GB2463682A - System and Method for Collection and Analysis of Sound Outdoors - Google Patents

Abstract

A system for the collection of audio from one or more locations comprises remote audio detection devices wirelessly connected to one or more collection points. The system uses ultra-low power circuitry in solar powered detectors to collect audio and transmit it to collection points, either directly or via intermediate detectors. A range of audio processing techniques is used to analyze the collected sound so that the sources can be identified and studied. The system also optionally provides for communication between users of the system and for the playback of audio into the monitored environment. Other features of the detection devices include position and orientation sensors, GPS and cellular phone functionality. The system can be used to monitor bird song.

Description

System and Method for Collection and Analysis of Sound The invention relates to a system and method for collecting and/or listening to bird-song or other sounds especially outdoors over wide areas.

There is increasing need to survey large areas of habitat in order to identify, locate, study and subsequently protect endangered species of birds, animals and insects. It is particularly difficult to determine the number and precise location of breeding pairs of birds. Counting the number of singing males holding territory is one approach to doing so but is very time consuming and labour intensive to do accurately over large areas. A low-cost, at least partially automated system for the collection and analysis of audio over wide areas is therefore required. The recent search for evidence of the continued existence of the Ivory-billed Woodpecker in the southern U.S.A. is a good example of this. The equipment used there and currently available (such as the "Song Meter" from Wildlife Acoustics Inc.) require that researchers visit the detectors to collect stored data. This is not only time consuming but also restricts the positioning of these. For example, it is difficult to justif' placing one at canopy level in a rainforest as the weekly collections of data from it would be impractical.

Audio recording devices for use in the field, such as RememBird� (subject of patent GB240 1240) have been available for over two years now and the continuing drop in size, cost and power requirements of both audio processing (as evidenced by portable MP3 players) and of wireless transmission protocols such as Zigbee� have the potential for cost effective audio collection and analysis over wide areas and long periods of time.

The present invention thus provides for one or more "Detectors" which can be placed or used outdoors. Each of these Contains a microphone, processing and storage means and means of transmitting the audio to one or more "Hubs" or audio collection points.

A hub may reproduce the sound via a loud-speaker andlor store it or feed it into an audio amplifier andlor computer system for further processing.

Significant side benefits of such a system are that it can also act (a) as a means for field scientists to communicate with each other while collecting audio in the field and (b) for audio playback.

The invention is described with reference to the following diagrams: Figure 1 -shows an example of a Detector Figure 2 -shows an example of a Hub.

Each Detector includes a means for detecting sound such as one or more microphones (1). This is typically connected to an amplifier (2) preferably with variable gain. Although a system can be constructed using purely analogue signals and wireless transmission, the preferred embodiment provides for an analogue to digital converter (ADC) (3) and subsequently processing the signal in a digital form.

The digital signal representing the audio signal is thus fed to a digital processor (4) or dedicated audio processing hardware.

This may optionally filter, compress andlor otherwise condition the audio signal according to a wide range of well-known audio processing algorithms. For example, a high pass filter may be used to reduce wind-noise; a notch filter may be used to reduce traffic or airplane noise; a squelch threshold may be applied to silence the output until audio above a threshold level is heard. More sophisticated algorithms may be employed to remove, for example, airplane noise which has very different characteristics from bird-song i.e. it is a sustained, gradually increasing or decreasing sound rather than a rapidly varying sound. Many integrated circuits designed for audio record/replay include graphic equalizer functions. The gain parameters for these may be used to temporarily attenuate e.g. airplane noise without the need for further filtering steps.

More sophisticated categorisation of the received audio may be needed, for example, to identify human speech. Even when detectors are placed far from paths, researchers may visit them from time to time; passers-by may come close to them etc. If detection devices worn or carried by field researchers are also to be used as part of the overall collection network then it is important that human speech is identified as such and does not hamper attempts to identify the target species. One approach to this is to use off-the-shelf speech recognition tools to process the audio. Many such tools provide an indication of the degree of certainty with which they have identified each word or phrase. When human speech is present, even if the speech recognizer does not interpret it correctly, it should (a) output some tentative words and (b) show a recognition likelihood significantly higher than that observed when other, non-human, sounds are present.

Optionally, automatic gain control may be applied to provide best possible resolution of the audio signal. Preferably, the gain factor applied shall be recorded so that the amplitude of the original signal can be calculated subsequently as this may be important in judging where the signal emanated from. Alternatively, the stereo processing capability present in many audio codec devices can be used to record two channels -one of which has had the automatic gain control applied and hence is easy to listen to while the other has had a fixed or much more slowly varying gain applied such that the signal level is a truer representation over the duration of an interesting song or call than would be the case with the channel subjected to rapidly varying gain.

