遗传算法(GA)选择能够区分鸭胚性别的波长点,将选取的特征波长转换为二维的光谱信息矩阵,二维光谱信息矩阵保留了一维光谱的有效信息,同时极大地方便了与卷积神经网络的结合。利用二维光谱信息矩阵和卷积神经网络相结合,实现孵化早期阶段鸭胚的雌雄识别。经检验,基于 SPA算法和CNN网络建立的模型效果较佳,其中训练集、开发集及测试集的准确率分别为93.36%,93.12%和93.83%;基于GA算法和CNN网络建立的模型效果次之,训练集、开发集及测试集的准确率分别为90.87%,93.12%和86.42%;基于CARS算法和CNN网络建立的模型的训练集、开发集及测试集的准确率分别为84.65%,83.75%和77.78%。研究结果表明基于可见/近红外光谱技术和卷积神经网络可以实现孵化早期鸭胚胎雌雄的无损鉴别,为后续相关自动化检测装置的研发提供了技术支撑。     

基于DEMON谱和LSTM网络的水下运动目标噪声基频检测

传统的船舶辐射噪声基频检测方法不仅依赖大量的先验知识,而且对背景噪声非常敏感.为了提高目标识别的稳定性和精确性,该文提出了一种基于深度神经网络的基频检测算法.首先从多通道水听器信号中提取DEMON谱,然后直接将二维谱特征矩阵输入由卷积神经网络和长短时记忆网络构成的级联网络,最后通过稠密层输出实现对基频的估计.仿真和外场...     

水下目标多模态深度学习分类识别研究

水下目标的分类识别对于水声探测具有重要意义。提出一种水下目标多模态深度学习分类识别方法。针对水声信号的一维时域模态和二维频域模态特征建立一种多模态特征融合的深度学习结构,结合长短时记忆网络和卷积神经网络的优点,对一维时域信号和二维频谱信号分别进行并行处理,对输出进行典型相关分析,形成特征融合表示,并利用相邻帧的相关性进行参数优化。利用实测水声信号对算法进行了验证。结果表明：提出的算法对于水下目标识别的精度有显著的提高。     

基于水样类型识别的光谱COD测量方法

基于紫外吸收光谱的COD测量方法,尽管具有快速、实时、免试剂、无污染等优势。但该方法对于组分多变的水样适应性不强,构建的单一计算模型不能适用于所有待测水样类型,导致其在复杂环境下测量准确度较低,从而限制了其应用领域。本研究提出一种基于水样类型识别的测量方法。其过程包括:动态识别水样类型→自动选择相应的"吸光度(Auv)-COD"算法模型→计算COD。该方法有效提高了紫外光谱法COD测量的准确度和适用性。该研究在传统的光谱识别技术的基础上,针对COD实际测量的特点加以改进。选取水样吸光度曲线的形貌特征作为水样类型的表征参数,利用LM-BP神经网络作为识别算法。并引入了"历史数据队列"、"历史识别因子"的概念,在此基础上形成了级联的神经网络结构。该算法实现了COD测量应用中的高准确度的光谱识别,进而提高了复杂环境下COD测量的精度。大量实验测试和结果表明,与传统的光谱识别技术相比,该方法在COD测量应用中具有更好的鲁棒性和准确性。水样类型识别准确率达98%以上。同时算法结构简单,计算量小,适用于资源受限的小型化COD测量仪。当仪器在复杂多变的水环境中进行测量时,采用该算法测量得到的COD精度有显著的提高。该方法的提出为光谱COD测量法在水体组分多变场合的应用及提高其测量精度提供了技术保证,可望解决传统紫外光谱COD测量法难以适应变化和复杂水环境应用的问题。     

基于三维荧光谱特征分析的油种鉴别技术的研究

介绍了运用神经网络进行模式识别的基本原理,将主成分分析法和BP神经网络相结合,提出矿物油三维荧光谱鉴别方案,并进行了系统设计,建立了基本的模型框架.选取矿物油三维荧光谱的特征参量,组成原始特征向量,采用主成分分析法进行预处理,而后选取主成分运用BP神经网络实现油种鉴别.该方法减少了输入变量的维数,消除了各输入变量的相关性,同时简化了网络结构,提高了程序运行的速度.通过实例进行了分析,结果证明该方法有效地实现了矿物油三维荧光谱的油种鉴别,同时该系统也可用于其它物质的光谱识别技术领域.     

Vocal production mechanisms in the budgerigar (Melopsittacus undulatus): the presence and implications of amplitude modulation.

In this paper acoustic evidence is presented for the presence of amplitude modulation in budgerigar (Melopsittacus undulatus) contact calls and learned English vocalizations. Previously, acoustic analyses of budgerigar vocalizations have consisted solely of visual inspection of spectrograms or power spectra (derived from Fourier transformation). Such analyses have led researchers to conclude that budgerigar vocalizations are primarily frequency-modulated, harmonic vocalizations. Although budgerigar calls have been shown to contain regions that are modulated in amplitude, the implications of this fact have been largely ignored. Amplitude modulation, the nonlinear interaction between two separate signals that results in the creation of new, heterodyne (sum and difference) frequencies, can produce a very complex Fourier spectrum that may resemble that produced by a harmonic vocalization. In this paper, the acoustic principles necessary for identifying amplitude modulation present in signals are outlined, and followed by data demonstrating that amplitude modulation is a prominent feature not only of natural budgerigar contact calls, but also of their learned English vocalizations. It is illustrated how analyzing a vocalization that contains amplitude modulation as if it were harmonic can result in misinterpretations of the acoustic and physical properties of the sound and sound source. The implications of amplitude modulation for studies of the ontogenetic, physical, and neural basis of budgerigar vocalizations are discussed, and a potential model for how the budgerigar syrinx may function to produce amplitude modulation is proposed.     

时频谱图和数据增强的水声信号深度学习目标识别方法

水声目标识别一直是水声领域研究的重点问题之一,深度学习方法可以有效地解决目标识别问题,然而,水声样本的稀少限制了该方法的应用.该文提出一种基于数据增强的水声信号深度学习目标识别方法,该方法以Mel功率谱作为网络的输入特征,通过对原始信号在时域和时频域的拉伸和掩蔽等变换,实现数据扩展和增加泛化性能的目的,最后,利用改进的...