Hi Martin,
Thanks for your detailed answer. I really appreciate the great effort
you put into explaining how things work. I am still on the learning curve.
I used your suggestions to the best of my understanding and it worked
beautifully for the one sync pattern test vector in the QA test.
Then, I took your suggestion for repeated sync patterns using an init
flag which I reset to 0 to restart the process. For the QA test, I
repeated the original input data twice (now I have 2 sync patterns), so
the expected QA output should be the original output repeated twice. I
modified the C code by adding an if statement at the end to check if
noutput_items == Buf_size+message_size (buf_size is the length of the
pattern, which I call preamble) and if it is, I reset the init flag to
zero as well as other params used in the initialization section of the
code. The QA test failed by producing an output with 3 repeated copies
of the original output rather than the expected 2 copies. I do not
expect you to send too much time looking at my code below, however, I
would appreciate it very much, if you would glance at it to see if you
can spot what I am doing wrong. The test to un-initialize (setting init
to 0) was done towards the end of the code block after the consume method.
int kk = 0;____
for (int i = 0; i < ninput_items[0]; i++)____
{____
if(!init){__
cnt += 1; // cnt number of passing bytes____
kk = initialize(in[i]);____
if (kk == 0){____
noutput_items = cnt;____
}else{____
memcpy((void*)out, (const void*)preamble, buf_size);____
noutput_items = cnt;____
}
} else {
out[i-cnt+buf_size] = in[i];____
noutput_items = buf_size + message_size;____
}____
} ____
consume_each (noutput_items);____
if (noutput_items == buf_size + message_size){____
init = 0; // re-initialize all____
cnt = 0;____
kk = 0;____
}____
return noutput_items;____
}
Thanks very much for the help.
Regards,
George
On Thu, Sep 10, 2020 at 12:06 AM Martin Luelf <mail@mluelf.de
<mailto:mail@mluelf.de>> wrote:
Dear George,
this also caused me a lot of headache when I started with GNURadio, so
here is what I learned about it.
Let's start with the forecast method. This tells GNURadio how many
input
samples you need on each input stream to generate the requested number
of output items. Usually GNURadio will run your forecast function a
couple of times with different output numbers to find a good data chunk
size to give to the general work function. Keep in mind that the number
of required input items you give here is a minimum number and GNURadio
might decide to give you more input data than you requested. It is also
important to know that the number of samples you request here is
just an
estimate. You are not forced to actually process that much data.
Now to the general_work function. noutput_items tells you how many
samples GNURadio would like you to produce (this is a maximum number
you
may produce less). It also tells you how much memory is allocated in
every array in the output_items vector. If you have only one output and
you used the default <out type> *out = (<out type> *) output_items[0];
definition this tells you how many items you can place (at most) into
the out array.
The ninput_items[] array tells you how many input items are
available in
the input arrays. Again if you just have one input and you use const
<in
type> *in = (const <in type> *) input_items[0]; ninput_items[0] is the
number of inputs available in the in array. You may not read more items
than that from the array.
Within the given input symbols you can start looking for your sync
pattern. If you generate output you have to write (in your case
copy) it
to the out array. At the end of general_work you call consume(0, K)
with
the number of input items K that you have consumed on input 0. That is
how many of the input items you have used and do not need to see again.
If you consume 0 symbols the next call to general_work will show you
the
exact same input samples again. If you consume ninput_items[0] you will
see completely new input samples on the first input the next time
general_work is called. And then you return the number of samples you
generated (i.e. how much samples you put into the out array). This
number must be smaller or equal noutput_items, because otherwise you
would have written out of the allocated memory which might result in a
segfault/core dump. Note that you don't have to call consume at the
very
end of work and there is also another way of telling GNURadio how many
samples you have produced, but let's leave that for another day.
So a very easy (but not the most efficient) setup for your problem
could be:
Add a boolean flag to your _impl class that both forecast and
general_work can read/write to. This flag will indicate whether or not
you have found the sync pattern or not. You initialize this flag
with false.
Assume you have a sync pattern of length L and a message with M data
symbols afterwards.
In forecast if the flag is set (meaning you have found the sync
pattern)
you need L+M symbols of input. If the flag is not set you need L input
samples, regardless of how many output samples GNURadio wants you to
generate.
In general work if the flag is false you search the input for the sync
pattern. If you found it at position i (counting from 0) you set the
flag to true, consume i samples (i.e. everything before the sync
marker). If the sync marker is not found you keep the flag to false and
consume the inputs that you have searched so far. In both cases you
return 0 since you have not generated any output yet.
If the flag is true you copy the first L+M samples from the input to
the
output, you set the flag to false (because after the data you have to
start searching for the sync marker again) you consume L+M samples and
return L+M samples.
Note: This is a very easy to understand scheme, but unfortunately not
very efficient. You only process a single block of either unwanted
spurious symbols, or one sync marker and data at a time. So once you
have a good understanding of how this works you should tweak that block
to be able to process multiple blocks of spurious symbols and sync
patterns/data within once call to general_work. It uses the same
kind of
logic, but requires more housekeeping of counters and indices.
If your input is symbols rather than bits/bytes you should also look at
the paper from J. Massey "Optimum Frame Synchronization" from 1972 on
how to perform optimum sync marker detection, which performs better
than
the intuitive correlation search.
Hope that gets you started.
Yours
Martin
On 10.09.20 04:34, George Edwards wrote:
> Hello,
>
> I am writing an OOT block in C++ that receives a sequence of
numbers and
> searches through for a sync pattern and passes to its output the
sync
> pattern and the bytes of data which follows it. The QA test
shows the
> block recognizes the pattern and will pass the pattern along with
the
> data that follow it, but there is a problem. The block does not
know a
> priori the number of spurious bytes preceding the sync pattern of
bytes,
> so I cannot set up the true relationship between the ninput_items
and
> noutput_items. However, the block can count the number of bytes that
> came in before the pattern. This is my problem:
>
> 1) In the OOT general_work method: If I set noutput_items =
> ninput_items[0], then in addition to passing the correct data to the
> output, it passes trailing zeros equal to the number of spurious
bytes
> that entered before the pattern.
>
> 2) If I set the return noutput_items = ninput_items - cnt (where
cnt is
> the number of spurious bytes before the pattern) depending on
where I
> put noutput_items in the code, it either throws a core dump or
cuts off
> the number of true data.
>
> Also, within the forecast method, I simply use a _for_ loop and set
> ninput_items_required[i]=noutput_items;
>
> I will appreciate any help to point me in the right direction in
dealing
> with this non-regular output/ inputs relationship.
>
> Thank you!
>
> George