| |
CSRE Features
Signal Recording and Playback
CSRE supports CD-quality (44.1 kHz, 16 bits/sample) recording and playback (mono
or stereo) on a number of readily available audio systems, in several price ranges.
The recording and playback of digital signals is done directly to/from disk so
signal length is limited only by hard disk space.
Entire signals or selected parts of any signal can be played back from all editing
and analysis windows. The editing environment is designed to allow quick segmentation
of signals. Signals can also be labelled or tagged to facilitate reference and
analysis.
Comprehensive Signal Editing Capabilities
Any digitized signal can be cut, spliced, and/or labeled. CSRE also lets you import
or export files between CSRE and other software packages, taper, and amplify or
attenuate signal segments.
 |
CSRE's
QuickwindowTM Editor provides an extremely efficient way to segment signals,
and measure the duration of segments and features. The zoom facility provides
precision down to the individual sample point. Informative displays provide
the duration, RMS amplitude and peak levels of any signal. Measured values
can be logged automatically to a disk file so they can be imported into
spreadsheet or wordprocessing packages for later analysis or reporting.
|
Full
View |
Pitch Analysis
 |
This
pitch contour (top window) is derived by processing the audio signal (middle
window) through Comb-Filter analysis. The filtered signal (bottom window)
shows the markers placed by the analysis between pitch cycles. |
Full
View |
CSRE lets you extract the fundamental frequency (pitch) of any voiced, digitized
speech sample, using your choice of the cepstral or comb filtering methods. CSRE
displays the pitch track, the time-domain signal and markers for individual glottal
cycles (comb filtering method only). Jitter and shimmer are measured automatically.
The comb filtering method allows visual confirmation that each pitch cycle has
been properly extracted, with manual correction if necessary.
Spectral Analysis
CSRE provides a choice of four powerful, modern spectral analysis methods, including:
- short-term Fourier transform
- Linear Predictive Coding (LPC) using autocorrelation
- Linear Predictive Coding (LPC) using modified covariance
- high-resolution time and frequency analysis using the Cone-kernel TFR
The results of these analysis can be displayed as full-color spectrograms or as
“waterfall” spectrograms. Available displays include a spectrogram plus a time-synchronized
waveform. Two spectrograms can also be displayed for comparison.
 |
CSRE's
dual spectrogram display which shows 2 independent signals analyzed using
different anlysis methods. Such a display is ideal for comparing different
analysis methods or to iluustrate the differences between normal and disordered
speech. |
Full
View |
Signals can be scrolled or zoomed facilitating detailed examination of even very
long signals. Any spectrogram or spectrogram segment can be replayed under mouse
control. Instantaneous measurement of the time, frequency and amplitude values
of any part of the spectrogram are available via the mouse-controlled cursor.
Any spectrogram or other display can be saved to disk as a PCX file for later,
publication-quality printing on a color or laser printer.
Analysis parameters such as the type and size of analysis window, degree of overlap,
and the analysis order can be changed via pop-up menus, so you can easily vary
the analysis parameters as the needs of your analysis application change. The
range of frequencies and amplitudes displayed can also be adjusted in this manner.
The system extracts spectral peaks automatically; with a simple mouse click, measured
values can be logged directly to disk files that can be imported into spreadsheet
and word processing programs for further analysis. Full tutorials are provided
to guide the new user through the analysis process.
Formant Extraction
CSRE’s formant extraction facility automatically extracts and displays speech
formants. Formant frequencies, amplitudes, and bandwidths are extracted using
LPC analysis and displayed as full-color parametric spectrograms.
 |
This
display shows the results of formant analysis using LPC (top window).
The display also shows time synchronized waveform and pitch contour. |
Full
View |
LPC analysis parameters can be controlled via pop-up menus. Any spectrogram or
spectrogram segment can be replayed under mouse control. Instantaneous measurement
of the time, frequency and amplitude values of any part of the display are available
via the mouse-controlled cursor. “Point and click” logging of formant parameter
data to disk produces files that can be imported into spreadsheet or word processing
programs for further analysis. Formant parameter files can also be imported directly
to CSRE’s parametric speech synthesizer, permitting “copy-synthesis” in which
the original signal is closely approximated.
CSRE Utilities
The CSRE package includes numerous special utilities targeting specific needs
identified by customers. One utility, Audiohdr, is a file format conversion program,
allowing you to import audio files created by other programs. It also converts
CSRE audio files to other standard formats so that they can be read by other programs.
Other utilities permit you to append a specified amount of silence to an audio
file, to add noise to a file at a specified signal to noise ratio, to interleave
two mono CSRE files to create a stereo file, to ramp the start and end of a file,
to replace labelled portions of a CSRE file with noise at a specified level, and
to perform third-octave analysis on a spectrogram file.
Measurement Capabilities
CSRE offers a comprehensive array of proven measurements, including signal and
segment duration, RMS amplitude, peak amplitude, formant frequencies, amplitudes
and bandwidths, fundamental frequency track, jitter, and shimmer. Segment labeling
and mouse-controlled playback of any segment of a displayed signal provide assurance
about just what is being measured. Extensive display, zoom, and scrolling capabilities
under mouse control help you find what you want and measure it fast. Automatic
logging to disk files that can be read by spreadsheet and wordprocessing packages
reduces clerical work and makes you more productive. Together, these features
give you an extremely powerful measurement capability.
CDROM databases
Interested in using standardized, CD-ROM databases? No problem. CSRE makes it
easy for you to read databases such as NIST’s TIMIT, the Oxford University Press
Acoustic-Phonetic Database, Kay’s database of disordered speech and many others.
Order CSRE | Download
CSRE Demo
|
|
