A site dedicated to the review and analysis of potential sasquatch vocalizations, Sasquatch Bioacoustic combines techniques from the domains of intelligence collection, audio analysis and bioacoustic studies to examine the evidence of sasquatch through their vocalizations. ~Monongahela

Monday, December 14, 2009

BFRO Olympic Peninsula 2009 Vocalizations

The BFRO held a well attended expedition to the Olympic Peninsula in Washington. The reports and discussion from that expedition can be seen here:

In that report there are a number of audio recordings that include whoops, grunts, various vocalizations and other audible noises. I took some time to analyze those recordings and made the following observations....

I gave this recording ( ) a good run through with head phones on, the volume up, no filters applied, and the spectrogram rolling. There's more to this recording than would meet the casual observer's ear. I found there are basically two speakers. Speaker 1 is in the foreground causing the breathing sounds (inhales and exhales) and might actually be more than one person, but it's hard to differentiate simple breaths of air that lack vocal characteristics.

The second speaker is in the middle distant background (as indicated by signal strength) and vocalizes in what I'll characterize as a "mewling" voice, or as others have described previously, a child's voice. I counted the number of syllables uttered by the second speaker and came up with a total of nine in the short 14 second clip (before it begins to loop). The first two are before the first speaker's first exhale (or inhale, hard to tell which). They are very faint and the second of the two is somewhat stepped on by the start of the first speaker's breath. I've labeled those utterances #1 and #2 on the annotated spectrogram at:

After the first speaker's first breath completes the second speaker emits three more sequential syllables. The #3 syllable is higher than the earlier two vocalizations, then the #4 syllable is lower than #1 or #2. Finally the #5 syllable is higher, but still lower than the earliest utterances on the recording.

Speaker 1 breathes again, including a loud exhale through the nose. Then speaker 2 utters syllables 6 & 7, more clearly than any of the earlier syllables. And finally syllables 8 and 9. These are the utterances that most clearly represent the words people seem to hear in the discussion above.

And now for a word about words...
I believe that audio analysis can be fraught with the same challenges that occur in image analysis. Specifically, the human mind when presented with information, either auditory or visual, will attempt to resolve that information into a pattern with which it is familiar or comfortable. And in the case of audio analysis, the harder you listen to something, the more likely your brain is to morph the signal in your ears into something you "think" you're hearing.

So just be cautious when you put the head phones on and don't "over-listen" what you're hearing. Sometimes the monkey's not saying Mommy.

That said, the possibility does exist that this could be an example of a clever mimicry. I posit this based upon the accumulating anecdotal evidence surrounding BF vocalizations mimicking other animals, reports of human voice mimicry, and the research of Mr. Scott Nelson which indicates distinct phonemes evident to his analysis of the Sierra Sounds CDs.

And if I were to allow speculation to run wild, I could conceive that this is a sasquatch attempting to lure the humans out of their tent by teasing them with the call of a human child, in the woods, at 2 a.m. But I'm not ready to go there, just yet.

It is also interesting to note that during the 2008 BFRO expedition to this same area, a voice that sounded like an infant was recorded under similar circumstances.


I took the day off work today to nurse a cold, and used that down time to go through the rest of these recordings in detail. I was surprised to find a good deal more in those recordings than is initially evident to the casual listener. Anyone who's really interested in these type of recordings should get themselves a good pair of head phones and spend some time listening slowly and closely. There's some cool stuff in there, and some of it's kinda weird, too.

Anyway, here's a summary with links to recordings and associated spectrograms. Note that I went though the full collection of recordings posted at:
There are more recordings there than are linked above.

 The first recording to look at is the "Departing Knock" recording:
The associated spectrogram is here:
The surprise from this file is the faint whoop just before the wood knock, and the even fainter screech (possibly an owl) after the wood knock. You can see the whoop in the spectrogram, but it takes a real eye to spot the screech.

Next is "Distant Knock and Vocal" at:
Spectrogram at:
The vocalization in this recording is a four tone howl, ascending from the first tone to the last.

