Are there any command-line tools to take a sound file (e.g. wav or mp3 format) and output a corresponding audio-range spectrogram
as one or more image files, either jpg,gif,png ?? Yes, it would be a very "long" image, or a series multiple images, each w/ a minute of spectrogram representation of sound.

Preferable would be something  like what http://www.sonicvisualiser.org/ outputs in "Layer->Add Melodic Range Spectrogram-> All Channels Mixed"

I'm aware of exellent tools like http://vamp-plugins.org/ && could certainly script something out of sonic-annotator, which would be too slow (the scripting and unnecessary I/O would be slow, not sonic-annotator itself).

Anybody have a http://vamp-plugins.org/vampy.html script that accomplishes this task before I go off and reinvent the wheel. Or better yet, an all-C or C++ solution?

One issue is how can a program automatically determine the best visualization of a spectrogram? 
Sonic-visualizer has controls like "Threshold" "Colour Rotation" and meter "Gain" that sometimes need to be tweaked. Anybody
know where to find such a file->spectrum-image converter that automatically adjusts for best contrast between beats, bass, melody, 
so as to produce something that lets people easily visualize music features, without needing to do any manual adjustments??


Thanks,

Niels
http://nielsmayer.com
 
PS: speaking of spectrograms:
Music: Broken Symmetry, Geometry, and Complexity 
by Gary W. Don, Karyn K. Muir, Gordon B. Volk, James S. Walker
Notices of the American Math Society, Jan 2010, Vol 57, Num.1 , pp. 30-49.
http://www.uwec.edu/walkerjs/MBSGC/