The audio signal and/or any derivatives of it may be stored locally (e.g. in memory, disk or other storage device (5)) -potentially for many hours or even weeks but often only for a few milliseconds before being transmitted via wire or, as in the case of the example in Figure 1, wireless transmission via a radio transmitted circuit (6) and antenna (7) to one or more hubs. It will be appreciated that using techniques and standards such as Zigbee� routing, the signal may "hop" from one detector to another -or via dedicated routing nodes -to one or more hubs at a distance further than the range of an individual transmission.

It will be appreciated that both detector and hub can largely be constructed from many of the components found in mp3 players as typified by Apple�'s iPod�. The very low power consumption demanded by such devices has led to the ready availability of codec and sound processing chips well suited to this application. Note that a single device may be constructed to be both a detector and a hub by the inclusion of both sound detection and sound production means.

Preferably, the circuitry within the Detector is powered by a renewable source such as a solar panel (9) connected so as to trickle charge a rechargeable battery (10).

Preferably said solar panel is protected by a transparent panel in the housing (8).

Preferably the entire detector circuitry is enclosed within a weatherproof housing (8) allowing it to remain operational outdoors in all weathers for many years.

It will be appreciated that a wide range of alternative power sources and stores may be used -including but not limited to, fuel cells, disposable batteries and wind turbines.

Optionally, the device may include a connector through which an additional power source may be provided to supplement its internal capabilities.

Preferably said housing (8) incorporates means for affixing the Detector permanently or semi-permanently to an external surface. This may include one or more instances of holes, lugs, ties, protrusions etc. allowing it to be affixed with nails, screws, straps, cable-ties, VelcroTM, adhesive or other fastenings. In rainforests, for example, where (a) sunlight for power is in short supply at ground level and (b) much of the sound of interest is generated high up in the canopy, the detector may be attached to a small grappling hook or parachute (depending on whether it is to be "fired" up into the canopy or dropped from above into it). In such cases, solar panels will typically be present on more than one face of the device so that it can collect sunlight whichever orientation it comes to rest in. In a further enhancement, flexible solar cells may be attached to or form part of said parachute. The attachment of the parachute to the detector is then at least partially the wiring connecting the solar cells to the detector beneath them. It will be appreciated that for such a scheme to be cost effective, the cost of each detector must be kept to a minimum since a proportion will fail to deploy properly. Note, however, that the radio transmission capability within them does allow them to be located should they fall out of site. The simple inclusion of a buzzer or other means of audio andlor light production can assist in the location of these "lost" devices. Of course, those containing a GPS location device (and able to "see" enough satellites can simply transmit their co-ordinates to the searchers).

The Detector is in communication, via wired or wireless transmission, with one or more "Hub" units. Figure 2 shows a typical Hub consisting of a transmission network capable of receiving information from one or more detectors (e.g. the antenna (16) and wireless sub-system (17)). The incoming data representing the audio or derivatives and/or summaries thereof is processed by processor (20) or dedicated audio hardware. It may be stored locally (15) and/or transmitted on to other computer systems e.g. via Universal Serial Bus circuit (22) or any one of a wide range of other connection mechanisms.

Optionally, the signals representing audio or derivatives thereof may be converted back to an audio signal using Digital to Analogue converter (18) and played via loudspeaker (12) or through external audio connector (21) to headphones, a separate audio amplifier or a computer's line input socket.

It will be appreciated that the hub may also contain one or more input/output devices that a user or external application may use to control the operation of the hub -and indirectly, of the remote detectors. For example, a simple push button switch may be provided to instruct the system to start or stop playing audio through the loudspeaker (12) which has been collected by a remote detector's micrqphone (1). Similarly a display device may be used to show a representation of the audio waveforms of collected audio and/or show the locations of detectors on a map.

It will be appreciated that any part, or indeed the whole of the functionality of the hub may be provided by the components within many home computers or laptops. Thus the hub may be implemented on a simple laptop with wireless connectivity or as a standalone unit capable of running from an internal battery (14) or external power connection (13). This latter type may therefore be connected to a general purpose computer digitally (e.g. via USB (22) and/or wireless network (17)) and/or by analogue audio connection (e.g. line out (21) to line in connector of the computer or hi-fl system.) The Hub and/or Detector may also be implemented on a wide range of computing platforms including but not limited to PC, Apple Mac, micro-controller based, Personal Digital Assistant (PDA), Mobile Phone or dedicated hardware. These hardware platforms may use any of a wide range of operating systems and storage systems.

