1lame(1) LAME audio compressor lame(1)
2
3
4
6 lame - create mp3 audio files
7
9 lame [options] <infile> <outfile>
10
12 LAME is a program which can be used to create compressed audio files.
13 (Lame ain't an MP3 encoder). These audio files can be played back by
14 popular MP3 players such as mpg123 or madplay. To read from stdin, use
15 "-" for <infile>. To write to stdout, use "-" for <outfile>.
16
18 Input options:
19
20 -r Assume the input file is raw pcm. Sampling rate and
21 mono/stereo/jstereo must be specified on the command line. For
22 each stereo sample, LAME expects the input data to be ordered
23 left channel first, then right channel. By default, LAME expects
24 them to be signed integers with a bitwidth of 16 and stored in
25 little-endian. Without -r, LAME will perform several fseek()'s
26 on the input file looking for WAV and AIFF headers.
27 Might not be available on your release.
28
29 -x Swap bytes in the input file (or output file when using
30 --decode).
31 For sorting out little endian/big endian type problems. If your
32 encodings sounds like static, try this first.
33 Without using -x, LAME will treat input file as native endian.
34
35 -s sfreq
36 sfreq = 8/11.025/12/16/22.05/24/32/44.1/48
37
38 Required only for raw PCM input files. Otherwise it will be
39 determined from the header of the input file.
40
41 LAME will automatically resample the input file to one of the
42 supported MP3 samplerates if necessary.
43
44 --bitwidth n
45 Input bit width per sample.
46 n = 8, 16, 24, 32 (default 16)
47
48 Required only for raw PCM input files. Otherwise it will be
49 determined from the header of the input file.
50
51 --signed
52 Instructs LAME that the samples from the input are signed (the
53 default for 16, 24 and 32 bits raw pcm data).
54
55 Required only for raw PCM input files.
56
57 --unsigned
58 Instructs LAME that the samples from the input are unsigned (the
59 default for 8 bits raw pcm data, where 0x80 is zero).
60
61 Required only for raw PCM input files and only available at
62 bitwidth 8.
63
64 --little-endian
65 Instructs LAME that the samples from the input are in little-
66 endian form.
67
68 Required only for raw PCM input files.
69
70 --big-endian
71 Instructs LAME that the samples from the input are in big-endian
72 form.
73
74 Required only for raw PCM input files.
75
76 --mp1input
77 Assume the input file is a MPEG Layer I (ie MP1) file.
78 If the filename ends in ".mp1" LAME will assume it is a MPEG
79 Layer I file. For stdin or Layer I files which do not end in
80 .mp1 you need to use this switch.
81
82 --mp2input
83 Assume the input file is a MPEG Layer II (ie MP2) file.
84 If the filename ends in ".mp2" LAME will assume it is a MPEG
85 Layer II file. For stdin or Layer II files which do not end in
86 .mp2 you need to use this switch.
87
88 --mp3input
89 Assume the input file is a MP3 file.
90 Useful for downsampling from one mp3 to another. As an example,
91 it can be useful for streaming through an IceCast server.
92 If the filename ends in ".mp3" LAME will assume it is an MP3.
93 For stdin or MP3 files which do not end in .mp3 you need to use
94 this switch.
95
96 --nogap file1 file2 ...
97 gapless encoding for a set of contiguous files
98
99 --nogapout dir
100 output dir for gapless encoding (must precede --nogap)
101
102 --out-dir dir
103 If no explicit output file is specified, a file will be written
104 at given path. Ignored when using piped/streamed input
105
106
107 Operational options:
108
109 -m mode
110 mode = s, j, f, d, m, l, r
111
112 Joint-stereo is the default mode for stereo files.
113
114 (s)imple stereo (Forced LR)
115 In this mode, the encoder makes no use of potentially existing
116 correlations between the two input channels. It can, however,
117 negotiate the bit demand between both channel, i.e. give one
118 channel more bits if the other contains silence or needs less
119 bits because of a lower complexity.
