Hearing in the Military Software Syrinx vibration dynamics Effect of air cavities in avian hearing X-ray recordings of singing birds Psychoacoustic testing 3D reconstruction of songbird syrinx Magnetic oriantation

Dissertation summary

Acoustic communication in the hooded crow (Corvus corone cornix)

The thesis covers acoustic communication with emphasis on long distance territorial signalling in the hooded crow. It includes aspects of the sender-receiver communication chain from (a) sound production in the sender, (b) over call propagation from sender to receiver over open field habitats, (c) to sensation and perception in the receiver.

Hooded crows live as pairs in territories sized up to 500 m2 (Coombs 1978). They advertise their territory acoustically by a series of 3-5 loud, rhythmically repeated, monotone calls (Coombs 1978; Cramp & Perrins 1994; Goodwin 1986; Madge & Burn 1994). For a sender territorial signals are ideally aimed at individuals outside the territorial boundary. Signalling is adaptive because it saves the owner time and energy to escalated aggressive behaviours (Bradbury & Vehrencamp 1998). For receivers like “floaters” (non-breeding birds) it should be adaptive to obtain information on which areas are occupied as territories and to have a choice of involving themselves in conflict or not. For receivers like territorial owners it should be adaptive to receive information about whether or not it is the well known neighbour(s) with whom the territory borders already have been settled that are present or whether neighbouring territories are open for incoming strangers. Furthermore, hooded crows are their own worst enemies, and it is presumably adaptive to detect groups of non-breeding predatory individuals outside the territory, as they present a threat to eggs and nestlings (Coombs 1978).

Together with the fact that hooded crows have very large territories this makes hooded crows a potential interesting case of long distance communication in birds. The objective of the thesis was then to investigate the hooded crow acoustic territorial signalling and possible adaptations to long distance communication by focussing primarily on detection issues.

Sound production

Crows are well known for their typical harsh quality calls. This comes about by the “pulsed nature” of the calls. On the basis of spectrographic observations together with envelope inspections these pulses seems to consist of amplitude and/or frequency modulations with modulation frequencies in the area of 75 – 100 Hz. Somewhat comparable to vibrato in the human voice (Laver 1994; Stevens 1998) only greatly exaggerated. To verify this and to make a basis for further research into the function of the signal design, we studied the sound production mechanism in situ by high-speed video recordings of crow phonations. Phonations were “forced” by gently pressing the abdomen and thereby forcing an air flow through the syrinx. A comparison between the fine structure of the forced phonations and spontaneous phonations within the same individuals revealed that the basic call structures were very comparable.

Crows, like other song birds, have a syrinx with two sound generating valves (Suthers & Zollinger 2004; Goller & Larsen 1997). Our results showed that the sound generating event was the opening of a syringeal valve. This generated a very brief (<1ms) high amplitude pressure pulse or "sound spike". These sound spikes were organized into short pulse trains of two to three sound spikes each giving the pulsed nature and harsh quality known from natural calls. Each single pulse train was produced in symphony by the two syrinx sides in patterns like "left-right, left-right ...”, or "left-right-left, left-right-left ...”. From these results and together with comparison to spontaneous calls we conclude that the typical crow call consists of a long series of brief high amplitude pressure pulses, or sound spikes, generally organized in pulses and that the call can be amplitude and/or frequency modulated depending on the temporal organization of the pulse trains and the sound spikes within them.

These results also provide the first high-speed direct filming of a song bird internal syrinx dynamics in situ not being under-sampled.

On the sound production part, we also did research into the signal composition, spectral profile, source level and directionality of close range (<1.5 m) recorded hooded crow calls. Calls were provoked by handling the birds. Much of the data still await further analysis.

Sound propagation

To investigate the influence of transmission on crow call composition and to assess over what distances crows are able to communicate in nature we performed transmission experiments. This was done over 40, 80, 160 and 320 m in open field habitats. We found in general, that besides geometrical spreading and frequency dependent attenuation, the ground effect is an important factor in determining how well different frequency ranges transmit. With sender receiver heights of 2.8 m (two habitats) above the ground these effects generated a so called "sound window" (Morton 1975; Marten & Marler 1977) of especially well transmitted frequencies between approximately 500 Hz and 2 kHz. Theoretical considerations and modelling (Attenborough pers. comm.) suggests that this “window” would probably exist at many other sender receiver heights relevant to crow communication.

Data from the close range (< 1.5 m) recordings of provoked crow calls (see above) shows a mean peak frequency of 1.6 ± 0.1 kHz (mean ± SD; N=4) and a 10 dB bandwidth of 520 ± 175 Hz (mean ± SD; N=4). This places the peak and the 10 dB bandwidth of crow calls in the upper half of the “sound window”. Hooded crows seems thus in a position where they can exploit this "natural amplifier". The ground effects have been largely neglected in the avian literature on transmission for sender receiver heights above approximately 1 m (Wiley & Richards 1982). We argue that the effect is not dependent on height and still very important well above ground.

