hvirtual/quicktime/docs/reading.html

   1 <TITLE>Reading</TITLE>
   2
   3 <H1>Reading a file</H1>
   4
   5 Most users simply want read access for a player toy or something.  A
   6 good place to start is before opening the file, making sure it is
   7 Quicktime with quicktime_check_sig().<P>
   8
   9 <CODE>
  10 quicktime_check_sig("path");<P>
  11 </CODE>
  12
  13 This returns 1 if it looks like a Quicktime file or 0 if it doesn't.
  14 Then you can open the file as described in <A
  15 HREF="opening.html">opening.html</A>.<P>
  16
  17 Next get the number of tracks for each media type in the file:<P>
  18
  19 <CODE>
  20 int quicktime_video_tracks(quicktime_t *file);<BR>
  21 int quicktime_audio_tracks(quicktime_t *file);
  22 </CODE>
  23 <P>
  24
  25
  26 While Quicktime can store multiple video tracks, the audio track count
  27 is a bit more complicated.  Usually you'll only encounter a single
  28 audio track.  Inside the audio track is a variable number of channels.
  29 To get the channel count call:<P>
  30
  31 <CODE>
  32 int quicktime_track_channels(quicktime_t *file, int track);
  33 </CODE>
  34 <P>
  35
  36 With the track parameter set to track 0.  Many routines require a
  37 <B>track</B> parameter to specify the track to operate on.  Tracks are
  38 always numbered from 0 to the total number of tracks - 1 for the
  39 particular media type.<P>
  40
  41 Audio tracks are numbered from 0 to the total number of audio tracks -
  42 1.  But like I said, you'll probably never encounter an audio track
  43 higher than 0.  Other routines you might find useful for getting audio
  44 information are:<P>
  45
  46 <CODE>
  47 long quicktime_sample_rate(quicktime_t *file, int track);<BR>
  48 long quicktime_audio_length(quicktime_t *file, int track);<BR>
  49 </CODE>
  50 <P>
  51
  52 quicktime_audio_length gives you the total number of samples.  The
  53 sample rate is samples per second.<P>
  54
  55 Routines you'll never use unless you want to write a codec are:<P>
  56
  57 <CODE>
  58 char* quicktime_audio_compressor(quicktime_t *file, int track);<BR>
  59 int quicktime_audio_bits(quicktime_t *file, int track);<BR>
  60 </CODE>
  61 <P>
  62
  63 The audio compressor call returns a 4 byte array identifying the data
  64 compression of the track.  These identifiers are 4 alphanumeric
  65 characters which go along with one of the #defines in quicktime.h.  The
  66 bits function returns the number of bits in a sample, usually
  67 meaningless.<P>
  68
  69 The most interesting contents of a Quicktime file are of course the
  70 video tracks.  Quicktime stores multiple video tracks.<P>
  71
  72 The available queries for each video track are:<P>
  73
  74
  75 <CODE>
  76 long quicktime_video_length(quicktime_t *file, int track);<BR>
  77 int quicktime_video_width(quicktime_t *file, int track);<BR>
  78 int quicktime_video_height(quicktime_t *file, int track);<BR>
  79 float quicktime_frame_rate(quicktime_t *file, int track);<BR>
  80 long quicktime_frame_size(quicktime_t *file, long frame, int track);<BR>
  81 int quicktime_video_depth(quicktime_t *file, int track);<BR>
  82 quicktime_reads_cmodel(quicktime_t *file,
  83                 int colormodel,
  84                 int track);<BR>
  85 </CODE>
  86 <P>
  87
  88
  89
  90 Tracks are numbered 0 to the total number of tracks - 1.  The video
  91 length is in frames.  The width and height are in pixels.  The frame
  92 rate is in frames per second.  Depth returns the total number of bits
  93 per pixel.  The only two values Quicktime for Linux returns are 24 and
  94 32 and the 32 bit depth is only returned when the format has an alpha
  95 channel.  There's no reason to use 16 or 8.<P>
  96
  97 <CODE>quicktime_reads_cmodel</CODE> allows you to determine the optimum
  98 color model for decompression output.  It requires a colormodel #define
  99 from colormodels.h.  If the codec can generate the desired colormodel
 100 without downsampling it returns 1.  If downsampling is required it
 101 returns 0.  You can assume all colormodels in colormodels.h are
 102 supported, whether they require downsampling or not.<P>
 103
 104 To get the four byte compressor type for the track issue:<P>
 105
 106 <CODE>
 107 char* quicktime_video_compressor(quicktime_t *file, int track);<BR>
 108 </CODE>
 109 <P>
 110
 111 Unless you get a really nihilistic file for reading, you can safely
 112 assume the encoding scheme for track 0 of audio or video is the same
 113 for all tracks.<P>
 114
 115 <A NAME="Decodingvideo">
 116 <H1>Decoding video</H1>
 117
 118 The library decodes compressed video frames into a buffer in whatever
 119 colormodel you desire but before then you should issue<P>
 120
 121 <CODE>
 122 int quicktime_supported_video(quicktime_t *file, int track);<BR>
 123 </CODE>
 124 <P>
 125
 126 to find out if the data for the track can be decoded by the library.
