history.html

LAME无疑是目前最优秀的MP3编码软件

  <li> Mark Taylor: Scalefactor preemphasis disabled for now.&nbsp;&nbsp; The 
    algorithm was often doing more harm than good.</li>
</ul>
<h3> LAME 3.56beta&nbsp; November 19 1999</h3>
<ul>
  <li> Kimmo Mustonen: portabilty code cleanup.</li>
  <li> Vladimir Marek: id3 genre patch.</li>
  <li> Conrad Sanderson: new applypatch script.</li>
  <li> Mark Taylor: Initial window type now "STOP_TYPE" to reduce initial attenuation.&nbsp; 
    This is needed because the new encoder delay is so short.&nbsp; With a NORM_TYPE, 
    the first 240 samples would be attenuated.</li>
  <li> Mark Taylor: Padding at end of file now adjusted (hopefully!) to produce 
    as little padding as possible while still guarantee all input samples are 
    encoded.</li>
  <li> <font color="#ff0000">Takehiro Tominaga:&nbsp; Reduced shortblock extra 
    bit allocation formulas by 10% since new huffman coding is at least 10% more 
    efficient.</font></li>
</ul>
<h3> LAME 3.55beta&nbsp; November 11 1999</h3>
<ul>
  <li> Albert Faber:&nbsp; updated BladeEnc.dll</li>
  <li> Mark Taylor: Simple lowpass filter added to linear downsampling routine.</li>
  <li> Nils Faerber: updated man page.</li>
  <li> Mark Taylor: All floating point variables are delcared FLOAT or&nbsp; FLOAT8.&nbsp; 
    Change the definition of FLOAT8 in machine.h to run at 32bit preceision.</li>
  <li> Mark Taylor: Bug (introduced in 3.54beta) in stereo-&gt;mono downsampling 
    fixed.</li>
</ul>
<h3> LAME 3.54beta&nbsp; November 8 1999</h3>
<ul>
  <li> Mark Taylor: Encoder delay is now 48 samples.&nbsp; Can be adjusted to 
    1160 to sync with FhG (see ENCDELAY in encoder.h)&nbsp; This is kind of amazing, 
    since if Takehiro put his MDCT/filterbank routine in a decoder, we could have 
    a total&nbsp; delay of only 96 samples.</li>
  <li> <font color="#ff0000">Mark Taylor: More inconstancies found and fixed in 
    MPEG2 tables.</font></li>
  <li> Mark Taylor: Resampling from an MP3 input file now works.&nbsp; But we 
    still dont have a lowpass filter so dont expect good results.</li>
</ul>
<h3> LAME 3.53beta&nbsp; November 8 1999</h3>
<ul>
  <li> <font color="#3366ff">Takehiro Tominaga:&nbsp; Fixed MPEG2 problem in new 
    MDCT routines.&nbsp; Takehiro's combined filterbank/MDCT routine is now the 
    default.&nbsp; Removes all buffering from psymodel.c and the filterbanks/MDCT 
    routines.</font></li>
</ul>
<h3> LAME 3.52beta&nbsp; November 8 1999</h3>
<ul>
  <li> By permission of copyright holders of all GPL code in LAME,&nbsp; all GPL 
    code is now released under a modified version of the LGPL (see the README 
    file)</li>
  <li> By popular demand, all C++ comments changed to C style comments</li>
  <li> Mark Taylor: Linear resampling now works.&nbsp; Use --resample to set an 
    output samplerate different from the input samplerate.&nbsp; (doesn't seem 
    to work with mp3 input files, and there is no lowpass filter, so dont expect 
    good results just yet)</li>
  <li> <font color="#3366ff">Takehiro Tominaga:&nbsp; Faster Huffman encoding 
    routines</font></li>
</ul>
<font color="#3366ff">The following changes are disabled because of MPEG2 problems.&nbsp; 
