Vocal glissandro as swept frequency source

Richard Owlett wrote:
> Rune Allnor wrote:
> 
>> On 14 Aug, 23:50, Richard Owlett <rowl...@atlascomm.net> wrote:
>>
>>> Rune Allnor wrote:
>>
>>
>> ...
>>
>>>>> Success *GRIN* 
>>>>> ;)http://www.virtual-bird.com/songs/gavia-arctica.mp3givesa very nice
>>>>> display. I'll post some samples as soon as my ISP solves a glitch.
>>
>>
>> ...
>>
>>>>> Rune, you mentioned having used a spectrogram to demonstrate 
>>>>> overtones.
>>>>> I didn't spot any with the loon recording I used. What did you use?
>>>
>>>
>>>> A standard spectrogram. The frequency range was 0-5kHz, if I remember
>>>> correctly, and that was sufficient bandwidth to see the overtones.
>>>
>>>
>>> I'll have to find a better recording. The background noise has a sharp
>>> drop-off ~3-3.5 kHz but there is relatively strong blip 7.1-7.3 kHz. I'm
>>> suspicious that it's an artifact rather than a signal.
>>
>>
>>
>> If I interpret the link you posted correctly, you've used an MP3
>> file.
>> That might not be the best choise, as MP3s have been beaten and
>> battered
>> in all sorts of ways to compress the sound.
>>
>> Try and see if you can find a .wav file.
>>
>> Rune
> 
> 
> Chuckle ;)  I had found a .wav file but chose to to use the .mp3 file 
> because it sounded "cleaner".
> 

Tried the .wav file I had. I was as bad as it sounded. But the noise did 
display nicely ;/

Rune Allnor wrote:

> On 14 Aug, 18:51, Richard Owlett <rowl...@atlascomm.net> wrote:
> 
>>Rune, you mentioned having used a spectrogram to demonstrate overtones.
>>I didn't spot any with the loon recording I used. What did you use?
> 
> 
> A standard spectrogram. The frequency range was 0-5kHz, if I remember
> correctly, and that was sufficient bandwidth to see the overtones.
> 

After a little more experimentation, I'm going to guess your spectrogram 
was based on log(intensity). I was doing things linear. A side effect of 
Scilab's contour() limited me to a 30:1 dynamic range.

I tried a log(intensity) approach and saw some additional structure at 
the expense of a horrible redisplay time. There was also an annoying 
amount of background clutter. That made me realize that my real 
objection to spectrograms wasn't to displaying intensity by shade/color 
but to their cluttered appearance.

I'm also getting some better quality recordings.
The local library has a CD with recordings by the Cornell Laboratory of 
Ornithology and the Borror Laboratory of Bioacoustics. 
www.google.com/search?q=birds%20recording%20wav%20-mp3 also gives some 
useful hits.