Merge remote branch 'jblum/wip/gr-audio' into wip/gr-audio

* jblum/wip/gr-audio:
  audio: high prio for platform specific audio osx
  audio: added windows and osx audio source files
  audio: added config checks for other audios, added jack and port
  audio: make prefs look like old audio, removed old audio.py
  audio: register arches with priorities, ex: prefer alsa over oss
  audio: added oss support to gr-audio
  audio: use conf files to specify default arch
  audio: moved the grc audio blocks into gr-audio dir
  audio: moved alsa support files into subdirectory
  audio: work on swig support, it imports
  audio: work to get alsa support in, lets see if this works before adding more
  audio: added audio factory registry and top level includes
  audio: skeleton for gr-audio component

5 files changed, 1685 insertions, 0 deletions

diff --git a/gr-audio/lib/osx/audio_osx.h b/gr-audio/lib/osx/audio_osx.h
new file mode 100644
index 000000000..79e79e36c
--- /dev/null
+++ b/gr-audio/lib/osx/audio_osx.h
@@ -0,0 +1,71 @@
+/* -*- c++ -*- */
+/*
+ * Copyright 2006 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_AUDIO_OSX_H
+#define INCLUDED_AUDIO_OSX_H
+
+#include <iostream>
+#include <string.h>
+
+#define CheckErrorAndThrow(err,what,throw_str)				\
+  if (err) {								\
+    OSStatus error = static_cast<OSStatus>(err);			\
+    char err_str[4];							\
+    strncpy (err_str, (char*)(&err), 4);				\
+    std::cerr << what << std::endl;					\
+    std::cerr << "  Error# " << error << " ('" << err_str		\
+	      << "')" << std::endl;					\
+    std::cerr << "  " << __FILE__ << ":" << __LINE__ << std::endl;	\
+    fflush (stderr);							\
+    throw std::runtime_error (throw_str);				\
+  }
+
+#define CheckError(err,what)						\
+  if (err) {								\
+    OSStatus error = static_cast<OSStatus>(err);			\
+    char err_str[4];							\
+    strncpy (err_str, (char*)(&err), 4);				\
+    std::cerr << what << std::endl;					\
+    std::cerr << "  Error# " << error << " ('" << err_str		\
+	      << "')" << std::endl;					\
+    std::cerr << "  " << __FILE__ << ":" << __LINE__ << std::endl;	\
+    fflush (stderr);							\
+  }
+
+#ifdef WORDS_BIGENDIAN
+#define GR_PCM_ENDIANNESS kLinearPCMFormatFlagIsBigEndian
+#else
+#define GR_PCM_ENDIANNESS 0
+#endif
+
+// Check the version of MacOSX being used
+#ifdef __APPLE_CC__
+#include <AvailabilityMacros.h>
+#ifndef MAC_OS_X_VERSION_10_6
+#define MAC_OS_X_VERSION_10_6 1060
+#endif
+#if MAC_OS_X_VERSION_MAX_ALLOWED < MAC_OS_X_VERSION_10_6
+#define GR_USE_OLD_AUDIO_UNIT
+#endif
+#endif
+
+#endif /* INCLUDED_AUDIO_OSX_H */
diff --git a/gr-audio/lib/osx/audio_osx_sink.cc b/gr-audio/lib/osx/audio_osx_sink.cc
new file mode 100644
index 000000000..f7aeb54f8
--- /dev/null
+++ b/gr-audio/lib/osx/audio_osx_sink.cc
@@ -0,0 +1,404 @@
+/* -*- c++ -*- */
+/*
+ * Copyright 2006-2011 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.
+ */
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
+#include "gr_audio_registry.h"
+#include <audio_osx_sink.h>
+#include <gr_io_signature.h>
+#include <stdexcept>
+#include <audio_osx.h>
+
+#define _OSX_AU_DEBUG_ 0
+
+AUDIO_REGISTER_SINK(REG_PRIO_HIGH, osx)(
+    int sampling_rate, const std::string &device_name, bool ok_to_block
+){
+    return audio_sink::sptr(new audio_osx_sink(sampling_rate, device_name, ok_to_block));
+}
+
+audio_osx_sink::audio_osx_sink (int sample_rate,
+				const std::string device_name,
+				bool do_block,
+				int channel_config,
+				int max_sample_count)
+  : audio_sink ("audio_osx_sink",
+		   gr_make_io_signature (0, 0, 0),
+		   gr_make_io_signature (0, 0, 0)),
+    d_sample_rate (0.0), d_channel_config (0), d_n_channels (0),
+    d_queueSampleCount (0), d_max_sample_count (0),
+    d_do_block (do_block), d_internal (0), d_cond_data (0),
+    d_OutputAU (0)
+{
+  if (sample_rate <= 0) {
+    std::cerr << "Invalid Sample Rate: " << sample_rate << std::endl;
+    throw std::invalid_argument ("audio_osx_sink::audio_osx_sink");
+  } else
+    d_sample_rate = (Float64) sample_rate;
+
+  if (channel_config <= 0 & channel_config != -1) {
+    std::cerr << "Invalid Channel Config: " << channel_config << std::endl;
+    throw std::invalid_argument ("audio_osx_sink::audio_osx_sink");
+  } else if (channel_config == -1) {
+// no user input; try "device name" instead
+    int l_n_channels = (int) strtol (device_name.data(), (char **)NULL, 10);
+    if (l_n_channels == 0 & errno) {
+      std::cerr << "Error Converting Device Name: " << errno << std::endl;
+      throw std::invalid_argument ("audio_osx_sink::audio_osx_sink");
+    }
+    if (l_n_channels <= 0)
+      channel_config = 2;
+    else
+      channel_config = l_n_channels;
+  }
+
+  d_n_channels = d_channel_config = channel_config;
+
+// set the input signature
+
+  set_input_signature (gr_make_io_signature (1, d_n_channels, sizeof (float)));
+
+// check that the max # of samples to store is valid
+
+  if (max_sample_count == -1)
+    max_sample_count = sample_rate;
+  else if (max_sample_count <= 0) {
+    std::cerr << "Invalid Max Sample Count: " << max_sample_count << std::endl;
+    throw std::invalid_argument ("audio_osx_sink::audio_osx_sink");
+  }
+
+  d_max_sample_count = max_sample_count;
+
+// allocate the output circular buffer(s), one per channel
+
+  d_buffers = (circular_buffer<float>**) new
+    circular_buffer<float>* [d_n_channels];
+  UInt32 n_alloc = (UInt32) ceil ((double) d_max_sample_count);
+  for (UInt32 n = 0; n < d_n_channels; n++) {
+    d_buffers[n] = new circular_buffer<float> (n_alloc, false, false);
+  }
+
+// create the default AudioUnit for output
+  OSStatus err = noErr;
+
+// Open the default output unit
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  AudioComponentDescription desc;
+#else
+  ComponentDescription desc;
+#endif
+
+  desc.componentType = kAudioUnitType_Output;
+  desc.componentSubType = kAudioUnitSubType_DefaultOutput;
+  desc.componentManufacturer = kAudioUnitManufacturer_Apple;
+  desc.componentFlags = 0;
+  desc.componentFlagsMask = 0;
+
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  AudioComponent comp = AudioComponentFindNext(NULL, &desc);
+  if (comp == NULL) {
+    std::cerr << "AudioComponentFindNext Error" << std::endl;
+    throw std::runtime_error ("audio_osx_sink::audio_osx_sink");
+  }
+#else
+  Component comp = FindNextComponent (NULL, &desc);
+  if (comp == NULL) {
+    std::cerr << "FindNextComponent Error" << std::endl;
+    throw std::runtime_error ("audio_osx_sink::audio_osx_sink");
+  }
+#endif
+
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  err = AudioComponentInstanceNew (comp, &d_OutputAU);
+  CheckErrorAndThrow (err, "AudioComponentInstanceNew", "audio_osx_sink::audio_osx_sink");
+#else
+  err = OpenAComponent (comp, &d_OutputAU);
+  CheckErrorAndThrow (err, "OpenAComponent", "audio_osx_sink::audio_osx_sink");
+#endif
+
+// Set up a callback function to generate output to the output unit
+
+  AURenderCallbackStruct input;
+  input.inputProc = (AURenderCallback)(audio_osx_sink::AUOutputCallback);
+  input.inputProcRefCon = this;
+
+  err = AudioUnitSetProperty (d_OutputAU,
+			      kAudioUnitProperty_SetRenderCallback, 
+			      kAudioUnitScope_Input,
+			      0, 
+			      &input, 
+			      sizeof (input));
+  CheckErrorAndThrow (err, "AudioUnitSetProperty Render Callback", "audio_osx_sink::audio_osx_sink");
+
+// tell the Output Unit what format data will be supplied to it
+// so that it handles any format conversions
+
+  AudioStreamBasicDescription streamFormat;
+  streamFormat.mSampleRate = (Float64)(sample_rate);
+  streamFormat.mFormatID = kAudioFormatLinearPCM;