120
121 (j)oint stereo
122 In this mode, the encoder can use (on a frame by frame basis)
123 either L/R stereo or mid/side stereo. In mid/side stereo, the
124 mid (L+R) and side (L-R) channels are encoded, and more bits are
125 allocated to the mid channel than the side channel. When there
126 isn't too much stereo separation, this effectively increases the
127 bandwidth, so having higher quality with the same amount of
128 bits.
129
130 Using mid/side stereo inappropriately can result in audible com‐
131 pression artifacts. Too much switching between mid/side and
132 regular stereo can also sound bad. To determine when to switch
133 to mid/side stereo, LAME uses a much more sophisticated algo‐
134 rithm than the one described in the ISO documentation.
135
136 (f)orced MS stereo
137 Forces all frames to be encoded with mid/side stereo. It should
138 be used only if you are sure that every frame of the input file
139 has very little stereo separation.
140
141 (d)ual channel
142 In this mode, the 2 channels will be totally independently
143 encoded. Each channel will have exactly half of the bitrate.
144 This mode is designed for applications like dual languages
145 encoding (for example: English in one channel and French in the
146 other). Using this encoding mode for regular stereo files will
147 result in a lower quality encoding.
148
149 (m)ono
150 The input will be encoded as a mono signal. If it was a stereo
151 signal, it will be downsampled to mono. The downmix is calcu‐
152 lated as the sum of the left and right channel, attenuated by 6
153 dB. Also note that, if using a stereo RAW PCM stream, you need
154 to use the -a parameter.
155
156 (l)eft channel only
157 The input will be encoded as a mono signal. If it was a stereo
158 signal, the left channel will be encoded only.
159
160 (r)ight channel only
161 The input will be encoded as a mono signal. If it was a stereo
162 signal, the right channel will be encoded only.
163
164
165 -a Mix the stereo input file to mono and encode as mono.
166 The downmix is calculated as the sum of the left and right chan‐
167 nel, attenuated by 6 dB.
168
169 This option is only needed in the case of raw PCM stereo input
170 (because LAME cannot determine the number of channels in the
171 input file). To encode a stereo RAW PCM input file as mono, use
172 lame -a -m m
173
174 For WAV and AIFF input files, using -m m will always produce a
175 mono .mp3 file from both mono and stereo input.
176
177 --freeformat
178 Produces a free format bitstream. With this option, you can use
179 -b with any bitrate higher than 8 kbps.
180
181 However, even if an mp3 decoder is required to support free
182 bitrates at least up to 320 kbps, many players are unable to
183 deal with it.
184
185 Tests have shown that the following decoders support free for‐
186 mat:
187 in_mpg123 up to 560 kbps
188 l3dec up to 310 kbps
189 LAME up to 640 kbps
190 MAD up to 640 kbps
191
192 --decode
193 Uses LAME for decoding to a wav file. The input file can be any
194 input type supported by encoding, including layer II files.
195 LAME uses a fork of mpglib known as HIP for decoding.
196
197 If -t is used (disable wav header), LAME will output raw pcm in
198 native endian format. You can use -x to swap bytes order.
199
200 This option is not usable if the MP3 decoder was explicitly dis‐
201 abled in the build of LAME.
202
203 -t Disable writing of the INFO Tag on encoding.
204 This tag is embedded in frame 0 of the MP3 file. It includes
205 some information about the encoding options of the file, and in
206 VBR it lets VBR aware players correctly seek and compute playing
207 times of VBR files.
208
209 When --decode is specified (decode to WAV), this flag will dis‐
210 able writing of the WAV header. The output will be raw pcm,
211 native endian format. Use -x to swap bytes.
212
213 --comp arg
214 Instead of choosing bitrate, using this option, user can choose
215 compression ratio to achieve.
216
217 --scale n
218 --scale-l n
219 --scale-r n
220 Scales input (every channel, only left channel or only right
221 channel) by n. This just multiplies the PCM data (after it has
222 been converted to floating point) by n.
223
224 n > 1: increase volume
225 n = 1: no effect
226 n < 1: reduce volume
227
228 Use with care, since most MP3 decoders will truncate data which
229 decodes to values greater than 32768.
230
231 --replaygain-fast
232 Compute ReplayGain fast but slightly inaccurately.
233
234 This computes "Radio" ReplayGain on the input data stream after
235 user‐specified volume‐scaling and/or resampling.
