Bird Song: How and Why Birds Sing

Bird song is a learned vocal behavior produced primarily by male songbirds during the breeding season to defend territory and attract mates. This article explains the biological functions of bird song, how birds acquire their songs through learning, the neural mechanisms that support song production and memory, and practical methods for identifying and studying bird songs using audio recordings and spectrograms. The content is written for students, researchers, life-science professionals, and informed general readers who want a rigorous but accessible account of current scientific understanding.

At a Glance

Bird song research spans multiple disciplines including neurobiology, behavioral ecology, evolutionary biology, and computational bioacoustics. The table below summarizes the primary functions, learning mechanisms, and study methods covered in this article.

Why Birds Sing

Song serves two primary functions in most songbird species. The first is territorial defense, where males use song to repel other males from a defended space. The second is mate attraction, where song functions to attract females and stimulate their courtship behavior. These functions are well established in the scientific literature on song evolution and communication Song function and the evolution of female preferences: why birds sing, why brains matter.

Understanding why birds sing provides the essential backdrop for understanding the physiological and neural mechanisms responsible for song learning, perception, and production. The reverse is also true. Understanding the mechanisms underlying song learning provides insight into how song has evolved as a communication signal. This reciprocal relationship between function and mechanism is a central theme in contemporary birdsong research Song function and the evolution of female preferences: why birds sing, why brains matter.

Territorial Defense

Males sing to advertise their presence and defend resources including nesting sites and food. Song functions as an acoustic signal that can be broadcast over long distances, allowing males to establish and maintain territories without direct physical confrontation. The amplitude of song, or how loud a bird sings, plays a decisive role in signal transmission. Birds can adjust how loud they sing in response to changes in their environment A comparative analysis of song amplitude across and within bird species.

Field measurements across 17 European songbird species found that song amplitude increased with increasing background noise, a phenomenon known as the Lombard effect. Song amplitude also increased when singing rival males were present and varied with the time of day. These findings highlight that birds can adjust how loud they sing in response to changes in the biotic and abiotic environments A comparative analysis of song amplitude across and within bird species.

Mate Attraction and Female Choice

Song also functions to attract females and stimulate their courtship. The developmental stress hypothesis proposes that learned features of song, including complexity and local dialect structure, can serve as indicators of male quality useful to females in mate choice. This hypothesis builds on the fact that brain structures underlying song learning largely develop during the first few months after hatching. During this same period, songbirds are likely to be subject to nutritional and other developmental stresses. Individuals that fare well in the face of stress are able to invest more resources to brain development and are expected to be correspondingly better at song learning. Learned features of song thus become reliable indicators of male quality, with reliability maintained by the developmental costs of song Song function and the evolution of female preferences: why birds sing, why brains matter.

Data from both field and laboratory studies provide broad support for the developmental stress hypothesis, illustrating the utility of connecting mechanistic and evolutionary analyses of song learning Song function and the evolution of female preferences: why birds sing, why brains matter.

How Birds Learn to Sing

Song learning in birds shares several parallels with human language learning. These analogies include an early sensitive period for learning, separation of sensory and motor phases of learning, innate knowledge of language or song, and specialized neural systems. However, song learning is usually viewed as a purely auditory process in distinction to human language learning. This view is implied in the typical experimental paradigm for studying song learning, in which the bird is isolated in a sound-proof chamber and exposed only to tape-recorded song Bird song learning as an adaptive strategy.

This paradigm remains the major method of studying song learning despite recent demonstrations of the importance of social variables. It contains the implicit assumption that non-auditory variables, especially social and ecological variables, play only a minor role in this process. Understanding the functions of song learning requires investigating it in the field, where social and ecological variables have full play. Field studies of song sparrows have shown that song learning takes on the clear outlines of an adaptive strategy when viewed in this perspective Bird song learning as an adaptive strategy.

Sensory and Motor Phases

Young songbirds memorize a tutor song and use the memory trace as a template to shape their own song by auditory feedback. This process involves two distinct phases. During the sensory phase, the juvenile listens to and forms auditory memories of adult tutor songs. During the sensorimotor phase, the juvenile uses these memories to shape its own vocalizations through auditory feedback The role of auditory feedback in birdsong.