+  streamFormat.mFormatFlags = (kLinearPCMFormatFlagIsFloat |
+			       GR_PCM_ENDIANNESS |
+			       kLinearPCMFormatFlagIsPacked |
+			       kAudioFormatFlagIsNonInterleaved);
+  streamFormat.mBytesPerPacket = 4;
+  streamFormat.mFramesPerPacket = 1;
+  streamFormat.mBytesPerFrame = 4;
+  streamFormat.mChannelsPerFrame = d_n_channels;
+  streamFormat.mBitsPerChannel = 32;
+
+  err = AudioUnitSetProperty (d_OutputAU,
+			      kAudioUnitProperty_StreamFormat,
+			      kAudioUnitScope_Input,
+			      0,
+			      &streamFormat,
+			      sizeof (AudioStreamBasicDescription));
+  CheckErrorAndThrow (err, "AudioUnitSetProperty StreamFormat", "audio_osx_sink::audio_osx_sink");
+
+// create the stuff to regulate I/O
+
+  d_cond_data = new gruel::condition_variable ();
+  if (d_cond_data == NULL)
+    CheckErrorAndThrow (errno, "new condition (data)",
+			"audio_osx_sink::audio_osx_sink");
+
+  d_internal = new gruel::mutex ();
+  if (d_internal == NULL)
+    CheckErrorAndThrow (errno, "new mutex (internal)",
+			"audio_osx_sink::audio_osx_sink");
+
+// initialize the AU for output
+
+  err = AudioUnitInitialize (d_OutputAU);
+  CheckErrorAndThrow (err, "AudioUnitInitialize",
+		      "audio_osx_sink::audio_osx_sink");
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "audio_osx_sink Parameters:" << std::endl;
+  std::cerr << "  Sample Rate is " << d_sample_rate << std::endl;
+  std::cerr << "  Number of Channels is " << d_n_channels << std::endl;
+  std::cerr << "  Max # samples to store per channel is " << d_max_sample_count << std::endl;
+#endif
+}
+
+bool audio_osx_sink::IsRunning ()
+{
+  UInt32 AURunning = 0, AUSize = sizeof (UInt32);
+
+  OSStatus err = AudioUnitGetProperty (d_OutputAU,
+				       kAudioOutputUnitProperty_IsRunning,
+				       kAudioUnitScope_Global,
+				       0,
+				       &AURunning,
+				       &AUSize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty IsRunning",
+		      "audio_osx_sink::IsRunning");
+
+  return (AURunning);
+}
+
+bool audio_osx_sink::start ()
+{
+  if (! IsRunning ()) {
+    OSStatus err = AudioOutputUnitStart (d_OutputAU);
+    CheckErrorAndThrow (err, "AudioOutputUnitStart", "audio_osx_sink::start");
+  }
+
+  return (true);
+}
+
+bool audio_osx_sink::stop ()
+{
+  if (IsRunning ()) {
+    OSStatus err = AudioOutputUnitStop (d_OutputAU);
+    CheckErrorAndThrow (err, "AudioOutputUnitStop", "audio_osx_sink::stop");
+
+    for (UInt32 n = 0; n < d_n_channels; n++) {
+      d_buffers[n]->abort ();
+    }
+  }
+
+  return (true);
+}
+
+audio_osx_sink::~audio_osx_sink ()
+{
+// stop and close the AudioUnit
+  stop ();
+  AudioUnitUninitialize (d_OutputAU);
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  AudioComponentInstanceDispose (d_OutputAU);
+#else
+  CloseComponent (d_OutputAU);
+#endif
+
+// empty and delete the queues
+  for (UInt32 n = 0; n < d_n_channels; n++) {
+    delete d_buffers[n];
+    d_buffers[n] = 0;
+  }
+  delete [] d_buffers;
+  d_buffers = 0;
+
+// close and delete control stuff
+  delete d_cond_data;
+  d_cond_data = 0;
+  delete d_internal;
+  d_internal = 0;
+}
+
+int
+audio_osx_sink::work (int noutput_items,
+		      gr_vector_const_void_star &input_items,
+		      gr_vector_void_star &output_items)
+{
+  gruel::scoped_lock l (*d_internal);
+
+  /* take the input data, copy it, and push it to the bottom of the queue
+     mono input are pushed onto queue[0];
+     stereo input are pushed onto queue[1].
+     Start the AudioUnit if necessary. */
+
+  UInt32 l_max_count;
+  int diff_count = d_max_sample_count - noutput_items;
+  if (diff_count < 0)
+    l_max_count = 0;
+  else
+    l_max_count = (UInt32) diff_count;
+
+#if 0
+  if (l_max_count < d_queueItemLength->back()) {
+//  allow 2 buffers at a time, regardless of length
+    l_max_count = d_queueItemLength->back();
+  }
+#endif
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "work1: qSC = " << d_queueSampleCount << ", lMC = "<< l_max_count
+	    << ", dmSC = " << d_max_sample_count << ", nOI = " << noutput_items << std::endl;
+#endif
+
+  if (d_queueSampleCount > l_max_count) {
+// data coming in too fast; do_block decides what to do
+    if (d_do_block == true) {
+// block until there is data to return
+      while (d_queueSampleCount > l_max_count) {
+// release control so-as to allow data to be retrieved;
+// block until there is data to return
+	d_cond_data->wait (l);
+// the condition's 'notify' was called; acquire control
+// to keep thread safe
+      }
+    }
+  }
+// not blocking case and overflow is handled by the circular buffer
+
+// add the input frames to the buffers' queue, checking for overflow
+
+  UInt32 l_counter;
+  int res = 0;
+  float* inBuffer = (float*) input_items[0];
+  const UInt32 l_size = input_items.size();
+  for (l_counter = 0; l_counter < l_size; l_counter++) {
+    inBuffer = (float*) input_items[l_counter];
+    int l_res = d_buffers[l_counter]->enqueue (inBuffer,
+					       noutput_items);
+    if (l_res == -1)
+      res = -1;
+  }
+  while (l_counter < d_n_channels) {
+// for extra channels, copy the last input's data
+    int l_res = d_buffers[l_counter++]->enqueue (inBuffer,
+						 noutput_items);
+    if (l_res == -1)
+      res = -1;
+  }
+
+  if (res == -1) {
+// data coming in too fast
+// drop oldest buffer
+    fputs ("aO", stderr);
+    fflush (stderr);
+// set the local number of samples available to the max
+    d_queueSampleCount = d_buffers[0]->buffer_length_items ();
+  } else {
+// keep up the local sample count
+    d_queueSampleCount += noutput_items;
+  }
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "work2: #OI = " << noutput_items << ", #Cnt = "
+	    << d_queueSampleCount << ", mSC = " << d_max_sample_count << std::endl;
+#endif
+
+  return (noutput_items);
+}
+
+OSStatus audio_osx_sink::AUOutputCallback
+(void *inRefCon, 
+ AudioUnitRenderActionFlags *ioActionFlags, 
+ const AudioTimeStamp *inTimeStamp, 
+ UInt32 inBusNumber, 
+ UInt32 inNumberFrames, 
+ AudioBufferList *ioData)
+{
+  audio_osx_sink* This = (audio_osx_sink*) inRefCon;
+  OSStatus err = noErr;
+
+  gruel::scoped_lock l (*This->d_internal);
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cb_in: SC = " << This->d_queueSampleCount
+	    << ", in#F = " << inNumberFrames << std::endl;
+#endif
+
+  if (This->d_queueSampleCount < inNumberFrames) {
+// not enough data to fill request
+    err = -1;
+  } else {
+// enough data; remove data from our buffers into the AU's buffers
+    int l_counter = This->d_n_channels;
+
+    while (--l_counter >= 0) {
+      size_t t_n_output_items = inNumberFrames;
+      float* outBuffer = (float*) ioData->mBuffers[l_counter].mData;
+      This->d_buffers[l_counter]->dequeue (outBuffer, &t_n_output_items);
+      if (t_n_output_items != inNumberFrames) {
+	throw std::runtime_error ("audio_osx_sink::AUOutputCallback(): "
+				  "number of available items changing "
+				  "unexpectedly.\n");
+      }
+    }
+
+    This->d_queueSampleCount -= inNumberFrames;
+  }
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cb_out: SC = " << This->d_queueSampleCount << std::endl;
+#endif
+
+// signal that data is available
+  This->d_cond_data->notify_one ();
+
+  return (err);
+}
diff --git a/gr-audio/lib/osx/audio_osx_sink.h b/gr-audio/lib/osx/audio_osx_sink.h
new file mode 100644
index 000000000..13bd95d53
--- /dev/null
+++ b/gr-audio/lib/osx/audio_osx_sink.h
@@ -0,0 +1,79 @@
+/* -*- c++ -*- */
+/*
+ * Copyright 2006-2011 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_AUDIO_OSX_SINK_H
+#define INCLUDED_AUDIO_OSX_SINK_H
+
+#include <gr_audio_sink.h>
+#include <string>
+#include <list>
+#include <AudioUnit/AudioUnit.h>
+#include <circular_buffer.h>
+
+/*!