 127 This returns 1 if it is and 0 if it isn't supported.<P>
 128
 129 Then use<P>
 130
 131 <CODE>
 132 <PRE>
 133 long quicktime_decode_video(quicktime_t *file,
 134         unsigned char **row_pointers,
 135         int track);<BR>
 136 </PRE>
 137 </CODE>
 138 <P>
 139
 140 to decompress a frame at the current position of the track into
 141 <CODE>**row_pointers</CODE> and advance the current position.  The
 142 array of rows must have enough space allocated for the entire frame,
 143 depending on the colormodel.  Planar colormodels use only the first 3
 144 row pointers, each pointing to one of the planes.<P>
 145
 146 Several parameters determine the decoder output.  They may be set
 147 before the call to <CODE>quicktime_decode_video</CODE>.<P>
 148
 149 <CODE>
 150 <PRE>
 151 void quicktime_set_cmodel(quicktime_t *file, int colormodel);
 152 </PRE>
 153 </CODE>
 154 <P>
 155
 156 Set the colormodel of the output frame to a value in colormodels.h.
 157 The default is <CODE>BC_RGB888</CODE>.<P>
 158
 159 <CODE>
 160 <PRE>
 161 void quicktime_set_row_span(quicktime_t *file, int row_span);
 162 </PRE>
 163 </CODE>
 164 <P>
 165
 166 Set the number of bytes in a row.  The default is the row width * bytes
 167 per pixel.<P>
 168
 169 <CODE>
 170 <PRE>
 171 void quicktime_set_window(quicktime_t *file,
 172         int in_x,                    /* Location of input frame to take picture */
 173         int in_y,
 174         int in_w,
 175         int in_h,
 176         int out_w,                   /* Dimensions of output frame */
 177         int out_h);
 178 </PRE>
 179 </CODE>
 180 <P>
 181
 182
 183
 184 The decoder "sees" a region of the movie screen defined by <CODE>in_x,
 185 in_y, in_w, in_h</CODE> and transfers it to the frame buffer defined by
 186 <CODE>**row_pointers</CODE>.  The size of the frame buffer is defined
 187 by <CODE>out_w, out_h</CODE>.  The default is a 1:1 transfer from the
 188 codec to the output frame.<P>
 189
 190
 191 For more about the track's current position go to <A
 192 HREF="positioning.html">positioning</A><P>
 193
 194
 195
 196
 197
 198
 199
 200 <H2>RAW ACCESS</H2><P>
 201
 202
 203 There are other routines for reading compressed data and chunks without
 204 a codec.  These allow you to perform direct copying of video data from
 205 one movie to the other after editing it, without recompressing it.<P>
 206
 207 <A NAME="Decodingaudio">
 208 <H1>Decoding audio</H1>
 209
 210 For reading audio, first use:<P>
 211
 212 <CODE>
 213 int quicktime_supported_audio(quicktime_t *file, int track);<BR>
 214 </CODE>
 215 <P>
 216
 217 To determine if the audio can be decompressed by the library.  This
 218 returns 1 if it is and 0 if it isn't supported.  Then use<P>
 219
 220 <CODE>
 221 int quicktime_decode_audio(quicktime_t *file, int16_t *output_i, float *output_f, long samples, int channel);<BR>
 222 </CODE>
 223 <P>
 224
 225
 226 To read a buffer's worth of samples for a single channel starting at
 227 the current position in the track.  Notice this command takes a channel
 228 argument not a track argument.  The channel argument is automatically
 229 converted into a track and channel.  Positioning information is