The higher-level auditory cortex, called the caudomedial nidopallium or NCM, is a potential storage site for tutor song memory. Recording single-neuron activity in the NCM of behaving juvenile zebra finches has identified the neuronal substrate for tutor song memory. After tutor song experience, a small subset of NCM neurons exhibit highly selective auditory responses to the tutor song. Blockade of GABAergic inhibition and sleep decrease their selectivity. These results suggest that experience-dependent recruitment of GABA-mediated inhibition shapes auditory cortical circuits, leading to sparse representation of tutor song memory in auditory cortical neurons Auditory experience-dependent cortical circuit shaping for memory formation in bird song learning.

Social Dimensions of Learning

Song learning is not purely auditory. Social variables play a significant role in how young birds acquire their songs. In a field study of three consecutive cohorts of song sparrows, researchers determined the older bird from whom the young bird learned most of his songs and measured how long both birds survived subsequently. The more songs a tutee learned from his primary tutor, the longer their mutual survival on their respective territories. This study provides the first evidence of a mutual benefit of bird song learning and teaching in nature Bird Song Learning is Mutually Beneficial for Tutee and Tutor.

This finding contrasts two mutually exclusive hypotheses about the tutor and tutee relationship. The cooperative hypothesis proposes that tutor and tutee mutually benefit from their relationship. The competitive hypothesis proposes that tutor and tutee compete over territory, so that one or the other suffers negative fitness consequences. The survival data support the cooperative hypothesis Bird Song Learning is Mutually Beneficial for Tutee and Tutor.

Neural Mechanisms of Song Production

The avian song system consists of specialized forebrain nuclei that control song learning and production. The neurobiological investigation of this system has largely focused on the unique neural features of vocal control systems that contribute to learned motor patterns in songbirds. The role of emotion has been disregarded in developing a theory of song learning and performance. Emerging evidence supports Darwin's observation that vocal communication is emotional expression. Neural pathways mediating emotional state remained integrated with the vocal control system as forebrain vocal control pathways evolved to support learned communication patterns. Vocalizations are therefore both a motor component of an emotional state and can influence emotional state via sensory feedback during vocal production. By acknowledging the importance of emotion in vocal communication, researchers propose that the song system and limbic brain are functionally linked in the production and reception of song Song and the limbic brain: a new function for the bird's own song.

Basal Ganglia Circuits

The songbird anterior forebrain pathway, a basal ganglia thalamocortical circuit, drives song motor learning. This pathway was previously thought not to play a role in song performance because early lesions showed no effect. This perspective has been revised by evidence that the pathway influences song syntax in some species. Bilateral excitotoxic lesions targeting the lateral and medial subdivisions of the anterior forebrain pathway in adult canaries produce a stuttering-like behavior, defined as prolonged, variable syllable repetition before transitions, resulting in a significant increase in phrase-duration variability. This effect was strongest in birds with medial and lateral lesions and was not observed in birds with lateral-only lesions. The increased variability persisted throughout the post-lesion recording period and was accompanied by small changes in the acoustic structure of syllables. These results implicate the medial anterior forebrain pathway in the ongoing control of phrase duration in adult canary song, challenging the view that this pathway is dispensable once song is learned Lesions Involving Medial Anterior Forebrain Pathway Circuitry Destabilize Phrase Timing in Adult Canary Song.

Synaptic Locus of Learning

Learning by imitation is the foundation for verbal and musical expression, but its neural basis remains unclear. A juvenile male zebra finch imitates the multisyllabic song of an adult tutor in a process that depends on a song-specialized cortico-basal ganglia circuit. Plasticity at a particular set of cortico-basal ganglia synapses is hypothesized to drive rapid learning-related changes in song before these changes are subsequently consolidated in downstream circuits. Combining a computational framework to quantify song learning with synapse-specific optogenetic and chemogenetic manipulations within and downstream of the cortico-basal ganglia circuit has identified the specific cortico-basal ganglia synapses that drive the acquisition and expression of rapid vocal changes during juvenile song learning. Transiently augmenting postsynaptic activity in the basal ganglia briefly accelerates learning rates and persistently alters song, demonstrating a direct link between basal ganglia activity and rapid learning. These results localize the specific cortico-basal ganglia synapses that enable a juvenile songbird to learn to sing and reveal the circuit logic and behavioral timescales of this imitative learning paradigm A synaptic locus of song learning.

Intrinsic Plasticity

Information processing and storage in the brain is thought to be primarily orchestrated by synaptic plasticity, but other neural mechanisms such as intrinsic plasticity are available. The intrinsic properties of a class of basal ganglia projecting song system neurons in zebra finch relate to each bird's unique learned song. These properties change over development and are maintained dynamically to rapidly change in response to auditory feedback perturbations. Exploring reciprocal interactions between intrinsic and network properties is a fruitful avenue for understanding mechanisms of birdsong learning Intrinsic plasticity and birdsong learning.

