Re: [Discuss-gnuradio] Questions regarding hierarchical polyphase channelizer

[Piotr Krysik]

Thanks Ben for the answer.

Regarding the last question, I want to clarify a bit what I meant.
Regular polyphase channelizer have oversample_rate option that lets you
set oversampling so you can get output sample rates:
 input_samp_rate/number_of_channels * oversample_rate

where oversample_rate = number_of_channels/i
and where i is an integer number from 1 to number_of_channels.

This is very usable in instances (like GSM or Tetra) when channel
spacing (200e3 Hz for GSM) is different than symbol rate (13e6/48 Hz for
GSM).
If you have sample rate at the input of the channelizer that is common
multiply of channel spacing and symbol rate, with this option you are
able to get signal at the output of the channelizer at exactly symbol
rate (or even its integer multiply in some cases).

For example for 13e6 GSM band I can extract all channels by setting
Channels parameter of the channelizer to 65 (13e6/0.2e6).
Then I can get output signal that is sampled with 4*gsm_symbol_rate by
setting oversampling to 65/12.

It would be great to have something like this in your block because it
is more efficient than the regular one. I can look into this if I will
find some time.

Best Regards,
Piotr Krysik

W dniu 28.10.2016 o 23:14, Ben Reynwar pisze:
> The algorithm should be identical in the hierarchical case.  The only
> difference is that because the hierarchical implementation splits the
> logic over several signal processing blocks, GNURadio is able to
> easily parallelize the implementation.  The n_filterbanks parameter is
> just the number of filterbank blocks that are used for the filtering.
> More filterbank blocks means more parallelizable, but also more
> overhead.  I'd have to spend some time reacquainting myself with the
> code to answer your last question properly, but it should be possible
> to do everything with the hierarchical implementation that you can
> with the single-block implementation.
>
> On Mon, Oct 24, 2016 at 6:30 AM, Piotr Krysik <address@hidden> wrote:
>> Hi all,
>>
>> Recently I did simple benchmark of hierarchical polyphase channelizer
>> block (written by Ben Reynwar according to git blame) vs regular
>> polyphase channelizer block. It seems that hierarchical version is 2-3
>> times faster.
>>
>> The documentation of the hierachical polyphase channelizer seems to be a
>> bit too brief though (at least for me). This topic has been touched on
>> this list before, but produced no more info about the block (in the
>> "hierarchical polyphase channelizer vs polyphase channelizer blocks in
>> GRC" thread).
>>
>> A bit more info is needed about the block's parameters and how it
>> capabilities relate to normal polyphase channelizer block. Especially:
>>
>> -what n_filterbanks parameter is for?
>>
>> -is it possible to do oversampling of the signal at the output channels
>> with this block the way it is possible with normal polyphase channelizer
>> block?
>>
>>