+ * \brief audio sink using OSX
+ *
+ * input signature is one or two streams of floats.
+ * Input samples must be in the range [-1,1].
+ */
+
+class audio_osx_sink : public audio_sink {
+
+  Float64             d_sample_rate;
+  int                 d_channel_config;
+  UInt32              d_n_channels;
+  UInt32              d_queueSampleCount, d_max_sample_count;
+  bool                d_do_block;
+  gruel::mutex*       d_internal;
+  gruel::condition_variable* d_cond_data;
+  circular_buffer<float>** d_buffers;
+
+// AudioUnits and Such
+  AudioUnit           d_OutputAU;
+
+public:
+  audio_osx_sink (int sample_rate = 44100,
+		  const std::string device_name = "2",
+		  bool do_block = true,
+		  int channel_config = -1,
+		  int max_sample_count = -1);
+
+  ~audio_osx_sink ();
+
+  bool IsRunning ();
+  bool start ();
+  bool stop ();
+
+  int work (int noutput_items,
+	    gr_vector_const_void_star &input_items,
+	    gr_vector_void_star &output_items);
+
+private:
+  static OSStatus AUOutputCallback (void *inRefCon, 
+				    AudioUnitRenderActionFlags *ioActionFlags, 
+				    const AudioTimeStamp *inTimeStamp, 
+				    UInt32 inBusNumber, 
+				    UInt32 inNumberFrames, 
+				    AudioBufferList *ioData);
+};
+
+#endif /* INCLUDED_AUDIO_OSX_SINK_H */
diff --git a/gr-audio/lib/osx/audio_osx_source.cc b/gr-audio/lib/osx/audio_osx_source.cc
new file mode 100644
index 000000000..e380156d6
--- /dev/null
+++ b/gr-audio/lib/osx/audio_osx_source.cc
@@ -0,0 +1,1016 @@
+/* -*- c++ -*- */
+/*
+ * Copyright 2006-2011 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.
+ */
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
+#include "gr_audio_registry.h"
+#include <audio_osx_source.h>
+#include <gr_io_signature.h>
+#include <stdexcept>
+#include <audio_osx.h>
+
+#define _OSX_AU_DEBUG_ 0
+#define _OSX_DO_LISTENERS_ 0
+
+AUDIO_REGISTER_SOURCE(REG_PRIO_HIGH, osx)(
+    int sampling_rate, const std::string &device_name, bool ok_to_block
+){
+    return audio_source::sptr(new audio_osx_source(sampling_rate, device_name, ok_to_block));
+}
+
+void PrintStreamDesc (AudioStreamBasicDescription *inDesc)
+{
+  if (inDesc == NULL) {
+    std::cerr << "PrintStreamDesc: Can't print a NULL desc!" << std::endl;
+    return;
+  }
+
+  std::cerr << "  Sample Rate        : " << inDesc->mSampleRate << std::endl;
+  char format_id[4];
+  strncpy (format_id, (char*)(&inDesc->mFormatID), 4);
+  std::cerr << "  Format ID          : " << format_id << std::endl;
+  std::cerr << "  Format Flags       : " << inDesc->mFormatFlags << std::endl;
+  std::cerr << "  Bytes per Packet   : " << inDesc->mBytesPerPacket << std::endl;
+  std::cerr << "  Frames per Packet  : " << inDesc->mFramesPerPacket << std::endl;
+  std::cerr << "  Bytes per Frame    : " << inDesc->mBytesPerFrame << std::endl;
+  std::cerr << "  Channels per Frame : " << inDesc->mChannelsPerFrame << std::endl;
+  std::cerr << "  Bits per Channel   : " << inDesc->mBitsPerChannel << std::endl;
+}
+
+// FIXME these should query some kind of user preference
+
+audio_osx_source::audio_osx_source (int sample_rate,
+				    const std::string device_name,
+				    bool do_block,
+				    int channel_config,
+				    int max_sample_count)
+  : audio_source ("audio_osx_source",
+		   gr_make_io_signature (0, 0, 0),
+		   gr_make_io_signature (0, 0, 0)),
+    d_deviceSampleRate (0.0), d_outputSampleRate (0.0),
+    d_channel_config (0),
+    d_inputBufferSizeFrames (0), d_inputBufferSizeBytes (0),
+    d_outputBufferSizeFrames (0), d_outputBufferSizeBytes (0),
+    d_deviceBufferSizeFrames (0), d_deviceBufferSizeBytes (0),
+    d_leadSizeFrames (0), d_leadSizeBytes (0),
+    d_trailSizeFrames (0), d_trailSizeBytes (0),
+    d_extraBufferSizeFrames (0), d_extraBufferSizeBytes (0),
+    d_queueSampleCount (0), d_max_sample_count (0),
+    d_n_AvailableInputFrames (0), d_n_ActualInputFrames (0),
+    d_n_user_channels (0), d_n_max_channels (0), d_n_deviceChannels (0),
+    d_do_block (do_block), d_passThrough (false),
+    d_internal (0), d_cond_data (0),
+    d_buffers (0),
+    d_InputAU (0), d_InputBuffer (0), d_OutputBuffer (0),
+    d_AudioConverter (0)
+{
+  if (sample_rate <= 0) {
+    std::cerr << "Invalid Sample Rate: " << sample_rate << std::endl;
+    throw std::invalid_argument ("audio_osx_source::audio_osx_source");
+  } else
+    d_outputSampleRate = (Float64) sample_rate;
+
+  if (channel_config <= 0 & channel_config != -1) {
+    std::cerr << "Invalid Channel Config: " << channel_config << std::endl;
+    throw std::invalid_argument ("audio_osx_source::audio_osx_source");
+  } else if (channel_config == -1) {
+// no user input; try "device name" instead
+    int l_n_channels = (int) strtol (device_name.data(), (char **)NULL, 10);
+    if (l_n_channels == 0 & errno) {
+      std::cerr << "Error Converting Device Name: " << errno << std::endl;
+      throw std::invalid_argument ("audio_osx_source::audio_osx_source");
+    }
+    if (l_n_channels <= 0)
+      channel_config = 2;
+    else
+      channel_config = l_n_channels;
+  }
+
+  d_channel_config = channel_config;
+
+// check that the max # of samples to store is valid
+
+  if (max_sample_count == -1)
+    max_sample_count = sample_rate;
+  else if (max_sample_count <= 0) {
+    std::cerr << "Invalid Max Sample Count: " << max_sample_count << std::endl;
+    throw std::invalid_argument ("audio_osx_source::audio_osx_source");
+  }
+
+  d_max_sample_count = max_sample_count;
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "source(): max # samples = " << d_max_sample_count << std::endl;
+#endif
+
+  OSStatus err = noErr;
+
+// create the default AudioUnit for input
+
+// Open the default input unit
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  AudioComponentDescription InputDesc;
+#else
+  ComponentDescription InputDesc;
+#endif
+  
+
+  InputDesc.componentType = kAudioUnitType_Output;
+  InputDesc.componentSubType = kAudioUnitSubType_HALOutput;
+  InputDesc.componentManufacturer = kAudioUnitManufacturer_Apple;
+  InputDesc.componentFlags = 0;
+  InputDesc.componentFlagsMask = 0;
+
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  AudioComponent comp = AudioComponentFindNext (NULL, &InputDesc);
+#else
+  Component comp = FindNextComponent (NULL, &InputDesc);
+#endif
+  
+  if (comp == NULL) {
+#ifndef GR_USE_OLD_AUDIO_UNIT
+    std::cerr << "AudioComponentFindNext Error" << std::endl;
+#else
+    std::cerr << "FindNextComponent Error" << std::endl;
+#endif
+    throw std::runtime_error ("audio_osx_source::audio_osx_source");
