1 <section id="vorbis-spec-codec">
4 $Id: 04-codec.xml,v 1.2 2002/10/14 23:47:54 giles Exp $
5 <emphasis>Last update to this document: September 20, 2002</emphasis>
8 <title>codec setup and packet decode</title>
11 <title>Overview</title>
14 This document serves as the top-level reference document for the
15 bit-by-bit decode specification of Vorbis I. This document assumes a
16 high-level understanding of the Vorbis decode process, which is
17 provided in the section <xref linkend="vorbis-spec-intro"/>. <xref
18 linkend="vorbis-spec-bitpacking"/> covers reading and writing bit fields from
19 and to bitstream packets.</para>
24 <title>Header decode and decode setup</title>
27 A Vorbis bitstream begins with three header packets. The header
28 packets are, in order, the identification header, the comments header,
29 and the setup header. All are required for decode compliance. An
30 end-of-packet condition during decoding the first or third header
31 packet renders the stream undecodable. End-of-packet decoding the
32 comment header is a non-fatal error condition.</para>
34 <section><title>Common header decode</title>
37 Each header packet begins with the same header fields.
41 1) [packet_type] : 8 bit value
42 2) 0x76, 0x6f, 0x72, 0x62, 0x69, 0x73: the characters 'v','o','r','b','i','s' as six octets
46 Decode continues according to packet type; the identification header
47 is type 1, the comment header type 3 and the setup header type 5
48 (these types are all odd as a packet with a leading single bit of '0'
49 is an audio packet). The packets must occur in the order of
50 identification, comment, setup.</para>
54 <section><title>Identification Header</title>
57 The identification header is a short header of only a few fields used
58 to declare the stream definitively as Vorbis, and provide a few externally
59 relevant pieces of information about the audio stream. The
60 identification header is coded as follows:</para>
63 1) [vorbis_version] = read 32 bits as unsigned integer
64 2) [audio_channels] = read 8 bit integer as unsigned
65 3) [audio_sample_rate] = read 32 bits as unsigned integer
66 4) [bitrate_maximum] = read 32 bits as signed integer
67 5) [bitrate_nominal] = read 32 bits as signed integer
68 6) [bitrate_minimum] = read 32 bits as signed integer
69 7) [blocksize_0] = 2 exponent (read 4 bits as unsigned integer)
70 8) [blocksize_1] = 2 exponent (read 4 bits as unsigned integer)
71 9) [framing_flag] = read one bit
75 <varname>[vorbis_version]</varname> is to read '0' in order to be compatible
76 with this document. Both <varname>[audio_channels]</varname> and
77 <varname>[audio_sample_rate]</varname> must read greater than zero. Allowed final
78 blocksize values are 64, 128, 256, 512, 1024, 2048, 4096 and 8192 in
79 Vorbis I. <varname>[blocksize_0]</varname> must be less than or equal to
80 <varname>[blocksize_1]</varname>. The framing bit must be nonzero. Failure to
81 meet any of these conditions renders a stream undecodable.</para>
84 The bitrate fields above are used only as hints. The nominal bitrate
85 field especially may be considerably off in purely VBR streams. The
86 fields are meaningful only when greater than zero.</para>
89 <listitem><simpara>All three fields set to the same value implies a fixed rate, or tightly bounded, nearly fixed-rate bitstream</simpara></listitem>
90 <listitem><simpara>Only nominal set implies a VBR or ABR stream that averages the nominal bitrate</simpara></listitem>
91 <listitem><simpara>Maximum and or minimum set implies a VBR bitstream that obeys the bitrate limits</simpara></listitem>
92 <listitem><simpara>None set indicates the encoder does not care to speculate.</simpara></listitem>
97 <section><title>Comment Header</title>
99 Comment header decode and data specification is covered in
100 <xref linkend="vorbis-spec-comment"/>.</para>
