|Status:||Archived||Start date:||03 Jun 2016|
|Target version:||Sprint 31|
use HexCylinders.py from gisas/support/015_AnatolyBerezkin and add the following line to create_simulation():
#2 Updated by herck almost 3 years ago
- Description updated (diff)
- Status changed from Sprint to Resolved
Negative values arise as a result of the finite precision in the fft's. In special cases, as in the given example, where big plateaus with zero values are interspersed with high and narrow peaks, the convolution using fft gives some quantization noise.
Current solution: truncate negative values.
Note that this noise is quite small compared to the signal, so this should not influence simulation results.