[Bug-gsl] Discrete Hankel Transform - unexpected behaviour

[Luca Fascione]

Hello,
my name is Lucas Fascione, I work for Weta Digital (the guys from The 
Lord of the Rings and King Kong).

I'm using your dht code to transform some data I need to convolve.

In trying to understand how many samples of my function I needed I 
conducted the following experiment:
- I generated a dht class with fixed maximum value and various sample 
counts (128, 256, 512, 1024, 2048)
- I transformed and plotted the various datasets

I was very surprised when I realized the transformed dataset doesn't 
seem to converge to a defined target value,
but keeps lowering in magnitude instead.

Here is a gnuplot expression to plot my transformed datasets so that 
they (almost) overlap:

plot "trans.2048.dht0.dat" using ($1):(log($2)) with linespoints,
      "trans.1024.dht0.dat" using ($1):(log($2/2.8)) with lines,
      "trans.512.dht0.dat" using ($1):(log($2/8)) with lines,
      "trans.256.dht0.dat" using ($1):(log($2/23)) with lines,
      "trans.128.dht0.dat" using ($1):(log($2/63)) with lines
     
Any insight you could spare on the situation would be highly appreciated.

Thank you
Luca Fascione

--
Luca Fascione
Software Engineer - Weta Digital
Phone:  +64  4 380 9864 (x4744)
Mobile: +64 21 0764 862