Spectrogram Correlation And Transformation Research Articles

A comprehensive computational model of animal biosonar signal processing.

Computational models of animal biosonar seek to identify critical aspects of echo processing responsible for the superior, real-time performance of echolocating bats and dolphins in target tracking and clutter rejection. The Spectrogram Correlation and Transformation (SCAT) model replicates aspects of biosonar imaging in both species by processing wideband biosonar sounds and echoes with auditory mechanisms identified from experiments with bats. The model acquires broadband biosonar broadcasts and echoes, represents them as time-frequency spectrograms using parallel bandpass filters, translates the filtered signals into ten parallel amplitude threshold levels, and then operates on the resulting time-of-occurrence values at each frequency to estimate overall echo range delay. It uses the structure of the echo spectrum by depicting it as a series of local frequency nulls arranged regularly along the frequency axis of the spectrograms after dechirping them relative to the broadcast. Computations take place entirely on the timing of threshold-crossing events for each echo relative to threshold-events for the broadcast. Threshold-crossing times take into account amplitude-latency trading, a physiological feature absent from conventional digital signal processing. Amplitude-latency trading transposes the profile of amplitudes across frequencies into a profile of time-registrations across frequencies. Target shape is extracted from the spacing of the object’s individual acoustic reflecting points, or glints, using the mutual interference pattern of peaks and nulls in the echo spectrum. These are merged with the overall range-delay estimate to produce a delay-based reconstruction of the object’s distance as well as its glints. Clutter echoes indiscriminately activate multiple parts in the null-detecting system, which then produces the equivalent glint-delay spacings in images, thus blurring the overall echo-delay estimates by adding spurious glint delays to the image. Blurring acts as an anticorrelation process that rejects clutter intrusion into perceptions.

How frequency hopping suppresses pulse-echo ambiguity in bat biosonar

Big brown bats transmit wideband FM biosonar sounds that sweep from 55 to 25 kHz (first harmonic, FM1) and from 110 to 50 kHz (second harmonic, FM2). FM1 is required to perceive echo delay for target ranging; FM2 contributes only if corresponding FM1 frequencies are present. We show that echoes need only the lowest FM1 broadcast frequencies of 25 to 30 kHz for delay perception. If these frequencies are removed, no delay is perceived. Bats begin echo processing at the lowest frequencies and accumulate perceptual acuity over successively higher frequencies, but they cannot proceed without the low-frequency starting point in their broadcasts. This reveals a solution to pulse-echo ambiguity, a serious problem for radar or sonar. In dense, extended biosonar scenes, bats have to emit sounds rapidly to avoid collisions with near objects. But if a new broadcast is emitted when echoes of the previous broadcast still are arriving, echoes from both broadcasts intermingle, creating ambiguity about which echo corresponds to which broadcast. Frequency hopping by several kilohertz from one broadcast to the next can segregate overlapping narrowband echo streams, but wideband FM echoes ordinarily do not segregate because their spectra still overlap. By starting echo processing at the lowest frequencies in frequency-hopped broadcasts, echoes of the higher hopped broadcast are prevented from being accepted by lower hopped broadcasts, and ambiguity is avoided. The bat-inspired spectrogram correlation and transformation (SCAT) model also begins at the lowest frequencies; echoes that lack them are eliminated from processing of delay and no longer cause ambiguity.

Noise sensitivity of a bio-inspired echolocation model

The Spectrogram Correlation And Transformation (SCAT) receiver is often cited as a model of bat auditory system that is responsible for echolocation signal processing and super resolution in range. We explore the response of a simplified baseband version of SCAT (BSCT) for two targets in white Gaussian noise. Knowing the loss of signal-to-noise ratio (SNR) relative to the matched filter (MF) is important for any practical applications of the algorithm. We show that this loss is 6~dB and increases if a critical noise level is surpassed.

