y=xsinx is known as the sinc function, the sine divided by x1. The unnormalized form is treated here.
Definition and singularity
sincx=xsinx
On the face of it the denominator vanishes at x=0, but the following limit exists.
x→0limxsinx=1
So x=0 is a removable singularity. Defining the value there to be 1 makes the graph continuous, passing through the point (0,1).
Domain, range and symmetry
With the value at x=0 filled in, the domain is all real numbers. Both sinx and x are odd, so their quotient is even and the graph is symmetric about the y-axis. The maximum value is 1, attained at x=0.
Zeros
The function vanishes where sinx=0 with x=0, that is at x=nπ for n=0. Note that x=0 is not a zero but the point where the maximum is attained.
Decay and the envelope
Since ∣sinx∣≤1, we have xsinx≤∣x∣1, so the oscillation decays along the envelope y=±x1. As x→±∞ the values approach 0, so the x-axis is a horizontal asymptote.
Extrema
The extrema sit at the solutions of tanx=x, and they shrink along the envelope as x moves away from the origin.
x
xsinx
0
1
≈4.4934
≈−0.2172
≈7.7253
≈0.1284
≈10.9041
≈−0.0913
The minimum is about −0.2172, so the values always lie between roughly −0.217 and 1.
Integral
The antiderivative cannot be written with elementary functions, but the integral out to infinity has a closed form2.
∫0∞xsinxdx=2π
This is the Dirichlet integral. The integral of xsinx diverges, so the convergence is conditional: the positive and negative swings cancel to leave a finite value.
Applications
The Fourier transform of a rectangular pulse
The interpolation kernel of the sampling theorem, an ideal low-pass filter
The intensity pattern of light diffracted by a single slit
The form πxsin(πx) is called the normalized sinc, adjusted so that its zeros fall exactly on the integers. In the sampling theorem it is the interpolation function that reconstructs a continuous signal from discrete samples3.