In relation to the results on sound transmission and results on hooded crow hearing abilities (see below) we estimated that maximal communication distances (active spaces) of crows probably lie between 500 - 700 m within a measured range of environmental noise levels. This is estimated based on simple detection, and factors like for instance binaural masking release, co-modulation masking release and co-modulation detection differences would theoretically increase the active space. Inter territory communication thus seems realistic.


It is also important to know something about the hearing abilities of the receiving crows to better understand the communication chain. To do that I have investigated two basic measures: hearing sensitivity in the quiet and critical ratios. Additionally I have done research into co-modulation detection differences which might be relevant for crow communication due to the special design of crow calls. Inspired by the modulations of crow calls, I also did research on the influence of amplitude and frequency modulations on the detectability of synthetic harmonic complexes using the well studied zebra finches as models.

For this research I build a complete psychoacoustic setup from scratch completely by myself. This consisted of two custom build sound proof booths (to be able to run two birds simultaneously). I Used Tucker-Davis System III components for stimuli generation etc. and own software programming (Delphi/Object Pascal) for running and controlling the two setups simultaneously. The two setups were built large enough to contain and test human subjects such that direct comparisons were possible.

Hearing sensitivity in the quiet

We found by direct comparison that the hearing sensitivity of the two tested crows was in magnitude and shape very much like the human one below approximately 6 kHz, while the hearing sensitivity cuts off abruptly up to 8 kHz as seen in many other birds (Dooling, Lohr, et al. 2000). Hooded crows have a 10-30 dB greater sensitivity at lower frequencies (<1.5 kHz) compared to other passerines. This is expected simply from their greater size. Their greater size also allows them to produce lower frequency sounds which transmit well in the environment, and altogether they are much better suited for long distance communication than other passerines.

Critical Ratios (CR)

Measuring the crows’ hearing in white noise (SL = 0 dB/Hz) we found an interesting and atypical distribution of their CRs. Instead of the more typical approximately 3 dB increase per octave of most other birds and mammals, the crow CR was roughly constant at 27 dB and showed an abrupt 5-6 dB decrease (better noise hearing) within a range above 500 Hz and below 2 kHz. This range of increased hearing ability in noise lies approximately 3 dB below the average bird CR in the same range (Dooling, Lohr, et al. 2000). It also contains the peak and the 10 dB bandwidth of crow calls (se above). Thus crows seem specially adapted for detecting crow calls in the noisy natural environment. Interestingly this range of low CRs is also matching the “sound window” found in our sound propagation study. Altogether this makes crows well suited for long distance communication.

Comodulation detection differences (CDD)

CDD is the difference in detectability of a narrowband noise signal masked by two flanking narrow band noises when the flanking bands are envelope modulated coherently with - or in the same way as - the signal compared to when it's not (McFadden 1987). The signal is usually easier to detect when it is modulated differently from the two flanking bands (McFadden 1987). In humans this can produce a difference in detectability of up to approximately 10 dB (Moore & Borrill 2002; Borrill & Moore 2002). Since 1) crow signals are somewhat narrow banded after long distance transmission, 2) their signals are usually heavily modulated in its envelope (see above) and 3) the environmental noise flanking (and covering) the signal is differently modulated due to air turbulence, the crow communication situation for the receiving crow is somewhat comparable to the CDD paradigm. Therefore I found it relevant and interesting to test if crows also show a CDD. My result showed that the two test crows and a human subject tested showed a CDD of approximately 10 dB when tested in the condition where humans usually show the greatest CDD (signal at 1.5 kHz, and flaking bands at 900 Hz and 2.1 kHz). This suggests that crows show a CDD effect similar to that of humans. It is possible that this improves detection of crow calls in natural communication.

Detection of modulated harmonic complexes in zebra finches

Phenomena like for instance detection of harmonic complexes (Buus 1985) but perhaps especially co-modulation masking release (CMR) in both birds and humans point to some degree of integration of information between auditory filters (see e.g. Klump & Langemann 1995; Verhey, Pressnitzer, et al. 2003; Hofer & Klump 2003). In CMR it is seemingly the coherent envelope modulations across maskers that enable correlation of information in different auditory filters and allow them to integrate across filters. Based upon this I speculated that the modulations coherent across frequencies of crow calls might allow auditory filters to integrate signal information and thus facilitate detection of the calls.

During a stay in Professor Robert J. Dooling's lab I used zebra finches as an initial model, which also has broad band harmonic calls, with the goal of later tests on crows. The birds were tested with harmonic complexes with F0 of 500 Hz and with all the harmonics up to 5 kHz which is comparable to natural zebra finch calls (see e.g. Lohr, Wright, et al. 2003). The whole complex was either (a) un-modulated, (b) amplitude modulated (modulation frequency = 75 Hz; modulation index = 0.75), (c) frequency modulated (modulation frequency = 75 Hz; frequency deviation = 100 Hz * harmonic number such that the fundamental would be modulated by ±50 Hz, the second harmonic by ±100 Hz, etc., as would be the case if the fundamental of a harmonic sound would modulate in frequency, or finally (d) both amplitude and frequency modulated. The modulations were approximated those measured in hooded crow calls. The results pointed to a possible effect of frequency modulations on detectability, but apparently none of amplitude modulations. However, the design turned out to be inappropriate and the results are only weakly indicative. A re-designed experimental series, preferably on crows, is necessary to confirm such an effect.