 230 automatically taken from the appropriate track and advanced for all the
 231 channels in the track.<P>
 232
 233 Notice the int16_t* and float* parameters.  This call can
 234 either return a buffer of int16 samples or float samples.  The argument
 235 for the data format you want should be passed a preallocated buffer big
 236 enough to contain the sample range while the undesired format should be
 237 passed NULL.  For a buffer of float samples you would say<P>
 238
 239 <CODE>
 240 result = quicktime_decode_audio(file, NULL, output_f, samples, channel);<BR>
 241 </CODE>
 242 <P>
 243
 244 For a buffer of signed int16 samples you would say<P>
 245
 246 <CODE>
 247 result = quicktime_decode_audio(file, output_i, NULL, samples, channel);<BR>
 248 </CODE>
 249 <P>
 250
 251 The data format you don't want should be passed a NULL.  The decoder
 252 automatically fills the appropriate buffer.  Floating point samples are
 253 from -1 to 0 to 1.<P>
 254
 255
 256 <A NAME="Readingrawvideo">
 257 <H1>Reading raw video</H1>
 258
 259 <CODE>
 260 long quicktime_read_frame(quicktime_t *file, unsigned char *video_buffer, int track);
 261 </CODE>
 262 <P>
 263
 264 <B>quicktime_read_frame</B> reads one frame worth of raw data from your
 265 current position on the specified video track and returns the number of
 266 bytes in the frame.  You have to make sure the buffer is big enough for
 267 the frame.   A return value of 0 means error.<P>
 268
 269 <CODE>
 270 long quicktime_frame_size(quicktime_t *file, long frame, int track);
 271 </CODE>
 272 <P>
 273
 274 gives up the number of bytes in the specified frame in the specified
 275 track even if you haven't read the frame yet.  Frame numbers start on
 276 0.<P>
 277
 278 <A NAME="Readingkeyframes">
 279 <H1>Accessing Keyframes</H1>
 280
 281 Quicktime offers very simple support for keyframes: a table of all the
 282 keyframe numbers in a track.  There are two things you can do with the
 283 keyframe table: insert keyframe numbers and retrieve keyframe
 284 numbers.<P>
 285
 286 <CODE>
 287 long quicktime_get_keyframe_before(quicktime_t *file, long frame, int track)
 288 </CODE>
 289 <P>
 290
 291 Gets the keyframe number before the <B>frame</B> argument.  The frames
 292 start on 0.<P>
 293
 294 <CODE>
 295 long quicktime_get_keyframe_after(quicktime_t *file, long frame, int track);
 296 </CODE>
 297 <P>
 298
 299 Gets the keyframe number after the <B>frame</B> argument.  The frames
 300 start on 0.<P>
 301
 302
 303 <CODE>
 304 void quicktime_insert_keyframe(quicktime_t *file, long frame, int track)
 305 </CODE>
 306 <P>
 307
 308 Inserts a keyframe into the table.  The frame argument starts on 0.<P>
 309
 310 <CODE>
 311 int quicktime_has_keyframes(quicktime_t *file, int track);
 312 </CODE>
 313 <P>
 314
 315 Returns TRUE if the track has keyframes.  The track starts on 0.<P>
 316
 317
 318 <A NAME="Readingrawaudio">
 319 <H1>Reading raw audio</H1>
 320
 321 There is no simple read or write access to raw audio.  Due to vagaries
 322 in the audio indexing and the lack of benefit in direct audio copying,
 323 you're better off using a codec.<P>