Temporal song structure is related to specific intrinsic properties of premotor basal ganglia projecting neurons. Neurons from birds who sang longer songs, including longer invariant vocalizations, had intrinsic properties that reflected increased post-inhibitory rebound. This suggests a rebound excitation mechanism underlying the ability of these neurons to integrate over long periods of time throughout the song and represent sequence information. A network model of realistic neurons shows how in vivo bursting properties link rebound excitation to network structure and behavior. These results demonstrate an explicit link between neuronal intrinsic properties and learned behavior. Sequential behaviors exhibiting temporal regularity require intrinsic properties to be included in realistic network-level descriptions Intrinsic properties link a network model to zebra finch song.

Adult Neurogenesis

One of the dogmas of neurobiology held that when nerve cells in the vertebrate brain die, they are not replaced by new ones. Research on adult canaries demonstrated the contrary. When the adult canary needs to learn new songs, it grows new neurons. This finding could eventually lead to the discovery of ways to repair lesions in the human brain From bird song to neurogenesis.

Song Amplitude and Acoustic Communication

Animals use acoustic signals to exchange information essential to their survival and reproduction. A crucial parameter in acoustic communication is the amplitude of the signal, as it plays a decisive role in signal transmission and can also encode information. Signal amplitude has been largely neglected in animal communication studies because it is difficult to assess. Field measurements of song amplitudes in 17 European songbird species investigated the sources of variation between and within species. Song amplitude increased with increasing background noise, in the presence of singing rival males, and varied with the time of day. These findings highlight that birds can adjust how loud they sing in response to changes in the biotic and abiotic environments A comparative analysis of song amplitude across and within bird species.

Phylogenetically informed analysis found no support for the long-standing hypotheses that song amplitude reflects body size or territory size across bird species. The variation of song amplitude between species is related to differences in ecology, in the strength of sexual selection, and in the costs of singing loudly instead of related to body size A comparative analysis of song amplitude across and within bird species.

Auditory Feedback and Song Maintenance

The brain song control system contains neurons with both premotor and auditory function. No evidence so far shows that these neurons respond to the bird's own song during singing. Also, no neurons have been found to respond to perturbation of auditory feedback in the brain area that is thought to be involved in the feedback control of song. The phenomenon of gating, in which neurons respond to playback of the bird's own song only during sleep or under anesthesia, is the sole known evidence for control of auditory input to the song system. It is not known whether the gating is involved in switching between the premotor and auditory function of neurons in the song control system The role of auditory feedback in birdsong.

Song Replay During Sleep

During sleep, sporadically, it is possible to find neural patterns of activity in areas of the avian brain that are activated during the generation of the song. In the vocal muscles of a sleeping bird, it is possible to detect activity patterns during these silent replays. A dynamical systems model for song production in suboscine birds translates the vocal muscle activity during sleep into synthetic songs. This approach poses the biomechanics as a unique window into the avian brain, with biophysical models as its probe Synthesizing avian dreams.

Studying Bird Song

Bird song research requires methods for recording, visualizing, and analyzing vocalizations. Spectrograms provide a visual representation of sound frequency over time and are the standard tool for song analysis. Recent advances in machine learning have produced automated tools for song segmentation, classification, and species identification.

Spectrogram Analysis

A spectrogram displays the acoustic structure of a song, showing frequency on the vertical axis, time on the horizontal axis, and amplitude as color intensity. Researchers use spectrograms to identify syllables, measure song duration, and compare songs across individuals and species. Human inspection of spectrograms remains common, but computational methods developed from human intuitions may not accurately reflect how birds perceive song similarity Bird song comparison using deep learning trained from avian perceptual judgments.

Automated Song Recognition

Transformer models have been applied to bird species classification using mel-spectrogram representations of bird songs. A study using the BirdClef 2024 dataset, consisting of 24,641 audio recordings across 182 bird species, evaluated Vision Transformer, Data-efficient Image Transformer, and Swin Transformer models. Each model was trained over 75 epochs with early stopping mechanisms. The results indicated strong training accuracy for all three models, however, validation accuracy plateaued between 78% and 84%, revealing overfitting issues. Various optimizations were implemented, including advanced learning rate scheduling, targeted dropout rates, weight decay adjustments, and data augmentation techniques. Despite these modifications, some overfitting persisted, particularly in the Data-efficient Image Transformer model, which showed decreased test accuracy after optimization. These findings highlight the effectiveness of transformer-based models for audio-based bird species classification, while also underscoring the need for further regularization techniques and fine-tuning A Comparative Analysis of Transformer Models for Bird Song Recognition Using Mel-Spectrograms.

