The sinc \text{sinc} sinc function is defined for x ≠ 0 x\neq0 x = 0 as
sinc ( x ) = sin ( x ) x \text{sinc}(x)=\frac{\sin(x)}{x} sinc ( x ) = x s i n ( x )
It can also be expressed as an infinite product:
sin ( x ) x = ∏ n = 1 ∞ ( 1 − x 2 π 2 n 2 ) \frac{\sin(x)}{x}=\prod_{n=1}^{\infty}\left(1-\frac{x^2}{\pi^2n^2}\right) x s i n ( x ) = ∏ n = 1 ∞ ( 1 − π 2 n 2 x 2 )
The normalised sinc \text{sinc} sinc function is
sin ( π x ) π x = ∏ n = 1 ∞ ( 1 − x 2 n 2 ) \frac{\sin(\pi x)}{\pi x}=\prod_{n=1}^{\infty}\left(1-\frac{x^2}{n^2}\right) π x s i n ( π x ) = ∏ n = 1 ∞ ( 1 − n 2 x 2 )
See this post on generating π \pi π using the prime numbers for a short, non-rigorous explanation on expressing functions in this form, or look into the Weierstrass factorisation theorem for a more rigorous explanation on expressing functions in this form. Letting x = 1 2 x=\frac{1}{2} x = 2 1 gives
2 π = ∏ n = 1 ∞ ( 1 − 1 4 n 2 ) = ∏ n = 1 ∞ ( 4 n 2 − 1 4 n 2 ) \frac{2}{\pi}=\prod_{n=1}^{\infty}\left(1-\frac{1}{4n^2}\right) =
\prod_{n=1}^{\infty}\left(\frac{4n^2-1}{4n^2}\right) π 2 = n = 1 ∏ ∞ ( 1 − 4 n 2 1 ) = n = 1 ∏ ∞ ( 4 n 2 4 n 2 − 1 )
This is Wallis’s formula for π \pi π . And taking the reciprocal:
π 2 = ∏ n = 1 ∞ ( 4 n 2 4 n 2 − 1 ) \frac{\pi}{2}=\prod_{n=1}^{\infty}\left(\frac{4n^2}{4n^2-1}\right) 2 π = ∏ n = 1 ∞ ( 4 n 2 − 1 4 n 2 )
If instead we let x = 1 4 x=\frac{1}{4} x = 4 1 we get
2 2 π = ∏ n = 1 ∞ ( 1 − 1 16 n 2 ) = ∏ n = 1 ∞ ( 16 n 2 − 1 16 n 2 ) \frac{2\sqrt{2}}{\pi}=\prod_{n=1}^{\infty}\left(1-\frac{1}{16n^2}\right)
=\prod_{n=1}^{\infty}\left(\frac{16n^2-1}{16n^2}\right) π 2 2 = n = 1 ∏ ∞ ( 1 − 16 n 2 1 ) = n = 1 ∏ ∞ ( 16 n 2 16 n 2 − 1 )
Again, taking the reciprocal gives
π 2 2 = ∏ _ n = 1 ∞ ( 16 n 2 16 n 2 − 1 ) \frac{\pi}{2\sqrt{2}}=\prod\_{n=1}^{\infty}\left(\frac{16n^2}{16n^2-1}\right) 2 2 π = ∏ _ n = 1 ∞ ( 16 n 2 − 1 16 n 2 )
π 2 4 = ∏ _ n = 1 ∞ ( 16 n 2 16 n 2 − 1 ) \frac{\pi\sqrt{2}}{4}=\prod\_{n=1}^{\infty}\left(\frac{16n^2}{16n^2-1}\right) 4 π 2 = ∏ _ n = 1 ∞ ( 16 n 2 − 1 16 n 2 )
These are both quite nice formulae for π \pi π . However, they converge VERY slowly. The first expression approximates π ≈ 3.1415848 \pi\approx3.1415848 π ≈ 3.1415848 after ten thousand terms.