Bio‐inspired processing of radar target echoes

Echolocating bats have evolved the ability to detect, resolve and discriminate targets in highly challenging environments using biological sonar. The way bats process signals in the receiving auditory system is not the same as that of radar and sonar and hence investigating differences and similarities might provide useful lessons to improve synthetic sensors. The Spectrogram Correlation And Transformation (SCAT) receiver is an existing model of the bat auditory system that takes into account the physiology and the neural organisation of bats that emit broadband signals. In this study, the authors present a baseband receiver equivalent to the SCAT that allows an analysis of target echoes at baseband. The baseband SCAT (BSCT) is used to investigate the output of the bat-auditory model for two closely spaced scatterers and to carry out an analysis of range resolution performance and a comparison with the conventional matched filter. Results firstly show that the BSCT provides improved resolution performance. It is then demonstrated that the output of the BSCT can be obtained with an equivalent matched-filter based receiver. The results are verified with a set of laboratory experiments at radio frequencies in a high signal-to-noise ratio.

Separation of MiPS/MaPS spectrogram transformations in biosonar

Echolocating bats and dolphins differ in signals, sound emission, and reception pathways, and important aspects of the acoustic medium. They do, however, receive broadcasts and echoes through similar parallel bandpass filters in the cochlea, which impose integration-times of several hundred microseconds. Echoes consisting of several overlapping reflections from target glints (insect body-parts, fish swim bladders) interfere upon reception to form complex echoes with spectral interference patterns that characterize the target’s shape. Bats transform these patterns into images that depict the glints themselves along the range axis, a process called Spectrogram Correlation and Transformation (SCAT). Experiments with dolphins suggest two different scales for recognizing shape from echo spectra—macro- vs micro-power spectral features for ST (MaPS for short glint separations of <80 µs; MiPS for longer separations >80 µs). We tested this finding in big brown bats trained to distinguish between 2-glint echoes with long, MiPS-like and short MaPS-like spectral features and found that MiPS covers delay separations of about 25-500 µs from the frequency separation of spectral nulls, which have to fit into FM1 for ST to occur. At a deep level, dolphins and bats appear to share a common processing strategy for forming images. [Work supported by ONR.]

Computational model of a bioinspired broadband receiver for sonar clutter reduction.

Echolocating bats have the ability to seamlessly navigate through dense foliage and other obstacles at flight velocity using only the information available in acoustic returns. The spatial resolution required to perform this feature cannot be explained by conventional beamforming and pulse design techniques. We describe a biologically inspired broadband sonar receiver that mimics parallel neural processing by echolocating bats to suppress clutter in complex acoustic environments. These results are incorporated into an improved version of the spectrogram correlation and transformation (SCAT) receiver by replacing the original spectrogram transformation block with a process that translates both harmonic coherence and spectral interference patterns into estimates of echo‐delay separations for closely spaced echo highlights. The model treats simple target echoes as highlight reconstructions, and broader lowpass filtered echoes—characteristic of most off‐axis clutter—as numerous overlapping poorly defined shapes. The monaural SCAT model for range‐only resolution is also expanded to a binaural, 2‐D model for high‐resolution sonar imaging in the range‐azimuth plane. Receiver performance is compared with conventional array processing methods through acoustic simulation and sonar image reconstruction amidst dense clutter. [Work supported by the ONR and internal investments by the NUWC, Division Newport.]

Pattern-matching analysis of fine echo delays by the spectrogram correlation and transformation receiver

Among a few previous attempts to model the outstanding echolocation capability of bats, the work by Saillant et al. [J. Acoust. Soc. Am. 94, 2691-2712 (1993)] is, arguably, one of the most frequently referenced studies in which the predictions of spectrogram correlation and transformation (SCAT) model were compared to the results of relevant behavioral experiments. The SCAT model consists of cochlear, spectrogram correlation and spectrogram transformation blocks, where the latter two processes estimate the overall and the fine time delays between the animal's call and the echoes, given the neural representation of the acoustic signals generated by the cochlear block. This paper first provides a rigorous account of the spectrogram transformation (ST) block. By approximating the neural signals in analytic forms, many aspects of the ST block are explained and discussed in relation to the predictive scope of the model. Furthermore, based on these analytical arguments, the ST block is investigated from a different point of view, interpreted as a pattern-matching process which may operate at the high level of the animal's auditory pathway.