In its simplest form of operation, the system merely relays audio collected from a microphone (1) on a Detector to the loudspeaker (12) on a hub. This allows one to listen to birds or indeed any other sound from a point where they would otherwise be inaudible or insufficiently audible.

Optionally, the user may start and stop the relaying of sound by pressing a switch or otherwise signalling to the hub.

Optionally, the Detector may process the audio for more efficient use of the network andlor to reduce power or storage requirements or to reduce the proportion of "uninteresting" audio transmitted. For example, a "squelch" or "vox detect" algorithm may be used to eliminate low-level background noise and only transmit audio when a signal is present above a threshold. This threshold may be dynamic so that, for example, gradually increasing wind strength leading to increasing background noise does not trigger transmission. The threshold and/or other rules for categorising audio may also be passed to the detector so as to refine over time the quantity and/or quality of audio transmitted.

By classifying the incoming audio into "interesting" versus "not interesting" -which may be as simple as a traditional "vox detect" algorithm (e.g. energy level greater than a threshold value for a threshold time) then each detector need only transmit for part of the time. Compressing the signal and sending a burst in less time than it took to record it also helps to reduce the duty cycle of transmission -reducing power drain and allowing others to use the bandwidth of the transmission channel.

More sophisticated analysis -into multiple levels of "interestingness" or simply differentiated sources -may determine how the audio is stored (if at all). At one extreme, the audio is determined to be "silence" and is discarded. At the other, it is kept uncompressed at the original sample size i.e. as unadulterated as possible.

Intermediate levels of "interestingness" may result in one or more intermediate compression levels (e.g. mp3 64kbps through 320kbps).

In one embodiment, the presence of an "interesting" signal at one Detector may be communicated to at least one and ideally all neighbouring Detectors. By instructing the neighbouring detectors to record audio at this time -even though the level of signal some or all of them picked up may not have itself been deemed sufficiently "interesting" -it is then possible to identify any common component. Calculating the correlation between the signals at different time intervals will identify the common component -and its relative amplitude and time shift -from that received at the other detectors. Thus the location of the source of the sound can be estimated. Repeating the calculation over a sliding time window allows the trajectory to be estimated too.

Optionally, Detectors in the flight-path may also be activated in advance of the anticipated traversal of the source into their locality.

The Doppler effect may also be used to compare the velocity component of the trajectory towards or away from each detector.

Sounds may be categorised according to key attributes of their "sonogram" (a graph of intensity of sound at each frequency over time).

Sounds may be further categorised according to whether they are from stationary or moving objects. For example, certain insects make sound by rubbing their legs together. They can only do this when at rest, not when in flight. Hence having identified a sustained buzzing as typical of, say, a cicada species known to inhabit the area, any fluctuations in volume of a single signal can be inferred to be due to increases or decreases in volume at the source rather than a source moving closer.

(Species of insect known to sing while perched on the back of a moving animal may cause confusion).

As many insect sounds are relatively simple, sustained signals, it is possible to identify the presence of multiple signals by the presence of a beat frequency since no two individuals are likely to have precisely the same frequency.

Step changes in amplitude are also a good indication of multiple sources. A single cicada will start and stop suddenly but within that burst of song the volume are changes should be gradual hence step changes in volume are another indication of a second source. When such signals are recorded at multiple locations, the relative volume before and after the step can be used to infer position location of the new source and, if the signal is of the same type, the approximate relative distance of the two sources.

Optionally, the above analysis and compression may take place immediately or after some time period. For example, a regularly repeating sound -such as a bird calling -may only be detectable with certainty after several cycles of song. At this point, the audio becomes "very interesting" and may be kept "raw" rather than compressed.

It will be appreciated that any division of processing between hub and detector is possible with the optimum split being determined by the cost and power of the processing resources in each.

Optionally, a detector may also include Digital to Analogue Converter, amplifier and loudspeaker -allowing it to output sound as well as to receive it. This may be of use in calling birds out by mimicking their calls previously recorded on this or any other system.

Optionally, the detector andlor the hub may record audio for future playback.

Preferably, each Detector has a unique identifier (e.g. the Media Access Control (MAC) layer address of its RF transmitter). This allows the base station to identify which detector each signal emanated from.