+  }
+
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  err = AudioComponentInstanceNew (comp, &d_InputAU);
+  CheckErrorAndThrow (err, "AudioComponentInstanceNew",
+		      "audio_osx_source::audio_osx_source");
+#else
+  err = OpenAComponent (comp, &d_InputAU);
+  CheckErrorAndThrow (err, "OpenAComponent",
+		      "audio_osx_source::audio_osx_source");
+#endif
+  
+
+  UInt32 enableIO;
+
+// must enable the AUHAL for input and disable output 
+// before setting the AUHAL's current device
+
+// Enable input on the AUHAL
+  enableIO = 1;
+  err = AudioUnitSetProperty (d_InputAU,
+			      kAudioOutputUnitProperty_EnableIO,
+			      kAudioUnitScope_Input,
+			      1, // input element
+			      &enableIO,
+			      sizeof (UInt32));
+  CheckErrorAndThrow (err, "AudioUnitSetProperty Input Enable",
+		      "audio_osx_source::audio_osx_source");
+
+// Disable output on the AUHAL
+  enableIO = 0;
+  err = AudioUnitSetProperty (d_InputAU,
+			      kAudioOutputUnitProperty_EnableIO,
+			      kAudioUnitScope_Output,
+			      0, // output element
+			      &enableIO,
+			      sizeof (UInt32));
+  CheckErrorAndThrow (err, "AudioUnitSetProperty Output Disable",
+		      "audio_osx_source::audio_osx_source");
+
+// set the default input device for our input AU
+
+  SetDefaultInputDeviceAsCurrent ();
+
+#if _OSX_DO_LISTENERS_
+// set up a listener if default hardware input device changes
+
+  err = AudioHardwareAddPropertyListener
+    (kAudioHardwarePropertyDefaultInputDevice,
+     (AudioHardwarePropertyListenerProc) HardwareListener,
+     this);
+
+  CheckErrorAndThrow (err, "AudioHardwareAddPropertyListener",
+		      "audio_osx_source::audio_osx_source");
+
+// Add a listener for any changes in the input AU's output stream
+// the function "UnitListener" will be called if the stream format
+// changes for whatever reason
+
+  err = AudioUnitAddPropertyListener
+    (d_InputAU,
+     kAudioUnitProperty_StreamFormat,
+     (AudioUnitPropertyListenerProc) UnitListener,
+     this);
+  CheckErrorAndThrow (err, "Adding Unit Property Listener",
+		      "audio_osx_source::audio_osx_source");
+#endif
+
+// Now find out if it actually can do input.
+
+  UInt32 hasInput = 0;
+  UInt32 dataSize = sizeof (hasInput);
+  err = AudioUnitGetProperty (d_InputAU,
+			      kAudioOutputUnitProperty_HasIO,
+			      kAudioUnitScope_Input,
+			      1,
+			      &hasInput,
+			      &dataSize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty HasIO",
+		      "audio_osx_source::audio_osx_source");
+  if (hasInput == 0) {
+    std::cerr << "Selected Audio Device does not support Input." << std::endl;
+    throw std::runtime_error ("audio_osx_source::audio_osx_source");
+  }
+
+// Set up a callback function to retrieve input from the Audio Device
+
+  AURenderCallbackStruct AUCallBack;
+
+  AUCallBack.inputProc = (AURenderCallback)(audio_osx_source::AUInputCallback);
+  AUCallBack.inputProcRefCon = this;
+
+  err = AudioUnitSetProperty (d_InputAU,
+			      kAudioOutputUnitProperty_SetInputCallback,
+			      kAudioUnitScope_Global,
+			      0,
+			      &AUCallBack,
+			      sizeof (AURenderCallbackStruct));
+  CheckErrorAndThrow (err, "AudioUnitSetProperty Input Callback",
+		      "audio_osx_source::audio_osx_source");
+
+  UInt32 propertySize;
+  AudioStreamBasicDescription asbd_device, asbd_client, asbd_user;
+
+// asbd_device: ASBD of the device that is creating the input data stream
+// asbd_client: ASBD of the client size (output) of the hardware device
+// asbd_user:   ASBD of the user's arguments
+
+// Get the Stream Format (device side)
+
+  propertySize = sizeof (asbd_device);
+  err = AudioUnitGetProperty (d_InputAU,
+			      kAudioUnitProperty_StreamFormat,
+			      kAudioUnitScope_Input,
+			      1,
+			      &asbd_device,
+			      &propertySize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty Device Input Stream Format",
+		      "audio_osx_source::audio_osx_source");
+
+#if _OSX_AU_DEBUG_
+  std::cerr << std::endl << "---- Device Stream Format ----" << std::endl;
+  PrintStreamDesc (&asbd_device);
+#endif
+
+// Get the Stream Format (client side)
+  propertySize = sizeof (asbd_client);
+  err = AudioUnitGetProperty (d_InputAU,
+			      kAudioUnitProperty_StreamFormat,
+			      kAudioUnitScope_Output,
+			      1,
+			      &asbd_client,
+			      &propertySize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty Device Ouput Stream Format",
+		      "audio_osx_source::audio_osx_source");
+
+#if _OSX_AU_DEBUG_
+  std::cerr << std::endl << "---- Client Stream Format ----" << std::endl;
+  PrintStreamDesc (&asbd_client);
+#endif
+
+// Set the format of all the AUs to the input/output devices channel count
+
+// get the max number of input (& thus output) channels supported by
+// this device
+  d_n_max_channels = asbd_client.mChannelsPerFrame;
+
+// create the output io signature;
+// no input siganture to set (source is hardware)
+  set_output_signature (gr_make_io_signature (1,
+					      d_n_max_channels,
+					      sizeof (float)));
+
+// allocate the output circular buffer(s), one per channel
+  d_buffers = (circular_buffer<float>**) new
+    circular_buffer<float>* [d_n_max_channels];
+  UInt32 n_alloc = (UInt32) ceil ((double) d_max_sample_count);
+  for (UInt32 n = 0; n < d_n_max_channels; n++) {
+    d_buffers[n] = new circular_buffer<float> (n_alloc, false, false);
+  }
+
+  d_deviceSampleRate = asbd_device.mSampleRate;
+  d_n_deviceChannels = asbd_device.mChannelsPerFrame;
+
+// create an ASBD for the user's wants
+
+  asbd_user.mSampleRate = d_outputSampleRate;
+  asbd_user.mFormatID = kAudioFormatLinearPCM;
+  asbd_user.mFormatFlags = (kLinearPCMFormatFlagIsFloat |
+			    GR_PCM_ENDIANNESS |
+			    kLinearPCMFormatFlagIsPacked |
+			    kAudioFormatFlagIsNonInterleaved);
+  asbd_user.mBytesPerPacket = 4;
+  asbd_user.mFramesPerPacket = 1;
+  asbd_user.mBytesPerFrame = 4;
+  asbd_user.mChannelsPerFrame = d_n_max_channels;
+  asbd_user.mBitsPerChannel = 32;
+
+  if (d_deviceSampleRate == d_outputSampleRate) {
+// no need to do conversion if asbd_client matches user wants
+    d_passThrough = true;
+    d_leadSizeFrames = d_trailSizeFrames = 0L;
+  } else {
+    d_passThrough = false;
+// Create the audio converter
+
+    err = AudioConverterNew (&asbd_client, &asbd_user, &d_AudioConverter);
+    CheckErrorAndThrow (err, "AudioConverterNew",
+			"audio_osx_source::audio_osx_source");
+
+// Set the audio converter sample rate quality to "max" ...