Mechanisms proposed for spectrogram correlation and transformation deconvolution in FM bat sonar

Big brown bats use time/frequency distributions to represent FM biosonar pulses and echoes as a consequence of reception through frequency tuned channels of the inner ear and subsequent processing by similarly tuned neural channels in the auditory pathway. Integration time is 350 μs, yet delay resolution is 2−10 μs, which must be based on detecting changes in the echo spectrum caused by interference between overlapping reflections inside the integration time. However, bats perceive not merely the echo interference spectrum but the numerical value of the delay separation from the spectrum, which requires deconvolution. Because spectrograms are the initial representation, this process is spectrogram correlation and transformation (SCAT). Proposed SCAT deconvolution mechanisms include extraction of echo envelope ripples for time-domain spectrometry, cepstral analysis of echoes, use of coherent or noncoherent reconstruction with basis functions, segmentation of onsets of overlapping replicas at moderate to long time separations, and localization of the occurrence of spectral interference ripples at specific times within dechirped spectrograms. Physiological evidence from single-unit recordings reveals a cepstral-like time–frequency process based on freqlets, both single-unit and multiunit responses reveal which may prove to be time-domain basis functions, and multiunit responses exhibit modulations by onset and envelope ripple. [Work supported by NIH and ONR.]

Transformation of external-ear spectral cues into perceived delays by the big brown bat, Eptesicus fuscus.

The external-ear transfer function for big brown bats (Eptesicus fuscus) contains two prominent notches that vary from 30 to 55 kHz and from 70 to 100 kHz, respectively, as sound-source elevation moves from -40 to +10 degrees. These notches resemble a higher-frequency version of external-ear cues for vertical localization in humans and other mammals. However, they also resemble interference notches created in echoes when reflected sounds overlap at short time separations of 30-50 micros. Psychophysical experiments have shown that bats actually perceive small time separations from interference notches, and here we used the same technique to test whether external-ear notches are recognized as a corresponding time separation, too. The bats' performance reveals the elevation dependence of a time-separation estimate at 25-45 micros in perceived delay. Convergence of target-shape and external-ear cues onto echo spectra creates ambiguity about whether a particular notch relates to the object or to its location, which the bat could resolve by ignoring the presence of notches at external-ear frequencies. Instead, the bat registers the frequencies of notches caused by the external ear along with notches caused by the target's structure and employs spectrogram correlation and transformation (SCAT) to convert them all into a family of delay estimates that includes elevation.

The spectrogram correlation and transformation receiver, revisited.

The spectrogram correlation and transformation (SCAT) receiver has been proposed as a model for the receiver structure used by fm bats. The main contribution of this paper lies in the analysis of which features of the proposed model are responsible for its high accuracy in estimating arrival times of overlapping echoes. Apart from providing an answer to this question, the analysis will also indicate the limitations of the SCAT receiver. In particular, it is shown that the temporal block of the SCAT receiver returns erroneous results for interecho delays < 20 microseconds and for interecho amplitude ratios > 6 dB. It is also shown that the spectral block of the SCAT receiver generates spurious arrival times if more than two overlapping echoes are present. Finally, it is discussed how the characteristic measurement errors predicted by this analysis can be used to further investigate the accuracy of the SCAT receiver as a model of the receiver structure used by fm bats.

