ex_record.c

/* ex_record
 * 
 * Press spacebar to begin and stop recording. Press 'p' to play back the 
 * previous recording. Up to five minutes of audio will be recorded.
 * 
 * The raw sample data is saved to a temporary file and played back via
 * an audio stream. Thus, minimal memory is used regardless of the length of
 * recording.
 * 
 * When recording, it's important to keep the distinction between bytes and
 * samples. A sample is only exactly one byte long if it is 8-bit and mono.
 * Otherwise it is multiple bytes. The recording functions always work with
 * sample counts. However, when using things like memcpy() or fwrite(), you 
 * will be working with bytes.
 */

#define ALLEGRO_UNSTABLE
#include "allegro5/allegro.h"
#include "allegro5/allegro_audio.h"
#include "allegro5/allegro_primitives.h"
#include "allegro5/allegro_acodec.h"

#include "common.c"

/* Comment out the following line to use 16-bit audio */
#define WANT_8_BIT_DEPTH 

/* The following constants are needed so that the rest of the code can
 * work independent from the audio depth. Both 8-bit and 16-bit should be
 * supported by all devices, so in your own programs you probably can get by
 * with just supporting one or the other.
 */

#ifdef WANT_8_BIT_DEPTH 

const ALLEGRO_AUDIO_DEPTH audio_depth = ALLEGRO_AUDIO_DEPTH_UINT8;
typedef uint8_t* audio_buffer_t;
const uint8_t sample_center = 128;
const int8_t min_sample_val = 0x80;
const int8_t max_sample_val = 0x7f;
const int sample_range = 0xff;
const int sample_size = 1;

#else

const ALLEGRO_AUDIO_DEPTH audio_depth = ALLEGRO_AUDIO_DEPTH_INT16;
typedef int16_t* audio_buffer_t;
const int16_t sample_center = 0;
const int16_t min_sample_val = 0x8000;
const int16_t max_sample_val = 0x7fff;
const int sample_range = 0xffff;
const int sample_size = 2;

#endif

/* How many samples do we want to process at one time?
   Let's pick a multiple of the width of the screen so that the
   visualization graph is easy to draw. */
const unsigned int samples_per_fragment = 320 * 4;

/* Frequency is the quality of audio. (Samples per second.) Higher
   quality consumes more memory. For speech, numbers as low as 8000
   can be good enough. 44100 is often used for high quality recording. */      
const unsigned int frequency = 22050;
const unsigned int max_seconds_to_record = 60 * 5;


/* The playback buffer specs don't need to match the recording sizes.
 * The values here are slightly large to help make playback more smooth.
 */
const unsigned int playback_fragment_count = 4;
const unsigned int playback_samples_per_fragment = 4096;

int main(int argc, char **argv)
{
   ALLEGRO_AUDIO_RECORDER *r;
   ALLEGRO_AUDIO_STREAM *s;
   
   ALLEGRO_EVENT_QUEUE *q;
   ALLEGRO_DISPLAY *d;
   ALLEGRO_FILE *fp = NULL;
   ALLEGRO_PATH *tmp_path = NULL;
      
   int prev = 0;
   bool is_recording = false;
   
   int n = 0; /* number of samples written to disk */
   
   (void) argc;
   (void) argv;

   if (!al_init()) {
      abort_example("Could not init Allegro.\n");
   }
   
   if (!al_init_primitives_addon()) {
      abort_example("Unable to initialize primitives addon");
   }
      
   if (!al_install_keyboard()) {
      abort_example("Unable to install keyboard");
   }
      
   if (!al_install_audio()) {
      abort_example("Unable to initialize audio addon");
   }
   
   if (!al_init_acodec_addon()) {
      abort_example("Unable to initialize acodec addon");
   }
   
   /* Note: increasing the number of channels will break this demo. Other
    * settings can be changed by modifying the constants at the top of the
    * file.
    */
   r = al_create_audio_recorder(1000, samples_per_fragment, frequency,
      audio_depth, ALLEGRO_CHANNEL_CONF_1);
   if (!r) {
      abort_example("Unable to create audio recorder");
   }
   
   s = al_create_audio_stream(playback_fragment_count,
      playback_samples_per_fragment, frequency, audio_depth,
      ALLEGRO_CHANNEL_CONF_1);      
   if (!s) {
      abort_example("Unable to create audio stream");
   }
      
   al_reserve_samples(0);
   al_set_audio_stream_playing(s, false);
   al_attach_audio_stream_to_mixer(s, al_get_default_mixer());
      
   q = al_create_event_queue();
   
   /* Note: the following two options are referring to pixel samples, and have
    * nothing to do with audio samples. */
   al_set_new_display_option(ALLEGRO_SAMPLE_BUFFERS, 1, ALLEGRO_SUGGEST);
   al_set_new_display_option(ALLEGRO_SAMPLES, 8, ALLEGRO_SUGGEST);
   
   d = al_create_display(320, 256);
   if (!d) {
      abort_example("Error creating display\n");
   }
      
   al_set_window_title(d, "SPACE to record. P to playback.");
   
   al_register_event_source(q, al_get_audio_recorder_event_source(r));
   al_register_event_source(q, al_get_audio_stream_event_source(s));
   al_register_event_source(q, al_get_display_event_source(d));
   al_register_event_source(q, al_get_keyboard_event_source());
   
   al_start_audio_recorder(r);
   
   while (true) {
      ALLEGRO_EVENT e;

      al_wait_for_event(q, &e);
       
      if (e.type == ALLEGRO_EVENT_AUDIO_RECORDER_FRAGMENT) {
         /* We received an incoming fragment from the microphone. In this
          * example, the recorder is constantly recording even when we aren't
          * saving to disk. The display is updated every time a new fragment
          * comes in, because it makes things more simple. If the fragments
          * are coming in faster than we can update the screen, then it will be
          * a problem.
          */          
         ALLEGRO_AUDIO_RECORDER_EVENT *re = al_get_audio_recorder_event(&e);
         audio_buffer_t input = (audio_buffer_t) re->buffer;
         int sample_count = re->samples; 
         const int R = sample_count / 320;
         int i, gain = 0;
         
