Bird Call Identifier
SMJ-A09BBird Call IdentifierIdentifying Songs of Bird Species through Digital SignalProcessing TechniquesA Major Qualifying Project submitted to the Faculty ofWORCESTER POLYTECHNIC INSTITUTEIn partial fulfillment of the requirements for the degree of Bachelor of ScienceSubmitted by:Tyler CarrollRose ColangeloTom StrottSubmitted to:Project Advisor:Professor Susan Jarvis29 Apr 2010This report represents work of WPI undergraduate students submitted to the faculty as evidence of a degreerequirement. WPI routinely publishes these reports on its web site without editorial or peer review. For moreinformation about the projects program at WPI, see http://www.wpi.edu/Academics/Projects.1
AbstractThe purpose of this Major Qualifying Project is to create a device that identifies bird callsin the wild. The aim of this project is to create a handheld device that will be able to interpretbird calls using a high quality microphone and various signal processing techniques and displaythe top matches on an LCD screen to the user. The main objective of this project is to be able toidentify the songs of several bird species in central Massachusetts on a lab development boardwhile additional work would include downloadable software that would add the bird songs ofmore bird species and the conversion of from a lab development board to a handheld device.2
AuthorshipThe three team members contributed equally to the work on this project and to the report.3
AcknowledgementsThe team would like to thanks the following people for their contributions to the project:Susan Jarvis, Adjunct Instructor, WPIMike Webster, Director of the Macaulay LibraryMichael Young, Sound Technician, Macaulay LibraryChristine Drew, Manager, Instruction & Outreach, Gordon Library4
Table of ContentsAbstract. 2Authorship . 3Acknowledgements . 4Table of Figures . 7Table of Tables . 8Table of Equations . 8Chapter 1: Introduction . 9Chapter 2: Background . 122. 1 Prior Art . 122.2 Signal Processing . 142. 3 Mel-frequency Cepstrum . 162.4 Bird Call Recognition . 18Chapter 3: Methodology. 203.1 Scope. 203.2 System Block Diagram. 213.3 Creation of an Algorithm. 233.4 Obtaining Samples . 243.5 Modular Design Choices. 263.5.1 Netbook Processor . 273.5.2 Front End Filter . 283.6.3 Digital Signal Processing Chip . 283.6.4 LCD Screen . 293.6.5 Microphone . 303.7 Algorithm Implementation. 303.7.1 Filtering and Frame Generation . 313.7.2 Windowing Concepts . 323.7.3 Mel-Scale Filtering. 333.7.4 Correlation and Database Implementation . 36Chapter 4: Results . 384.1 MATLAB Testing Results. 384.2 Testing the Algorithm in C . 435
4.2.1 Fast Fourier Transform Test. 434.2.2 Discrete Cosine Transform Test . 454.2.3 Correlation Function . 474.2.4 Hamming Window Function . 504.2.5 Testing the MFCC Algorithm . 524.3 Obstacles . 544.3.1 Record and Playback Function . 544.3.2 Front End Filtering. 554.3.3 Complex Numbers. 564.3.4 Fast Fourier Transform . 574.3.5 Memory Problems. 574.3.6 Noise. 59Chapter 5: Conclusions . 615.1 Discussion of MATLAB Results. 615.2 Discussion of C Results . 615.3 Future Work Recommendations . 625.3.1 MATLAB Future Work. 625.3.2 C Implementation Future Work . 625.3.3 Hardware Design Future Work . 63Appendices. 65Appendix A: MATLAB Test Results . 65Appendix B: MATLAB Source Code. 85Bird Finder with Database. 85MFCC Comparison Code. 85Appendix C: DSP Source Code . 87source.c . 87kannumfcc1.2.c . 98fft.c . 102corr.c . 102mfcc bank.c . 105References . 1156
Table of FiguresFigure 1: System Block Diagrams . 21Figure 2: Cape May Warbler Song Comparison With Background Noise in Signal 1 . 26Figure 3: Magnitude Response of Lowpass Filter. 28Figure 4: Block Diagram of C6713 DSK . 29Figure 5: Algorithm Flow Chart . 32Figure 6: Frequency Domain Filtering Flow Chart . 35Figure 7: Comparison of Two Calls from the Same Carolina Wren . 39Figure 8: Comparison of Two Human Whistle Trills . 40Figure 9: Magnolia Warbler Comparison with One "Unknown" Song . 41Figure 10: Carolina Wren Comparison with One "Unknown" Song . 42Figure 11: 2 kHz Sine Wave. 44Figure 12: Fast Fourier Transform of 2 kHz Sine Wave . 45Figure 13: DCT of Input Samples in C . 46Figure 14: DCT of Input Samples in MATLAB . 47Figure 15: Original Input Signal . 48Figure 16: Time Shifted Input Signal. 48Figure 17: Rectangular Windowed Cross Correlation of Input Signals. 