Convolutional neural networks have also been used for bird song detection. The structure of a convolutional neural network consists of an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. Data augmentation is done with the help of pitch shifting, volume adjustment, and masking of frequency. The technologies used are spectrograms and mel-frequency cepstral coefficients Convolutional Neural Network-Based Bird Sound Detection in a Forest Area with Reference to the Argumentation of Data.

Real-Time Syllable Classification

Closed-loop interventions require online recognition of a specific target syllable while the bird is singing, for example, for manipulation of auditory feedback, song-triggered neuronal microstimulation, or optogenetics. Existing tools for closed-loop interventions can recognize only single syllables through manually created templates, with limited flexibility to adapt to new experiments. A novel neural network approach called Moove, which stands for Marking Online using only the Onsets of Vocal Elements, performs real-time syllable segmentation and classification of Bengalese finch songs. Moove's two-stage architecture detects syllable onsets and offsets and classifies syllables using acoustic information only from the first part of the syllable, enabling precise temporal contingency between behavior and feedback. Moove correctly segments and classifies Bengalese finch syllables in real time, with speed and reliability that allows effective operant conditioning experiments. Moove could be used for other closed-loop experiments on vocal signals Real-Time Segmentation and Classification of Birdsong Syllables for Learning Experiments.

Perceptual Validation of Similarity Measures

Most assessments of song similarity are based on human inspection of spectrograms or computational methods developed from human intuitions. Using a novel automated operant conditioning system, researchers collected a large corpus of zebra finch decisions about song syllable similarity. This dataset was used to compare and externally validate similarity algorithms in widely used publicly available software including Raven, Sound Analysis Pro, and Luscinia. Although these methods all perform better than chance, they do not closely emulate the avian assessments. A novel deep learning method that produces perceptual similarity judgments trained on avian decisions outperforms the established methods in accuracy and more closely approaches the avian assessments. Inconsistent and hence ambiguous decisions are a common occurrence in animal behavioral data. A modification of the deep learning training that accommodates these leads to the strongest performance. This approach is the best way to validate methods to compare song similarity Bird song comparison using deep learning trained from avian perceptual judgments.

Practical Workflow for Identifying Bird Songs

Field identification of bird songs requires a systematic approach that combines listening skills, recording equipment, and analysis tools. The following workflow provides a practical method for students and researchers.

Step 1: Listen and Describe

Begin by listening carefully to the song and describing its basic features. Note the rhythm, pitch range, tempo, and whether the song consists of repeated phrases or continuous sound. Describe the habitat and time of day, as these contextual cues narrow the range of possible species. Record the date, location, weather conditions, and any visible behavior of the singing bird.

Step 2: Record the Song

Use a directional microphone and a portable recorder capable of capturing high-frequency sounds. Position yourself at a distance that allows a clear recording without disturbing the bird. Record for at least one minute to capture the full range of song types. Note the recording level to avoid clipping, which distorts the amplitude information that can be important for analysis.

Step 3: Generate a Spectrogram

Transfer the recording to a computer and generate a spectrogram using audio analysis software. Set the frequency range to cover the expected song range of the target species, typically 1 to 10 kilohertz for most songbirds. Adjust the time and frequency resolution to clearly resolve individual syllables. Examine the spectrogram for syllable types, repetition patterns, and overall song structure.

Step 4: Compare with Reference Recordings

Compare your spectrogram with reference recordings from field guides or online databases. Focus on diagnostic features such as syllable shape, frequency range, and temporal pattern. Be aware that individual variation exists within species and that regional dialects can differ. Use multiple reference recordings to account for this variation.

Step 5: Confirm with Multiple Observations

Confirm your identification by observing the bird visually when possible and by hearing the song on multiple occasions. Note any song types that vary with context, such as dawn song versus daytime song. Record your identification in a field notebook with the date, location, and supporting evidence.

Records and Measurements

Systematic record keeping is essential for bird song studies. The following table summarizes the measurements that researchers commonly collect and the tools used to obtain them.

Common Failure Patterns in Song Identification

Several recurring errors occur when identifying bird songs. Recognizing these patterns improves accuracy and prevents misidentification.

