Context and Patterns with Audio
- Details
- Category: Articles
- Published: Saturday, 09 January 2021 03:02
- Written by Super User
- Hits: 1456
I am often asked by people to listen to some audio they collected for an opinion from people in the MABRC and other independent researchers. Over the years, I've listened to a lot of my own audio and have learned that context has a lot to do when trying to accurately identifying audio. By no means am I am expert as there is no such thing as an expert in this field but I have compared a bunch of unknown audio to try and find what critter made it.
I have received a lot of clips with 1-3 seconds of a vocalization that is questioned. It is so cool hearing the possible howls and screams that it is easy to just clip those out and say what is it. I've been guilty of that myself. The biggest thing we deal with in the woods is distortion of the audio. Whether it is the hills and hollers that echo, wind noise, insect or frog sounds, the quality of the audio is not the best for making proper Identification. We all have recorded that howl or scream way off in the distance that barely registers on the spectrogram, which is the best tool for identification. So we are left wondering. Even some times when we do get close audio and a good spectrogram signature there can be distortion by mic hum, traffic in the distance, a whole bunch of things that doesn't make the audio crystal clear.
That is where context comes into play. Without knowing what was going on before and after the vocalization was recorded makes it tougher for identification. Let's make a scenario of audio in the woods. Over the course of a five minute period you record 2 howls, 3 knocks and the neighborhood dogs go off. In research we have all recorded stuff like that. None of it alone is evidence by itself in my opinion. But when it is reviewed as it was recorded the sounds can make a possible picture of what is going on at the time. If the dogs weren't barking before the knocks and stopped when after the howls, it could say that whatever did the knocks alerted the dogs and the howls scared them. Is that what went on for sure? We don't know of course but it could be an indicator of some activity that may be what we are looking for, an apex predator. Using context can also be very important when there is repeated sounds later in the same area. Is there a pattern to it? Do we still get knocks, then the dogs go wild, then the howls and the dogs go quiet? If so that is a pattern and has more meaning than if it only happens once. Repeatability is one of the primary keys to collecting evidence. If the same unknown sounds are repeated at different times it really makes easier to say that it is not a distortion.
I know that when we made the protocols for the MABRC Evidence Review Board, one of the requirement was to have at least 2 hours of audio before and after the suspected audio. That way it is fairly easy to help weed out deliberate hoaxes, but more importantly gives us a good context on what is in the environment. We've all seen those YouTube videos where someone goes out in the woods, sets up a video camera, and within seconds get a whole array of the different purported Bigfoot vocals. The strange thing is that they are all so clear, they are all really quick and flip from one type of vocal to another. Sorry folks, I've listened to a lot of audio collected in the woods and recorded some that I suspect is bigfoot and stuff like that in my opinion doesn't happen. If it did we could track them easily find them because they wouldn't be quiet. That is not the way any critters in the woods live. One big red flag for me is when you ask these folks for the audio before and after you only get silence or the famous lines to the effect "if you can't tell this is bigfoot then you don't know anything". Or "if you spent time in the woods you'd know what it is". There is also some very good stuff out there but unlike the ones above there is usually a lot of subtle vocals mixed with regular woods noises. Those are the ones where the average joe thinks is boring, or is mixed with known critters and because it is slow, or it must be a _____ because it happens at the same time, so they don't spend much time to listening to them.
The key to trying to understand the context of unknown sounds is a lot more than just listening to it happen one time. When patterns start to be seen it is easier to compare to known animals in the woods. Knowing where domestic critters are in your area and how they normally act, which dog bark is where, where the livestock animals are, all play in understanding when something happens that doesn't fit. If you hear a cow or a dog where there is no dogs or cows you may want to take a better look at the sound to make sure your mind isn't filling in the blanks. Keeping a detailed notebook of your audio you've reviewed including dates and times, keeping the raw audio files all are important if you are trying to see patterns and try to pick out behaviors as well as gives a better idea where you could set up the audio for clearer better results that could help identify what you are hearing. This is very important for long term research as it allows us to better know when and where to set up to get a better chance of being in the right place at the right time. I think we all would much rather spend time in the woods than sitting at the computer reviewing audio, comparing it to known animals, and looking for patterns, but by doing the review and research properly we can spend our time in the woods more effectively, have less review of the just normal known critters and regular woods sounds. We all know how boring it can be listening to audio when nothing is going on and how exciting it can be when things are happening. Context and finding patterns can help a lot when it comes to trying to limit uneventful recordings.
Also context and patterns can help with the known critters. What coyotes, foxes, and other critters sound like. What is the pattern of their vocals? How long do the vocalizations last? Is those odd sounding vocals that mix with coyotes actually coyotes? Do they happen regularly or just occasionally? Does it match the spectrogram of coyote vocals? All that information makes it easier to hear when things are off or added vocals by another known or possibly unknown critter. Listening to how normal critters travel through the woods can help with identifying foot/hoof falls. Do they seem to move more when ambient noise is present or doesn't it seem to make a difference? What can you hear happening during the movement? Can you hear them eating as they move along? Do they seem to linger in one area or just slowly travel through? That can be some great information to have when trying to figure out what is moving around camp or recorders. When you hear a predator call or an owl do the little critters of the forest give their alarm calls? Or do they stay quiet? How do the sounds change when a predator is in the area? Again it is all stuff that we can use to put things into context to paint a better picture of how life in the woods works. I think we all have experienced of how things can be so quiet when we come into an area. But if we sit down by a tree and stay silent the birds, squirrels, and even the bugs start to move around again and there is quite a bit of activity and sounds out there. I think there are patterns to all the critters out there like that which we can hear changes in activity and it makes it easier to understand what is moving through the woods.