Università degli Studi di Pavia

Centro Interdisciplinare di Bioacustica e Ricerche Ambientali

Software for sound analysis - Windows

The development of dsp techniques and of low-cost high-speed computer hardware with large hard disks has made the computer analysis of bioacoustical signals an every-day invaluable tool for ethological research and for monitoring the underwater environment.
The Windows environment offers many advantages over DOS. The drawback is that Windows requires a lot of hardware resources: mainly RAM and CPU power. A same task, on a same machine, may be 10 times slower in Windows than in DOS. Fortunately, CPU speed continues to increase while prices decrease. This fact induced to begin the development of a completely new analysis software able to run in windows. Capabilities and features now available on todays's machines were a dream only few years ago; visit the DSPW pages to know more about the evolution of the hardware and the development of this software.

The new software was released for internal use in 1997 with the name wSpecGram but was then definitively named SEA (Sound Emission Analyzer). It runs in any windows version and uses any installed windows-compliant sound device, including digital I/O boards and boards with up to 8 channels at 192 kHz. The software is also able to fully exploit the capabilities of the new Dodotronic UltraMic200k and UltraMic250k, ultrasonic microphones with incorporated AD and USB port able to receive ultrasounds up to 120 kHz. The software works well on notebooks and tablet pcs too and this makes it a great tool for field research. The software is available in two versions, SeaWave that is free for personal non profit use, and SeaPro for professional and institutional use. Advanced versions for specific research are also available for CIBRA and its partners.

As it is based on Windows it provides ease of use and full compatibility with the sound devices and sound analysis software available on the market.
The software includes analysis, recording and display tools, with real-time spectrogram and cepstrogram, spectral averaging, frequency tracking, event counting, scheduled recording, etc..
Other features include: 1 or 2 channels, real-time spectrogram display, real-time cepstrogram display, wrap-around or scrolling display, wide control on all analysis parameters, frequency-time cursor while in real-time mode, FFT size up to 16k points, frequency zoom capabilities with real-time spanning, frequency tracking, file analysis, file play with real-time display, drag & drop file play, play list management with real-time file concatenation, file recording with real-time display, display up to 3840x1200 pixels and compatibility with rotated and multiple displays, screen save in .bmp format. Sound files and spectrograms are saved in a standard format to allow further processing with other software.
To fullfill the requirements of acoustic surveys, recording capabilities have been expanded by adding scheduled recording capabilities, on event recording, gps position logging, and a continuous survey mode to record a file each hour until all the available disks space is filled. By using a RAID controller with, say, 4x80GB disks, it records for 17 days (stereo, 16 bit wav, 48kHz sampling). HD support is now upgraded to manage multiple disks up to 2TB each. More functions are in development, in particular for detecting and classifying acoustic events.
Depending on the installed sound acquisition devices, analog I/O is allowed in the audio frequency range and/or in the ultrasonic range up to 48 and 96kHz (96 and 192 kHz sampling respectively). Digital I/O by means of PCI boards, USB devices or FireWire devices is also possible to provide direct transfer from DAT recorders or other digital systems to the PC.

An advanced version, named SEADAQ, has been developed to support very fast sampling rates for studying ultrasounds extending to more than 200 kHz. This version uses National Instruments DAQ devices that include PCI boards, PCMCIA cards and FireWire data acquisition peripherals with sampling rates up to MHz sampoling. By using a PCMCIA card in a fast notebook it is possible to acquire, record and display in real time ultrasounds with sampling rates up to 500 kHz sampling.

High resolution real-time capabilities, typically available in much more expensive instruments, are very useful in field experiments to monitor the acoustic activities of the emitting subjects (immediate correlation among observed behaviours and emitted/received signals) and to optimise the instrumental setup (minimisation of noise, transducer placement, verification of the recording chain). These facilities allow to immediately evaluate the results of an experiment instead of waiting for later analyses on the recordings, and they make easier to analyse long periods. A portable version based on a notebook can be easily moved across laboratories or used in on-field applications, for example those requiring real-time visualization and recording of acoustic events. Real-time capabilities and the continuous recording to disk with GPS position logging are particularly suitable for the continuous monitoring of underwater environment and for carrying out wide area acoustic surveys for both scientific and mitigation needs. By using a GIS it is then possible to map survey tracks and to plot the detected acoustic events in real-time.

The classification of received acoustic events, either biological or artificial sounds, is still a challenge. A trained observer is normally able to correctly classify basic sound categories in real-time by join listening with spectrogram observation; though this is a very demanding task and requires skilled operators. Doing the job 24/24h for long periods may be prohibitive. This is why it is required to develop reliable sound classification algorithms for working in real-time or for post processing long recordings. Such algorithms have been developed for very specific uses and can't be applied to generic tasks yet.

To know more visit the page dedicated to the SeaPro software suite and see spectrogram samples on the pages on Soundscapes, on birds and on Marine Mammals' voices.