Overlapping Frequency Ranges

Many species produce songs that overlap in frequency range and temporal structure. Relying on a single feature such as pitch or duration leads to errors. Use multiple diagnostic features and consider the habitat and geographic location.

Regional Dialects

Song dialects vary geographically within species. A recording from one region may not match reference recordings from another region. Consult regional field guides and local expert knowledge when working in unfamiliar areas.

Mimicry

Some species incorporate sounds from other species into their songs. This mimicry can confuse identification, especially when the mimicked elements dominate the song. Listen for the species-specific syntax and delivery pattern instead of individual elements.

Degraded Recordings

Background noise, distance, and recording equipment limitations degrade the acoustic signal. Low-amplitude elements may be lost, and frequency information may be distorted. Record under favorable conditions and note the quality limitations in your records.

Individual Variation

Individual birds within a species vary in their song performance. Young birds may sing incomplete or variable versions of the species typical song. Compare multiple songs from the same individual and account for age related variation.

Welfare and Safety Context

Bird song research involving live animals requires attention to welfare standards and legal regulations. Observational studies that do not disturb birds have minimal welfare impact. Recording and playback experiments can affect bird behavior and should be designed to minimize stress. Captive studies require appropriate housing, nutrition, and veterinary oversight.

Ethical Considerations for Playback Experiments

Playback experiments expose birds to simulated territorial intrusions. These experiments can elevate stress hormones and alter behavior. Limit playback duration and intensity, and avoid repeated exposure of the same individuals. Monitor birds for signs of distress and terminate experiments if adverse effects are observed.

Permits and Regulations

Many jurisdictions require permits for capturing, handling, or recording protected bird species. Check local regulations before beginning any field study. Institutional animal care and use committees review and approve research involving vertebrate animals. Compliance with these requirements is a professional responsibility.

Data Quality and Reproducibility

The quality of bird song recordings affects the reliability of analysis. Document recording equipment, settings, and conditions in your records. Archive raw recordings and analysis scripts to support reproducibility. Unsupervised classification methods can improve the quality of a bird song recording dataset by identifying and removing poor quality recordings Unsupervised classification to improve the quality of a bird song recording dataset.

Limitations and Professional Escalation

Bird song research has inherent limitations that researchers should acknowledge. Automated classification tools show promise but currently plateau in validation accuracy due to overfitting A Comparative Analysis of Transformer Models for Bird Song Recognition Using Mel-Spectrograms. Computational similarity measures do not closely emulate avian perceptual judgments Bird song comparison using deep learning trained from avian perceptual judgments. These limitations affect both research conclusions and practical identification tools.

When to Seek Expert Assistance

Consult an expert ornithologist or bioacoustician when you encounter songs that resist identification despite systematic analysis, when you need to confirm a rare or unusual species record, when your research requires quantitative comparison of songs across populations, or when you plan to use playback or capture methods that require permits. Professional escalation ensures that your observations contribute reliably to the scientific record.

Emerging Research Directions

Current research is exploring how early environmental insults affect vocal development. Embryonic exposure to valproic acid in zebra finches disrupts song acquisition, resulting in impaired tutor imitation, delayed song consolidation, and increased motif variability. These abnormalities co-occurred with social deficits, including reduced modulation of female-directed song and diminished social engagement, and frequency-specific hyperacusis. These findings demonstrate that embryonic valproic acid exposure disrupts the development of learned vocalizations and social behavior in a vocal learner Vocal and social impairments in songbirds embryonically exposed to an autism risk factor.

The pineal gland plays a role in the photoperiodic control of bird song frequency and repertoire in the house sparrow The role of the pineal gland in the photoperiodic control of bird song frequency and repertoire in the house sparrow, Passer domesticus. This research connects seasonal changes in day length to song production and may inform understanding of how environmental cues regulate vocal behavior.

Frequently Asked Questions

What is the difference between a bird song and a bird call?

Bird songs are typically longer, more complex vocalizations produced mainly by males during the breeding season to defend territory and attract mates. Bird calls are shorter, simpler sounds used for alarm, contact, and other immediate functions. The distinction is not always sharp, and some species produce calls that are structurally complex.

Do female birds sing?

In many temperate songbird species, only males sing. However, female song is common in tropical species and is increasingly recognized in temperate species. Research on female song has expanded in recent years, and the functions of female song include territory defense and mate attraction similar to male song.

How do birds learn their songs?