+// requires more samples, but should sound nicer
+
+    UInt32 ACQuality = kAudioConverterQuality_Max;
+    propertySize = sizeof (ACQuality);
+    err = AudioConverterSetProperty (d_AudioConverter,
+				     kAudioConverterSampleRateConverterQuality,
+				     propertySize,
+				     &ACQuality);
+    CheckErrorAndThrow (err, "AudioConverterSetProperty "
+			"SampleRateConverterQuality",
+			"audio_osx_source::audio_osx_source");
+
+// set the audio converter's prime method to "pre",
+// which uses both leading and trailing frames
+// from the "current input".  All of this is handled
+// internally by the AudioConverter; we just supply
+// the frames for conversion.
+
+//   UInt32 ACPrimeMethod = kConverterPrimeMethod_None;
+    UInt32 ACPrimeMethod = kConverterPrimeMethod_Pre;
+    propertySize = sizeof (ACPrimeMethod);
+    err = AudioConverterSetProperty (d_AudioConverter, 
+				     kAudioConverterPrimeMethod,
+				     propertySize,
+				     &ACPrimeMethod);
+    CheckErrorAndThrow (err, "AudioConverterSetProperty PrimeMethod",
+			"audio_osx_source::audio_osx_source");
+
+// Get the size of the I/O buffer(s) to allow for pre-allocated buffers
+      
+// lead frame info (trail frame info is ignored)
+
+    AudioConverterPrimeInfo ACPrimeInfo = {0, 0};
+    propertySize = sizeof (ACPrimeInfo);
+    err = AudioConverterGetProperty (d_AudioConverter, 
+				     kAudioConverterPrimeInfo,
+				     &propertySize,
+				     &ACPrimeInfo);
+    CheckErrorAndThrow (err, "AudioConverterGetProperty PrimeInfo",
+			"audio_osx_source::audio_osx_source");
+
+    switch (ACPrimeMethod) {
+    case (kConverterPrimeMethod_None):
+      d_leadSizeFrames =
+	d_trailSizeFrames = 0L;
+      break;
+    case (kConverterPrimeMethod_Normal):
+      d_leadSizeFrames = 0L;
+      d_trailSizeFrames = ACPrimeInfo.trailingFrames;
+      break;
+    default:
+      d_leadSizeFrames = ACPrimeInfo.leadingFrames;
+      d_trailSizeFrames = ACPrimeInfo.trailingFrames;
+    }
+  }
+  d_leadSizeBytes = d_leadSizeFrames * sizeof (Float32);
+  d_trailSizeBytes = d_trailSizeFrames * sizeof (Float32);
+
+  propertySize = sizeof (d_deviceBufferSizeFrames);
+  err = AudioUnitGetProperty (d_InputAU,
+			      kAudioDevicePropertyBufferFrameSize,
+			      kAudioUnitScope_Global,
+			      0,
+			      &d_deviceBufferSizeFrames,
+			      &propertySize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty Buffer Frame Size",
+		      "audio_osx_source::audio_osx_source");
+
+  d_deviceBufferSizeBytes = d_deviceBufferSizeFrames * sizeof (Float32);
+  d_inputBufferSizeBytes = d_deviceBufferSizeBytes + d_leadSizeBytes;
+  d_inputBufferSizeFrames = d_deviceBufferSizeFrames + d_leadSizeFrames;
+
+// outBufSizeBytes = floor (inBufSizeBytes * rate_out / rate_in)
+// since this is rarely exact, we need another buffer to hold
+// "extra" samples not processed at any given sampling period
+// this buffer must be at least 4 floats in size, but generally
+// follows the rule that
+// extraBufSize =  ceil (rate_in / rate_out)*sizeof(float)
+
+  d_extraBufferSizeFrames = ((UInt32) ceil (d_deviceSampleRate
+					    / d_outputSampleRate)
+			     * sizeof (float));
+  if (d_extraBufferSizeFrames < 4)
+    d_extraBufferSizeFrames = 4;
+  d_extraBufferSizeBytes = d_extraBufferSizeFrames * sizeof (float);
+
+  d_outputBufferSizeFrames = (UInt32) ceil (((Float64) d_inputBufferSizeFrames)
+					    * d_outputSampleRate
+					    / d_deviceSampleRate);
+  d_outputBufferSizeBytes = d_outputBufferSizeFrames * sizeof (float);
+  d_inputBufferSizeFrames += d_extraBufferSizeFrames;
+
+// pre-alloc all buffers
+
+  AllocAudioBufferList (&d_InputBuffer, d_n_deviceChannels,
+			d_inputBufferSizeBytes);
+  if (d_passThrough == false) {
+    AllocAudioBufferList (&d_OutputBuffer, d_n_max_channels,
+			  d_outputBufferSizeBytes);
+  } else {
+    d_OutputBuffer = d_InputBuffer;
+  }
+
+// create the stuff to regulate I/O
+
+  d_cond_data = new gruel::condition_variable ();
+  if (d_cond_data == NULL)
+    CheckErrorAndThrow (errno, "new condition (data)",
+			"audio_osx_source::audio_osx_source");
+
+  d_internal = new gruel::mutex ();
+  if (d_internal == NULL)
+    CheckErrorAndThrow (errno, "new mutex (internal)",
+			"audio_osx_source::audio_osx_source");
+
+// initialize the AU for input
+
+  err = AudioUnitInitialize (d_InputAU);
+  CheckErrorAndThrow (err, "AudioUnitInitialize",
+		      "audio_osx_source::audio_osx_source");
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "audio_osx_source Parameters:" << std::endl;
+  std::cerr << "  Device Sample Rate is " << d_deviceSampleRate << std::endl;
+  std::cerr << "  User Sample Rate is " << d_outputSampleRate << std::endl;
+  std::cerr << "  Max Sample Count is " << d_max_sample_count << std::endl;
+  std::cerr << "  # Device Channels is " << d_n_deviceChannels << std::endl;
+  std::cerr << "  # Max Channels is " << d_n_max_channels << std::endl;
+  std::cerr << "  Device Buffer Size is Frames = " << d_deviceBufferSizeFrames << std::endl;
+  std::cerr << "  Lead Size is Frames = " << d_leadSizeFrames << std::endl;
+  std::cerr << "  Trail Size is Frames = " << d_trailSizeFrames << std::endl;
+  std::cerr << "  Input Buffer Size is Frames = " << d_inputBufferSizeFrames << std::endl;
+  std::cerr << "  Output Buffer Size is Frames = " << d_outputBufferSizeFrames << std::endl;
+#endif
+}
+
+void
+audio_osx_source::AllocAudioBufferList (AudioBufferList** t_ABL,
+					UInt32 n_channels,
+					UInt32 bufferSizeBytes)
+{
+  FreeAudioBufferList (t_ABL);
+  UInt32 propertySize = (offsetof (AudioBufferList, mBuffers[0]) +
+			 (sizeof (AudioBuffer) * n_channels));
+  *t_ABL = (AudioBufferList*) calloc (1, propertySize);
+  (*t_ABL)->mNumberBuffers = n_channels;
+
+  int counter = n_channels;