236
237 The ReplayGain analysis does not affect the content of a com‐
238 pressed data stream itself, it is a value stored in the header
239 of a sound file. Information on the purpose of ReplayGain and
240 the algorithms used is available from http://www.replay‐
241 gain.org/.
242
243 Only the "RadioGain" Replaygain value is computed, it is stored
244 in the LAME tag. The analysis is performed with the reference
245 volume equal to 89dB. Note: the reference volume has been
246 changed from 83dB on transition from version 3.95 to 3.95.1.
247
248 This switch is enabled by default.
249
250 See also: --replaygain-accurate, --noreplaygain
251
252 --replaygain-accurate
253 Compute ReplayGain more accurately and find the peak sample.
254
255 This computes "Radio" ReplayGain on the decoded data stream,
256 finds the peak sample by decoding on the fly the encoded data
257 stream and stores it in the file.
258
259 The ReplayGain analysis does not affect the content of a com‐
260 pressed data stream itself, it is a value stored in the header
261 of a sound file. Information on the purpose of ReplayGain and
262 the algorithms used is available from http://www.replay‐
263 gain.org/.
264
265
266 By default, LAME performs ReplayGain analysis on the input data
267 (after the user‐specified volume scaling). This behavior might
268 give slightly inaccurate results because the data on the output
269 of a lossy compression/decompression sequence differs from the
270 initial input data. When --replaygain-accurate is specified the
271 mp3 stream gets decoded on the fly and the analysis is performed
272 on the decoded data stream. Although theoretically this method
273 gives more accurate results, it has several disadvantages:
274
275 * tests have shown that the difference between the ReplayGain
276 values computed on the input data and decoded data is usu‐
277 ally not greater than 0.5dB, although the minimum volume
278 difference the human ear can perceive is about 1.0dB
279
280 * decoding on the fly significantly slows down the encoding
281 process
282
283 The apparent advantage is that:
284
285 * with --replaygain-accurate the real peak sample is deter‐
286 mined and stored in the file. The knowledge of the peak
287 sample can be useful to decoders (players) to prevent a
288 negative effect called 'clipping' that introduces distor‐
289 tion into the sound.
290
291 Only the "RadioGain" ReplayGain value is computed, it is stored
292 in the LAME tag. The analysis is performed with the reference
293 volume equal to 89dB. Note: the reference volume has been
294 changed from 83dB on transition from version 3.95 to 3.95.1.
295
296 This option is not usable if the MP3 decoder was explicitly dis‐
297 abled in the build of LAME. (Note: if LAME is compiled without
298 the MP3 decoder, ReplayGain analysis is performed on the input
299 data after user-specified volume scaling).
300
301 See also: --replaygain-fast, --noreplaygain --clipdetect
302
303 --noreplaygain
304 Disable ReplayGain analysis.
305
306 By default ReplayGain analysis is enabled. This switch disables
307 it.
308
309 See also: --replaygain-fast, --replaygain-accurate
310
311 --clipdetect
312 Clipping detection.
313
314 Enable --replaygain-accurate and print a message whether clip‐
315 ping occurs and how far in dB the waveform is from full scale.
316
317 This option is not usable if the MP3 decoder was explicitly dis‐
318 abled in the build of LAME.
319
320 See also: --replaygain-accurate
321
322 --preset type | [cbr] kbps
323 Use one of the built-in presets.
324
325 Have a look at the PRESETS section below.
326
327 --preset help gives more infos about the the used options in
328 these presets.
329
330 --noasm type
331 Disable specific assembly optimizations ( mmx / 3dnow / sse ).
332 Quality will not increase, only speed will be reduced. If you
333 have problems running Lame on a Cyrix/Via processor, disabling
334 mmx optimizations might solve your problem.
335
336
337 Verbosity:
338
339 --disptime n
340 Set the delay in seconds between two display updates.
341
342 --nohist
343 By default, LAME will display a bitrate histogram while produc‐
344 ing VBR mp3 files. This will disable that feature.