But to try them, set MDCTDELAY=48 in encoder.h, instead of MDCTDELAY=528.:</font> 
<ul>
  <li> <font color="#3366ff">Takehiro Tominaga:&nbsp; New MDCT routines with shorter 
    delay (48 samples instead of 528) and even faster than the old routines.</font></li>
  <li> <font color="#3366ff">Takehiro Tominaga:&nbsp; Removed extra buffering 
    in psymodel.c</font></li>
</ul>
<h3> LAME 3.51&nbsp; November 7 1999</h3>
<ul>
  <li> Takehiro Tominaga: Bug in quantize.c absolute threshold of hearing calculation 
    for non-44.1 kHz input files.</li>
</ul>
<h3> LAME 3.50&nbsp; November 1 1999</h3>
<ul>
  <li> LAME 3.37beta becomes official LAME 3.50 release</li>
</ul>
<h3> LAME 3.37beta&nbsp; November 1 1999</h3>
<ul>
  <li> <font color="#ff0000">Lionel Bonnet:&nbsp; Found severe bug in MPEG2 Short 
    block SNR.</font></li>
  <li> Sergey Sapelin:&nbsp; VBR Toc improvement.</li>
  <li> Sergey Dubov: fskip() routine</li>
  <li> Conrad Sanderson: replacement for filterbank.c.&nbsp;&nbsp; Not much faster 
    but amazingly simpler.</li>
</ul>
<h3> LAME 3.36beta&nbsp; October 25 1999</h3>
<ul>
  <li> Albert Faber:&nbsp; more MSVC and BladeDLL updates</li>
  <li> Kimmo Mustonen:&nbsp; Much code cleanup and Amiga updates</li>
  <li> Anton Oleynikov: Borland C updates</li>
  <li> Mark Taylor: More stdin fixes:&nbsp; For some reason, forward fseek()'s 
    would fail when used on pipes even though it is okay with redirection from 
    "&lt;". So I changed all the forward fseek()'s to use fread().&nbsp; This 
    should improve stdin support for wav/aiff files.&nbsp; If you know the input 
    file is raw pcm, you can still use the '-r' option to avoid *all* seeking 
    of any kind.</li>
  <br>
  &nbsp; 
</ul>
<h3> LAME 3.35beta&nbsp; October 21 1999</h3>
<ul>
  <li> <font color="#ff0000">Leonid Kulakov:&nbsp; Serious bug in MPEG2 scalefactor 
    band tables fixed.</font></li>
  <li> Portability patches from:&nbsp; Anton Oleynikov, Sigbj&oslash;rn Skj&aelig;ret, 
    Mathew Hendry, Richard Gorton</li>
  <li> Alfred Weyers: compiler options, updated timestatus.</li>
  <li> Albert Faber:&nbsp; BladeDll and other updates (new machine.h).</li>
  <li> Monty:&nbsp; updated Makefile to fix gcc inline math bug.</li>
  <br>
  &nbsp; 
</ul>
<h3> LAME 3.34beta&nbsp; October 12 1999</h3>
<ul>
  <li> <font color="#ff0000">Mark Taylor: Bug fixed:&nbsp; minimum bitrate in 
    VBR mode could be ignored for a few frames.</font></li>
  <li> <font color="#ff0000">Mark Taylor: New (minor) VBR tunings.</font></li>
  <li> Tim Ruddick: New wav/aiff header parsing routines.&nbsp; Better parsing 
    and fewer fseek()'s.</li>
  <li> Anton Oleynikov:&nbsp; patches to work with Borland C</li>
  <li> <font color="#ff0000">Gabriel Bouvigne:&nbsp; Experimental voice option 
    enabled with --voice</font></li>
</ul>
<h3> LAME 3.33beta&nbsp; October 11 1999</h3>
<ul>
  <li> <font color="#ff0000">Robert Hegemann: RH VBR mode now the default and 
    only VBR mode.&nbsp; The new code will always quantize to 0 distortion and 
    the quality is increased by reducing the masking from the psy-model.&nbsp; 
    -X0 is still the default for now.</font></li>
  <li> <font color="#ff0000">Robert Hegemann: new -X5 mode</font></li>