+
+  while (--counter >= 0) {
+    (*t_ABL)->mBuffers[counter].mNumberChannels = 1;
+    (*t_ABL)->mBuffers[counter].mDataByteSize = bufferSizeBytes;
+    (*t_ABL)->mBuffers[counter].mData = calloc (1, bufferSizeBytes);
+  }
+}
+
+void
+audio_osx_source::FreeAudioBufferList (AudioBufferList** t_ABL)
+{
+// free pre-allocated audio buffer, if it exists
+  if (*t_ABL != NULL) {
+    int counter = (*t_ABL)->mNumberBuffers;
+    while (--counter >= 0)
+      free ((*t_ABL)->mBuffers[counter].mData);
+    free (*t_ABL);
+    (*t_ABL) = 0;
+  }
+}
+
+bool audio_osx_source::IsRunning ()
+{
+  UInt32 AURunning = 0, AUSize = sizeof (UInt32);
+
+  OSStatus err = AudioUnitGetProperty (d_InputAU,
+				       kAudioOutputUnitProperty_IsRunning,
+				       kAudioUnitScope_Global,
+				       0,
+				       &AURunning,
+				       &AUSize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty IsRunning",
+		      "audio_osx_source::IsRunning");
+
+  return (AURunning);
+}
+
+bool audio_osx_source::start ()
+{
+  if (! IsRunning ()) {
+    OSStatus err = AudioOutputUnitStart (d_InputAU);
+    CheckErrorAndThrow (err, "AudioOutputUnitStart",
+			"audio_osx_source::start");
+  }
+
+  return (true);
+}
+
+bool audio_osx_source::stop ()
+{
+  if (IsRunning ()) {
+    OSStatus err = AudioOutputUnitStop (d_InputAU);
+    CheckErrorAndThrow (err, "AudioOutputUnitStart",
+			"audio_osx_source::stop");
+    for (UInt32 n = 0; n < d_n_user_channels; n++) {
+      d_buffers[n]->abort ();
+    }
+  }
+
+  return (true);
+}
+
+audio_osx_source::~audio_osx_source ()
+{
+  OSStatus err = noErr;
+
+// stop the AudioUnit
+  stop();
+
+#if _OSX_DO_LISTENERS_
+// remove the listeners
+
+  err = AudioUnitRemovePropertyListener
+    (d_InputAU,
+     kAudioUnitProperty_StreamFormat,
+     (AudioUnitPropertyListenerProc) UnitListener);
+  CheckError (err, "~audio_osx_source: AudioUnitRemovePropertyListener");
+
+  err = AudioHardwareRemovePropertyListener
+    (kAudioHardwarePropertyDefaultInputDevice,
+     (AudioHardwarePropertyListenerProc) HardwareListener);
+  CheckError (err, "~audio_osx_source: AudioHardwareRemovePropertyListener");
+#endif
+
+// free pre-allocated audio buffers
+  FreeAudioBufferList (&d_InputBuffer);
+
+  if (d_passThrough == false) {
+    err = AudioConverterDispose (d_AudioConverter);
+    CheckError (err, "~audio_osx_source: AudioConverterDispose");
+    FreeAudioBufferList (&d_OutputBuffer);
+  }
+
+// remove the audio unit
+  err = AudioUnitUninitialize (d_InputAU);
+  CheckError (err, "~audio_osx_source: AudioUnitUninitialize");
+
+#ifndef GR_USE_OLD_AUDIO_UNIT
+  err = AudioComponentInstanceDispose (d_InputAU);
+  CheckError (err, "~audio_osx_source: AudioComponentInstanceDispose");
+#else
+  err = CloseComponent (d_InputAU);
+  CheckError (err, "~audio_osx_source: CloseComponent");
+#endif
+
+// empty and delete the queues
+  for (UInt32 n = 0; n < d_n_max_channels; n++) {
+    delete d_buffers[n];
+    d_buffers[n] = 0;
+  }
+  delete [] d_buffers;
+  d_buffers = 0;
+
+// close and delete the control stuff
+  delete d_cond_data;
+  d_cond_data = 0;
+  delete d_internal;
+  d_internal = 0;
+}
+
+bool
+audio_osx_source::check_topology (int ninputs, int noutputs)
+{
+// check # inputs to make sure it's valid
+  if (ninputs != 0) {
+    std::cerr << "audio_osx_source::check_topology(): number of input "
+	      << "streams provided (" << ninputs
+	      << ") should be 0." << std::endl;
+    throw std::runtime_error ("audio_osx_source::check_topology()");
+  }
+
+// check # outputs to make sure it's valid
+  if ((noutputs < 1) | (noutputs > (int) d_n_max_channels)) {
+    std::cerr << "audio_osx_source::check_topology(): number of output "
+	      << "streams provided (" << noutputs << ") should be in [1,"
+	      << d_n_max_channels << "] for the selected audio device."
+	      << std::endl;
+    throw std::runtime_error ("audio_osx_source::check_topology()");
+  }
+
+// save the actual number of output (user) channels
+  d_n_user_channels = noutputs;
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "chk_topo: Actual # user output channels = "
+	    << noutputs << std::endl;
+#endif
+
+  return (true);
+}
+
+int
+audio_osx_source::work
+(int noutput_items,
+ gr_vector_const_void_star &input_items,
+ gr_vector_void_star &output_items)
+{
+  // acquire control to do processing here only
+  gruel::scoped_lock l (*d_internal);
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "work1: SC = " << d_queueSampleCount
+	    << ", #OI = " << noutput_items
+	    << ", #Chan = " << output_items.size() << std::endl;
+#endif
+
+  // set the actual # of output items to the 'desired' amount then
+  // verify that data is available; if not enough data is available,
+  // either wait until it is (is "do_block" is true), return (0) is no
+  // data is available and "do_block" is false, or process the actual
+  // amount of available data.
+
+  UInt32 actual_noutput_items = noutput_items;
+
+  if (d_queueSampleCount < actual_noutput_items) {
+    if (d_queueSampleCount == 0) {
+      // no data; do_block decides what to do
+      if (d_do_block == true) {
+	while (d_queueSampleCount == 0) {
+	  // release control so-as to allow data to be retrieved;
+	  // block until there is data to return
+	  d_cond_data->wait (l);
+	  // the condition's 'notify' was called; acquire control to
+	  // keep thread safe
+	}
+      } else {
+	// no data & not blocking; return nothing
+	return (0);
+      }
+    }
+    // use the actual amount of available data
+    actual_noutput_items = d_queueSampleCount;
+  }
+
+  // number of channels
+  int l_counter = (int) output_items.size();
+
+  // copy the items from the circular buffer(s) to 'work's output buffers
+  // verify that the number copied out is as expected.