Other software developed at CIBRA

SEAcommander
a software to remotely control a SEAPro/SEAUltra running on a networked PC

CatWav
a program to catalog .wav files

SlotLogger
classification of received signals & distribution of data by UDP; works on 1 minute time slots

EventLogger
software to log acoustic detections and other events

NMEAManager
management of serial NMEA data and broadcast by UDP, conversion of serial to UDP and back; it is required to make GPS information available to SeaPro for georeferenced filenaming

UDPMonitor
display & log UDP messages; useful to monitor GPS data and other information transmitted through UDP

OziManager
feed OziExplorer with NMEA data received on UDP port (for example NMEA data distributed by NMEAManager)

ArcUDP
VBA extension to ArcView 8.2 to receive and plot UDP data generated by NMEAManager, SlotLogger and EventLogger

WhaleSizer
a software to assess in real-time the size of sperm whales by displaying and measuring the Inter Pulse Interval (IPI) computed by means of cepstrum analysis.

WaveBeamformer
based on the WaveRecorder, this software allows to perform beamforming in the time domain. It allows to record either the original channels or the beamformed ones. It also allows to select any 2 beams for listening and/or for recording. It works on up to 96 channels with sampling rate up to 192ks/s. Spacing of the sensors must fit the beamformer parameters.

WaveRecorder (developed for the NEMO-ONDE Project)
a multichannel recorder (1-96 channels; up to 192ks/s; 16, 20 or 24 bits; stores data in 16bits, 32bits, or Float format; scheduled recording)

WaveRecorder2x2 (developed for the NEMO-ONDE Project)
4 channels recorder able to get in synch two stereo boards (4 channels; up to 192ks/s; 16, 20 or 24 bits; stores data in 16bits, 32bits, or Float format; scheduled recording)

SeaPro4ch (developed for the NEMO-ONDE Project)
special version of SeaPro/SeaUltra able to read and play 4 channels wav files recorded with WaveRecorder. It allows to select two channels for play/display any 2 channels of the 4 stored in the file. This ability will be incorporated soon in the SeaPro/SeaUltra.

SeaDSP
this software works in post-processing only to manipulate files; it allows many processing functions including filtering, downsampling, and band shifting to shift ultrasonic components into audible range. It also performs spectral compression to compress a wideband spectrum into the audible range.

SeaNET
still in experimental stage, this SeaPro/Ultra option allows to analyze/record sounds distributed on the network by the software AudioGem.

Examples

The wSpecgram display in 2 channels mode.

Screenshot of SEADaq showing pulse trains emitted by an ant.

SEADaq showing mice calls in the range 48 to 60 kHz (frequency zoom mode).

Whistles, clicks and bangs emitted by a bottlenose dolphin.

Software for sound analysis - DOS

DOS programs were developed to fully exploit the capabilities of CPUs and sound boards available until few years ago. Visit the DSPW and History pages to know more about the development of this software and the constraints imposed by hardware. The software was developed to support the following signal acquisition boards:

Download the software, install it, run it and then email us for registration (see below) and further info.

Hardware requirements: a PC with at least a 486DX2/66 CPU (faster CPUs allow smoother display and wider control on analysis parameters), a VGA graphic board, a Creative SoundBlaster 16/32/64 or a true compatible (Vibra16 chipset seems to work well too).
The sound board must be correctly installed with the SET BLASTER command string in the autoexec.bat file. It is strictly required the high DMA channel set to 5, 6 or 7.
The software requires to run in MS-DOS mode; it can be run in Win95/98 in full screen DOS mode as well.
Installation: create a directory just below the root, unzip the files in that directory, read carefully the help file (sbrta.hlp - a plain text file), run the sbrta program: a short beep will say that the board has been recognized. If something goes wrong, please report to the Author all the information on the welcome screen.
Registration: when you run it for the first time, the program will check for a properly configured suitable sound board; if everything is ok it sounds a beep and then displays a SiteCode, unique for you PC, and it prompts for a Password. Email the SiteCode to gianni.pavan@unipv.it to get your own Password. The registration is free! It is required to keep track of the circulating copies of the software.
Software features: Hard Disk recording and/or real-time spectrographic display up to 44100 samples/sec. Wide control on all the analysis parameters (FFT 256, 512 and 1024 points; several wheighting functions with zero padding and overlap up to 93.75%; frame skipping, frame packing, and frame averaging). Three display modes (wrap-around and scrolling horizontal display; four strips scrolling display; vertical scrolling display). Spectrogram and envelope display. Real-time cepstrogram display. Frequency tracking. On screen cursor. Several colour tables. Screen save to .bmp file.

Examples

Repeated bird calls as shown by the Duetto software.
The envelope plot (black trace) is overlaid on the 0-14kHz colour spectrogram.

A bird song shown in four black & white spectrogram strips.