103 <section><title>Setup Header</title>
106 Vorbis codec setup is configurable to an extreme degree:
110 <imagedata fileref="components.png" format="PNG"/>
113 <phrase>[decoder pipeline configuration]</phrase>
119 The setup header contains the bulk of the codec setup information
120 needed for decode. The setup header contains, in order, the lists of
121 codebook configurations, time-domain transform configurations
122 (placeholders in Vorbis I), floor configurations, residue
123 configurations, channel mapping configurations and mode
124 configurations. It finishes with a framing bit of '1'. Header decode
125 proceeds in the following order:</para>
127 <section><title>codebooks</title>
130 <listitem><simpara><varname>[vorbis_codebook_count]</varname> = read eight bits as unsigned integer and add one</simpara></listitem>
131 <listitem><simpara>Decode <varname>[vorbis_codebook_count]</varname> codebooks in order as defined
132 in <xref linkend="vorbis-spec-codebook"/>. Save each configuration, in
133 order, in an array of
134 codebook configurations <varname>[vorbis_codebook_configurations]</varname>.</simpara></listitem>
139 <section><title>time domain transforms</title>
142 These hooks are placeholders in Vorbis I. Nevertheless, the
143 configuration placeholder values must be read to maintain bitstream
147 <listitem><simpara><varname>[vorbis_time_count]</varname> = read 6 bits as unsigned integer and add one</simpara></listitem>
148 <listitem><simpara>read <varname>[vorbis_time_count]</varname> 16 bit values; each value should be zero. If any value is nonzero, this is an error condition and the stream is undecodable.</simpara></listitem>
153 <section><title>floors</title>
156 Vorbis uses two floor types; header decode is handed to the decode
157 abstraction of the appropriate type.</para>
160 <listitem><simpara><varname>[vorbis_floor_count]</varname> = read 6 bits as unsigned integer and add one</simpara></listitem>
161 <listitem><para>For each <varname>[i]</varname> of <varname>[vorbis_floor_count]</varname> floor numbers:
163 <listitem><simpara>read the floor type: vector <varname>[vorbis_floor_types]</varname> element <varname>[i]</varname> =
164 read 16 bits as unsigned integer</simpara></listitem>
165 <listitem><simpara>If the floor type is zero, decode the floor
166 configuration as defined in <xref linkend="vorbis-spec-floor0"/>; save
168 configuration in slot <varname>[i]</varname> of the floor configuration array <varname>[vorbis_floor_configurations]</varname>.</simpara></listitem>
169 <listitem><simpara>If the floor type is one,
170 decode the floor configuration as defined in <xref
171 linkend="vorbis-spec-floor1"/>; save this configuration in slot <varname>[i]</varname> of the floor configuration array <varname>[vorbis_floor_configurations]</varname>.</simpara></listitem>
172 <listitem><simpara>If the the floor type is greater than one, this stream is undecodable; ERROR CONDITION</simpara></listitem>
179 <section><title>residues</title>
182 Vorbis uses three residue types; header decode of each type is identical.
186 <listitem><simpara><varname>[vorbis_residue_count]</varname> = read 6 bits as unsigned integer and add one
187 </simpara></listitem>
188 <listitem><para>For each of <varname>[vorbis_residue_count]</varname> residue numbers:
190 <listitem><simpara>read the residue type; vector <varname>[vorbis_residue_types]</varname> element <varname>[i]</varname> = read 16 bits as unsigned integer</simpara></listitem>
191 <listitem><simpara>If the residue type is zero,
192 one or two, decode the residue configuration as defined in <xref
193 linkend="vorbis-spec-residue"/>; save this configuration in slot <varname>[i]</varname> of the residue configuration array <varname>[vorbis_residue_configurations]</varname>.</simpara></listitem>