"Distant Whoop" is at:
Spectrogram at:
A couple surprises here. There are very faint vocals just before and just after the whoop, but you probably won't hear them without head phones. They barely made a mark on the spectrogram. Interestingly, this whoop sounds very similar to the whoops recently recorded by RCM944 outside Richmond Virginia and posted/discussed in this BFRO Forum thread:

Then we have the second "Momma" recording from this collection here:
and the spectrogram is at:
The surprise here was the appearance of three faint mmmm-like hums. The third hum steps on the second "momma" vocalization, indicating these are probably two different speakers.

This next one is a strange whoop recording:
Spectrogram at:
I say strange because it sounds more like a whistle than a vocalization, to me. There's also a deep breath of air in seeming sync with the whistle-whoop which might be from the same source as the whoop. It sounds very close to the microphone, and I would have been very startled to hear a call this nearby.

Next is a dog bark, pair of whoops, and a wood knock:
spectrogram at:
Nothing extra about this one, but it is a great recording. And if you're really into signal recognition, the first whoop demonstrates "reversed" comma shape that I've seen in a growing number of whoop calls. The descending whoop at the end is a little unusual, not a call I've heard before (but I'm still new to this).

And finally we get to what I think is probably the weirdest, coolest, most interesting recording in this collection. This segment is a little over a minute long and it has a LOT going on in it. But if you don't use head phones you'll probably miss most of the action. The file is here:
And to do this recording justice I had to make 4 spectrograms to capture all the annotations and notes I made.

In the first spectrogram:
the recording starts out with a speaker making some canine sounding barks, but before this spectrogram finishes the speaker takes a strange turn. It's fourth vocalization sounds more like a descending howl that trails off in a faint warbling tone. Hmmmm.

On to the second spectrogram:
In this section the speaker issues a vocal similar to a whimper, then a very soft waivering tone before a two note vocal (you won't hear this without headphones). Still later another soft vocal just before a possible bird call and then ending with a descending moan. Bird twitters appear high in the spectrogram as they start up at this point and continue through the rest of the recording.

The third spectrogram:
I should mention that traffic noise has been audible all throughout this recording, and it is easily distinguishable at the very lowest edge of the spectrogram. At the 40 second mark we have a strange and complex call. It begins with a descending whine/moan, to a short grunt, to a drawn out waivering note. I think these are all produced by the same speaker because this call is, or variants of it, repeat over the rest of the recording. Immediately after this there's a very breathy grunt, as if produced by an animal with a large lung capacity. A few seconds later some very faint vocals, a soft grunt and simultaneous soft vocals again.

Moving on to the fourth and last spectrogram:
At the 54 second point the waivering vocalization returns, again with its simultaneous grunt, and followed shortly by a low murmuring grunt. Then some kind of mechanical click (sounded like a camera shutter to me), and a very weak repeat of the waivering vocal with a simultaneous, descending grunt. At 1:01 we see the start of the strongest waivering vocal signal in this recording. On my very first play through I thought "canine", but then I heard the simultaneous grunt and the manner in which the waiver drops off steeply in its very next utterance. At 1:04 a descending moan starts and slides down in frequency until it transitions into the waivering vocal. This time the simultaneous grunt appears to be absent. But there is a very weak and strange sounding bird warble at the beginning of the call (but I think that's probably a bird in the background).

So like I said, this is one busy and weird recording. If there is more to this recording I would really enjoy examining it. With all that noise going on it seems like the vocalizations weren't going to end any time soon.

BFRO Pennsylvania 2009 Howls

The notes from the 2009 BFRO expedition to Pennsylvania were finally published on their discussion forum:

and in that thread the expedition organizer shared a couple of howl recordings he describes in the notes. I made a couple of spectrograms for these two recordings and noted a couple of things worth pointing out.