345 Histogram display might not be available on your release.
346
347 -S
348 --silent
349 --quiet
350 Do not print anything on the screen.
351
352 --verbose
353 Print a lot of information on the screen.
354
355 --help Display a list of available options.
356
357
358 Noise shaping & psycho acoustic algorithms:
359
360 -q qual
361 0 <= qual <= 9
362
363 Bitrate is of course the main influence on quality. The higher
364 the bitrate, the higher the quality. But for a given bitrate,
365 we have a choice of algorithms to determine the best scalefac‐
366 tors and Huffman encoding (noise shaping).
367
368 For CBR and ABR, the following table applies:
369
370 -q 0:
371 Use the best algorithms (Best Huffman coding search, full outer
372 loop, and the highest precision of several parameters).
373
374 -q 1 to q 4:
375 Similar to -q 0 without the full outer loop and decreasing pre‐
376 cision of parameters the further from q0. -q 3 is the default.
377
378 -q 5 and -q 6:
379 Same as -q 7, but enables noise shaping and increases subblock
380 gain
381
382 -q 7 to -q 9:
383 Same as -f. Very fast, OK quality. Psychoacoustics are used for
384 pre-echo and mid/side stereo, but no noise-shaping is done.
385
386 For the default VBR mode since LAME 3.98, the following table
387 applies :
388
389 -q 0 to -q 4:
390 include all features of the other modes and additionally use the
391 best search when applying Huffman coding.
392
393 -q 5 and -q 6:
394 include all features of -q7, calculate and consider actual quan‐
395 tisation noise, and additionally enable subblock gain.
396
397 -q 7 to -q 9
398 This level uses a psymodel but does not calculate quantisation
399 noise when encoding: it takes a quick guess.
400
401
402
403 -h Alias of -q 2
404
405 -f Alias of -q 7
406
407
408
409 CBR (constant bitrate, the default) options:
410
411 -b n For MPEG-1 (sampling frequencies of 32, 44.1 and 48 kHz)
412 n = 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256,
413 320
414
415 For MPEG-2 (sampling frequencies of 16, 22.05 and 24 kHz)
416 n = 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160
417
418 For MPEG-2.5 (sampling frequencies of 8, 11.025 and 12 kHz)
419 n = 8, 16, 24, 32, 40, 48, 56, 64
420
421 Default is 128 for MPEG1 and 64 for MPEG2 and 32 for MPEG2.5
422 (64, 32 and 16 respectively in case of mono).
423
424 --cbr enforce use of constant bitrate. Used to disable VBR or ABR
425 encoding even if their settings are enabled.
426
427
428 ABR (average bitrate) options:
429
430 --abr n
431 Turns on encoding with a targeted average bitrate of n kbits,
432 allowing to use frames of different sizes. The allowed range of
433 n is 8 - 310, you can use any integer value within that range.
434
435 It can be combined with the -b and -B switches like: lame --abr
436 123 -b 64 -B 192 a.wav a.mp3 which would limit the allowed frame
437 sizes between 64 and 192 kbits.
438
439 The use of -B is NOT RECOMMENDED. A 128 kbps CBR bitstream,
440 because of the bit reservoir, can actually have frames which use
441 as many bits as a 320 kbps frame. VBR modes minimize the use of
442 the bit reservoir, and thus need to allow 320 kbps frames to get
443 the same flexibility as CBR streams.
444
445
446 VBR (variable bitrate) options:
447
448 -v use variable bitrate (--vbr-new)
449
450 --vbr-old
451 Invokes the oldest, most tested VBR algorithm. It produces very
452 good quality files, though is not very fast. This has, up
453 through v3.89, been considered the "workhorse" VBR algorithm.
454
455 --vbr-new
456 Invokes the newest VBR algorithm. During the development of
457 version 3.90, considerable tuning was done on this algorithm,
458 and it is now considered to be on par with the original --vbr-
459 old. It has the added advantage of being very fast (over twice
460 as fast as --vbr-old ). This is the default since 3.98.
461
462 -V n 0 <= n <= 9.999
463 Enable VBR (Variable BitRate) and specifies the value of VBR
464 quality (default = 4). Decimal values can be specified, like
465 4.51.
