pfb_channelizer_ccf.h

Go to the documentation of this file.

/* -*- c++ -*- */
/*
 * Copyright 2009,2010,2012 Free Software Foundation, Inc.
 *
 * This file is part of GNU Radio
 *
 * GNU Radio is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 3, or (at your option)
 * any later version.
 *
 * GNU Radio is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with GNU Radio; see the file COPYING.  If not, write to
 * the Free Software Foundation, Inc., 51 Franklin Street,
 * Boston, MA 02110-1301, USA.
 */


#ifndef INCLUDED_FILTER_PFB_CHANNELIZER_CCF_H
#define INCLUDED_FILTER_PFB_CHANNELIZER_CCF_H

#include <gnuradio/block.h>
#include <gnuradio/filter/api.h>

namespace gr {
namespace filter {

/*!
 * \brief Polyphase filterbank channelizer with
 *        gr_complex input, gr_complex output and float taps
 * \ingroup channelizers_blk
 *
 * \details
 * This block takes in complex inputs and channelizes it to <EM>M</EM>
 * channels of equal bandwidth. Each of the resulting channels is
 * decimated to the new rate that is the input sampling rate
 * <EM>fs</EM> divided by the number of channels, <EM>M</EM>.
 *
 * The PFB channelizer code takes the taps generated above and builds
 * a set of filters. The set contains <EM>M</EM>filters
 * and each filter contains ceil(taps.size()/decim) taps.
 * Each tap from the filter prototype is sequentially inserted into
 * the next filter. When all of the input taps are used, the remaining
 * filters in the filterbank are filled out with 0's to make sure each
 * filter has the same number of taps.
 *
 * Each filter operates using the gr::blocks::fir_filter_XXX
 * class of GNU Radio, which takes the input stream at <EM>i</EM>
 * and performs the inner product calculation to <EM>i+(n-1)</EM>
 * where <EM>n</EM> is the number of filter taps. To efficiently
 * handle this in the GNU Radio structure, each filter input must
 * come from its own input stream. So the channelizer must be
 * provided with <EM>M</EM> streams where the input stream has
 * been deinterleaved. This is most easily done using the
 * gr::blocks::stream_to_streams block.
 *
 * The output is then produced as a vector, where index <EM>i</EM>
 * in the vector is the next sample from the <EM>i</EM>th
 * channel. This is most easily handled by sending the output to a
 * gr::blocks::vector_to_streams block to handle the conversion
 * and passing <EM>M</EM> streams out.
 *
 * The input and output formatting is done using a hier_block2 called
 * pfb_channelizer_ccf. This can take in a single stream and outputs
 * <EM>M</EM> streams based on the behavior described above.
 *
 * The filter's taps should be based on the input sampling rate.
 *
 * For example, using the GNU Radio's firdes utility to building
 * filters, we build a low-pass filter with a sampling rate of
 * <EM>fs</EM>, a 3-dB bandwidth of <EM>BW</EM> and a transition
 * bandwidth of <EM>TB</EM>. We can also specify the out-of-band
 * attenuation to use, <EM>ATT</EM>, and the filter window
 * function (a Blackman-harris window in this case). The first input
 * is the gain of the filter, which we specify here as unity.
 *
 *    <B><EM>self._taps = filter.firdes.low_pass_2(1, fs, BW, TB,
 *       attenuation_dB=ATT, window=filter.firdes.WIN_BLACKMAN_hARRIS)</EM></B>
 *
 * The filter output can also be oversampled. The oversampling rate
 * is the ratio of the the actual output sampling rate to the normal
 * output sampling rate. It must be rationally related to the number
 * of channels as N/i for i in [1,N], which gives an outputsample rate
 * of [fs/N, fs] where fs is the input sample rate and N is the number
 * of channels.
 *
 * For example, for 6 channels with fs = 6000 Hz, the normal rate is
 * 6000/6 = 1000 Hz. Allowable oversampling rates are 6/6, 6/5, 6/4,
 * 6/3, 6/2, and 6/1 where the output sample rate of a 6/1 oversample
 * ratio is 6000 Hz, or 6 times the normal 1000 Hz. A rate of 6/5 = 1.2,
 * so the output rate would be 1200 Hz.
 *
 * The theory behind this block can be found in Chapter 6 of
 * the following book:
 *
 *    <B><EM>f. harris, "Multirate Signal Processing for Communication
 *       Systems," Upper Saddle River, NJ: Prentice Hall, Inc. 2004.</EM></B>
 *
 * When dealing with oversampling, the above book is still a good
 * reference along with this paper:
 *
 *    <B><EM>E. Venosa, X. Chen, and fred harris, “Polyphase analysis
 *       filter bank down-converts unequal channel bandwidths with
 *       arbitrary center frequencies - design I,” in SDR’10-WinnComm,
 *       2010.</EM></B>
 */
class FILTER_API pfb_channelizer_ccf : virtual public block
{
public:
    // gr::filter::pfb_channelizer_ccf::sptr
    typedef boost::shared_ptr<pfb_channelizer_ccf> sptr;