+
+  while (--l_counter >= 0) {
+    size_t t_n_output_items = actual_noutput_items;
+    d_buffers[l_counter]->dequeue ((float*) output_items[l_counter],
+				   &t_n_output_items);
+    if (t_n_output_items != actual_noutput_items) {
+      std::cerr << "audio_osx_source::work(): ERROR: number of "
+		<< "available items changing unexpectedly; expecting "
+		<< actual_noutput_items << ", got "
+		<< t_n_output_items << "." << std::endl;
+      throw std::runtime_error ("audio_osx_source::work()");
+    }
+  }
+
+  // subtract the actual number of items removed from the buffer(s)
+  // from the local accounting of the number of available samples
+
+  d_queueSampleCount -= actual_noutput_items;
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "work2: SC = " << d_queueSampleCount
+	    << ", act#OI = " << actual_noutput_items << std::endl
+	    << "Returning." << std::endl;
+#endif
+
+  return (actual_noutput_items);
+}
+
+OSStatus
+audio_osx_source::ConverterCallback
+(AudioConverterRef inAudioConverter,
+ UInt32* ioNumberDataPackets,
+ AudioBufferList* ioData,
+ AudioStreamPacketDescription** ioASPD,
+ void* inUserData)
+{
+  // take current device buffers and copy them to the tail of the
+  // input buffers the lead buffer is already there in the first
+  // d_leadSizeFrames slots
+
+  audio_osx_source* This = static_cast<audio_osx_source*>(inUserData);
+  AudioBufferList* l_inputABL = This->d_InputBuffer;
+  UInt32 totalInputBufferSizeBytes = ((*ioNumberDataPackets) * sizeof (float));
+  int counter = This->d_n_deviceChannels;
+  ioData->mNumberBuffers = This->d_n_deviceChannels;
+  This->d_n_ActualInputFrames = (*ioNumberDataPackets);
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cc1: io#DP = " << (*ioNumberDataPackets)
+	    << ", TIBSB = " << totalInputBufferSizeBytes
+	    << ", #C = " << counter << std::endl;
+#endif
+
+  while (--counter >= 0)  {
+    AudioBuffer* l_ioD_AB = &(ioData->mBuffers[counter]);
+    l_ioD_AB->mNumberChannels = 1;
+    l_ioD_AB->mData = (float*)(l_inputABL->mBuffers[counter].mData);
+    l_ioD_AB->mDataByteSize = totalInputBufferSizeBytes;
+  }
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cc2: Returning." << std::endl;
+#endif
+
+  return (noErr);
+}
+
+OSStatus
+audio_osx_source::AUInputCallback (void* inRefCon,
+				   AudioUnitRenderActionFlags* ioActionFlags,
+				   const AudioTimeStamp* inTimeStamp,
+				   UInt32 inBusNumber,
+				   UInt32 inNumberFrames,
+				   AudioBufferList* ioData)
+{
+  OSStatus err = noErr;
+  audio_osx_source* This = static_cast<audio_osx_source*>(inRefCon);
+
+  gruel::scoped_lock l (*This->d_internal);
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cb0: in#F = " << inNumberFrames
+	    << ", inBN = " << inBusNumber
+	    << ", SC = " << This->d_queueSampleCount << std::endl;
+#endif
+
+// Get the new audio data from the input device
+
+  err = AudioUnitRender (This->d_InputAU,
+			 ioActionFlags,
+			 inTimeStamp,
+			 1, //inBusNumber,
+			 inNumberFrames,
+			 This->d_InputBuffer);
+  CheckErrorAndThrow (err, "AudioUnitRender",
+		      "audio_osx_source::AUInputCallback");
+
+  UInt32 AvailableInputFrames = inNumberFrames;
+  This->d_n_AvailableInputFrames = inNumberFrames;
+
+// get the number of actual output frames,
+// either via converting the buffer or not
+
+  UInt32 ActualOutputFrames;
+
+  if (This->d_passThrough == true) {
+    ActualOutputFrames = AvailableInputFrames;
+  } else {
+    UInt32 AvailableInputBytes = AvailableInputFrames * sizeof (float);
+    UInt32 AvailableOutputBytes = AvailableInputBytes;
+    UInt32 AvailableOutputFrames = AvailableOutputBytes / sizeof (float);
+    UInt32 propertySize = sizeof (AvailableOutputBytes);
+    err = AudioConverterGetProperty (This->d_AudioConverter,
+		   kAudioConverterPropertyCalculateOutputBufferSize,
+				     &propertySize,
+				     &AvailableOutputBytes);
+    CheckErrorAndThrow (err, "AudioConverterGetProperty CalculateOutputBufferSize", "audio_osx_source::audio_osx_source");
+
+    AvailableOutputFrames = AvailableOutputBytes / sizeof (float);
+
+#if 0
+// when decimating too much, the output sounds warbly due to
+// fluctuating # of output frames
+// This should not be a surprise, but there's probably some
+// clever programming that could lessed the effect ...
+// like finding the "ideal" # of output frames, and keeping
+// that number constant no matter the # of input frames
+    UInt32 l_InputBytes = AvailableOutputBytes;
+    propertySize = sizeof (AvailableOutputBytes);
+    err = AudioConverterGetProperty (This->d_AudioConverter,
+		     kAudioConverterPropertyCalculateInputBufferSize,
+				     &propertySize,
+				     &l_InputBytes);
+    CheckErrorAndThrow (err, "AudioConverterGetProperty CalculateInputBufferSize", "audio_osx_source::audio_osx_source");
+
+    if (l_InputBytes < AvailableInputBytes) {
+// OK to zero pad the input a little
+      AvailableOutputFrames += 1;
+      AvailableOutputBytes = AvailableOutputFrames * sizeof (float);
+    }
+#endif
+
+#if _OSX_AU_DEBUG_
+    std::cerr << "cb1:  avail: #IF = " << AvailableInputFrames
+	      << ", #OF = " << AvailableOutputFrames << std::endl;
+#endif
+    ActualOutputFrames = AvailableOutputFrames;
+
+// convert the data to the correct rate
+// on input, ActualOutputFrames is the number of available output frames
+
+    err = AudioConverterFillComplexBuffer (This->d_AudioConverter,
+	   (AudioConverterComplexInputDataProc)(This->ConverterCallback),
+					   inRefCon,
+					   &ActualOutputFrames,
+					   This->d_OutputBuffer,
+					   NULL);
+    CheckErrorAndThrow (err, "AudioConverterFillComplexBuffer",
+			"audio_osx_source::AUInputCallback");
+
+// on output, ActualOutputFrames is the actual number of output frames
+
+#if _OSX_AU_DEBUG_
+    std::cerr << "cb2: actual: #IF = " << This->d_n_ActualInputFrames
+	      << ", #OF = " << AvailableOutputFrames << std::endl;
+    if (This->d_n_ActualInputFrames != AvailableInputFrames)
+      std::cerr << "cb2.1: avail#IF = " << AvailableInputFrames
+		<< ", actual#IF = " << This->d_n_ActualInputFrames << std::endl;
+#endif
+  }
+
+// add the output frames to the buffers' queue, checking for overflow
+
+  int l_counter = This->d_n_user_channels;
+  int res = 0;
+
+  while (--l_counter >= 0) {
+    float* inBuffer = (float*) This->d_OutputBuffer->mBuffers[l_counter].mData;
+
+#if _OSX_AU_DEBUG_
+    std::cerr << "cb3: enqueuing audio data." << std::endl;
+#endif
+
+    int l_res = This->d_buffers[l_counter]->enqueue (inBuffer, ActualOutputFrames);
+    if (l_res == -1)
+      res = -1;
+  }
+
+  if (res == -1) {
+// data coming in too fast
+// drop oldest buffer
+    fputs ("aO", stderr);
+    fflush (stderr);
+// set the local number of samples available to the max
+    This->d_queueSampleCount = This->d_buffers[0]->buffer_length_items ();
+  } else {
+// keep up the local sample count
+    This->d_queueSampleCount += ActualOutputFrames;