466 0 = highest quality.
467
468
469 ABR and VBR options:
470
471 -b bitrate
472 For MPEG-1 (sampling frequencies of 32, 44.1 and 48 kHz)
473 n = 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256,
474 320
475
476 For MPEG-2 (sampling frequencies of 16, 22.05 and 24 kHz)
477 n = 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160
478
479 For MPEG-2.5 (sampling frequencies of 8, 11.025 and 12 kHz)
480 n = 8, 16, 24, 32, 40, 48, 56, 64
481
482 Specifies the minimum bitrate to be used. However, in order to
483 avoid wasted space, the smallest frame size available will be
484 used during silences.
485
486 -B bitrate
487 For MPEG-1 (sampling frequencies of 32, 44.1 and 48 kHz)
488 n = 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256,
489 320
490
491 For MPEG-2 (sampling frequencies of 16, 22.05 and 24 kHz)
492 n = 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160
493
494 For MPEG-2.5 (sampling frequencies of 8, 11.025 and 12 kHz)
495 n = 8, 16, 24, 32, 40, 48, 56, 64
496
497 Specifies the maximum allowed bitrate.
498
499 Note: If you own an mp3 hardware player build upon a MAS 3503
500 chip, you must set maximum bitrate to no more than 224 kpbs.
501
502 -F Strictly enforce the -b option.
503 This is mainly for use with hardware players that do not support
504 low bitrate mp3.
505
506 Without this option, the minimum bitrate will be ignored for
507 passages of analog silence, i.e. when the music level is below
508 the absolute threshold of human hearing (ATH).
509
510
511 Experimental options:
512
513 -X n 0 <= n <= 7
514
515 When LAME searches for a "good" quantization, it has to compare
516 the actual one with the best one found so far. The comparison
517 says which one is better, the best so far or the actual. The -X
518 parameter selects between different approaches to make this
519 decision, -X0 being the default mode:
520
521 -X0
522 The criteria are (in order of importance):
523 * less distorted scalefactor bands
524 * the sum of noise over the thresholds is lower
525 * the total noise is lower
526
527 -X1
528 The actual is better if the maximum noise over all scalefactor
529 bands is less than the best so far.
530
531 -X2
532 The actual is better if the total sum of noise is lower than the
533 best so far.
534
535 -X3
536 The actual is better if the total sum of noise is lower than the
537 best so far and the maximum noise over all scalefactor bands is
538 less than the best so far plus 2dB.
539
540 -X4
541 Not yet documented.
542
543 -X5
544 The criteria are (in order of importance):
545 * the sum of noise over the thresholds is lower
546 * the total sum of noise is lower
547
548 -X6
549 The criteria are (in order of importance):
550 * the sum of noise over the thresholds is lower
551 * the maximum noise over all scalefactor bands is lower
552 * the total sum of noise is lower
553
554 -X7
555 The criteria are:
556 * less distorted scalefactor bands
557 or
558 * the sum of noise over the thresholds is lower
559
560 -Y lets LAME ignore noise in sfb21, like in CBR
561
562
563 MP3 header/stream options:
564
565 -e emp emp = n, 5, c
566
567 n = (none, default)
568 5 = 0/15 microseconds
569 c = citt j.17
570
571 All this does is set a flag in the bitstream. If you have a PCM
572 input file where one of the above types of (obsolete) emphasis
573 has been applied, you can set this flag in LAME. Then the mp3
574 decoder should de-emphasize the output during playback, although
575 most decoders ignore this flag.
576
577 A better solution would be to apply the de-emphasis with a
578 standalone utility before encoding, and then encode without -e.
579
580 -c Mark the encoded file as being copyrighted.
581
582 -o Mark the encoded file as being a copy.
583
584 -p Turn on CRC error protection.
585 It will add a cyclic redundancy check (CRC) code in each frame,
586 allowing to detect transmission errors that could occur on the
587 MP3 stream. However, it takes 16 bits per frame that would oth‐
588 erwise be used for encoding, and then will slightly reduce the
589 sound quality.
590
591 --nores
592 Disable the bit reservoir. Each frame will then become indepen‐
593 dent from previous ones, but the quality will be lower.