194 <listitem><simpara>If the the residue type is greater than two, this stream is undecodable; ERROR CONDITION</simpara></listitem>
201 <section><title>mappings</title>
204 Mappings are used to set up specific pipelines for encoding
205 multichannel audio with varying channel mapping applications. Vorbis I
206 uses a single mapping type (0), with implicit PCM channel mappings.</para>
209 <listitem><simpara><varname>[vorbis_mapping_count]</varname> = read 6 bits as unsigned integer and add one</simpara></listitem>
210 <listitem><simpara>For each <varname>[i]</varname>
211 of <varname>[vorbis_mapping_count]</varname> mapping numbers:</simpara>
213 <listitem><simpara>read the mapping type: 16 bits as unsigned integer. There's no reason to save the mapping type in Vorbis I.</simpara></listitem>
214 <listitem><simpara>If the mapping type is nonzero, the stream is undecodable</simpara></listitem>
215 <listitem><para>If the mapping type is zero:
217 <listitem><simpara>read 1 bit as a boolean flag</simpara>
219 <listitem><simpara>if set, <varname>[vorbis_mapping_submaps]</varname> = read 4 bits as unsigned integer and add one</simpara></listitem>
220 <listitem><simpara>if unset, <varname>[vorbis_mapping_submaps]</varname> = 1</simpara></listitem>
223 <listitem><simpara>read 1 bit as a boolean flag; if set, square polar channel mapping is in use:</simpara>
225 <listitem><simpara><varname>[vorbis_mapping_coupling_steps]</varname>= read 8 bits as unsigned integer and add one</simpara></listitem>
226 <listitem><simpara>for <varname>[j]</varname> each of <varname>[vorbis_mapping_coupling_steps]</varname>
229 <listitem><simpara>vector <varname>[vorbis_mapping_magnitude]</varname> element
230 <varname>[j]</varname>= read <link linkend="vorbis-spec-ilog">ilog</link>([audio_channels] - 1) bits as
231 unsigned integer</simpara></listitem>
232 <listitem><simpara>vector <varname>[vorbis_mapping_angle]</varname> element <varname>[j]</varname>=
233 read <link linkend="vorbis-spec-ilog">ilog</link>([audio_channels] - 1)
234 bits as unsigned integer</simpara></listitem>
235 <listitem><simpara>the numbers read in the above two steps are channel numbers representing the channel to treat as
236 magnitude and the channel to treat as angle, respectively. If for any coupling step the angle channel number equals the
237 magnitude channel number, the magnitude channel number is greater than <varname>[audio_channels]</varname>-1, or the angle
238 channel is greater than <varname>[audio_channels]</varname>-1, the stream is undecodable.</simpara></listitem>
243 <listitem><simpara>read 2 bits (reserved field); if the value is nonzero, the stream is undecodable</simpara></listitem>
244 <listitem><simpara>if <varname>[vorbis_mapping_submaps]</varname> is greater than one, we read channel multiplex
245 settings. For each <varname>[j]</varname> of <varname>[audio_channels]</varname> channels:</simpara>
247 <listitem><simpara>vector <varname>[vorbis_mapping_mux]</varname> element <varname>[j]</varname> = read 4 bits as
248 unsigned integer</simpara></listitem>
249 <listitem><simpara>if the value is greater than the highest numbered submap
250 (<varname>[vorbis_mapping_submaps]</varname> - 1), this in an error condition rendering the stream
251 undecodable</simpara></listitem>
254 <listitem><simpara>for each submap <varname>[j]</varname> of <varname>[vorbis_mapping_submaps]</varname> submaps, read
255 the floor and residue numbers for use in decoding that submap:</simpara>
257 <listitem><simpara>read and discard 8 bits (the unused time configuration placeholder)</simpara></listitem>
258 <listitem><simpara>read 8 bits as unsigned integer for the floor number; save in vector
259 <varname>[vorbis_mapping_submap_floor]</varname> element <varname>[j]</varname></simpara></listitem>
260 <listitem><simpara>verify the floor number is not greater than the highest number floor configured for the bitstream.