The sound production studies indicated that the assumption of a “vibrato like” amplitude- and/or frequency modulation design of hooded crow calls seems reasonable. Crows shows a CDD effect which means they could potentially use the modulated call design in call detection or possibly in perceptual grouping like in humans. Furthermore, although only working as a pilot experiment, the investigations on zebra finches weakly indicated that the frequency modulations might increase the detectability of a harmonic complex. We found that crows in general hear rather well compared to other passerines in the best transmitting low frequencies, probably simply due to their greater size, and that they might show special adaptations to detect crow calls in noise indicated by the results on critical ratios. The sound propagation study indicated a range of best sound transmission between roughly 500 Hz and 2 kHz in crow habitats. Peak energy of long distance transmitted crow calls (up to 320) lies within the range 1.2 to 1.8 kHz. The special sensitive range of hearing in noise (CR) lies in the range between 500 Hz and 2 kHz. The overlapping of these ranges is probably not coincidental and provides altogether the hooded crows with optimal transmitting and receiving conditions.

All the different experiments in my Ph.D. thesis has put light on important and basic elements of the "territorial communication chain" in hooded crows and make a firm basis for further investigations. They show that hooded crows are well suited for long distance communication and that they may also to some degree be specially adapted for it.

Download thesis (pdf)



Borrill, S. J. and Moore, B. C. J. (2002). Evidence that co-modulation detection differences depend on within-channel mechanisms. J. Acoust. Soc. Am. 111, 309-319.

Bradbury, J. W. and Vehrencamp, S. L. (1998). 'Principles of Animal Communication.' (Sinauer Associates, Inc: Sunderland, Massachusetts.)

Coombs, C. J. F. (1978). 'The Crows: A Study of the Corvids of Europe.' (B. T. Batsford Ltd.: London.)

Cramp, S. and Perrins, C. M. (1994). 'Handbook of the Birds of Europe, the Middle East, and North Africa: The Birds of the Western Palaearctic - Volume VIII: Crows to Finches.' (Oxford University Press: Oxford.)

Dooling, R. J., Lohr, B., and Dent, M. L. (2000). Hearing in birds and reptiles. In 'Comparative hearing: Birds and Reptiles. Vol. 13.' (Eds. R. J. Dooling, R. R. Fay, and A. N. Popper.) pp. 309-359. (Springer: New York.)

Goller, F. and Larsen, O. N. (1997). A new mechanism of sound generation in songbirds. PNAS 94, 14787-14791.

Goodwin, D. (1986). 'Crows of the World.' (The British Museum of Natural History: London, UK.)

Hofer, S. B. and Klump, G. M. (2003). Within- and Across-Channel Processing in Auditory Masking: A Physiological Study in the Songbird Forebrain. J. Neurosci. 23, 5732.

Klump, G. M. and Langemann, U. (1995). Co-modulation masking release in a songbird. Hear. Res. 87, 157-164.

Laver, J. (1994). 'Principles of Phonetics.' (Cambridge University Press: Cambridge.)

Lohr, B., Wright, T. F., and Dooling, R. J. (2003). Detection and discrimination of natural calls in masking noise by birds; estimating the active space of a signal. Anim. Behav. 65, 763-777.

Madge, S. and Burn, H. (1994). 'Crows and Jays: A Guide to the Crows, Jays and Magpies of the World.' (A & C Black: London.)

Marten, K. and Marler, P. (1977). Sound-transmission and its significance for animal vocalization. 1. Temperate habitats. Behav. Ecol. Sociobiol. 2, 271-290.

McFadden, D. (1987). Co-modulation detection differences using noise-band signals. J. Acoust. Soc. Am. 81, 1519-1527.

Moore, B. C. J. and Borrill, S. J. (2002). Tests of a within-channel account of co-modulation detection differences. J. Acoust. Soc. Am. 112, 2099-2109.

Morton, E. S. (1975). Ecological sources of selection on avian sounds. American Naturalist 109, 17-34.

Stevens, K. N. (1998). 'Acoustic Phonetics.' (The MIT Press: Cambridge, Massachusetts.)

Suthers, R. A. and Zollinger, S. A. (2004). Producing song - The vocal apparatus. Behavioral Neurobiology of Birdsong 1016, 109-129.

Verhey, J. L., Pressnitzer, D., and Winter, I. M. (2003). The psychophysics and physiology of co-modulation masking release. Experimental Brain Research 153, 405-417.

Wiley, R. H. and Richards, D. G. (1982). Adaptations for acoustic communication in birds: sound transmission and signal detection. In 'Acoustic communication in birds: Communication and behavior.' (Ed D. E. Kroodsma.) pp. 131-181. (Academic Press: New York.)