AM demodulation question

[amohar]

Hello everyone,

I'm trying to understand the demodulation theory and thus understand the GnuRadio graphs implementing software demodulation. I'm still a bit stuck on the bare basics, so here are my questions:

1. I understand that the rational resampler can be used to lower the number of received samples by using the decimation parameter. This is helpful for speeding up the software operations while still keeping the quality. Low-level filter also has the decimation parameter, why isn't that one usually used for decimation? In what case would we use filter's decimation?

2. I understand that the AM demod (for instance) usually has two steps, rectifying the signal to baseband signal and then applying low-pass filter to get rid of unwanted high frequencies (leftovers from the modulation) and get a nice wave. My question is why do we need the low-pass filter before the AM demod (in fact WBFM as well)? Some say that it is for band selection, but shouldn't it then be the band-pass filter?

3. Is there any good information about Frequency xlating FIR filter? I know it can use firdes Python filters, but still haven't got to grips with them. What are the Taps (I know this should be the place for firdes Python call to low_pass_2, but what does it actually mean)? I guess the center frequency is the one I want to tune into... Is this the same frequency as set on SDR input source? Why is this center frequency missing in the low-pass filter block? If this frequency is translated to 0Hz (as docs say), do I still get my left channel?

Just by writing these questions down I've made some things clearer to myself. I would still be very grateful if anyone would clarity on these things even more. Also, although I've read anything and everything online that I could find regarding this topic, I would appreciate any resources that you might think of as helpful.

Thank you for any info, Andrej

[KD6W]

GNU SDR expert I am not, I know enough to get into trouble. But as the name of two objects implies, 1. The rational resampler is just that, the "Low Pass" filter also has a decimation process as a function to create a filter to make a cut off hence, a low pass filter. 2. AM demod can be swamped with alias energy coming from above the desired tuning range so a low pass in front of the highest tuneable range (pass-band) is applied to knock down unwanted signals from above the area of interest. Signal energy from below doesn't matter since it won't introduce an alias and a low pass filter is much easier to design and have sharper slopes than a pass-band filter. 3. Just like you i have those SAME questions and wait for soul to appear on this planet to write the book and teach the courses on GNU radio! HNY - 73