What is the optimal sample rate for p25 capture?

[nycap]

Hello everyone. Does anyone know what sample rate should i use to capture p25 iq baseband for later playback? some say oversampling is good. i have heard that oversampling is not good. i want to get intelligable audio while letting my machine run as slowly as possible so i would think that i would like to get as low as possible. however i am using the newsky e4k dongle and the test says the minimum sample rate is 900k! this cannot be right. can it?

Thanks!

[DSheirer]

Originally Posted by nycap

however i am using the newsky e4k dongle and the test says the minimum sample rate is 900k! this cannot be right. can it?

Thanks!

If you're using gnuradio, try the Frequency Translating FIR filter to get the sample rate down from 900k down to 2-4.5 times the highest frequency component, in addition to translating your signal from the left or right of the LO frequency over to baseband (0 Hz). If you're writing your own code, you can try a cascaded series of half-band FIR filters to cut the sample rate in half with each filter.

I think someone already wrote a P25 decoder app or a 4FSK demodulator in gnuradio that you could study to see how they set it up. If the sample app is fed from a USRP, the source sample rates can be even higher than 900k.

[nycap]

Thanks DSheirer! I know these are real amateur questions but I know nothing of signal processing and have to learn so. . . So your saying that the goal of the filters is to lower the sample rate to between 2 and 4.5 times the highest frequency component. What is the "frequency component"?

[DSheirer]

Originally Posted by nycap

Thanks DSheirer! I know these are real amateur questions but I know nothing of signal processing and have to learn so. . . So your saying that the goal of the filters is to lower the sample rate to between 2 and 4.5 times the highest frequency component. What is the "frequency component"?

So, let your learning begin ... Smith posted his Scientists and Engineer's Guide to DSP online ... and you can read it at:

The Scientist and Engineer's Guide to Digital Signal Processing

Read the whole thing, but specifically read chapter three where he describes sampling rates.

If you find that you really like DSP, I've learned a lot from Lyon's book: Understanding Digital Signal Processing. I've been studying it for 6 or 7 months and find that I keep learning something new every time I reread each chapter 7 - 15 times over ... it eventually sinks in :-)

To answer your question ... filtering reduces/eliminates the higher frequency components before you decimate, so that through decimating (reducing) the sample rate, those higher frequency components don't fold down (alias) on top of your lower frequency components, where your P25 signal frequency components are ... you'll have trouble decoding the signal if it's corrupted by aliasing.

Denny

[nycap]

Thanks for the link Denny! This book looks great becease I never took topology and I don't want to dust off my old calc3 book. And the other books are upwards of $200 some of them. I will definetly be using this book. But for now the quick and dirty will do. So I think p25 goes at 9600 right? So then the channel rate in the decoder should be 19.2k up to 38.4k?