[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Problem in using trellis Error, correction en- an

From: Francisco Llaryora
Subject: Re: [Discuss-gnuradio] Problem in using trellis Error, correction en- and decoder
Date: Fri, 5 Nov 2010 20:39:56 -0300

in the past Achilleas wrote:-------
There are a couple of potential problems with your flowgraph:

1) the trellis encoder outputs unpacked bytes. For example, if your FSM has output cardinality 4 then it outputs 2 useful bits per byte.
Are you sure you are handling this the right way at the packet encoder?
I do not know how that block works but you should check it.

however, the main problem is the following

2) You need to substitute the "trellis viterbi" by the "trellis viterbi combo" block and input as parameters the kind of "distance" metric you want and the corresponding constellation, etc. To do that you need to know what is the output of the packet decoder...
I have never worked with packet en/decoder so i cannot help with these.

-------------end of quote.
well, I first want to thank you, because it success. (The problem is solved).
I did simplify the example: a source, a trellis and a viterbi combo.
Next, change the types for bytes.
Next use 'packed to unpacked' in the head.and 'unpacked to packed' in the tail of the flow.

i test with noise source the trellis, trellis-D8psk and trellis-DBPSK.
 trellis-DBPSK in noise is better than trellis in noise .
but  trellis in noise  is better than  trellis-D8psk in noise.

DBPSK Hamming distance=2.
trellis Hamming distance=sqrt(2).
D8pskHamming distance<sqrt(2).

the snapshot...where are the snapshot?


reply via email to

[Prev in Thread] Current Thread [Next in Thread]