Re: [Discuss-gnuradio] Removing Correlator False Tags

[Richard Bell]

OK I think I understand your suggestion. Thanks for the help. I'll let you know how it goes.

Rich

On Mon, Jul 20, 2015 at 8:33 AM, Marcus Müller <address@hidden> wrote:

Hi Rich,

this sounds like you shouldn't calculate the correlation for the part of the samples where you can't verify the validity. Let l_in == ninput_items[0] be the amount of items in your first call of work(), then just check the correlation for 0...(l_in - N) , N being the length you need to be positive that the correlation tag should be set, still consuming l_in.

You'd use set_history(N+1) to ensure the N last items of the last call are still present in this call to work. This way, you'd see the N "end" items twice.

Best regards,
Marcus

On 07/20/2015 04:20 PM, Richard Bell wrote:

Hi all,

 

I need some help solving this problem. To recap, myself and Andy Wall have come to the conclusion and agreed that the Correlation Estimator block will insert false correlation tags into a stream of data if the correlation peak happens to exist at the end or beginning of a block of notuput_items. I am writing a block that proceeds the Correlation Estimator block whose job it is to remove these false tags.

 

The scenario I'm having trouble dealing with is when a false tag is placed at the end of a block, because I won't know it's a false tag until I can check the first item of the next block. As far as I'm aware of, once I'm in the next block of noutput_items, I've lost the ability to change tags in the previous block (i.e. I can't effect tags in a previous call to work). How would you go about handling this, is there already a mechanism I'm not aware of?

 

Rich

 

 

 

_______________________________________________
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