|
Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.naic.edu/alfa/galfa/docs/galspect/signalproc/lpfcoefficients.html
Дата изменения: Mon Oct 4 23:19:44 2004 Дата индексирования: Sun Dec 23 06:27:44 2007 Кодировка: Поисковые слова: п п п п п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п |
The file lpf_coefficients.txt, which is created automatically as part of the filter generator's output, contains the actual values of the low-pass filter coefficients. The values are all real numbers converted from the fixed-point value used in hardware, which results in the quantizing effect of the bit length. Taking a DFT of these coefficients will give a more realistic jagged noise floor.
The file contains a few comments about the filter in lines beginning with a "#". The bit width of the coefficients, the decimation of the filter, width of the filter and such are stored here. The current filter has a width of 24 and a decimation of 14, so there are 14*24=336 lines of filter coefficients in the rest of the file.
It is a symmetric even filter so the first 168 values are the same as the last 168 values in reverse order.
In the hardware, the gain of the filter is scaled by the "gain" comment at the beginning of the file (about 1.05) to accommodate for the window function and the truncated sin(x)/x to give a mean gain of 1.0 for the filter. This scaling is not built into the coefficients and is done after the filter.