This question is posted in response to a recent one seeking the volume of $y =(a^{2/3}-x^{2/3})^{3/2}$ rotated about the x-axis. I wondered why people don't seek a more general solution when posed with such problems. Suppose tomorrow's problem was a power different than 2/3, of worse yet, what if there were two different powers? Or if it required the rotation about the y-axis?
So, the question is, what is the volume of revolution of $y =(1-x^q)^p$ about the x,y-axes, for arbitrary (p,q)?
What I tried: The equation in question is in the form a general superconics that I have described on the MSE here. I have expanded that to include bodies of revolution about the x,y-axes using Pappus's $2^{}$ Centroid Theorem. The results are shown in the Answer posted below.
The simplest approach to finding the volume of a body of revolution is through Pappus's $2^{}$ Centroid Theorem. Essentially, the volume of a body of revolution is the area multiplied by the path length of the centroid (call it $R=(R_x,R_y)$, depending on the axis of revolution), Thus, $V=2\pi RA$. Now, we have shown previously that the area under the curve $y =(1-|x|^q)^p$, $x\in[-1,1]$ is given by
$$A=\Psi(p,q)=\frac{2\Gamma(p+1)\Gamma(1+1/q)}{\Gamma(p+1+1/q)}$$
This was readily derived from a formal integration leading in turn to the beta and then gamma functions. For the present problem (positive x only), the area would be half that. Now, we can similarly derive equations for the centroid of the curve, and that turns out to be as follows,
$$R=\frac{1}{2A}\biggr(\Psi(2/q,1,p),\Psi(2p,q)\biggr)$$
So, finally we can determine the volume of revolution of $y =(1-x^q)^p$, $x\in[0,1]$ to be
$$V=\frac{\pi}{2}\biggr(\Psi(2p,q),\Psi(2/q,1,p)\biggr)$$
for revolution about the x and y-axes, respectively.
We have determined the volume for the canonical case. In the referenced link we show how to scale the results to the more general case given by $y =b\biggr(1-(x/a)^q\biggr)^p$.