1 - the howls come three at a time and start high, then drop in frequency, and
2 - the howls start at or just below 600 hertz
3 - they lack the bark or warble common to many coyote howls

600 hertz is the low end of the coyote howl spectrum, with only a few recordings in my collection that go lower. In fact coyote howls in these same recordings are above 600 hertz. Calls below 600 hertz are more frequently associated with humans, wolves, dogs, or suspected sasquatch howls.

The howls are here:

and the associated spectrograms are here:

Bad Roads Early

Last Saturday was the last day of doe season in West Virginia, so things will start to wind down and the woods will be emptying out soon. Back roads at elevation already have a couple inches of ice on them after that heavy snow a week ago. Tried to visit a remote stand of virgin timber to look for snow tracks, but no dice, roads just too treacherous.

Michigan Recording Project Analysis

Over the course of several days in early December 2009 I turned my attention to the detailed review of the audio recordings published on the MRP website. As I worked through the recordings, I made the following notes of my observations:

December 1
As part of some ongoing audio analysis, I looked at the frequency ranges of the calls in these audio recordings and noted four interesting things:

1 - The calls are squarely within the range of frequencies most humans are capable of, but also in the range of other purported BF howls.

2 - The calls exhibit significantly more power than any of the human vocalizations on the recordings, either the hellos, or the casual conversation. Whatever makes those calls has a powerful vocal capability.

3 - The couple instances of muttering sounds are very low in the frequency range, mostly below 300 hertz, which is quite deep.

4 - In one of the recordings from 2008 there is one call that is immediately preceded by two rapid hoots, very similar to the type heard from other apes. These are particularly interesting because they happen very quickly and transition into a howl so rapidly that I believe a human would be extremely challenged to produce these sounds (not impossible with significant training, maybe, but certainly not easy).
December 5
After some more off and on listening to these recordings I've found three unusual calls that I thought were worth mentioning. I don't have permission to share snippets of their audio, but I can share some spectrograms that I've generated from these recordings.

The first call came from the August 30, 2008 recording and caught my attention because it is very reminiscent of the unique four tone howl that was recorded during the 2009 BFRO expedition to Minnesota (see: ).

The Michigan four tone howl is visible in this spectrogram:

And you can compare it to the Minnesota four tone howl in this spectrogram:

The differences are; tone's 3 and 4 are reversed from one recording to the other, and in the Michigan example the 4th tone is carried out as a long howl, where in Minnesota it cuts off abruptly.

What I find compelling about the Michigan four tone howl (and the Minnesota as well) is the lingual dexterity required to produce such a rapid change in frequencies, while simultaneously generating so much power. I don't feel that this is within the normal human's ability to accomplish (if you know someone who can do this I'd love to have a recording of it).

The next MRP spectrogram worth highlighting is this one from August 30, 2008:

I mentioned this one in the December 1 notes above as containing a couple hoots, but after looking at it more closely I can report even more. In this call, it starts out with a high keening howl. But, AT THE SAME TIME, the vocalizer generates a grunt and two ape-like hoots.

The grunt-hoot-hoot is simultaneous to the keening howl, and both sequences are only about a half second long (very fast). They start at the same time, and end at the same time, and transition into a long whooping howl (similar to the many other calls in the recordings).

This is a very complex lingual accomplishment, and I'm hard pressed to believe a human has the ability to generate it (but again, if you can prove that assumption wrong, please do). And I'll admit I was still wondering if this could be a "manufactured" sound, but on a second pass through the recording I realized a similar "grunt-hoot-howl" call was issued about five minutes earlier in the same recording (only not so clearly or noticeably). I'll develop a spectrogram of that call later for comparison.

The third unique call which I spectrogrammed is not outside of the human vocal range, but its a very unique call and its "human-ness" was part of what makes it stand out. Here's the spectrogram of this call captured in the August 2, 2008 recording:

This call starts with a voice warble that sounds just like the old "George of the jungle" cartoon where George yells "AH ah AH ah AH ahhh". Then the howl changes into the normal form for these recordings and ends with a unique "oooowWAAAAH"! Again, this is within the range of a human, assuming that human had some really powerful lungs, but it's a very interesting sounding call.