594
595 --strictly-enforce-ISO
596 With this option, LAME will enforce the 7680 bit limitation on
597 total frame size.
598 This results in many wasted bits for high bitrate encodings but
599 will ensure strict ISO compatibility. This compatibility might
600 be important for hardware players.
601
602
603 Filter options:
604
605 --lowpass freq
606 Set a lowpass filtering frequency in kHz. Frequencies above the
607 specified one will be cutoff.
608
609 --lowpass-width freq
610 Set the width of the lowpass filter. The default value is 15%
611 of the lowpass frequency.
612
613 --highpass freq
614 Set an highpass filtering frequency in kHz. Frequencies below
615 the specified one will be cutoff.
616
617 --highpass-width freq
618 Set the width of the highpass filter in kHz. The default value
619 is 15% of the highpass frequency.
620
621 --resample sfreq
622 sfreq = 8, 11.025, 12, 16, 22.05, 24, 32, 44.1, 48
623 Select output sampling frequency (only supported for encoding).
624 If not specified, LAME will automatically resample the input
625 when using high compression ratios.
626
627
628 ID3 tag options:
629
630 --tt title
631 audio/song title (max 30 chars for version 1 tag)
632
633 --ta artist
634 audio/song artist (max 30 chars for version 1 tag)
635
636 --tl album
637 audio/song album (max 30 chars for version 1 tag)
638
639 --ty year
640 audio/song year of issue (1 to 9999)
641
642 --tc comment
643 user-defined text (max 30 chars for v1 tag, 28 for v1.1)
644
645 --tn track[/total]
646 audio/song track number and (optionally) the total number of
647 tracks on the original recording. (track and total each 1 to
648 255. Providing just the track number creates v1.1 tag, providing
649 a total forces v2.0).
650
651 --tg genre
652 audio/song genre (name or number in list)
653
654 --tv id=value
655 Text or URL frame specified by id and value (v2.3 tag). User
656 defined frame. Syntax: --tv "TXXX=description=content"
657
658 --add-id3v2
659 force addition of version 2 tag
660
661 --id3v1-only
662 add only a version 1 tag
663
664 --id3v2-only
665 add only a version 2 tag
666
667 --id3v2-latin1
668 add following options in ISO-8859-1 text encoding.
669
670 --id3v2-utf16
671 add following options in unicode text encoding.
672
673 --space-id3v1
674 pad version 1 tag with spaces instead of nulls
675
676 --pad-id3v2
677 same as --pad-id3v2-size 128
678
679 --pad-id3v2-size num
680 adds version 2 tag, pad with extra "num" bytes
681
682 --genre-list
683 print alphabetically sorted ID3 genre list and exit
684
685 --ignore-tag-errors
686 ignore errors in values passed for tags, use defaults in case an
687 error occurs
688
689
690 Analysis options:
691
692 -g run graphical analysis on <infile>. <infile> can also be a .mp3
693 file. (This feature is a compile time option. Your binary may
694 for speed reasons be compiled without this.)
695
696
698 LAME is able to embed ID3 v1, v1.1 or v2 tags inside the encoded MP3
699 file. This allows to have some useful information about the music
700 track included inside the file. Those data can be read by most MP3
701 players.
702
703 Lame will smartly choose which tags to use. It will add ID3 v2 tags
704 only if the input comments won't fit in v1 or v1.1 tags, i.e. if they
705 are more than 30 characters. In this case, both v1 and v2 tags will be
706 added, to ensure reading of tags by MP3 players which are unable to
707 read ID3 v2 tags.
708
709
711 LAME is able to encode your music using one of its 3 encoding modes:
712 constant bitrate (CBR), average bitrate (ABR) and variable bitrate
713 (VBR).
714
715 Constant Bitrate (CBR)
716 This is the default encoding mode, and also the most basic. In
717 this mode, the bitrate will be the same for the whole file. It
718 means that each part of your mp3 file will be using the same
719 number of bits. The musical passage being a difficult one to
720 encode or an easy one, the encoder will use the same bitrate, so
721 the quality of your mp3 is variable. Complex parts will be of a
722 lower quality than the easiest ones. The main advantage is that
723 the final files size won't change and can be accurately pre‐
724 dicted.
725
726 Average Bitrate (ABR)
727 In this mode, you choose the encoder will maintain an average
728 bitrate while using higher bitrates for the parts of your music
729 that need more bits. The result will be of higher quality than
730 CBR encoding but the average file size will remain predictable,
731 so this mode is highly recommended over CBR. This encoding mode
732 is similar to what is referred as vbr in AAC or Liquid Audio (2
733 other compression technologies).
734
735 Variable bitrate (VBR)
736 In this mode, you choose the desired quality on a scale from 9
737 (lowest quality/biggest distortion) to 0 (highest quality/lowest
738 distortion). Then encoder tries to maintain the given quality
739 in the whole file by choosing the optimal number of bits to
740 spend for each part of your music. The main advantage is that
741 you are able to specify the quality level that you want to
742 reach, but the inconvenient is that the final file size is
743 totally unpredictable.
744
745
747 The --preset switches are aliases over LAME settings.
748
749 To activate these presets:
750
751 For VBR modes (generally highest quality):
752
753 --preset medium
754 This preset should provide near transparency to most people on
755 most music.
756
757 --preset standard
758 This preset should generally be transparent to most people on
759 most music and is already quite high in quality.
760
761 --preset extreme
762 If you have extremely good hearing and similar equipment, this
763 preset will generally provide slightly higher quality than the
764 standard mode.
765
766 For CBR 320kbps (highest quality possible from the --preset switches):
767
768 --preset insane
769 This preset will usually be overkill for most people and most
770 situations, but if you must have the absolute highest quality
771 with no regard to filesize, this is the way to go.
772
773 For ABR modes (high quality per given bitrate but not as high as VBR):
774
775 --preset kbps
776 Using this preset will usually give you good quality at a speci‐
777 fied bitrate. Depending on the bitrate entered, this preset
778 will determine the optimal settings for that particular situa‐
779 tion. While this approach works, it is not nearly as flexible
780 as VBR, and usually will not attain the same level of quality as
781 VBR at higher bitrates.
782
783 cbr If you use the ABR mode (read above) with a significant bitrate
784 such as 80, 96, 112, 128, 160, 192, 224, 256, 320, you can use
785 the --preset cbr kbps option to force CBR mode encoding instead
786 of the standard ABR mode. ABR does provide higher quality but
787 CBR may be useful in situations such as when streaming an MP3
788 over the Internet may be important.
789
790
791
793 Fixed bit rate jstereo 128kbs encoding:
794
795 lame -b 128 sample.wav sample.mp3
796
797
798 Fixed bit rate jstereo 128 kbps encoding, highest quality:
799
800 lame -q 0 -b 128 sample.wav sample.mp3
801
802
803 To disable joint stereo encoding (slightly faster, but less quality at
804 bitrates <= 128 kbps):
805
806 lame -m s sample.wav sample.mp3
807
808
809 Variable bitrate (use -V n to adjust quality/filesize):
810
811 lame -V 2 sample.wav sample.mp3
812
813
814 Streaming mono 22.05 kHz raw pcm, 24 kbps output:
815
816 cat inputfile | lame -r -m m -b 24 -s 22.05 - - > output
817
818
819 Streaming mono 44.1 kHz raw pcm, with downsampling to 22.05 kHz:
820
821 cat inputfile | lame -r -m m -b 24 --resample 22.05 - - > output
822
823
824 Encode with the standard preset:
825
826 lame --preset standard sample.wav sample.mp3
827
828
830 Probably there are some.
831
833 mpg123(1), madplay(1), sox(1)
834
836 LAME originally developed by Mike Cheng and now maintained by
837 Mark Taylor, and the LAME team.
838
839 GPSYCHO psycho-acoustic model by Mark Taylor.
840 (See http://www.mp3dev.org/).
841
842 mpglib by Michael Hipp
843
844 Manual page by William Schelter, Nils Faerber, Alexander Leidinger,
845 and Rogério Brito.
846
847
848
849LAME 3.99 December 08, 2013 lame(1)