Optionally the detector and/or the hub may include a real-time clock so that recordings can be time-stamped and this information can be stored in a database allowing the user to retrieve recordings from a specific date/time and detector.

Alternatively, the Hub may transmit a clock or timing signal to the Detectors, allowing them to synchronize with it and hence with each other. This precise synchronization allows the source of sounds relative to the detectors to be determined by measuring the time difference between them arriving at different Detectors and triangulating the resulting differences in distance. Such techniques are well known in the art.

Preferably, each detector also incorporates position and/or motion detection components eg. Global Positioning System. It may then transmit its location andlor velocity to the huh, allowing the source of the audio signal to be noted and marked on a map. Alternatively, the location of each detector may be marked on a map or otherwise entered into the software running on the Hub or an attached computer.

Optionally a detector may have multiple microphones. These may be omni-directional or directional. Knowing the location and orientation of these microphones, the individual audio signals can be compared for strength and phase shift to determine the location and/or trajectory of the source of the sound. A particular arrangement of microphones and associated housings (e.g. parabolic reflectors around four at ninety degrees to each other) may be tested in an anechoic chamber to determine the actual response of each microphone to a sound emanating from a given position. This calibration can then be used to determine more accurately the location of sounds subsequently heard in a real deployment.

Similarly, a hub may use the known locations and/or orientations of one or more detectors to estimate the location and trajectory of the source of sounds.

Note that, being outdoors, the microphones will also be exposed to the weather and hence are preferably sheltered from the rain and snow. However, a simple cover will result in considerable noise being generated as the raindrops fall on it. The upper surfaces of the detector, especially around the microphones are therefore preferably covered with an absorbent material into which the raindrop dissipates rather than impacts heavily. This allows the unit to continue working to at least some degree during rain-showers.

In the same way that speech recognition software can be used to identify words and speaker profiles can be compared to identify speakers, so the audio recorded can be compared against recordings of known bird or animal species to identify (albeit not perfectly) the species calling. Such algorithms are typically very processor intensive and hence the hub may offload some or all of this processing to a connected computer or -in the extreme -to a network of attached computers such as used by distributed computing projects like SETI (Search for Extra-Terrestrial Intelligence).

Optionally, one user's system may contribute audio and/or derivatives thereof e.g. statistics about the audio to large scale monitoring programs. The audio may be analysed and/or annotated manually by the user or remotely by others e.g. those more expert in bird-song.

Optionally, the hub allows the user to play audio from previous recordings -whether selected items from the detectors' earlier inputs or recordings provided by others e.g. a reference library of birdsong. In this way the user may be trained to recognise specific species prior to them being heard for real. Optionally, a spoken announcement of the species and/or instruction on how to identify the particular sound is played.

Optionally the selection of sounds played is randomised with announcements following the song. This provides a useful training exercise.

Optionally, the system may be configured such that live sound from the detectors overrides any recording that may be playing. Preferably an audible indication is given of the switch from recorded to live audio and back again.

In addition to being of use to those wishing to hear birds which would otherwise be out of their current audio range, it will be appreciated that the system has value to anyone attempting to survey an area or locate a specific creature. Hence the Hub may be used in the field. Given sufficient storage capacity, one hub may be left unattended for long periods of time collecting the audio from a set of much cheaper Detectors. In this way, full audio coverage of a wide area can be achieved more cost effectively than with current record and store devices, each of which would require its own bulk storage. Furthermore, the real-time or near real-time transmission of audio to the Hub makes the system more valuable to the field-worker since he/she can be alerted to a sound immediately rather than having to wait till the data is analysed later as is the case with standalone recording devices.

In a number of remote areas, cellular phone coverage is provided yet the bandwidth available to rural base stations is never or very rarely fully occupied. These areas are also often difficult to study as access is difficult and the manpower is not available to provide the degree of coverage needed to determine populations of animals.

The present invention lends itself to deployment of large numbers of the Detectors across a wide area. Using existing cellular base stations to receive and store the audio over otherwise unused channels would allow monitoring of several square kilometres around a base station at very marginal cost. Adding a control mechanism by which the hub can monitor or be advised of the current level of genuine cellular use, it may instruct some or all of the detectors to reduce or stop their transmissions such that genuine cellular telephone users are never prevented from getting a channel because of the loading due to the detectors.

