The bitsync is not random data. Especially with random data, the filtering starts working, reducing bandwidth. This is why with GMSK it is preferred that data be as random as possible, and why the scrambler is used (not JUST for reducing bit errors..) --f
________________________________ From: [email protected] [mailto:[email protected]] On Behalf Of Jonathan Naylor Sent: Thursday, May 14, 2009 9:24 PM To: [email protected] Subject: [DSTAR_DIGITAL] Re: External D-Star encoding. > I'm not sure the 4800hz statement is correct. The data stream is 4800 > bits per second which is put through GMSK before hitting the radio > which requires a pretty flat response, which is why it is injected > into the modulator directly (not through the audio chain) and pulled > directly from the discriminator. I don't think the 4800bps directly > maps to 4800hz. I would like to point you to the bit synchronisation sequence before the frame synchronisation pattern which consists of nothing but a 10101010101010... sequence which at 4800Bd and which will map to 4800 Hz, or very close. The actual data will be a bit less unless the 10101010.. sequence crops up. I can assure you that sampling the data at 48000 Hz and taking the bit value every ten samples works beautifully! Jonathan G4KLX