    /*!
     * Build the polyphase filterbank decimator.
     * \param numchans (unsigned integer) Specifies the number of
     *                 channels <EM>M</EM>
     * \param taps (vector/list of floats) The prototype filter to
     *             populate the filterbank.
     * \param oversample_rate (float) The oversampling rate is the
     *                                ratio of the the actual output
     *                                sampling rate to the normal
     *                                output sampling rate.  It must
     *                                be rationally related to the
     *                                number of channels as N/i for
     *                                i in [1,N], which gives an
     *                                outputsample rate of [fs/N,
     *                                fs] where fs is the input
     *                                sample rate and N is the
     *                                number of channels.
     *
     *                          For example, for 6 channels
     *                          with fs = 6000 Hz, the normal
     *                          rate is 6000/6 = 1000
     *                          Hz. Allowable oversampling
     *                          rates are 6/6, 6/5, 6/4, 6/3,
     *                          6/2, and 6/1 where the output
     *                          sample rate of a 6/1
     *                          oversample ratio is 6000 Hz,
     *                          or 6 times the normal 1000 Hz.
     */
    static sptr
    make(unsigned int numchans, const std::vector<float>& taps, float oversample_rate);

    /*!
     * Resets the filterbank's filter taps with the new prototype filter
     * \param taps (vector/list of floats) The prototype filter to populate the
     * filterbank.
     */
    virtual void set_taps(const std::vector<float>& taps) = 0;

    /*!
     * Print all of the filterbank taps to screen.
     */
    virtual void print_taps() = 0;

    /*!
     * Return a vector<vector<>> of the filterbank taps
     */
    virtual std::vector<std::vector<float>> taps() const = 0;

    /*!
     * Set the channel map. Channels are numbers as:
     * <pre>
     *  N/2+1 | ... | N-1 | 0 | 1 |  2 | ... | N/2
     * <------------------- 0 -------------------->
     *                     freq
     * </pre>
     *
     * So output stream 0 comes from channel 0, etc. Setting a new
     * channel map allows the user to specify which channel in frequency
     * he/she wants to got to which output stream.
     *
     * The map should have the same number of elements as the number
     * of output connections from the block. The minimum value of
     * the map is 0 (for the 0th channel) and the maximum number is
     * N-1 where N is the number of channels.
     *
     * We specify M as the number of output connections made where M
     * <= N, so only M out of N channels are driven to an output
     * stream. The number of items in the channel map should be at
     * least M long. If there are more channels specified, any value
     * in the map over M-1 will be ignored. If the size of the map
     * is less than M the behavior is unknown (we don't wish to
     * check every entry into the work function).
     *
     * This means that if the channelizer is splitting the signal up
     * into N channels but only M channels are specified in the map
     * (where M <= N), then M output streams must be connected and
     * the map and the channel numbers used must be less than
     * N-1. Output channel number can be reused, too. By default,
     * the map is [0...M-1] with M = N.
     */
    virtual void set_channel_map(const std::vector<int>& map) = 0;

    /*!
     * Gets the current channel map.
     */
    virtual std::vector<int> channel_map() const = 0;
};

} /* namespace filter */
} /* namespace gr */

#endif /* INCLUDED_FILTER_PFB_CHANNELIZER_CCF_H */