[Lowfer] 136 Dial freq for OP-OPDS 32 for tonight...,

[JD]

>>>So, is it true that OPDS can only decode callsigns that it has in its 
>>>database? In other words, if a new station started transmitting 136 kHz 
>>>Op32, there is zero chance that OPDS would decode it, though "vanilla" 
>>>Opera could?

No, to the extent I understand it, OPDS can decode anything new that's got 
enough signal to noise for regular OPERA to decode.

It appears to me to work something similar to WSPR's deep search.  The 
difference is in the _partial_ decodes, which the detection software would 
not be able to declare positively as this or that station on its own. 
Marcus himself describes the improvement as both a function of detection AND 
IDENTIFICATION.  He adds, "It attempts to do a correlation between received 
signals and templates from a predefined search list of callsigns."  In other 
words, by comparing the recognizable pieces of actual decoded signal to an 
expected pattern, it can identify the _probability_ of it being this or that 
station.  Hence, the correlation factor in the display.

Personally, the lower the correlation, the more I would be inclined to call 
that a question of pattern recognition rather than a decode.

But before anyone flames me for saying anything beyond what I actually did, 
however, let me state clearly that pattern recognition algorithms are a 
valid tool in some types of communication, and I'm not saying anybody 
shouldn't pursue any tool they like.  All I've been saying is that 
terminology matters, and a lot of misimpressions...and even wishful 
thinking...seem to arise from loose use thereof.

John