261 If it is, the bitstream is undecodable</simpara></listitem>
262 <listitem><simpara>read 8 bits as unsigned integer for the residue number; save in vector
263 <varname>[vorbis_mapping_submap_residue]</varname> element <varname>[j]</varname></simpara></listitem>
264 <listitem><simpara>verify the residue number is not greater than the highest number residue configured for the
265 bitstream. If it is, the bitstream is undecodable</simpara></listitem>
268 <listitem><simpara>save this mapping configuration in slot <varname>[i]</varname> of the mapping configuration array
269 <varname>[vorbis_mapping_configurations]</varname>.</simpara></listitem>
270 </orderedlist></para>
278 <section><title>modes</title>
281 <listitem><simpara><varname>[vorbis_mode_count]</varname> = read 6 bits as unsigned integer and add one</simpara></listitem>
282 <listitem><simpara>For each of <varname>[vorbis_mode_count]</varname> mode numbers:</simpara>
284 <listitem><simpara><varname>[vorbis_mode_blockflag]</varname> = read 1 bit</simpara></listitem>
285 <listitem><simpara><varname>[vorbis_mode_windowtype]</varname> = read 16 bits as unsigned integer</simpara></listitem>
286 <listitem><simpara><varname>[vorbis_mode_transformtype]</varname> = read 16 bits as unsigned integer</simpara></listitem>
287 <listitem><simpara><varname>[vorbis_mode_mapping]</varname> = read 8 bits as unsigned integer</simpara></listitem>
288 <listitem><simpara>verify ranges; zero is the only legal value in Vorbis I for
289 <varname>[vorbis_mode_windowtype]</varname>
290 and <varname>[vorbis_mode_transformtype]</varname>. <varname>[vorbis_mode_mapping]</varname> must not be greater than the highest number mapping in use. Any illegal values render the stream undecodable.</simpara></listitem>
291 <listitem><simpara>save this mode configuration in slot <varname>[i]</varname> of the mode configuration array
292 <varname>[vorbis_mode_configurations]</varname>.</simpara></listitem>
295 <listitem><simpara>read 1 bit as a framing flag. If unset, a framing error occurred and the stream is not
296 decodable.</simpara></listitem>
300 After reading mode descriptions, setup header decode is complete.
310 <title>Audio packet decode and synthesis</title>
313 Following the three header packets, all packets in a Vorbis I stream
314 are audio. The first step of audio packet decode is to read and
315 verify the packet type; <emphasis>a non-audio packet when audio is expected
316 indicates stream corruption or a non-compliant stream. The decoder
317 must ignore the packet and not attempt decoding it to audio</emphasis>.
320 <section><title>packet type, mode and window decode</title>
323 <listitem><simpara>read 1 bit <varname>[packet_type]</varname>; check that packet type is 0 (audio)</simpara></listitem>
324 <listitem><simpara>read <link linkend="vorbis-spec-ilog">ilog</link>([vorbis_mode_count]-1) bits
325 <varname>[mode_number]</varname></simpara></listitem>
326 <listitem><simpara>decode blocksize <varname>[n]</varname> is equal to <varname>[blocksize_0]</varname> if
327 <varname>[vorbis_mode_blockflag]</varname> is 0, else <varname>[n]</varname> is equal to <varname>[blocksize_1]</varname>.</simpara></listitem>
328 <listitem><simpara>perform window selection and setup; this window is used later by the inverse MDCT:</simpara>
330 <listitem><simpara>if this is a long window (the <varname>[vorbis_mode_blockflag]</varname> flag of this mode is
333 <listitem><simpara>read 1 bit for <varname>[previous_window_flag]</varname></simpara></listitem>
334 <listitem><simpara>read 1 bit for <varname>[next_window_flag]</varname></simpara></listitem>
335 <listitem><simpara>if <varname>[previous_window_flag]</varname> is not set, the left half
336 of the window will be a hybrid window for lapping with a
337 short block. See <xref
338 linkend="vorbis-spec-window"/> for an illustration of overlapping
340 windows. Else, the left half window will have normal long
341 shape.</simpara></listitem>
342 <listitem><simpara>if <varname>[next_window_flag]</varname> is not set, the right half of
343 the window will be a hybrid window for lapping with a short
344 block. See <xref linkend="vorbis-spec-window"/> for an
345 illustration of overlapping dissimilar