I came across another expression whilst writing the π \pi π and primes post . In that post we showed that
π x cot ( π x ) = 1 − 2 ∑ n = 1 ∞ ζ ( 2 n ) x 2 n \pi x\cot(\pi x)=1-2\sum_{n=1}^{\infty}\zeta(2n)x^{2n} π x cot ( π x ) = 1 − 2 ∑ n = 1 ∞ ζ ( 2 n ) x 2 n
π x 2 cot ( π x ) = 1 2 − ∑ n = 1 ∞ ζ ( 2 n ) x 2 n \frac{\pi x}{2}\cot(\pi x)=\frac{1}{2}-\sum_{n=1}^{\infty}\zeta(2n)x^{2n} 2 π x cot ( π x ) = 2 1 − ∑ n = 1 ∞ ζ ( 2 n ) x 2 n
Letting x = 1 4 x=\frac{1}{4} x = 4 1
π 8 = 1 2 − ∑ n = 1 ∞ ζ ( 2 n ) 4 2 n \frac{\pi}{8}=\frac{1}{2}-\sum_{n=1}^{\infty}\frac{\zeta(2n)}{4^{2n}} 8 π = 2 1 − ∑ n = 1 ∞ 4 2 n ζ ( 2 n )
We’ve shown that the Riemann zeta function at even positive integers is given by the following expression:
ζ ( 2 n ) = ( − 1 ) k + 1 ( 2 π ) 2 n β 2 n 2 ( 2 n ) ! \zeta(2n)=\frac{(-1)^{k+1}(2\pi)^{2n}\beta_{2n}}{2(2n)!} ζ ( 2 n ) = 2 ( 2 n )! ( − 1 ) k + 1 ( 2 π ) 2 n β 2 n
The first few even valued Bernoulli numbers, β 2 n \beta_{2n} β 2 n , are
β 2 = 1 6 ; β 4 = − 1 30 ; β 6 = 1 42 ; β 8 = − 1 30 \beta_2=\frac{1}{6}; \ \ \beta_4=-\frac{1}{30}; \ \ \beta_6=\frac{1}{42}; \ \ \beta_8=-\frac{1}{30} β 2 = 6 1 ; β 4 = − 30 1 ; β 6 = 42 1 ; β 8 = − 30 1
And so, ζ ( 2 n ) \zeta(2n) ζ ( 2 n ) for the first few values of n n n is
ζ ( 2 ) = π 2 6 ; ζ ( 4 ) = π 4 90 ; ζ ( 6 ) = π 6 945 ; ζ ( 8 ) = π 8 9450 \zeta(2)=\frac{\pi^2}{6}; \ \ \zeta(4)=\frac{\pi^4}{90}; \ \ \zeta(6)=\frac{\pi^6}{945}; \ \ \zeta(8)=\frac{\pi^8}{9450} ζ ( 2 ) = 6 π 2 ; ζ ( 4 ) = 90 π 4 ; ζ ( 6 ) = 945 π 6 ; ζ ( 8 ) = 9450 π 8
Expanding the sum out:
π 8 = 1 2 − π 2 6 ⋅ 4 2 − π 4 90 ⋅ 4 4 − π 6 945 ⋅ 4 6 − π 8 9450 ⋅ 4 8 − . . . \frac{\pi}{8}=\frac{1}{2}-\frac{\pi^2}{6\cdot4^2}-\frac{\pi^4}{90\cdot4^4}-\frac{\pi^6}{945\cdot4^6}-\frac{\pi^8}{9450\cdot4^8}-.\ .\ . 8 π = 2 1 − 6 ⋅ 4 2 π 2 − 90 ⋅ 4 4 π 4 − 945 ⋅ 4 6 π 6 − 9450 ⋅ 4 8 π 8 − . . .
Although not particulary useful for actually generating or approximating π \pi π ,
it’s still pretty cool. We can simplify it visually by defining a constant a n a_n a n :
a n = ( − 1 ) n + 1 4 n β 2 n 2 ( 2 n ) ! a_n=\frac{(-1)^{n+1}4^n\beta_{2n}}{2(2n)!} a n = 2 ( 2 n )! ( − 1 ) n + 1 4 n β 2 n
So the expression becomes
π 8 = 1 2 − a 1 π 2 4 2 − a 2 π 4 4 4 − a 3 π 6 4 6 − a 4 π 8 4 8 − . . . \frac{\pi}{8}=\frac{1}{2}-a_1\frac{\pi^2}{4^2}-a_2\frac{\pi^4}{4^4}-a_3\frac{\pi^6}{4^6}-a_4\frac{\pi^8}{4^8}-.\ .\ . 8 π = 2 1 − a 1 4 2 π 2 − a 2 4 4 π 4 − a 3 4 6 π 6 − a 4 4 8 π 8 − . . .
