Viterbi traceback length

[tzushky]

:| Hello Hello,

I've been trying to find out the exact reason for needing 5L traceback length in the case of Viterbi decoding ( of unpunctured codes, as is stated generally). I just want a formula, a reason, an explanation why 5L and not for example less, because I'm using a Viterbi decoder for frequency modulation and I get the same result for 5L as for L+2 (L=3)... Does anyone know what book or article or reference can explain it?? (I've searched everywhere and all I get is " at least 5L has been proved to work...)

I'd appreciate it very much, thanks in advance.....

Bye:!:

 

[tzushky]

Nobody knows...? Pretty Plz

 

[bulx]

havent ever seen any proof of this. It must be an empirical result. Just like you find good convol codes through computer search.
-b

 

[spweda]

After 5 time constrain lengths, the output of traceback merges to one paths, no formal proof on this observation.