346 windows. Else, the left right window will have normal long
347 shape.</simpara></listitem>
350 <listitem><simpara> if this is a short window, the window is always the same
351 short-window shape.</simpara></listitem>
357 Vorbis windows all use the slope function y=sin(2*PI*sin^2(x/n)),
358 but dissimilar lapping requirements can affect overall shape. Window
359 generation proceeds as follows:</para>
362 <listitem><simpara> <varname>[window_center]</varname> = <varname>[n]</varname> / 2</simpara></listitem>
363 <listitem><simpara> <varname>[left_window_start]</varname></simpara></listitem>
364 <listitem><para> if (<varname>[vorbis_mode_blockflag]</varname> is set and <varname>[previous_window_flag]</varname> is
367 <listitem><simpara><varname>[left_window_start]</varname> = <varname>[n]</varname>/4 -
368 <varname>[blocksize_0]</varname>/4</simpara></listitem>
369 <listitem><simpara><varname>[left_window_end]</varname> = <varname>[n]</varname>/4 + <varname>[blocksize_0]</varname>/4</simpara></listitem>
370 <listitem><simpara><varname>[left_n]</varname> = <varname>[blocksize_0]</varname>/2</simpara></listitem>
374 <listitem><simpara><varname>[left_window_start]</varname> = 0</simpara></listitem>
375 <listitem><simpara><varname>[left_window_end]</varname> = <varname>[window_center]</varname></simpara></listitem>
376 <listitem><simpara><varname>[left_n]</varname> = <varname>[n]</varname>/2</simpara></listitem>
377 </orderedlist></para>
379 <listitem><para> if (<varname>[vorbis_mode_blockflag]</varname> is set and <varname>[next_window_flag]</varname> is not
382 <listitem><simpara><varname>[right_window_start]</varname> = <varname>[n]*3</varname>/4 -
383 <varname>[blocksize_0]</varname>/4</simpara></listitem>
384 <listitem><simpara><varname>[right_window_end]</varname> = <varname>[n]*3</varname>/4 +
385 <varname>[blocksize_0]</varname>/4</simpara></listitem>
386 <listitem><simpara><varname>[right_n]</varname> = <varname>[blocksize_0]</varname>/2</simpara></listitem>
390 <listitem><simpara><varname>[right_window_start]</varname> = <varname>[window_center]</varname></simpara></listitem>
391 <listitem><simpara><varname>[right_window_end]</varname> = <varname>[n]</varname></simpara></listitem>
392 <listitem><simpara><varname>[right_n]</varname> = <varname>[n]</varname>/2</simpara></listitem>
393 </orderedlist></para>
395 <listitem><simpara> window from range 0 ... <varname>[left_window_start]</varname>-1 inclusive is zero</simpara></listitem>
396 <listitem><simpara> for <varname>[i]</varname> in range <varname>[left_window_start]</varname> ...
397 <varname>[left_window_end]</varname>-1, window(<varname>[i]</varname>) = sin(2*PI*sin^2((<varname>[i]</varname>-<varname>[left_window_start]</varname>+.5)/<varname>[left_n]</varname>*PI/2))</simpara></listitem>
398 <listitem><simpara> window from range <varname>[left_window_end]</varname> ... <varname>[right_window_start]</varname>-1
399 inclusive is one</simpara></listitem><listitem><simpara> for <varname>[i]</varname> in range <varname>[right_window_start]</varname> ... <varname>[right_window_end]</varname>-1, window(<varname>[i]</varname>) = sin(2*PI*sin^2((<varname>[i]</varname>-<varname>[right_window_start]</varname>+.5)/<varname>[right_n]</varname>*PI/2.+PI/2.))</simpara></listitem>
400 <listitem><simpara> window from range <varname>[rigth_window_start]</varname> ... <varname>[n]</varname>-1 is
401 zero</simpara></listitem>
405 An end-of-packet condition up to this point should be considered an
406 error that discards this packet from the stream. An end of packet
407 condition past this point is to be considered a possible nominal
412 <section><title>floor curve decode</title>
415 From this point on, we assume out decode context is using mode number
416 <varname>[mode_number]</varname> from configuration array
417 <varname>[vorbis_mode_configurations]</varname> and the map number
418 <varname>[vorbis_mode_mapping]</varname> (specified by the current mode) taken
419 from the mapping configuration array
420 <varname>[vorbis_mapping_configurations]</varname>.</para>
423 Floor curves are decoded one-by-one in channel order.</para>
426 For each floor <varname>[i]</varname> of <varname>[audio_channels]</varname>