In a further variant of the system, a collection of detectors may be of use to one or more observers in the area. If each observer carries a Detector and/or "hub" e.g. in the form of a device attached to their binoculars (such as a wireless equipped "RememBird�") the detectors may be configured to establish contact with any hub -not just the one to which they have been "bonded" as is normal practice with e.g. blue-tooth earpieces. In this way, an observer moving into range of a detector may be provided with early warning of sound emanating from the area of that detector or indeed any beyond it which it is able to relay messages from. This is of immense help in tracking down wildlife in dense jungle e.g. where birds flock together but flocks may be far apart and moving rapidly.

Many areas requiring detailed study are sparsely populated and remote -hence do not receive coverage from cellular telephone networks. Therefore, in a further enhancement to the system, communication between the observers may be provided by the Detectors positioned andlor carried within the area. This allows users of the mobile hubs to stay in touch verbally and/or via text or other messaging while they are in the vicinity of the detectors. Use of the synchronized beaconing of the Zigbee� protocol allows nodes to respond rapidly when needed without being on and using excessive power at all times.

A further benefit of integrating the audio collection system with the communication system is that the spoken communications between field staff is known to be such and can therefore be more easily categorised as such by the system. This avoids having to review audio recordings that turn out to be the staff talking to each other while present in the study area.

Further, by having each such portable detector record throughout the duration of the user/wearer being in the study area, the audio "pollution" caused by his/her movements and speech can be accurately recorded. This can then be used as a further input to the overall audio analysis -helping to eliminate false positive categorisations from the detectors when in fact these are only picking up noise made by the staff in the area.

By building the detection and/or collection mechanisms into the RememBird� device, users are able to conimunicate with each other while watching the wildlife since RememBird is designed to be used without looking at the controls and even while looking through binoculars -to which it can be attached.

The further inclusion of a Global Positioning Module (GPS) either directly into the RememBird� or by communication with one (e.g. via blue-tooth to a GPS within a cellular phone also being carrie) allows the device's position to be tracked over time.

The use of a wired or wireless (e.g. blue-tooth) earphone ensures that sound pollution is minimized. However, it is uncomfortable to wear such an earphone at all times. A feature of the communications system and of RememBird� in general is that when audio needs to be generated it is played at a very low volume. This necessitates placing the device close to ones head. To alert the user to this, an audible signal is used but rather than a standard "beep" or warning tone, the system may preferably use a natural sound e.g. a non-threatening insect or bird contact call that will (a) not alarm wildlife that hears it and (b) is itself easily distinguishable from other sounds and hence can easily be categorised correctly as a system generated sound should it be heard by any of the collection devices. Optionally the time and location at which it was generated can also be stored for better discrimination later.

Field scientists also make use of "playback" to attract species to them. This is especially prevalent when banding (U.S. term) or ringing (European term) but is also widely and increasingly used simply to observe birds.

It will be appreciated that a wireless device capable of receiving and playing audio on command from a wirelessly connected device has a number of benefits over the traditional wired loudspeaker approach. The user can more easily retire behind cover without having to lay out a wire; the wire adds weight to the user's load and is often too short for effective use.

As more and more birders carry with them the ability to lure out birds using playback, so noise pollution due to this is increasing. This can distort counts and lead to records of birds out of range that were never actually there. Someone out of site playing a distinctive species could easily be misinterpreted and reported as a rarity! There is therefore a further benefit of the proposed system. By integrating playback control into the system -and noting the time, location and what audio was played back automatically, this can be taken into account when analyzing the recordings made by all detectors in the area.

It will be appreciated that the output of the system may be displayed using mapping techniques such as GoogleMaps� or GoogleEarth to show the location and times of recordings thought to be of interest. Manual, automated or semi-automated categorisation of the recordings allows representations of the types of sound heard, their locations, frequency etc.

Claims (38)

