|
From: | Nemanja Savic |
Subject: | Re: [Discuss-gnuradio] synchronization between blocks |
Date: | Tue, 4 Feb 2014 09:59:21 +0100 |
On Feb 3, 2014, at 10:14 AM, Nemanja Savic <address@hidden> wrote:IIRC: Because of the way GNU Radio internally handles data processing (it waits for data upstream to become available before -starting- processing for a given block), and since -your specific graph- contains only sync blocks (yes?), then the sample delay difference between the 2 branches of -your specific graph- should be constant. Further, it should be the same constant for every graph execution.
> at the moment I thought that the problem was solved, but actually something strange is going on. I would only like to ask whether delay between two paths can roll out somehow (i don't know if this verb exists). The point is that at the beginig of the execution, it looks like the script is doing fine, but then I realize signs of unalligned signals which I can't in some other way.
If any blocks within a specific path are not sync blocks, e.g., a packet decoder, then the sample delay difference is not guaranteed (constant or otherwise).
The wall-clock delay difference will not be constant; it can be any non-negative (and, generally > 0) time.
All of the above said: If you're into compiling GNU Radio, you can go into gnuradio-runtime/lib/block_executor.cc and set "ENABLE_LOGGING" to 1. Recompile/reinstall GNU Radio, then when you execute your graph you'll get a printout dump of what each block is doing. That printout might help you debug this issue; or, not. You might also add LOG messages in that file which could help ... maybe; or, not.
Hope this helps! - MLD
_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Prev in Thread] | Current Thread | [Next in Thread] |