  <li> Mathew Hendry: New timing code, removes the need for HAVETIMES</li>
  <li> <font color="#3366ff">Mathew Hendry:&nbsp; assembler quantize_xrpow for 
    Windows</font></li>
  <li> Iwasa Kazmi:&nbsp; stdin/stdout patch for Windows</li>
  <li> Mark Taylor: New option: "--athonly" will ignore the psy-model output and 
    use only the absolute threshold of hearing for the masking.</li>
  <br>
  &nbsp; 
</ul>
<h3> LAME 3.32beta&nbsp; October 8 1999</h3>
<ul>
  <li> <font color="#3366ff">Takehiro Tominaga:&nbsp; faster long block spreading 
    function convolution for non 44.1 kHz sampling frequencies, and faster short 
    block spreading function convolution for all sampling frequencies.</font></li>
  <li> <font color="#ff0000">Takehiro Tominaga:&nbsp; Completly rewritten huffman 
    table selection and count_bits().&nbsp; More efficient table selection results 
    in many more bits per frame.</font></li>
  <li> <font color="#ff0000">Takehiro Tominaga:&nbsp; More efficient scalefac 
    compress setting.</font></li>
  <li> <font color="#3366ff">Mike Cheng: new calc_noise2()</font></li>
  <li> Alfred Weyers: patch for timestatus() seconds rollover</li>
  <br>
  &nbsp; 
</ul>
<h3> LAME 3.31beta&nbsp; September 28 1999</h3>
<ul>
  <li> Albert Faber:&nbsp; updated his BladeDLL code.&nbsp; This allows LAME to 
    be compiled into a BladeEnc compatiable .dll.</li>
  <li> <font color="#3366ff">Mike Cheng: faster l3psycho_ener() routine.</font></li>
  <li> Sigbj&oslash;rn Skj&aelig;ret: more code cleanup.</li>
</ul>
<h3> LAME&nbsp; 3.30beta&nbsp; September 27 1999</h3>
<ul>
  <li> Conrad Sanderson:&nbsp; ID3 tag code added (type 'lame' for instructions)</li>
  <li> new mdct.c from Mike Cheng (no faster, but much cleaner code)</li>
  <li> Mathew Hendry: Microsoft nmake makefile and a couple other changes for 
    MSVC</li>
  <li> More modulization work:&nbsp; One input sound file interface handles mp3's, 
    uncompressed audio, with or without LIBSNDFILE.&nbsp; Fixes (hopefully) a 
    bunch of file I/O bugs introduced in 3.29 (Mark Taylor)</li>
  <li> LAME will now print valid samplerate/bitrate combinations (Mark Taylor)</li>
  <li> stdin/stdout fix for OS/2 (Paul Hartman)</li>
  <li> For mp3 input files, totalframes estimated based on filesize and first 
    frame bitrate. (Mark Taylor)</li>
  <li> Updated all functions with new style prototypes.&nbsp; (Sigbj&oslash;rn 
    Skj&aelig;ret)</li>
  <br>
  &nbsp; 
</ul>
<h3> LAME 3.29beta&nbsp; September 21 1999</h3>
<ul>
  <li> <font color="#ff0000">Bug in bigv_bitcount fixed.&nbsp; Loop.c was overestimating 
    the number of bits needed, resulting in wasted bits every frame.&nbsp; (Leonid 
    A. Kulakov)</font></li>
  <li> <font color="#ff0000">Bug in *_choose_table() fixed&nbsp;&nbsp; These routines 
    would not sellect the optimal Huffman table in some cases.&nbsp;&nbsp; (Leonid 
    A. Kulakov)</font></li>
  <li> <font color="#ff0000">Tuning of ATH normalization (macik)</font></li>
  <li> Removed unused variables and fixed function prototypes (Sigbj&oslash;rn 
    Skj&aelig;ret)</li>
  <li> Sami Farin sent a&nbsp; .wav&nbsp; file that&nbsp; LAME built in support 
    choked on.&nbsp;&nbsp; I added a slightly more sophisticated wav header parsing 