Young songbirds memorize a tutor song during an early sensitive period and use auditory feedback to shape their own vocalizations. The process involves a sensory phase where the bird forms an auditory memory of the tutor song and a sensorimotor phase where the bird matches its own output to that memory The role of auditory feedback in birdsong.

Can birds learn new songs as adults?

Most songbird species have a limited sensitive period for song learning, but some species, including canaries, continue to modify their songs as adults. Adult canaries grow new neurons when learning new songs, demonstrating that the adult brain retains plasticity for vocal learning From bird song to neurogenesis.

Why do birds sing at dawn?

The dawn chorus is a period of intense singing activity around sunrise. Several hypotheses explain this behavior, including optimal acoustic transmission conditions, low foraging activity at dawn, and the need to reaffirm territory boundaries after the night. Song amplitude also varies with the time of day A comparative analysis of song amplitude across and within bird species.

How do researchers measure song similarity?

Researchers use spectrograms and computational algorithms to compare songs. Widely used software includes Raven, Sound Analysis Pro, and Luscinia. However, these methods do not closely emulate avian perceptual judgments. Deep learning methods trained on avian decisions outperform established methods in accuracy Bird song comparison using deep learning trained from avian perceptual judgments.

What is the developmental stress hypothesis?

The developmental stress hypothesis proposes that learned features of song, including complexity and local dialect structure, serve as indicators of male quality. Brain structures underlying song learning develop during the first few months after hatching, a period when songbirds are subject to nutritional and other developmental stresses. Individuals that fare well in the face of stress invest more resources to brain development and are better at song learning Song function and the evolution of female preferences: why birds sing, why brains matter.

How is bird song research relevant to human health?

Bird song is a model system

Related Articles

• Elements Of Statistical Learning
• Elements Of Statistical Learning
• Elements Of Statistical Learning
• Introduction To Statistical Learning
• Introduction To Statistical Learning

References and Further Reading

• NCBI Literature Resources. National Center for Biotechnology Information.
• PubMed. National Library of Medicine.
• Song and the limbic brain: a new function for the bird's own song.. Annals of the New York Academy of Sciences, 2004.
• From bird song to neurogenesis.. Scientific American, 1989.
• Synthesizing avian dreams.. Chaos (Woodbury, N.Y.), 2024.
• Song function and the evolution of female preferences: why birds sing, why brains matter.. Annals of the New York Academy of Sciences, 2004.
• Intrinsic plasticity and birdsong learning.. Neurobiology of learning and memory, 2021.
• A comparative analysis of song amplitude across and within bird species.. Proceedings. Biological sciences, 2025.
• The role of auditory feedback in birdsong.. Annals of the New York Academy of Sciences, 2004.
• Bird song learning as an adaptive strategy.. Ciba Foundation symposium, 1997.
• Bird Song Learning is Mutually Beneficial for Tutee and Tutor. 2019.
• Real-Time Segmentation and Classification of Birdsong Syllables for Learning Experiments.. 2026.
• Auditory experience-dependent cortical circuit shaping for memory formation in bird song learning.. 2016.
• Intrinsic properties link a network model to zebra finch song.. 2026.
• Vocal and social impairments in songbirds embryonically exposed to an autism risk factor.. 2026.
• Lesions Involving Medial Anterior Forebrain Pathway Circuitry Destabilize Phrase Timing in Adult Canary Song. 2026.
• A synaptic locus of song learning.. 2026.
• A Comparative Analysis of Transformer Models for Bird Song Recognition Using Mel-Spectrograms. 2024 7th International Conference on Information and Communications Technology (ICOIACT), 2024.
• Bird song comparison using deep learning trained from avian perceptual judgments. bioRxiv, 2022.
• STFT Transformers for Bird Song Recognition. Conference and Labs of the Evaluation Forum, 2021.
• Radar Target Classification Using Enhanced Doppler Spectrograms with ResNet34_CA in Ubiquitous Radar. Remote Sensing, 2024.
• A Classification Algorithm of UAV and Bird Target Based on L/K Dual-Band Micro-Doppler and Mamba. Drones, 2026.
• Convolutional Neural Network-Based Bird Sound Detection in a Forest Area with Reference to the Argumentation of Data. 2025 13th International Conference on Intelligent Embedded, MicroElectronics, Communication and Optical Networks (IEMECON), 2025.
• Unsupervised classification to improve the quality of a bird song recording dataset. Ecological Informatics, 2023.
• The role of the pineal gland in the photoperiodic control of bird song frequency and repertoire in the house sparrow, Passer domesticus. Hormones and Behavior, 2014.

This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.