49Figure 18: Wrap Around Cross Correlation of Input Signals . 50Figure 19: Hamming Window Function in C . 51Figure 20: Hamming Window Function in MATLAB . 52Figure 21: Cape May Warbler song comparison . 66Figure 22: Cape May Warbler song comparison . 67Figure 23: Cape May Warbler song comparison . 68Figure 24: Cape May Warbler song comparison . 69Figure 25: Cape May Warbler song comparison . 70Figure 26: Magnolia Warbler song comparison . 71Figure 27: Magnolia Warbler song comparison . 72Figure 28: Magnolia Warbler song comparison . 73Figure 29: Magnolia Warbler song comparison . 74Figure 30: Magnolia Warbler song comparison . 75Figure 31: Mourning Warbler song comparison . 76Figure 32: Mourning Warbler song comparison . 77Figure 33: Mourning Warbler song comparison . 78Figure 34: Mourning Warbler song comparison . 79Figure 35: Carolina Wren song comparison. 80Figure 36: Carolina Wren song comparison. 81Figure 37: Carolina Wren song comparison. 82Figure 38: Carolina Wren song comparison. 83Figure 39: Carolina Wren song comparison. 847
Table of TablesTable 1: System Requirements for 32-Bit MATLAB . 27Table 2: Bird song average correlations and percentages choosing the correct bird . 43Table 3: Cape May Correlations . 53Table 4: Carolina Wren Correlations . 53Table 5: Magnolia Warbler Correlations. 53Table 6: Mourning Warbler Correlations . 53Table of EquationsEquation 1: Discrete Fourier Transform . 15Equation 2: Cross Correlation . 16Equation 3: Conversion from Hertz into Mel . 178
Chapter 1: IntroductionVarious species of birds have unique bird calls. These bird calls are distinct based oninflection, length, and context, meaning the same bird may have more than one call. A devicethat would analyze the signal and identify the bird based on the bird call could be of tremendoushelp to an ornithologist. This project proposed the development of this device using signalprocessing and embedded design. The first task was to find or create a database of high-qualitybird calls to use for identification. Using this database, the team compared various features of thebird calls of a certain species and ascertained the features which distinguish that species fromother species. Using these features, a recorded bird call was identifiable as a species of bird.This project is an important effort because ornithology is not always an exact science inthe field; it is based on the interpretation of the scientist hearing the bird’s song. A device thatcould quantitatively match signal waveforms would make the science more exact. Furthermore,bird watching is a hobby that many people enjoy. The ability to identify birds could increase theenjoyment of bird watching enthusiasts everywhere.Before discussing the project further, it is important to delve into the prior art related tothe project to gain an understanding of how current products work and what needs improvement.Several products which identify bird songs already exist, but none of them currently inproduction do exactly what was attempted in this project. There is a wealth of information on theinternet as well as some handheld devices that require the user to match the bird calls. Adiscontinued product exists that does digital signal processing of the bird call and displays likelyspecies, and the team intends to produce something in that vein with some improvements.In order to create a bird call identification device, the team needed to correctly utilizeseveral signal processing techniques. Some of these techniques included filters, discrete Fourier9
transforms, cross-correlation, wavelets, cepstral analysis, and audio spectrograms (Cai et al.,2007; Lee et al., 2006). Filters were necessary to improve the quality of the bird songs andremove any unwanted noise. Bird songs cover a wide frequency range and discrete Fouriertransforms allowed the team to analyze the different frequencies in each call. Cross-correlationallowed the team to compare recorded bird calls with the bird call database, both in time andfrequency. Next, the group used discrete wavelet transforms. An advantage that discrete wavelettransforms (DWT) have over Fourier transforms is temporal resolution; a DWT captures bothfrequency and time information. Additionally, the team created audio spectrograms using bothFourier and discrete wavelet transforms to examine each bird’s song. Lastly, the group used melfrequency cepstral coefficients.
