Help with reading sonogram please (1 Viewer)

Greetings All,

In the wee early hours of the morning, in the dark, my brother recorded a bird singing its best. Please note, the links ending in html take you to a streaming audio player with the file on line, the second of each pair of links will allow you to download the file to play on your own system. I have put four links, the first two are to the full recording, 15 Mb / 15minutes long. The second two links are to a 3 minute clip of the original file, 2.8mb for the sake of brevity...
The sound files are here

Full file = 15 minutes
http://www.4shared.com/audio/IPFiUX5w/Mocker2a.html
http://dc251.4shared.com/download/327524294/1aae5b9c/Mocker2a.mp3

3 Minute file:
http://www.4shared.com/audio/72PFnNVU/Mocker2a1.html
http://dc192.4shared.com/download/327541698/ac8fbaf9/Mocker2a1.mp3

Sonogram image in Jpeg format here:
http://www.4shared.com/photo/RBC8SkwM/firsttrack001.html
http://dc189.4shared.com/download/328150409/e1cb9962/firsttrack001.jpg


The bird is a Northern Mocking bird, the recording was made in South West Florida at 1:00 AM EDT (local time) in the morning June 27th, 2010. The quality of the recording is excellent especially considering my brother made it with a Nikon D90 camera in movie mode and converted the file to MP3 format to send it to me. To make this recording, he was sitting on his doorstep hand holding the camera and I think the quality of this recording is amazing considering the conditions...

What is fascinating to both of us, when looking at this file as a sonogram, is that the 2nd and 3rd harmonics are mirror image of the prime sound and 1st harmonic.

My first reaction was that, as this was recorded in a space between two trailers, was that what appears as 2nd and 3rd harmonics, is this some kind of an echo?

However there is no time delay, the change is in frequency, not time, and the image shows as true harmonics excepting the upper two harmonics in the image are reversed. It seems everything above 5.5Khz is a mirror of the primary sound by the bird as he makes it.

I can only assume the bird is doing this however I've never seen this mirroring of harmonics before that I recall.

Am I reading this sonogram correctly?

Is this real, or possibly a software glitch or a problem with the camera audio digitizer?

Your comments and advise most needed and welcome,

fmhill said:

Greetings All,

Sonogram image in Jpeg format here:
http://www.4shared.com/photo/RBC8SkwM/firsttrack001.html
http://dc189.4shared.com/download/328150409/e1cb9962/firsttrack001.jpg

...with a Nikon D90 camera in movie mode and converted the file to MP3 format to send it to me. To make this recording, he was sitting on his doorstep hand holding the camera and I think the quality of this recording is amazing considering the conditions...

What is fascinating to both of us, when looking at this file as a sonogram, is that the 2nd and 3rd harmonics are mirror image of the prime sound and 1st harmonic.
...
Your comments and advise most needed and welcome,

Click to expand...

Hi,
what you are seeing is aliasing. This kind of phenomenon is very common in cheap microphones connected to expensive cameras/videos/dictating devices, which are mainly aimed at recording human voices. Based on the sonogram the camera is sampling at a rate of 11KHz, but the software sampling is at a higher rate. In true life the second and higher harmonics follow the shape of the fundamental, but at a frequency of exactly 2x, 3x etc. Aliasing on the other hand takes place at frequencies higher than 0.5X sampling rate, Nykvist frequency, which in this case is 5.5KHz. How aliasing shows up is as a mirror image symmetric to the nyquist frequency. In your example you can even see a second order aliasing taking place above 11KHz, which turns a copy of the "inverse" frequency image upright again. These are not harmonics, although they look a little bit like ones.

Regards
Harry J

Mitch & Shadow,
let me make a small refinement to my answer. Yes, you are actually seeing some parts of the second harmonics, but the overall symmetry of aliasing kind of hides this out.. In each phrase the first tone is changing from about 5 to 2 Hz. This does not a strong harmonics, but the second tone and third tones at 2 to 2.5KHz do show the 2nd harmonics between 4 and 5KHz - note the same shape (but a twice as steep and twoce the frequency of the fundamental). Above this you get the aliasing mirror images. If your sound frequency is very close to the aliasing you can get a "leak" aliased down below the nyqvist frequency and that is exactly what you get in the very last notes of each phrase. As an example, counting down from the lowest frequency in the last note of the last phase you get: 2KHz fundamental, 2.2 and 3KHz reverberation (=echoes) from previous sounds, 4KHz second harmonic, 5.2KHz alias of the 3rd harmonic, 5.7KHz 3rd harmonic, and above this pure aliases.

Harry J

HarryJ,

Thank you so much for your reply... I think you've hit the nail square on the head...

It should have dawned on me seeing the reversal of the image at 5.5khz but it never crossed my mind that the camera would have such a low digitizing rate as 11Khz...

Thanks for your help...

Doing a sonagram directly from the mp3, I can see that the original was precisely 11025 Hz. A lot of space was wasted converting this into a 44.1 kHz MP3... a 11025 Hz 16-bit PCM WAV would've been 20.1 MB, and the MP3 is 14.6 MB, only about 27% smaller.

I've had aliasing sometimes even in a 48 kHz recording, like once when I recorded an Anna's Hummingbird dive sound at close range. There were harmonics the microphone picked up higher than 24 kHz (going as high as 25.5 kHz), which showed up mirrored going down from 24 kHz (except in this case, it was a mirror of something otherwise not present in the sonogram), for example 25.5 kHz was mirrored to 22.5 kHz. Because of this I reconfigured my recorder to do 96 kHz.


I listened to the entire Mocker2a.mp3, because I'm fascinated by mockingbirds. It's a good-quality recording, helped a lot by being done at night with no distracting noises!

Much of it sounded pretty familiar, but there were some interesting bits like:

0:12.31 - kind of interesting 3-syllable sound.
3:17.32, then again at 14:39.46 - I like this sound.
7:11.33-7:21.46, anomaly at 7:17.014 (change in sound with no pause)
7:50.76-7:53.05 - silence during which another bird is heard
11.56.4-12:10.4 - at 12:00.82, this sound gets a nice trill/rattle added to it, then this extra is discarded again
14:50.25 - I like this sound too.

I heard a 0.77 second echo in parts. That's a pretty long echo. Corresponds to about 880 feet.

There wasn't a single sound that I recognized as an imitation of a bird I'm familiar with. I think of this kind of mockingbird as an "abstract singer" who hasn't yet learned to do realistic imitations; maybe a young bird. OTOH, this is in Florida, and I live in California.