src/spandsp/playout.h

/*
 * SpanDSP - a series of DSP components for telephony
 *
 * playout.h
 *
 * Written by Steve Underwood <steveu@coppice.org>
 *
 * Copyright (C) 2005 Steve Underwood
 *
 * All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU Lesser General Public License version 2.1,
 * as published by the Free Software Foundation.
 *
 * This program 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 Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 * $Id: playout.h,v 1.11 2008/04/17 14:27:00 steveu Exp $
 */

#if !defined(_SPANDSP_PLAYOUT_H_)
#define _SPANDSP_PLAYOUT_H_

/*! \page playout_page Play-out (jitter buffering)
\section playout_page_sec_1 What does it do?
The play-out module provides a static or dynamic length buffer for received frames of
audio or video data. It's goal is to maximise the receiver's tolerance of jitter in the
timing of the received frames.

Dynamic buffers are generally good for speech, since they adapt to provide the smallest delay
consistent with a low rate of packets arriving too late to be used. For things like FoIP and
MoIP, a static length of buffer is normally necessary. Any attempt to elastically change the
buffer length would wreck a modem's data flow.
*/

/* Return codes */
enum
{
    PLAYOUT_OK = 0,
    PLAYOUT_ERROR,
    PLAYOUT_EMPTY,
    PLAYOUT_NOFRAME,
    PLAYOUT_FILLIN,
    PLAYOUT_DROP
};

/* Frame types */
#define PLAYOUT_TYPE_CONTROL    0
#define PLAYOUT_TYPE_SILENCE    1
#define PLAYOUT_TYPE_SPEECH     2

typedef int timestamp_t;

typedef struct playout_frame_s
{
    /*! The actual frame data */
    void *data;
    /*! The type of frame */
    int type;
    /*! The timestamp assigned by the sending end */
    timestamp_t sender_stamp;
    /*! The timespan covered by the data in this frame */
    timestamp_t sender_len;
    /*! The timestamp assigned by the receiving end */
    timestamp_t receiver_stamp;
    /*! Pointer to the next earlier frame */
    struct playout_frame_s *earlier;
    /*! Pointer to the next later frame */
    struct playout_frame_s *later;
} playout_frame_t;

/*!
    Playout (jitter buffer) descriptor. This defines the working state
    for a single instance of playout buffering.
*/
typedef struct
{
    /*! TRUE if the buffer is dynamically sized */
    int dynamic;
    /*! The minimum length (dynamic) or fixed length (static) of the buffer */
    int min_length;
    /*! The maximum length (dynamic) or fixed length (static) of the buffer */
    int max_length;
    /*! The target filter threshold for adjusting dynamic buffering. */
    int dropable_threshold;

    int start;

    /*! The queued frame list */
    playout_frame_t *first_frame;
    playout_frame_t *last_frame;
    /*! The free frame pool */
    playout_frame_t *free_frames;

    /*! The total frames input to the buffer, to date. */
    int frames_in;
    /*! The total frames output from the buffer, to date. */
    int frames_out;
    /*! The number of frames received out of sequence. */
    int frames_oos;
    /*! The number of frames which were discarded, due to late arrival. */
    int frames_late;
    /*! The number of frames which were never received. */
    int frames_missing;
    /*! The number of frames trimmed from the stream, due to buffer shrinkage. */
    int frames_trimmed;

    timestamp_t latest_expected;
    /*! The present jitter adjustment */
    timestamp_t current;
    /*! The sender_stamp of the last speech frame */
    timestamp_t last_speech_sender_stamp;
    /*! The duration of the last speech frame */
    timestamp_t last_speech_sender_len;

    int not_first;
    /*! The time since the target buffer length was last changed. */
    timestamp_t since_last_step;
    /*! Filter state for tracking the packets arriving just in time */
    int32_t state_just_in_time;
    /*! Filter state for tracking the packets arriving late */
    int32_t state_late;
    /*! The current target length of the buffer */
    int target_buffer_length;
    /*! The current actual length of the buffer, which may lag behind the target value */
    int actual_buffer_length;
} playout_state_t;

#if defined(__cplusplus)
extern "C"
{
#endif

/*! Queue a frame
    \param s The play-out context.
    \param data The frame data.
    \param sender_len Length of frame (for voice) in timestamp units.
    \param sender_stamp Sending end's time stamp.
    \param receiver_stamp Local time at which packet was received, in timestamp units.
    \return One of
        PLAYOUT_OK:  Frame queued OK.
        PLAYOUT_ERROR: Some problem occured - e.g. out of memory. */
int playout_put(playout_state_t *s, void *data, int type, timestamp_t sender_len, timestamp_t sender_stamp, timestamp_t receiver_stamp);

/*! Get the next frame.
    \param s The play-out context.
    \param frame The frame.
    \param sender_stamp The sender's timestamp.
    \return One of
        PLAYOUT_OK:  Suitable frame found.
        PLAYOUT_DROP: A frame which should be dropped was found (e.g. it arrived too late).
                      The caller should request the same time again when this occurs.
        PLAYOUT_NOFRAME: There's no frame scheduled for this time.
        PLAYOUT_FILLIN: Synthetic signal must be generated, as no real data is available for
                        this time (either we need to grow, or there was a lost frame).
        PLAYOUT_EMPTY: The buffer is empty.
 */
int     playout_get(playout_state_t *s, playout_frame_t *frame, timestamp_t sender_stamp);

/*! Unconditionally get the first buffered frame. This may be used to clear out the queue, and free
    all its contents, before the context is freed.
    \param s The play-out context.
    \return The frame, or NULL is the queue is empty. */
playout_frame_t *playout_get_unconditional(playout_state_t *s);

/*! Find the current length of the buffer.
    \param s The play-out context.
    \return The length of the buffer. */
timestamp_t playout_current_length(playout_state_t *s);

/*! Find the time at which the next queued frame is due to play.
    Note: This value may change backwards as freshly received out of order frames are
          added to the buffer.
    \param s The play-out context.
    \return The next timestamp. */
timestamp_t playout_next_due(playout_state_t *s);

/*! Create a new instance of play-out buffering.
    \param min_length Minimum length of the buffer, in samples.
    \param max_length Maximum length of the buffer, in samples. If this equals min_length, static
           length buffering is used.
    \return The new context */
playout_state_t *playout_new(int min_length, int max_length);

/*! Destroy an instance of play-out buffering.
    \param s The play-out context to be destroyed */
void playout_free(playout_state_t *s);

/*! Reset an instance of play-out buffering.
    NOTE:  The buffer should be empty before you call this function, otherwise
           you will leak queued frames, and some internal structures
    \param s The play-out context.
    \param min_length Minimum length of the buffer, in samples.
    \param max_length Maximum length of the buffer, in samples. If this equals min_length, static
           length buffering is used. */
void playout_restart(playout_state_t *s, int min_length, int max_length);

#if defined(__cplusplus)
}
#endif

#endif
/*- End of file ------------------------------------------------------------*/