Euler famously proved that ζ ( 2 ) = π 2 6 \zeta(2) = \frac{\pi^{2}}{6} ζ ( 2 ) = 6 π 2 . He started with the
function sin x x \frac{\sin x}{x} x s i n x , i.e. the sinc \text{sinc} sinc function, which has the
Taylor series expansion
sinc x = sin x x = ∑ n = 0 ∞ ( − 1 ) n ( 2 n + 1 ) ! x 2 n = 1 − x 2 3 ! + x 4 5 ! − x 6 7 ! + … \text{sinc }x = \frac{\sin x}{x} = \sum_{n=0}^{\infty}
\frac{(-1)^{n}}{(2n+1)!}x^{2n} = 1 - \frac{x^{2}}{3!} + \frac{x^{4}}{5!} -
\frac{x^{6}}{7!} + \ldots sinc x = x sin x = n = 0 ∑ ∞ ( 2 n + 1 )! ( − 1 ) n x 2 n = 1 − 3 ! x 2 + 5 ! x 4 − 7 ! x 6 + …
The sinc \text{sinc} sinc function can also be expressed as the following infinite product:
sinc x = ∏ n = 1 ∞ ( 1 − x 2 π 2 n 2 ) = ( 1 − x 2 π 2 ) ( 1 − x 2 4 π 2 ) ( 1 − x 2 9 π 2 ) … \text{sinc }x = \prod_{n=1}^{\infty} \left( 1 - \frac{x^{2}}{\pi^{2}n^{2}}
\right) = \left( 1 - \frac{x^{2}}{\pi^{2}} \right) \left( 1 -
\frac{x^{2}}{4\pi^{2}} \right) \left( 1 - \frac{x^{2}}{9\pi^{2}} \right) \ldots sinc x = n = 1 ∏ ∞ ( 1 − π 2 n 2 x 2 ) = ( 1 − π 2 x 2 ) ( 1 − 4 π 2 x 2 ) ( 1 − 9 π 2 x 2 ) …
= 1 − x 2 ( 1 π 2 + 1 4 π 2 + 1 9 π 2 + … ) + … = 1 - x^{2} \left( \frac{1}{\pi^{2}} + \frac{1}{4\pi^{2}} +
\frac{1}{9\pi^{2}} + \ldots \right) + \ldots = 1 − x 2 ( π 2 1 + 4 π 2 1 + 9 π 2 1 + … ) + …
Equating the x 2 x^{2} x 2 coefficients of the above to the Taylor series expansion
gives the desired result:
− ( 1 π 2 + 1 4 π 2 + 1 9 π 2 + … ) = − 1 3 ! -\left( \frac{1}{\pi^{2}} + \frac{1}{4\pi^{2}} + \frac{1}{9\pi^{2}} + \ldots
\right) = - \frac{1}{3!} − ( π 2 1 + 4 π 2 1 + 9 π 2 1 + … ) = − 3 ! 1
ζ ( 2 ) = ∑ _ n = 1 ∞ 1 n 2 = π 2 6 \zeta(2) = \sum\_{n=1}^{\infty} \frac{1}{n^{2}} = \frac{\pi^{2}}{6} ζ ( 2 ) = ∑ _ n = 1 ∞ n 2 1 = 6 π 2
A similar result can be derived for the odd squares. Instead of sinc x \text{sinc } x sinc x
we start with cos x \cos x cos x . The series expansion for cos x \cos x cos x is
cos x = ∑ n = 0 ∞ ( − 1 ) n ( 2 n ) ! x 2 n = 1 − x 2 2 ! + x 4 4 ! − x 6 6 ! + … \cos x = \sum_{n=0}^{\infty} \frac{(-1)^{n}}{(2n)!}x^{2n} = 1 -
\frac{x^{2}}{2!} + \frac{x^{4}}{4!} - \frac{x^{6}}{6!} + \ldots cos x = n = 0 ∑ ∞ ( 2 n )! ( − 1 ) n x 2 n = 1 − 2 ! x 2 + 4 ! x 4 − 6 ! x 6 + …
cos x \cos x cos x can also be expressed as an infinite product:
cos x = ∏ n = 1 ∞ ( 1 − 4 x 2 π 2 ( 2 n − 1 ) 2 ) = 1 − x 2 ( 4 π 2 + 4 9 π 2 + 4 25 π 2 + … ) + … \cos x = \prod_{n=1}^{\infty} \left( 1 - \frac{4x^{2}}{\pi^{2}(2n - 1)^{2}}