+  }
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cb4: #OI = " << ActualOutputFrames
+	    << ", #Cnt = " << This->d_queueSampleCount
+	    << ", mSC = " << This->d_max_sample_count << std::endl;
+#endif
+
+// signal that data is available, if appropraite
+  This->d_cond_data->notify_one ();
+
+#if _OSX_AU_DEBUG_
+  std::cerr << "cb5: returning." << std::endl;
+#endif
+
+  return (err);
+}
+
+void
+audio_osx_source::SetDefaultInputDeviceAsCurrent
+()
+{
+// set the default input device
+  AudioDeviceID deviceID;
+  UInt32 dataSize = sizeof (AudioDeviceID);
+  AudioHardwareGetProperty (kAudioHardwarePropertyDefaultInputDevice,
+			    &dataSize,
+			    &deviceID);
+  OSStatus err = AudioUnitSetProperty (d_InputAU,
+				       kAudioOutputUnitProperty_CurrentDevice,
+				       kAudioUnitScope_Global,
+				       0,
+				       &deviceID,
+				       sizeof (AudioDeviceID));
+  CheckErrorAndThrow (err, "AudioUnitSetProperty Current Device",
+		      "audio_osx_source::SetDefaultInputDeviceAsCurrent");
+}
+
+#if _OSX_DO_LISTENERS_
+OSStatus
+audio_osx_source::HardwareListener
+(AudioHardwarePropertyID inPropertyID, 
+ void *inClientData)
+{
+  OSStatus err = noErr;
+  audio_osx_source* This = static_cast<audio_osx_source*>(inClientData);
+
+  std::cerr << "a_o_s::HardwareListener" << std::endl;
+
+// set the new default hardware input device for use by our AU
+
+  This->SetDefaultInputDeviceAsCurrent ();
+
+// reset the converter to tell it that the stream has changed
+
+  err = AudioConverterReset (This->d_AudioConverter);
+  CheckErrorAndThrow (err, "AudioConverterReset",
+		      "audio_osx_source::UnitListener");
+
+  return (err);
+}
+
+OSStatus
+audio_osx_source::UnitListener
+(void *inRefCon,
+ AudioUnit ci,
+ AudioUnitPropertyID inID,
+ AudioUnitScope inScope,
+ AudioUnitElement inElement)
+{
+  OSStatus err = noErr;
+  audio_osx_source* This = static_cast<audio_osx_source*>(inRefCon);
+  AudioStreamBasicDescription asbd;			
+
+  std::cerr << "a_o_s::UnitListener" << std::endl;
+
+// get the converter's input ASBD (for printing)
+
+  UInt32 propertySize = sizeof (asbd);
+  err = AudioConverterGetProperty (This->d_AudioConverter,
+				   kAudioConverterCurrentInputStreamDescription,
+				   &propertySize,
+				   &asbd);
+  CheckErrorAndThrow (err, "AudioConverterGetProperty "
+		      "CurrentInputStreamDescription",
+		      "audio_osx_source::UnitListener");
+
+  std::cerr << "UnitListener: Input Source changed." << std::endl
+	    << "Old Source Output Info:" << std::endl;
+  PrintStreamDesc (&asbd);
+
+// get the new input unit's output ASBD
+
+  propertySize = sizeof (asbd);
+  err = AudioUnitGetProperty (This->d_InputAU,
+			      kAudioUnitProperty_StreamFormat,
+			      kAudioUnitScope_Output, 1,
+			      &asbd, &propertySize);
+  CheckErrorAndThrow (err, "AudioUnitGetProperty StreamFormat",
+		      "audio_osx_source::UnitListener");
+
+  std::cerr << "New Source Output Info:" << std::endl;
+  PrintStreamDesc (&asbd);
+
+// set the converter's input ASBD to this
+
+  err = AudioConverterSetProperty (This->d_AudioConverter,
+				   kAudioConverterCurrentInputStreamDescription,
+				   propertySize,
+				   &asbd);
+  CheckErrorAndThrow (err, "AudioConverterSetProperty "
+		      "CurrentInputStreamDescription",
+		      "audio_osx_source::UnitListener");
+
+// reset the converter to tell it that the stream has changed
+
+  err = AudioConverterReset (This->d_AudioConverter);
+  CheckErrorAndThrow (err, "AudioConverterReset",
+		      "audio_osx_source::UnitListener");
+
+  return (err);
+}
+#endif
diff --git a/gr-audio/lib/osx/audio_osx_source.h b/gr-audio/lib/osx/audio_osx_source.h
new file mode 100644
index 000000000..754f0d928
--- /dev/null
+++ b/gr-audio/lib/osx/audio_osx_source.h
@@ -0,0 +1,115 @@
+/* -*- c++ -*- */
+/*
+ * Copyright 2006-2011 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_AUDIO_OSX_SOURCE_H
+#define INCLUDED_AUDIO_OSX_SOURCE_H
+
+#include <gr_audio_source.h>
+#include <string>
+#include <AudioToolbox/AudioToolbox.h>
+#include <AudioUnit/AudioUnit.h>
+#include <circular_buffer.h>
+
+/*!
+ * \brief audio source using OSX
+ *
+ * Input signature is one or two streams of floats.
+ * Samples must be in the range [-1,1].
+ */
+
+class audio_osx_source : public gr_sync_block {
+
+  Float64             d_deviceSampleRate, d_outputSampleRate;
+  int                 d_channel_config;
+  UInt32              d_inputBufferSizeFrames, d_inputBufferSizeBytes;
+  UInt32              d_outputBufferSizeFrames, d_outputBufferSizeBytes;
+  UInt32              d_deviceBufferSizeFrames, d_deviceBufferSizeBytes;
+  UInt32              d_leadSizeFrames, d_leadSizeBytes;
+  UInt32              d_trailSizeFrames, d_trailSizeBytes;
+  UInt32              d_extraBufferSizeFrames, d_extraBufferSizeBytes;
+  UInt32              d_queueSampleCount, d_max_sample_count;
+  UInt32              d_n_AvailableInputFrames, d_n_ActualInputFrames;
+  UInt32              d_n_user_channels, d_n_max_channels, d_n_deviceChannels;
+  bool                d_do_block, d_passThrough, d_waiting_for_data;
+  gruel::mutex*       d_internal;
+  gruel::condition_variable* d_cond_data;
+  circular_buffer<float>** d_buffers;
+
+// AudioUnits and Such
+  AudioUnit           d_InputAU;
+  AudioBufferList*    d_InputBuffer;
+  AudioBufferList*    d_OutputBuffer;
+  AudioConverterRef   d_AudioConverter;
+
+public:
+  audio_osx_source (int sample_rate = 44100,
+		    const std::string device_name = "",
+		    bool do_block = true,
+		    int channel_config = -1,
+		    int max_sample_count = -1);
+
+  ~audio_osx_source ();
+
+  bool start ();
+  bool stop ();
+  bool IsRunning ();
+
+  bool check_topology (int ninputs, int noutputs);
+
+  int work (int noutput_items,
+	    gr_vector_const_void_star &input_items,
+	    gr_vector_void_star &output_items);
+
+private:
+  void SetDefaultInputDeviceAsCurrent ();
+
+  void AllocAudioBufferList (AudioBufferList** t_ABL,
+			     UInt32 n_channels,
+			     UInt32 inputBufferSizeBytes);
+
+  void FreeAudioBufferList (AudioBufferList** t_ABL);
+
+  static OSStatus ConverterCallback (AudioConverterRef inAudioConverter,
+				     UInt32* ioNumberDataPackets,
+				     AudioBufferList* ioData,
+				     AudioStreamPacketDescription** outASPD,
+				     void* inUserData);
+
+  static OSStatus AUInputCallback (void *inRefCon,
+				   AudioUnitRenderActionFlags *ioActionFlags,
+				   const AudioTimeStamp *inTimeStamp,
+				   UInt32 inBusNumber,
+				   UInt32 inNumberFrames,
+				   AudioBufferList *ioData);
+#if _OSX_DO_LISTENERS_
+  static OSStatus UnitListener (void *inRefCon,
+				AudioUnit ci,
+				AudioUnitPropertyID inID,
+				AudioUnitScope inScope,
+				AudioUnitElement inElement);
+
+  static OSStatus HardwareListener (AudioHardwarePropertyID inPropertyID, 
+				    void *inClientData);
+#endif
+};
+
+#endif /* INCLUDED_AUDIO_OSX_SOURCE_H */