428 <listitem><simpara><varname>[submap_number]</varname> = element <varname>[i]</varname> of vector [vorbis_mapping_mux]</simpara></listitem>
429 <listitem><simpara><varname>[floor_number]</varname> = element <varname>[submap_number]</varname> of vector
430 [vorbis_submap_floor]</simpara></listitem>
431 <listitem><simpara>if the floor type of this
432 floor (vector <varname>[vorbis_floor_types]</varname> element
433 <varname>[floor_number]</varname>) is zero then decode the floor for
434 channel <varname>[i]</varname> according to the
435 <xref linkend="vorbis-spec-floor0-decode"/>floor 0 decode
436 algorithm</simpara></listitem>
437 <listitem><simpara>if the type of this floor
438 is one then decode the floor for channel <varname>[i]</varname> according
439 to the <xref linkend="vorbis-spec-floor1-decode"/>floor 1 decode
440 algorithm</simpara></listitem>
441 <listitem><simpara>save the needed decoded floor information for channel for later synthesis</simpara></listitem>
442 <listitem><simpara>if the decoded floor returned 'unused', set vector <varname>[no_residue]</varname> element
443 <varname>[i]</varname> to true, else set vector <varname>[no_residue]</varname> element <varname>[i]</varname> to
444 false</simpara></listitem>
449 An end-of-packet condition during floor decode shall result in packet
450 decode zeroing all channel output vectors and skipping to the
451 add/overlap output stage.</para>
455 <section><title>nonzero vector propagate</title>
458 A possible result of floor decode is that a specific vector is marked
459 'unused' which indicates that that final output vector is all-zero
460 values (and the floor is zero). The residue for that vector is not
461 coded in the stream, save for one complication. If some vectors are
462 used and some are not, channel coupling could result in mixing a
463 zeroed and nonzeroed vector to produce two nonzeroed vectors.</para>
466 for each <varname>[i]</varname> from 0 ... <varname>[vorbis_mapping_coupling_steps]</varname>-1
469 <listitem><simpara>if either <varname>[no_residue]</varname> entry for channel
470 (<varname>[vorbis_mapping_magnitude]</varname> element <varname>[i]</varname>) or (channel
471 <varname>[vorbis_mapping_angle]</varname> element <varname>[i]</varname>) are set to false, then both
472 must be set to false. Note that an 'unused' floor has no decoded floor
473 information; it is important that this is remembered at floor curve
474 synthesis time.</simpara></listitem>
480 <section><title>residue decode</title>
483 Unlike floors, which are decoded in channel order, the residue vectors
484 are decoded in submap order.</para>
487 for each submap <varname>[i]</varname> in order from 0 ... <varname>[vorbis_mapping_submaps]</varname>-1</para>
490 <listitem><simpara><varname>[ch]</varname> = 0</simpara></listitem>
491 <listitem><simpara>for each channel <varname>[j]</varname> in order from 0 ... <varname>[audio_channels]</varname></simpara>
493 <listitem><simpara>if channel <varname>[j]</varname> is in submap <varname>[i]</varname> (vector <varname>[vorbis_mapping_mux]</varname> element <varname>[j]</varname> is equal to <varname>[i]</varname>)</simpara>
495 <listitem><para>if vector <varname>[no_residue]</varname> element <varname>[j]</varname> is true
497 <listitem><simpara>vector <varname>[do_not_decode_flag]</varname> element <varname>[channels_in_bundle]</varname> is set</simpara></listitem>
501 <listitem><simpara>vector <varname>[do_not_decode_flag]</varname> element <varname>[channels_in_bundle]</varname> is unset</simpara></listitem>
502 </orderedlist></para>
504 <listitem><simpara>increment <varname>[ch]</varname></simpara></listitem>
508 </listitem><listitem><simpara><varname>[residue_number]</varname> = vector <varname>[vorbis_mapping_submap_residue]</varname> element <varname>[i]</varname></simpara></listitem>
509 <listitem><simpara><varname>[residue_type]</varname> = vector <varname>[vorbis_residue_types]</varname> element <varname>[residue_number]</varname></simpara></listitem>
510 <listitem><simpara>decode <varname>[ch]</varname> vectors using residue <varname>[residue_number]</varname>, according to type <varname>[residue_type]</varname>, also passing vector <varname>[do_not_decode_flag]</varname> to indicate which vectors in the bundle should not be decoded. Correct per-vector decode length is <varname>[n]</varname>/2.</simpara></listitem>
511 <listitem><simpara><varname>[ch]</varname> = 0</simpara></listitem>
512 <listitem><simpara>for each channel <varname>[j]</varname> in order from 0 ... <varname>[audio_channels]</varname></simpara>
514 <listitem><simpara>if channel <varname>[j]</varname> is in submap <varname>[i]</varname> (vector <varname>[vorbis_mapping_mux]</varname> element <varname>[j]</varname> is equal to <varname>[i]</varname>)</simpara>
516 <listitem><simpara>residue vector for channel <varname>[j]</varname> is set to decoded residue vector <varname>[ch]</varname></simpara></listitem>
517 <listitem><simpara>increment <varname>[ch]</varname></simpara></listitem>
526 <section><title>inverse coupling</title>
529 for each <varname>[i]</varname> from <varname>[vorbis_mapping_coupling_steps]</varname>-1 descending to 0
532 <listitem><simpara><varname>[magnitude_vector]</varname> = the residue vector for channel
533 (vector <varname>[vorbis_mapping_magnitude]</varname> element <varname>[i]</varname>)</simpara></listitem>
534 <listitem><simpara><varname>[angle_vector]</varname> = the residue vector for channel (vector
535 <varname>[vorbis_mapping_angle]</varname> element <varname>[i]</varname>)</simpara></listitem>
536 <listitem><simpara>for each scalar value <varname>[M]</varname> in vector <varname>[magnitude_vector]</varname> and the corresponding scalar value <varname>[A]</varname> in vector <varname>[angle_vector]</varname>:</simpara>
538 <listitem><para>if (<varname>[M]</varname> is greater than zero)
540 <listitem><para>if (<varname>[A]</varname> is greater than zero)
542 <listitem><simpara><varname>[new_M]</varname> = <varname>[M]</varname></simpara></listitem>
543 <listitem><simpara><varname>[new_A]</varname> = <varname>[M]</varname>-<varname>[A]</varname></simpara></listitem>
547 <listitem><simpara><varname>[new_A]</varname> = <varname>[M]</varname></simpara></listitem>
548 <listitem><simpara><varname>[new_M]</varname> = <varname>[M]</varname>+<varname>[A]</varname></simpara></listitem>
554 <listitem><para>if (<varname>[A]</varname> is greater than zero)
556 <listitem><simpara><varname>[new_M]</varname> = <varname>[M]</varname></simpara></listitem>
557 <listitem><simpara><varname>[new_A]</varname> = <varname>[M]</varname>+<varname>[A]</varname></simpara></listitem>
561 <listitem><simpara><varname>[new_A]</varname> = <varname>[M]</varname></simpara></listitem>
562 <listitem><simpara><varname>[new_M]</varname> = <varname>[M]</varname>-<varname>[A]</varname></simpara></listitem>
567 <listitem><simpara>set scalar value <varname>[M]</varname> in vector <varname>[magnitude_vector]</varname> to <varname>[new_M]</varname></simpara></listitem>
568 <listitem><simpara>set scalar value <varname>[A]</varname> in vector <varname>[angle_vector]</varname> to <varname>[new_A]</varname></simpara></listitem>
576 <section><title>dot product</title>
579 For each channel, synthesize the floor curve from the decoded floor
580 information, according to packet type. Note that the vector synthesis
581 length for floor computation is <varname>[n]</varname>/2.</para>
584 For each channel, multiply each element of the floor curve by each
585 element of that channel's residue vector. The result is the dot
586 product the floor and residue vectors for each channel; the produced
587 vectors are the length <varname>[n]</varname>/2 audio spectrum for each
591 One point is worth mentioning about this dot product; a common mistake
592 in a fixed point implementation might be to assume that a 32 bit
593 fixed-point representation for floor and residue and direct
594 multiplication of the vectors is sufficient for acceptable spectral
595 depth in all cases because it happens to mostly work with the current
596 Xiph.Org reference encoder. </para>
599 However, floor vector values can span ~140dB (~24 bits unsigned), and
600 the audio spectrum vector should represent a minimum of 120dB (~21
601 bits with sign), even when output is to a 16 bit PCM device. For the
602 residue vector to represent full scale if the floor is nailed to
603 -140dB, it must be able to span 0 to +140dB. For the residue vector
604 to reach full scale if the floor is nailed at 0dB, it must be able to
