- Spectral slope
In
astrophysics andplanetary science , spectral slope is a measure of dependence of thereflectance on thewavelength .Indigital signal processing , it is a measure of how quickly the spectrum of an audio sound tails off towards the high frequencies, calculated using alinear regression G. Peeters, [http://recherche.ircam.fr/equipes/analyse-synthese/peeters/ARTICLES/Peeters_2003_cuidadoaudiofeatures.pdf A large set of audio features for sound description] , tech. rep., IRCAM, 2004.] .pectral slope in Astrophysics / Planetary Science
The visible and
infrared spectrum of the reflected sunlight is used to infer physical and chemical properties of the surface of a body. Some objects are brighter (reflect more) in longer wavelengths (red). Consequently, in visible light they will appear redder than objects showing no dependence of reflectance on the wavelength.The diagram illustrates three slopes:
*a "red slope", the reflectance is increasing with the wavelengths
*"flat spectrum" (in black)
*And a "blue slope", the reflectance actually diminishing with the wavelengthsThe slope is typically expressed in percentage increase of reflectance (i.e. reflexivity) per unit of wavelength: %/100 nm (or % /1000 Å)
The slope is mostly used in near infrared part of the spectrum while colour indices are commonly used in the visible part of the spectrum.
The
trans-Neptunian object Sedna is a typical example of a body showing a steep red slope (20%/100nm) while Orcus' spectrum appears flat in near infra-red.pectral slope in audio
The spectral "slope" of many natural audio signals (their tendency for to have less energy at high frequencies) has been known for many years [D. B. Fry, The Physics of Speech, Cambridge Textbooks in Linguistics, Cambridge University Press, 1996.] , and the fact that this slope is related to the nature of the sound source. One way to quantify this is by applying
linear regression to the Fourier magnitude spectrum of the signal, which produces a single number indicating the slope of the line-of-best-fit through the spectral data .Alternative ways to characterise a sound signal's distribution of energy vs frequency include
spectral rolloff ,spectral centroid .References
Wikimedia Foundation. 2010.
