1: \begin{abstract}

2: This paper presents a new technique for achieving spectral accuracy and 

3: fast computational performance in a class of black hole perturbation and 

4: gravitational self-force calculations involving extreme mass ratios and 

5: generic orbits.  Called \emph{spectral source integration} (SSI), this 

6: method should see widespread future use in problems that entail (i) 

7: point-particle description of the small compact object, (ii) frequency 

8: domain decomposition, and (iii) use of the background eccentric geodesic 

9: motion.  Frequency domain approaches are widely used in both perturbation 

10: theory flux-balance calculations and in local gravitational self-force 

11: calculations.  Recent self-force calculations in Lorenz gauge, using the 

12: frequency domain and method of extended homogeneous solutions, have been 

13: able to accurately reach eccentricities as high as $e \simeq 0.7$.  We 

14: show here SSI successfully applied to Lorenz gauge.  In a double precision 

15: Lorenz gauge code, SSI enhances the accuracy of results and makes a factor 

16: of three improvement in the overall speed.  The primary initial application of 

17: SSI--for us its \emph{raison d'\^{e}tre}--is in an arbitrary precision 

18: \emph{Mathematica} code that computes perturbations of eccentric orbits 

19: in the Regge-Wheeler gauge to extraordinarily high accuracy (e.g., 200 

20: decimal places).  These high accuracy eccentric orbit calculations would not 

21: be possible without the exponential convergence of SSI.  We believe the 

22: method will extend to work for inspirals on Kerr, and will be the subject of 

23: a later publication.  SSI borrows concepts from discrete-time signal 

24: processing and is used to calculate the mode normalization coefficients 

25: in perturbation theory via sums over modest numbers of points around an 

26: orbit.  A variant of the idea is used to obtain spectral accuracy in 

27: solution of the geodesic orbital motion.

28: 

29: \end{abstract}

30: