Sound Recording - Widening the Detectable Frequency Range

A Think Tank of Techniques and Technology.
Post Reply
User avatar
TrevorPeters
Silver Status
Posts: 236
Joined: Wed Jan 06, 2016 10:06 am
Position: Believer
Location: Ipswich. Qld

Sound Recording - Widening the Detectable Frequency Range

Unread post by TrevorPeters » Sun Mar 27, 2016 5:29 pm

So Sasquatch Ontario pushes the boundaries of credibility for many and I don't want to weigh into that discussion here, however, when I was watching this recent video by Thinker Thunker it reminded me of some thoughts I had back when I was writing my blog posts about Gearing Up.

A Closer Look (or listen) to Sasquatch Ontario (ThinkerThunker)

What sort of equipment would you need to record sound frequencies not within the normal human vocal range?

Thinker Thunker makes some interesting points during his comparison of Sasquatch Ontario (SO) audio and his own mimicking effort (even though he got the word wrong).
The frequency plot shows that the SO recording seems to saturate at 40kHz.

This is interesting as this is the upper frequency response threshold for most decent hand held audio recording equipment. That's because the equipment is design with human audio range sounds in mind (i.e. 20Hz to 20kHz plus or minus a bit). For an example check out the frequency response specifications under Audio Performance for the TASCAM DR-05

It has been mentioned before about infrasound (below 20Hz) and we all know bats can echo-locate at frequencies much higher than 20kHz.

Just as Rusty has been able to use frequency in his recordings to differentiate between hopping wallabies and the elusive and much heavier stompers that he is trying to video, perhaps we have a means here to differentiate between real and faked vocalisations. If a vocalisation was recorded showing frequencies outside the human range then it would stand to reason that it was not made by a human voice box.

But every endeavour has its problems and specialist equipment would no doubt be needed at an associated expense.

Wildlife Acoustics sell a range of products but I have not yet had time to look at all the microphone specifications.

What would probably be needed is a wideband (infrasound to ultrasonic) microphone similar to Hydrophones (2Hz to 192kHz) and a recorder with a matching frequency response.

If anyone knows of any equipment that fits the bill, please post it.

User avatar
TrevorPeters
Silver Status
Posts: 236
Joined: Wed Jan 06, 2016 10:06 am
Position: Believer
Location: Ipswich. Qld

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by TrevorPeters » Sun Mar 27, 2016 5:48 pm

So I found the QTC50 from Earthworks.

This is probably the most cost effective solution I have seen so far.

Having a frequency range of 3Hz to 50kHz it gives a bit of frequency band extension in either direction and has an omnidirectional capture pattern, which is good because you can never tell which direction the hairy fellas are going to come from.

Unfortunately it is 48V phantom powered, so you will need a recorder capable of supplying that and it costs around US$1400 (on special).

User avatar
Rusty2
Long Time Contributor
Posts: 1407
Joined: Mon Aug 23, 2010 5:30 pm
Position: Believer
Location: East Coast

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by Rusty2 » Tue May 10, 2016 5:31 pm

These are the results I came up with Trevor .
I've included my Jah - ah recording for comparison . There seems to be quite a difference .
You do not have the required permissions to view the files attached to this post.

User avatar
Rusty2
Long Time Contributor
Posts: 1407
Joined: Mon Aug 23, 2010 5:30 pm
Position: Believer
Location: East Coast

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by Rusty2 » Tue May 10, 2016 6:40 pm

I feel like I should clarify something .
You may notice the lines above Thunker Thunkers voice and the S . O vocalisation . These are octaves which are normal in a voice . Sometimes there are many octaves .
The difference between the 3 different recordings is that the Jah ah vocalisation is a lot further away and the audio recorder wasn't able to pick up enough information to detect octaves .
If it were , the picture would show the same lines .
In essence , the main part of each recording has been highlighted and you can see the difference .
Sorry for any confusion .

User avatar
Rusty2
Long Time Contributor
Posts: 1407
Joined: Mon Aug 23, 2010 5:30 pm
Position: Believer
Location: East Coast

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by Rusty2 » Thu May 12, 2016 4:57 pm

My sincere apologies , I feel like a real doof .

The Jah - ah vocalisation has nothing to do with Sasquatch Ontario or Thinker Thunker . It's a higher voice/tone/pitch/note .

This is a problem with Thinker Thunkers actual recording of himself . Even if it is a little out of pitch compared to S.O. then the difference will be noticable making it difficult to calculate anything
You do not have the required permissions to view the files attached to this post.

User avatar
TrevorPeters
Silver Status
Posts: 236
Joined: Wed Jan 06, 2016 10:06 am
Position: Believer
Location: Ipswich. Qld

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by TrevorPeters » Thu May 12, 2016 6:01 pm

Hi Rusty,

As always I have some questions. I can be a pain that way. (claps)

So I'm making some assumptions about your last image here:

1. You used the audio from the Thinker Thunker You Tube clip.
2. You used a linear frequency scale with some range limits.

I can't tell from the images what sort of FFT profile you used or if you tried them all and this was the sharpest result.

So my questions are:

1. What are we trying to see here? (I assume you were showing the lack of Infra-sound?)
2. What factors would prevent us from seeing anything?

Its here I start to waffle, so you can switch off now....

