Poles of AR estimator

Hi,
I am workin on a LPC AR estimator and i need to compute each of its poles.
The denominator is a complex coefficients polynom of 120 degrees one.
Does anyone knows a way to find the roots of a complex coef. polynomial ?

Thx to those who can help

Can you use Muller's Method?
Richard
"Arthur Herbert" <herbert.arthur@caramail.com> wrote in message 
news:417e4b4c$0$3812$8fcfb975@news.wanadoo.fr...
> Hi,
> I am workin on a LPC AR estimator and i need to compute each of its poles.
> The denominator is a complex coefficients polynom of 120 degrees one.
> Does anyone knows a way to find the roots of a complex coef. polynomial ?
>
> Thx to those who can help
> 

Arthur Herbert <herbert.arthur@caramail.com> wrote in message news:<417e4b4c$0$3812$8fcfb975@news.wanadoo.fr>...
> Hi,
> I am workin on a LPC AR estimator and i need to compute each of its poles.
> The denominator is a complex coefficients polynom of 120 degrees one.
> Does anyone knows a way to find the roots of a complex coef. polynomial ?
> 
> Thx to those who can help

I used Laguerre's method for complex polynomials of orders in the range 
10 to 20. I don't know how well the numerics would work for polynomials 
of orders 100+. When I read up on those things, I saw a lot of refernces 
to something called the "Bulrich (spelling?) & Stoer (spelling?)" method.
It seemed to be the method of choise for anybody who was seriuos about 
numerics, but I never found any material that showed how the method works.

Rune

Arthur Herbert wrote:
> Hi,
> I am workin on a LPC AR estimator and i need to compute each of its poles.
> The denominator is a complex coefficients polynom of 120 degrees one.
> Does anyone knows a way to find the roots of a complex coef. polynomial ?

This might help, or might be complete overkill (they factored
polynomials up to a degree of 1 000 000):

http://www.dsp.rice.edu/software/fvhdp.shtml

Regards,
Andor