Skip to content

Peter with Assignment 3

September 9, 2009

Musical instrument sound: Horn

A very interesting sound. At first, it may seemed like there weren’t as many harmonics as other instruments but actually it is just such a low sound that the vertical linear scale on the spectrogram hides this. Given this, the first spectrogram shows the whole sound in its full range. There isn’t much information to see from this graphic except that it is obviously a low sound, as the horn crescendos more higher harmonics are added and the clear change in tone after the user stops blowing. The second spectrogram is much more interesting. I used a high window sampling size (16384) to improve pitch clarity and zoomed in vertically on the bottom of the sound (since it is so low). Using the measurement tool, all the harmonics are very clear. The spectrum does not provide much information but it is a linear measurement so presumably cannot get much information from this low sound.

Sharp Attacks: Ping pong ball bouncing

I liked this sound because the ball was left to drop on the table making the hits get progressively faster and quieter, a cool effect. It was hard to get much information for the spectrum since it is such a short sound and I did not want the program to try and sample the silence. However, once I changed the scale to linear, I got enough information to create a reasonable graphic. The spectrogram is a much more useful graphic. I put the sample size way down (to the min, 32) to get as accurate temporal information as possible (since pitch is irrelevant). As the sound progresses, both the diminuendo and speeding up of the bounces is very clear.

Inharmonic sound: fingernail striking a goblet

This sound had an incredible recording and the long decay of the sound was very well preserved. The spectrum for this sound came out very nicely. I chose to sample it nearer the end of the sound to get a clearer spectrum, and I was in fact able to get very clear peaks at the fundamental and the few harmonics, which sustained this long. For the spectrum, I chose a relatively high window size since pitch was more important, but not high enough to distort the temporal quality of the decays. The image shown highlights the many (and quite complex) harmonics at the start and how they slowly decay finally leaving just the fundamental frequency.

Bird sound: Rooster

I decided to pick a rooster call since it’s a much more interesting sound than your average chickadee. The spectrum was relatively uninteresting for this sound since there are so many frequencies. But the lower frequencies that remain pretty constant throughout the call (shown in orange on the spectrogram) have some nice clear peaks. The spectrum gives both good temporal and frequency information, so I chose a window size that gave information about both these aspects. First, the distinction between the two parts of the call is very clear. Also, I was surprised to see how many clear harmonics there were. Although these harmonics move up and down a lot as the sound is very pitchy, they are well defined at most parts in the sound.

2 Comments leave one →
  1. peterlivesey permalink
    September 9, 2009 10:00 pm

    I forgot to link the pictures…
    so here it is

  2. September 13, 2009 5:25 pm

    The trumpet is not that low, its fundamental frequency is around 230Hz and thus the window size should not be that big, with a window size of 2048 you can see all the harmonics quite well. It would also had been better to plot the Spectrum with a linear frequency scale to see how the harmonics are equally spaced.

    Good with the Ping Pong, the Goblet and the Rooster sounds.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: