Sonograms (= sound spectrograms or audiospectrograms) are frequency (= pitch, Y axis) versus time (X axis) displays of the energy contained in sounds. Amplitude (loudness) variation may be represented by the density ("blackness") of the image.
There are many ways of producing sound spectrograms. Output varies dramatically depending on the equipment and settings used. All of the spectrograms used on the diagnosis page were produced manually on late1970s versions of Kay Elemetrics machines, then scanned and made into GIF files.
Although once an exclusive province of professional researchers, most computer users can now produce high quality sonograms using inexpensive software such as SoundEdit (Macromedia) and C-Canary. For example, here are some crossbill calls (Type 5 "toops") produced with different software and settings:
Other than output quality, the shapes of "objects" in the 2-dimensional acoustic space will change depending on the scale of the axes. The ones depicted on the diagnosis page all have the same scale: the X axis (time) for each box is equivalent to 100 milliseconds (one tenth of a second), and the Y axis goes from 1000 to 9000 Hertz (1 - 9 kHz). This is exactly twice as broad as most bird sonograms found in the ornithological literature. The differences among crossbill call notes are more apparent when stretched in the time dimension (relative to the "standard display", which has the benefit of fitting more information, such as an entire 2-second bird song, on a rectangular piece of sonograph paper).