Yes you threw me a bit with the first couple of posts but that's OK.
I'm thinking it might be good to see if the woodape/hominid/what-ever-you-want-to-call-it-oid, has a vocal range higher and or lower than the normal human vocal range.
To do that however, it begins with the equipment. If you don't have a microphone that has reasonable gain below 20Hz and above 20-30kHz then you won't capture anything significant even if it is there.
So my curiosity in this regard is not easy to satisfy without spending big bucks that I currently don't have.

So until I hear otherwise I must assume that most, if not all researchers in the field are using your run-of-the-mill digital recorders with human audio range microphones. Unfortunately these don't have the frequency range required to help answer this sort of question, or not that I have seen. Generally, the cheaper the instrument the weaker the performance at the edges of the frequency range and this happens to be just where we need the best effort. Life's like that sometimes.

This is probably why your last image shows pretty much nothing below 60Hz. It tells me that the combined losses of inadequate microphone choice and compression/reprocessing by putting it on You Tube, etc., have combined to give some unacceptable losses. 20Hz to 60Hz is technically still in the human audio range, so if the capture mechanism can't even pick that up we have no hope of detecting any Infra-sound that might be present at frequencies below 20Hz. We could speculate that there were no low frequencies present but we would not know for sure because it is unlikely that suitable equipment was used in the first instance. Although standard microphones often go a bit higher than 20kHz I would not be surprised to see similar losses at the top end if compared to an audio clip captured with a real wideband microphone.

I have looked into this a few times and the best I could come up with was to maybe put a system together using individual components (microphone cartridge and pre-amplifier). Bruel & Kjaer have some microphones that may do the job once matched with an appropriate preamp but as with any design there is always a problem or two. For one, my previous experience with B & K stuff tells me they will not be cheap. Secondly they may all require 48V Phantom power or some other voltage that will complicate things when you really want the resulting instrument to be portable. No doubt these problems could be solved with the right amount of funds (says Mr. Empty-Pockets).

Bruel & Kjaer Acoustic Microphone Cartridges

The QTC50 from Earthworks was a bit of a trade-off and I would speculate will not have the overall range of a B&K system.

Anyway, thanks for having a go at it Rusty and keep up the good field work you do. We all appreciate your efforts mate.

User avatar
TrevorPeters
Silver Status
Posts: 236
Joined: Wed Jan 06, 2016 10:06 am
Position: Believer
Location: Ipswich. Qld

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by TrevorPeters » Thu May 12, 2016 6:44 pm

M K Davis discusses possible infrasound

Yep it sure annoyed me, even through cheap PC speakers.

User avatar
Rusty2
Long Time Contributor
Posts: 1407
Joined: Mon Aug 23, 2010 5:30 pm
Position: Believer
Location: East Coast

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by Rusty2 » Fri May 13, 2016 3:59 pm

Hey Trevor , yes I used Thinker Thunkers own voice and S.O recording .

The frequencies are vertical and the time is linear . The two recordings amount to almost 2 seconds in length . You could say that the above picture is a magnified view at the very bottom end of the scale between 0 and 140 Hz but still the entire recording in length .

The representation is an auto adjustable short time fourier transform . (STFT) There are others that I could use like Regular STFT , Multi-resolution or Adaptively sparse . At this magnification there isn't a whole lot of difference .

Yes . I'm showing that there seems to be little difference between Thinker Thinkers voice and S.O. and I'm seeing little to no audio below 60 Hz without the original recordings which may be using inferior equipment anyway .

Distance from the microphone would stop us from seeing infrasound and octaves . Pitch would have a lot to do with it as well between TT an S.O . I had a look at my own voice (picture attached) from one of my commentaries and noticed infrasound down to zero hz . If I can produce it then how can we tell the difference between us and them ?

Your right , if we're not using the right equipment we don't know what we're missing . I have looked at the earthworks mic's as well but couldn't justify the expense . What your after is specialised equipment and audio programs which means $$$ . I wonder if there is an audio program that actually measures below 0 Hz . If Hz's are cycles per second than there is no such thing as a noise which is below 0 Hz , or am I wrong ?
You do not have the required permissions to view the files attached to this post.

User avatar
TrevorPeters
Silver Status
Posts: 236
Joined: Wed Jan 06, 2016 10:06 am
Position: Believer
Location: Ipswich. Qld

Re: Sound Recording - Widening the Detectable Frequency Rang

Unread post by TrevorPeters » Fri May 13, 2016 7:23 pm

Rusty2 wrote: I had a look at my own voice (picture attached) from one of my commentaries and noticed infrasound down to zero hz . If I can produce it then how can we tell the difference between us and them ?
That's hard to answer when we have zero audio samples recorded with proper equipment. Maybe we won't ever be able to do so. Generally it would take analysis of a large number of recordings to be confident of identifying any characteristic differences that might be present, but then we might find no differences too. It will no doubt become an academic argument once other methods have produced enough proof to get the scientific community involved.

Rusty2 wrote: I wonder if there is an audio program that actually measures below 0 Hz . If Hz's are cycles per second than there is no such thing as a noise which is below 0 Hz , or am I wrong ?
Yes and no.

In practical terms you are correct that you can't go out, record and listen to a -1kHz tone for example. At least not that I am aware of.

The concept of negative frequency does exist in mathematical evaluation of sinusoidal waveforms or what is called complex sinusoids. It has to do with one waveform leading or lagging the other so the negative values are relative rather than absolute. There is a lot of mathematics with greek letters and cos, sin, etc. that just hurts my old brain. I hated trigonometry and algebra the first few times I had to learn and use it, now I just hope to forget it. Ignorance can be bliss sometimes.

Best not to lose any sleep over it. I know I won't. :D

Post Reply