1. CLAIMS1. A system for the collection of audio from one of more locations characterised in that the audio detection devices are in wireless communication with one or more collection points.
2. 2. A system of Claim I in which said wireless transmission includes forwarding of messages between detection and/or collection devices.
3. 3. A system of Claim 2 in which said wireless transmission is achieved using the IEEE 802.15.4 ("Zigbee�") protocol.
4. 4. A system of Claim 1 in which said devices are at least partially powered by one or more renewable sources including but not limited to solar power and wind power.
5. 5. A system of Claim I in which said detection and/or collection points are synchronized to a received time signal
6. 6. A system of Claim 1 in which said detection andlor collection points include position and/or orientation sensors.
7. 7. A system of Claim 6 in which said position sensor is a global positioning system.
8. 8. A system of Claim 1 in which multiple microphones are deployed at one or more detection points.
9. 9. A system of Claim 1 in which the source of a sound is estimated by comparing the received signal from one or more audio detection devices.
10. 10. A system of Claim 8 or 9 in which the difference in distance of the sound from each audio detection point is estimated by comparing the delay in reception at each detection point.
11. 11. A system of Claim 8 or 9 in which the difference in distance of the sound from each audio detection point is estimated by comparing the amplitude received at each detection point.
12. 12. A system of claim I in which the trajectory of a sound source is estimated by comparing the frequency of the received signal over time.
13. 13. A system of Claim 8 or 9 in which the direction of the source of a sound is estimated by processing the received audio from multiple microphones.
14. 14. A system of Claim 1 in which the frequency change over time is used to estimate the trajectory of the sound source, ii
15. 15. A system of any combination of Claims 8 through 12 in which more than one of these factors are combined to provide a better estimate of location andlor trajectory.
16. 16. A system of Claim 1 in which the received audio is subject to automatic gain control.
17. 17. A system of Claim I in which the received audio is categorised into two or more categories.
18. 18. A system of Claim 17 in which said categorisation is performed by comparing the received audio against a threshold volume, amplitude or energy level.
19. 19. A system of Claim 18 in which said threshold further includes time constraints or trends.
20. 20. A system of Claim 17 in which said categorisation is performed by comparing the intensity of sound at different frequencies over time against known and/or unknown patterns.
21. 21. A system of Claim 17 in which the subsequent actions of the system are determined by the categorisation of said audio.
22. 22. A system of Claim 21 in which said audio is only stored and/or forwarded when it is deemed to be in one or more specific categories.
23. 23. A system of Claim 17 in which the changes in amplitude of a signal over time are categorised into gradual versus step changes according to the rate of change of volume or energy in any or all frequency bands in a sliding time window.
24. 24. A system of Claim 1 in which the audio is subject to band-pass filtering.
25. 25. A system of Claim 17 in which at least one of said categories is that of traffic noise.
26. 26. A system of Claim 17 in which at least one of said categories is that of airplane noise.
27. 27. A system of Claim 17 in which at least one of said categories is that of human speech.
28. 28. A system of Claim 27 in which speech recognition software is used to determine the categorisation.
29. 29. A system of Claim 28 in which the accuracy measure output from the speech recognition software is used in the decision to categorise the audio as human speech.
30. 30. A system of Claim 12 in which a second recording is made of the signal, to which different or no automatic gain control parameters are applied.
31. 31. A system of Claim 13 in which only sound in one or more specific categories is proactively or routinely transmitted to the collection point(s).
32. 32. A system of Claim 24 in which the sound not transmitted to the collection points is retained for a period during which any or all of it may be transmitted on request to the collection point(s).
33. 33. A system of Claim 1 in which the detection devices are only active on request from the collection point.
34. 34. A system of Claim 26 in which the detection devices activate periodically to communicate with one or more collection points and hence determine whether or not to remain active.
35. 35. A system of Claim 1 in which the functionality of detector and/or collection point is provided by a cellular phone or derivative thereof.
36. 36. A system of Claim 1 in which transmission occurs over a pre-existing wireless communication link.
37. 37. A system of Claim 29 in which the use of bandwidth is controlled so as to use only available bandwidth on said pre-existing communication link.
38. 38. A system of Claim I in which the detection and/or collection devices arecarried by users in the field39. A system of Claim 38 in which detection and/or collection devices automatically establish communication with said portable devices as they come into range.40. A system of Claim 1 in which the detection and/or collection devices are further used to facilitate communication between those using said devices.41. A system of Claim 40 in which the time and location of said communications between those using said devices is used to further categorise audio received by the devices used and/or neighbouring devices that may have recorded said communications.42. A system of Claim 1 in which said detection and/or collection devices are further used to play audio into the environment within which audio is being monitored by said devices.43. A system of Claim 42 in which the time and/or location and/or volume of said playing of audio is noted.44. A system of Claim 43 in which said play details are used in analysing the audio recorded near that point at that time.45. A system of claim 38 in which said carried devices record audio for at least a portion of the time.46. A system of Claim 45 in which said recordings from said carried detectors are used to analyse the recordings made by the static detectors nearby.47. A system of Claim 38 in which the location of the device is tracked and noted.48. A system of Claim 38 in which audible alerts are indicated by means of one or more sounds derived from those naturally occurring in the wild.49. A system of Claim 1 in which said detectors are attached to parachutes.50. A system of Claim 1 in which said detectors are attached to grappling devices 51. A system of Claim I in which said detectors include means for production of audible and/or visual indication