As mentioned above, I noted a second howl in the MRP recordings that came earlier than, and appeared similar to, the complex grunt-hoot-howl previously described. While this second howl is not quite as clear, it is very similar in form, including the complex beginning with a high pitched howl in sync with two low frequency grunts (not quite hoots). Kind of interesting to see the two spectrograms side by side. They're very similar.

Here's the one from 1:32 in the August 30, 2008 recording:

And here's the one from 3:52 in the same recording:

December 7
So far I haven't had any luck isolating the lowest frequency sounds in the August 30, 2008 recording (where I'm focusing most of my attention at the moment). There's a lot of noise at the low end of these recordings, and it may cover up any sounds that were emitted that low.

I have had some luck with a low-pass filter though that attenuated a lot of the high frequencies and allowed me to bring out the low end mumbles and grumbles a little more clearly. For me, the big question is "what's that warble sound" at 7:05. It sounds somewhat mechanical, yet biological at the same time. The nearest thing I've heard to it is an unbalanced tire at long distance on a roadway, or maybe a horse exhaling through its mouth, flapping its lips.

I started from the beginning and paid attention to the grumbles and mumbles. They all sound like they're being emitted by the same speaker throughout the recording. Early on, they sound louder, more energetic. Toward the end they become quieter and more relaxed. It could be a reflection of tension easing in the speaker over time. The first "mumble" is at 0:56 and if I loop it I hear it's composed of 5 syllables. They're not intelligible words, but they are distinctly uttered pulses of sound.

The second mumble is more like a growl, at 2:50, just after the dogs and coyotes go quiet. It's two parts. The first part of the growl rises in frequency, then the second part drops in frequency. Next at 2:54 theres a single moaning mumble and at 2:56 another moaning mumble, just a little stronger. At 3:02 a very soft moaning mumble and one more at 3:10.

At 3:33 there's another low mumble and this one has the first hint of repetition, somewhat similar to the warble heard at 7:05. There are at least 4 rhythmic sound pulses in this mumble. Another moaning mumble at 3:37 seems time synchronized with four breaths right after that, first time I've noted them synced up so cleanly (indicating they could be coming from the same source).

Another moaning mumble at 3:50, just before the complex grunt-hoot-howl call. And at 4:02 a more aggressive growling mumble in the midst of several wood knocks. At 4:36, a faint breathy mumble and at 4:48 two more moaning mumbles, but this time with a hint of warble creeping in.

At 6:03 there's a very deep breathy exhale, that sounds like it comes from a large chested animal. I just mention it because it's worthy of attention. It's the third in a series of breaths, each progressively deeper than the one before.

Then at 6:14 theres a mumble that sounds like a series of muttered, but unintelligible syllables. Altogether I hear nine syllables, but what I find more striking is the wavering pattern of the last seven. They too seem to suggest a warble.

Another growling mumble at 6:25.

At 6:54 there's a very faint warble. It sounds different from the earlier moaning mumbles with warbling characteristics. Less vocal, more mechanical or airy sounding, and very faint.

Finally, starting at about 7:03, there's an extended warble sound. Still faint in relationship to other sounds on the recording, but clearly among the loudest of these low frequency sounds. It sounds like a louder and longer version of the warble at 6:54, slightly mechanical, but still seemingly biological in tone.

GUT MODE KICKING IN: I will venture to postulate that this sound too is coming from the source of all the earlier moaning mumbles. But in this case it has somehow lots it's mumbling quality. Maybe its an example of exhaled air, as people commonly do when releasing tension. I don't know, but that's just where my gut takes me on that one. Unfortunately there's not enough information to nail it down any better, but I will point out that a similar warble exists on an earlier recording. Which in my mind does a lot to correlate this warble with the source of all the other mumbles and howls. GUT MODE OFF.