\right) = 1 - x^{2}\left( \frac{4}{\pi^{2}} + \frac{4}{9\pi^{2}} +
\frac{4}{25\pi^{2}} + \ldots \right) + \ldots cos x = n = 1 ∏ ∞ ( 1 − π 2 ( 2 n − 1 ) 2 4 x 2 ) = 1 − x 2 ( π 2 4 + 9 π 2 4 + 25 π 2 4 + … ) + …
And again, equating the x 2 x^{2} x 2 coefficients:
− 4 π 2 ( 1 + 1 9 + 1 25 + … ) = − 1 2 ! - \frac{4}{\pi^{2}}\left( 1 + \frac{1}{9} + \frac{1}{25} + \ldots \right) = -
\frac{1}{2!} − π 2 4 ( 1 + 9 1 + 25 1 + … ) = − 2 ! 1
∑ _ n = 1 ∞ 1 ( 2 n − 1 ) 2 = π 2 8 \sum\_{n=1}^{\infty} \frac{1}{(2n - 1)^{2}} = \frac{\pi^{2}}{8} ∑ _ n = 1 ∞ ( 2 n − 1 ) 2 1 = 8 π 2
Going back to the infinite product form of sin x x \frac{\sin x}{x} x s i n x ,
sinc x = ∏ n = 1 ∞ ( 1 − x 2 π 2 n 2 ) = ( 1 − x 2 π 2 ) ( 1 − x 2 4 π 2 ) ( 1 − x 2 9 π 2 ) … , \text{sinc }x = \prod_{n=1}^{\infty} \left( 1 - \frac{x^{2}}{\pi^{2}n^{2}}
\right) = \left( 1 - \frac{x^{2}}{\pi^{2}} \right) \left( 1 -
\frac{x^{2}}{4\pi^{2}} \right) \left( 1 - \frac{x^{2}}{9\pi^{2}} \right) \ldots, sinc x = n = 1 ∏ ∞ ( 1 − π 2 n 2 x 2 ) = ( 1 − π 2 x 2 ) ( 1 − 4 π 2 x 2 ) ( 1 − 9 π 2 x 2 ) … ,
and noting that each term is the difference of two squares, we can separate them out:
sin x = x ( 1 − x π ) ( 1 + x π ) ( 1 − x 2 π ) ( 1 + x 2 π ) ( 1 − x 3 π ) ( 1 + x 3 π ) … \sin x = x\left( 1 - \frac{x}{\pi} \right)\left( 1 + \frac{x}{\pi} \right) \left( 1 - \frac{x}{2\pi} \right)\left( 1 + \frac{x}{2\pi} \right) \left( 1 - \frac{x}{3\pi} \right)\left( 1 + \frac{x}{3\pi} \right) \ldots sin x = x ( 1 − π x ) ( 1 + π x ) ( 1 − 2 π x ) ( 1 + 2 π x ) ( 1 − 3 π x ) ( 1 + 3 π x ) …
Taking the log \log log of both sides:
log sin x = log x + log ( 1 − x π ) + log ( 1 + x π ) + log ( 1 − x 2 π ) + … , \log \sin x = \log x + \log \left( 1 - \frac{x}{\pi} \right) + \log \left( 1 +
\frac{x}{\pi} \right) + \log \left( 1 - \frac{x}{2\pi} \right) + \ldots, log sin x = log x + log ( 1 − π x ) + log ( 1 + π x ) + log ( 1 − 2 π x ) + … ,
and then taking the derivative:
cos x sin x = 1 x − 1 π − x + 1 π + x − 1 2 π − x + … \frac{\cos x}{\sin x} = \frac{1}{x} - \frac{1}{\pi - x} + \frac{1}{\pi + x} - \frac{1}{2\pi - x} + \ldots s i n x c o s x = x 1 − π − x 1 + π + x 1 − 2 π − x 1 + …
Letting x = π 4 x = \frac{\pi}{4} x = 4 π gives Leibniz’s formula for π \pi π :
π 4 = 1 − 1 3 + 1 5 − 1 7 + 1 9 − … . \frac{\pi}{4} = 1 - \frac{1}{3} + \frac{1}{5} - \frac{1}{7} + \frac{1}{9} -
\ldots . 4 π = 1 − 3 1 + 5 1 − 7 1 + 9 1 − … .