    to handle this, but if you have trouble, use libsndfile.</li>
  <li> Resampling hooks and options added.&nbsp; Buffering and resampling routines 
    need to be written.</li>
  <li> LAME will now take an mp3 file as input.&nbsp; When resampling code is 
    working, LAME will be able to (for example) convert a high bitrate stereo 
    mp3 to a low bitrate mono mp3 for streaming.</li>
</ul>
<h3> LAME 3.28beta&nbsp; September 15 1999</h3>
<ul>
  <li> <font color="#ff0000">Serious bug fixed in high frequency MDCT coefficients.&nbsp; 
    Huffman coding was reversing the order of the count1 block quadruples.&nbsp;&nbsp; 
    (Leonid A. Kulakov)</font></li>
  <li> nint() problems under Tru64 unix fixed and preprocessor variable HAVE_NINT 
    removed.&nbsp; (Bob Bell)</li>
  <li> Compiler warning fixes and code cleanup&nbsp;&nbsp; (Sigbj&oslash;rn Skj&aelig;ret, 
    Lionel Bonnet)</li>
  <li> USAGE file now includes suggestions for downsampling.&nbsp; For low bitrate 
    encodings, proper downsampling can give dramatically better results.&nbsp; 
    (John Hayward-Warburton)</li>
</ul>
<h3> LAME 3.27beta&nbsp; September 12 1999</h3>
<ul>
  <li> Several bugs in encode.c and l3bitstream.c fixed by Lionel Bonnet.</li>
  <li> Bugs in new VBR (#define RH) formula for mono input file and mid/side encoding 
    fixed.</li>
</ul>
<h3> LAME 3.26beta&nbsp; September 10 1999</h3>
<ul>
  <li> The "-m m" option (mono .mp3 file) will automatically mix left and right 
    channels if the input file is stereo.&nbsp; (Alfred Weyers)</li>
  <li> <font color="#ff0000">New quant_compare algorithm (method for deciding 
    which of two quantizations is better) enabled with -X4 (Greg Maxwell)</font></li>
  <li> <font color="#ff0000">New mid/side VBR bit allocation formula.&nbsp; Mid 
    channel bits are set by the quality requirements, and then the side channel 
    uses a reduced number of bits (in a proportion coming from the fixed bitrate 
    code).&nbsp; This might not be optimal, but it should be pretty good and no 
    one knows what the optimal solution should be.&nbsp; (Greg Maxwell)</font></li>
  <li> <font color="#ff0000">New VBR (#define RH) tunings based on detailed listening 
    tests by Macik and Greg Maxwell.</font></li>
  <li> Sigbj&oslash;rn Skj&aelig;ret fixed several compiler warnings (which turned 
    out to be potential bugs)</li>
  <li> Takehiro Tominaga fixed a low bitrate bug in reduce_side()</li>
  <li> Alfred Weyers fixed some buffer overflows.</li>
  <li> <font color="#ff0000">New ATH (absolute threshold of hearing) formula replaces 
    buggy ISO code, and adds&nbsp; analog silence treatment&nbsp; (removal of 
    coefficients below below ATH).&nbsp;&nbsp; These are turned on by default 
    but have not been fully tested.&nbsp; (Robert Hegemann)</font></li>
  <li> <font color="#ff0000">Bug in short block spreading function fixed.&nbsp; 
    (Robert Hegemann)</font></li>
</ul>
<h3> LAME 3.25beta&nbsp; August 22 1999</h3>
<ul>
  <li> Sigbj&oslash;rn Skj&aelig;ret fixed a zero byte malloc call.&nbsp; This 
    bug&nbsp; was introduced in 3.24 and causes problems on non Linux systems.</li>
  <li> Bit allocation routines would sometimes allocate more than 4095 bits to