Various species of birds have unique bird calls. These bird calls are distinct based on inflection, length, and context, meaning the same bird may have more than one call. A device that would analyze the signal and identify the bird based on the bird call could be of tremendous help to an ornithologist.
Tibetan Buddhist Wheel of Life Identifier: 141748 Yarmulke Identifier: 2004.53.4 Hebrew Bible (Tanakh) Identifier: 2004.23.1 Prayer Rug Identifier: 2005.31.1 Tibetan Lama’s Crown Identifier: 113860 Rama Figurine Identifier: 121085 Ganesh Figurine Identifier: 116859 Qu’ran Identifier: 2005.31.2 Creche Identifier: E2006.1.25a-b Torii .
BeLux October 2015 vat incl. Little Bird 1x Little Bird piece 100 Bird 1x Bird piece 150 Super Bird 1x Super Bird piece 250 Sub AIR Wireless subwoofer piece 600 Bird pack 2 stands L&B 2x stand for Little or Bird pack 140 Bird pack 2 stands Super 2x stand for Super pack 180 iTransmitter High definition wireless piece 90 USB Transmitter Wireless transmitter piece
Base Controller Brand Base Controller Model Add-on Brand Add-on Model Based Controllers Add-on Qualifying Products List as of Aug 01, 2021 . Rain Bird ESP-LXME Rain Bird IQ4G-USA Rain Bird ESP-LXME Rain Bird IQNCC4G Rain Bird ESP-LXME Rain Bird IQNCCEN Rain Bird ESP-LXME Rain Bird IQNCCRS. Page 5 of 5
Bird Care Tips Keep the bird in a warm room. Feed your bird food it is used to eating. Give your bird twelve hours of quiet and darkness each day. Do not handle your bird for the first few weeks. Except during playtime, keep the bird in its cage. Avoid loud noises around your bird.
Behavior: How a bird acts can be a great clue for identification. Being a careful observer of bird behavior can help you take your bird identification to the next level. How is the bird sitting, eating, or flying? For example, when looking at the bird's posture, note if the bird is upright or horizontal. These traits can help
the bird-scarer must be justified and you must be able to demonstrate you are complying with best practice at all times. National Farmers Union leaflet: 'Bird Deterrents and Bird Scarers Bird-scarers must only be used to protect crops and be used only where there is a reasonable expectation of being effective. Although the use of bird-
fessional bird watchers sometimes disagree on the species given an image of a bird. It is a difficult problem that pushes the limits of the visual abilities for both humans and computers. Although dif-ferent bird species share the same basic set of parts, different bird species can vary dramatically in shape and appearance.
astrophysics, solar- terrestrial physics and efforts to harness fusion power using tokamaks. The syllabus comprises: Solar-like magnetic activity on other stars. The basic equations of magneto-hydrodynamics. Stellar coronae: X-ray properties and energetics of coronal loops. Energetics of magnetic field configurations. MHD waves and propagation of information. Solar and stellar dynamos: mean .