         /* Calculate the volume, and display it regardless if we are actively
          * recording to disk. */
         for (i = 0; i < sample_count; ++i) {
            if (gain < abs(input[i] - sample_center))
               gain = abs(input[i] - sample_center);
         }
        
         al_clear_to_color(al_map_rgb(0,0,0));
        
         if (is_recording) {
            /* Save raw bytes to disk. Assumes everything is written
             * succesfully. */
            if (fp && n < frequency / (float) samples_per_fragment * 
               max_seconds_to_record) {
               al_fwrite(fp, input, sample_count * sample_size);
               ++n;
            }

            /* Draw a pathetic visualization. It draws exactly one fragment
             * per frame. This means the visualization is dependent on the 
             * various parameters. A more thorough implementation would use this
             * event to copy the new data into a circular buffer that holds a
             * few seconds of audio. The graphics routine could then always
             * draw that last second of audio, which would cause the
             * visualization to appear constant across all different settings.
             */
            for (i = 0; i < 320; ++i) {
               int j, c = 0;
               
               /* Take the average of R samples so it fits on the screen */
               for (j = i * R; j < i * R + R && j < sample_count; ++j) {
                  c += input[j] - sample_center;
               }
               c /= R;
               
               /* Draws a line from the previous sample point to the next */
               al_draw_line(i - 1, 128 + ((prev - min_sample_val) /
                  (float) sample_range) * 256 - 128, i, 128 +
                  ((c - min_sample_val) / (float) sample_range) * 256 - 128,
                  al_map_rgb(255,255,255), 1.2);
               
               prev = c;
            }
         }
         
         /* draw volume bar */
         al_draw_filled_rectangle((gain / (float) max_sample_val) * 320, 251,
            0, 256, al_map_rgba(0, 255, 0, 128));
            
         al_flip_display();
      }
      else if (e.type == ALLEGRO_EVENT_AUDIO_STREAM_FRAGMENT) {
         /* This event is received when we are playing back the audio clip.
          * See ex_saw.c for an example dedicated to playing streams.
          */
         if (fp) {
            audio_buffer_t output = al_get_audio_stream_fragment(s);
            if (output) {
               /* Fill the buffer from the data we have recorded into the file.
                * If an error occurs (or end of file) then silence out the
                * remainder of the buffer and stop the playback.
                */
               const size_t bytes_to_read =
                  playback_samples_per_fragment * sample_size;
               size_t bytes_read = 0, i;
               
               do {
                  bytes_read += al_fread(fp, (uint8_t *)output + bytes_read,
                     bytes_to_read - bytes_read);                  
               } while (bytes_read < bytes_to_read && !al_feof(fp) &&
                  !al_ferror(fp));
               
               /* silence out unused part of buffer (end of file) */
               for (i = bytes_read / sample_size;
                  i < bytes_to_read / sample_size; ++i) {
                     output[i] = sample_center;
               }
               
               al_set_audio_stream_fragment(s, output);
               
               if (al_ferror(fp) || al_feof(fp)) {
                  al_drain_audio_stream(s);
                  al_fclose(fp);
                  fp = NULL;
               }
            }
         }
      }      
      else if (e.type == ALLEGRO_EVENT_DISPLAY_CLOSE) {
         break;
      }
      else if (e.type == ALLEGRO_EVENT_KEY_CHAR) {
         if (e.keyboard.unichar == 27) {
            /* pressed ESC */
            break;
         }
         else if (e.keyboard.unichar == ' ') {
            if (!is_recording) {
               /* Start the recording */
               is_recording = true;
               
               if (al_get_audio_stream_playing(s)) {
                  al_drain_audio_stream(s);
               }
               
               /* Reuse the same temp file for all recordings */
               if (!tmp_path) {
                  fp = al_make_temp_file("alrecXXX.raw", &tmp_path);
               }
               else {
                  if (fp) al_fclose(fp);
                  fp = al_fopen(al_path_cstr(tmp_path, '/'), "w");
               }
               
               n = 0;
            }
            else {
               is_recording = false;
               if (fp) {
                  al_fclose(fp);
                  fp = NULL;
               }
            }
         }
         else if (e.keyboard.unichar == 'p') {
            /* Play the previously recorded wav file */
            if (!is_recording) {
               if (tmp_path) {
                  fp = al_fopen(al_path_cstr(tmp_path, '/'), "r");
                  if (fp) {
                     al_set_audio_stream_playing(s, true);
                  }
               }
            }
         }
      }
   }
   
   /* clean up */
   al_destroy_audio_recorder(r);
   al_destroy_audio_stream(s);
      
   if (fp)
      al_fclose(fp);
      
   if (tmp_path) {
      al_remove_filename(al_path_cstr(tmp_path, '/'));
      al_destroy_path(tmp_path);
   }
   
   return 0;
}