Biologically inspired SCAT sonar receiver for 2-D imaging

The SCAT (spectrogram correlation and transformation) model of biosonar in bats represents FM broadcasts and echoes as spectrographs with an integration time of 300–400 μs. The model then mimics signal processing by bats with two parallel computational paths: (1) determination of the delay of echo spectrograms using delay lines and coincidence detectors, and (2) deconvolution of the echo spectrum to determine delay separations smaller than the integration-time window [Saillant et al., J. Acoust. Soc. Am. (1993)] to produce high-resolution echo-delay images. The new, binaural version of the SCAT model receives broadcasts and echoes at two ‘‘ears’’ and uses binaural algorithms to produce range/azimuth sonar images. The model incorporates several stages of binaural interactions that facilitate different functions including determination of azimuth itself, enhancement of target detection, and separation of adaptive imaging from these other functions. The model produces images that locate the principal glints in complex targets from only a few sonar broadcasts without having to sample many different target aspect angles. [Work supported by ONR and NSF.]

Composition of biosonar images for target recognition by echolocating bats

Echolocating bats can recognize flying insects as sonar targets in a variety of different acoustic situations ranging from open spaces to dense clutter. Target classification must depend on perceiving images whose dimensions can tolerate intrusion of additional echoes from other objects, even echoes arriving at about the same time as those from the insect, without disrupting image organization. The big brown bat, Eptesicus fuscus, broadcasts FM sonar sounds in the 15–100 kHz band and perceives the arrival-time of echoes with an accuracy of 10–15 ns and a two-point resolution of 2 μs, which suggests that perception of fine detail on the dimension of echo delay or target range is the basis for reconstructing complex acoustic scenes and recognizing targets that are embedded in these scenes. The directionality of the bat's sonar sound is very broad, making it impossible to isolate echoes from individual targets merely by aiming the head and ears at one object instead of another. Consequently, segregation of targets must depend on isolating their echoes as discrete events along the axis of delay. That is, the bat's images must correspond to impulse responses of target scenes. However, the bat's sonar broadcasts are several milliseconds long, and the integration time of echo reception is about 350 μs, so perception of separate delays for multiple echoes only a few microseconds apart requires deconvolution of spectrally-complex echoes that overlap and interfere with each other within the 350-μs integration time. The bat's auditory system encodes the FM sweeps of transmissions and echoes as half-wave-rectified, magnitude-unsquared spectrograms, and then registers the time that elapses between each frequency in the broadcast and the echo, effectively correlating the spectrograms. The interference patterns generated by overlap of multiple echoes are then used to modify these delay estimates by adding fine details of the delay structure of echoes. This is equivalent to transformation of the spectrograms into the time domain, or deconvolution of echo spectra by spectrogram correlation and transformation (SCAT). However, while deconvolution overcomes integration time, the bat's receiving antennas reverberate for about 100 μs, smearing the echoes upon arrival. The bat overcomes this problem by receiving echoes from different directions than the transmitted sound, which radiates from the mouth. The broad range of antenna reverberations common to the emission and echoes thus cancel out, leaving only narrow elevation-dependent differences, which in fact appear in the bat's images. The SCAT algorithms successfully recreate images comparable to those perceived by the bat and provide for classification of targets from their glint structure in different situations.

A computational model of echo processing and acoustic imaging in frequency- modulated echolocating bats: The spectrogram correlation and transformation receiver

The spectrogram correlation and transformation (SCAT) model of the sonar receiver in the big brown bat (Eptesicus fuscus) consists of a cochlear component for encoding the bat's frequency modulated (FM) sonar transmissions and multiple FM echoes in a spectrogram format, followed by two parallel pathways for processing temporal and spectral information in sonar echoes to reconstruct the absolute range and fine range structure of multiple targets from echo spectrograms. The outputs of computations taking place along these parallel pathways converge to be displayed along a computed image dimension of echo delay or target range. The resulting image depicts the location of various reflecting sources in different targets along the range axis. This series of transforms is equivalent to simultaneous, parallel forward and inverse transforms on sonar echoes, yielding the impulse responses of targets by deconvolution of the spectrograms. The performance of the model accurately reproduces the images perceived by Eptesicus in a variety of behavioral experiments on two-glint resolution in range, echo phase sensitivity, amplitude-latency trading of range estimates, dissociation of time- and frequency-domain image components, and ranging accuracy in noise.