605 represent -140dB to +0dB. Thus, in order to handle full range
606 dynamics, a residue vector may span -140dB to +140dB entirely within
607 spec. A 280dB range is approximately 48 bits with sign; thus the
608 residue vector must be able to represent a 48 bit range and the dot
609 product must be able to handle an effective 48 bit times 24 bit
610 multiplication. This range may be achieved using large (64 bit or
611 larger) integers, or implementing a movable binary point
612 representation.</para>
616 <section><title>inverse MDCT</title>
619 Convert the audio spectrum vector of each channel back into time
620 domain PCM audio via an inverse Modified Discrete Cosine Transform
621 (MDCT). A detailed description of the MDCT is available in the paper
622 <citetitle pubwork="article"><ulink
623 url="http://www.iocon.com/resource/docs/ps/eusipco_corrected.ps">The
624 use of multirate filter banks for coding of high quality digital
625 audio</ulink></citetitle>, by T. Sporer, K. Brandenburg and B. Edler. The window
626 function used for the MDCT is the window determined earlier.</para>
630 <section><title>overlap_add</title>
633 Windowed MDCT output is overlapped and added with the right hand data
634 of the previous window such that the 3/4 point of the previous window
635 is aligned with the 1/4 point of the current window (as illustrated in
636 <xref linkend="vorbis-spec-window"/>. The overlapped portion
637 produced from overlapping the previous and current frame data is
638 finished data to be returned by the decoder. This data spans from the
639 center of the previous window to the center of the current window. In
640 the case of same-sized windows, the amount of data to return is
641 one-half block consisting of and only of the overlapped portions. When
642 overlapping a short and long window, much of the returned range is not
643 actually overlap. This does not damage transform orthogonality. Pay
644 attention however to returning the correct data range; the amount of
645 data to be returned is:
647 <programlisting>window_blocksize(previous_window)/4+window_blocksize(current_window)/4</programlisting>
649 from the center (element windowsize/2) of the previous window to the
650 center (element windowsize/2-1, inclusive) of the current window.</para>
653 Data is not returned from the first frame; it must be used to 'prime'
654 the decode engine. The encoder accounts for this priming when
655 calculating PCM offsets; after the first frame, the proper PCM output
656 offset is '0' (as no data has been returned yet).</para>
660 <section><title>output channel order</title>
663 Vorbis I specifies only a channel mapping type 0. In mapping type 0,
664 channel mapping is implicitly defined as follows for standard audio
669 <term>one channel</term>
671 <simpara>the stream is monophonic</simpara>
673 </varlistentry><varlistentry>
674 <term>two channels</term><listitem>
675 <simpara>the stream is stereo. channel order: left, right</simpara>
677 </varlistentry><varlistentry>
678 <term>three channels</term><listitem>
679 <simpara>the stream is a 1d-surround encoding. channel order: left,
680 center, right</simpara>
682 </varlistentry><varlistentry>
683 <term>four channels</term><listitem>
684 <simpara>the stream is quadraphonic surround. channel order: front left,
685 front right, rear left, rear right</simpara>
687 </varlistentry><varlistentry>
688 <term>five channels</term><listitem>
689 <simpara>the stream is five-channel surround. channel order: front left,
690 front center, front right, rear left, rear right</simpara>
692 </varlistentry><varlistentry>
693 <term>six channels</term><listitem>
694 <simpara>the stream is 5,1 surround. channel order: front left, front
695 center, front right, rear left, rear right, LFE</simpara>
697 </varlistentry><varlistentry>
698 <term>greater than six channels</term><listitem>
699 <simpara>channel use and order is defined by the application</simpara>
705 Applications using Vorbis for dedicated purposes may define channel
706 mapping as seen fit. Future channel mappings (such as three and four
707 channel <ulink url="http://www.ambisonic.net/">Ambisonics</ulink>) will
708 make use of channel mappings other than mapping 0.</para>