|
| From: | Richard Bell |
| Subject: | Re: [Discuss-gnuradio] DSSS sync question |
| Date: | Fri, 5 Feb 2016 15:46:05 -0800 |
The last thing I wonder is, can a receiver just pick up a DSSS signal and start applying the despreading code? I watched a YouTube video about this and the example involved multiplying the spreading code by the voltages of the composite waveform and averaging them. My system takes 16 chips to express 1 bit. Let’s say my demodulator starts on the 10th chip and goes on for 16 chips, getting 6 chips from this cycle and 10 from the next. If it keeps on like this, it will never fall into sync, and without being in sync it can’t get any real bits to help itself align.Sent from Windows MailFrom: Richard Bell
Sent: Friday, February 5, 2016 5:59 PM
To: Henry Barton, address@hiddenIt will depend on how the rest of the radio is built up. I'm not familiar with VP9, but can I assume it's a spec on bits in a higher layer then Layer 1? Another words, you are assuming you have bits to correlate with, as opposed to wave shapes?You're getting into the difficulties of radio design now. You need to fully understand the needs of your system to make decisions like this. You don't have bits until you've synchronized and demodulated your signal. If you require some sort of FEC, it will need block alignment before you can decode it, so the correlator will need to be in the waveshape domain. You can still use the known VP9 headers to correlate to in this cas, but you wouldn't correlate to the bit version, you would correlate to the modulated version of those headers.P.S. Please reply to the mailing list, so others can see and reply if need be.RichOn Fri, Feb 5, 2016 at 2:53 PM, Henry Barton <address@hidden> wrote:I'm hoping to transmit a VP9 transport stream, so perhaps the predictable headers will be enough?Sent from Windows MailFrom: Richard Bell
Sent: Friday, February 5, 2016 5:51 PM
To: Henry Barton, address@hiddenSo long as you know what you're looking for in any given scenario, you can use that to correlate to. It can be data or a preamble. If your receiver knows the data will always be a certain way ahead of time though, it's hard to call that data. Semantics at that point.RichOn Fri, Feb 5, 2016 at 2:44 PM, Henry Barton <address@hidden> wrote:That sounds great, Richard. But I wonder, what if the useful payload contains that sequence by chance?Sent from Windows MailFrom: Richard Bell
Sent: Friday, February 5, 2016 5:27 PM
To: Henry Barton
Cc: address@hiddenTypically a correlator is used to look for a known sequence of bits, so the radio can align the rest of the processing from the end of this known sequence. This is referred to as frame synchronization. You could use the correlation estimation block to implement something like this. It would place a tag on the stream when it finds your known sequence and you would then know how everything is aligned from then on.RichOn Fri, Feb 5, 2016 at 1:04 PM, Henry Barton <address@hidden> wrote:Hi all. I've successfully written a DSSS modulator and demodulator in Windows with a chip rate of 16x. It writes samples to a file that the demodulator can read and despread. Before I try any practical implementations, I need to know how a DSSS stream would be synchronized. Assuming the transmitter and receiver were perfectly clocked in unison, what stops the receiver from tuning in in the middle of a byte, thus getting a nibble from the current byte and a nibble from the next?
_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
| [Prev in Thread] | Current Thread | [Next in Thread] |