history.html

LAME uses the MPGLIB decoding engine, from the MPG123 package, written by: Michael Hipp (www.mpg123

Albert Faber did the work for this, including getting libsndfile running
under win32.</li>
  
  <li> CRC checksum now working!&nbsp;&nbsp; (Thanks to Johannes Overmann
)</li>
  
  <li> frame analyzer will now work with mono .mp3 files</li>
  
  <li> <font color="#3366ff">more code tweeks from Jan Peman.</font></li>
  
  <li> <font color="#3366ff">Compaq-Alpha(Linux) fixes and speedups from
Nils Faerber.</font></li>
  
  <li> <font color="#3366ff">Faster bin_search_StepSize from Juha Laukala.</font></li>
  
  <li> <font color="#3366ff">Faster quantize() from Mike Cheng</font></li>
  
  <li> <font color="#3366ff">Faster quantize_xrpow() from Chris Matrakidis.&nbsp; 
xrpow_flag removed since this option is now on by default.</font></li>
  
  <li> Fixed .wav header parsing from Nils Faerber.</li>
  
  <li> Xing VBR frame info header code from Albert Faber.&nbsp;&nbsp; "Xing"
and "LAME 3.12" embedded in first frame.</li>
  
  <li> <font color="#ff0000">Bug in VBR bit allocation based on "over" value
fixed.</font></li>
 
</ul>
  
<h3> LAME 3.11&nbsp; June 3 1999</h3>
  
<blockquote> 
  <li> Almost all warnings (-Wall) now fixed!&nbsp; (Thanks to Jan Peman)</li>
  
  <li> More coding improvements from Gabriel Bouvigne and Warren Toomey.</li>
  
  <li> <font color="#ff0000">VBR&nbsp; (variable bit rate).&nbsp;&nbsp; Increases 
bit rate for short blocks and for frames where the number of bands containing 
audible distortion is greater than a given value.&nbsp; Much tuning needs 
to be done.</font></li>
  
  <li> Patch to remove all atan() calls from James Droppo.</li>
 </blockquote>
  
  <h3> LAME 3.10 May 30 1999</h3>
  
  <ul>
 
    <li> <font color="#3366ff">Fast mode (-f)&nbsp; disables psycho-acoustic
model for real time encoding on older machines.&nbsp; Thanks to Lauri Ahonen 
who first sent a patch for this.</font></li>
  
    <li> <font color="#ff0000">New bit reservoir usage scheme to accommodate
the new pre-echo detection formulas.</font></li>
  
    <li> <font color="#ff0000">Tuning of AWS and ENER_AWS pre-echo formulas
by Gabriel Bouvigne and myself.&nbsp; They work great!&nbsp; now on by default.</font></li>
  
    <li> In jstereo, force blocktypes for left &amp; right channels to be
identical.&nbsp; FhG seems to do this.&nbsp; It can be disabled with "-d".</li>
  
    <li> Patches to compile MP3x under win32 (Thanks to Albert Faber).</li>
  
    <li> <font color="#3366ff">bin_serach_stepsize limited to a quantizationStepSize 
of&nbsp; -210 through 45.</font></li>
  
    <li> <font color="#ff0000">outer_loop()&nbsp; will now vary Mid &amp;
Side scalefactors independently.&nbsp; Can lead to better quantizations,
but it is slower (twice as many quantizations to look at).&nbsp; Running
with "-m f" does not need this and will run at the old speed</font></li>
  
    <li> <font color="#ff0000">Bug in inner_loop would allow quantizations
larger than allowed.&nbsp; (introduced in lame3.04, now fixed.)</font></li>
  
    <li> Updated HTML documentation from Gabriel Bouvigne.</li>
  
    <li> Unix&nbsp; man page from William Schelter.</li>
  
    <li> <font color="#ff0000">numlines[] bug fixed.&nbsp; (Thanks to Rafael
Luebbert, MPecker author).</font></li>
  
    <li> <font color="#3366ff">Quantization speed improvements from Chirs
Matrakidis.</font></li>
  
    <li> <font color="#ff0000">When comparing quantizations with the same
number of bands with audible distortion, use the one with the largest scalefactors, 
not the first one outer_loop happened to find.</font></li>
  
    <li> Improved defination of best quantization when using -f (fast mode).</li>
  
    <li> subblock code now working.&nbsp; But no algorithm to choose subblock
gains yet.</li>
  
    <li> Linux now segfaults on floating point exceptions.&nbsp; Should prevent 
me from releasing binaries that crash on other operating systems.</li>
 
  </ul>
  
  <h4> May 22 1999</h4>
  
  <ul>
 
    <li> Version 3.04 released.</li>
  
    <li> Preliminary documentation from Gabriel Bouvigne.</li>
  
    <li> <font color="#3366ff">I wouldn't have thought it was possible, but
now there are even more speed improvements from Chris Matrakidis!&nbsp; Removed 
one FFT when using joint stereo, and many improvements in loop.c.</font></li>
  
    <li> "Fake" ms_stereo mode renamed "Force" ms_stereo since it forces
mid/side stereo on all frames.&nbsp; For some music this is said to be a
problem, but for most music mode is probably better than the default jstereo
because it uses specialized mid/side channel masking thresholds.</li>
  
    <li> Small bugs in Force ms_stereo mode fixed.</li>
  
    <li> Compaq Alpha fixes from Nathan Slingerland.</li>
  
    <li> <font color="#ff0000">Some new experimental pre-echo detection formulas 
in l3psy.c (#ifdef AWS and #ifdef ENER_AWS, both off by default.&nbsp; Thanks
to Gabriel Bouvigne and Andre Osterhues)</font></li>
  
    <li> Several bugs in the syncing of data displayed by mp3x (the frame
analyzer) were fixed.</li>
  
    <li> highq (-h) option added.&nbsp; This turns on things (just one so
far) that should sound better but slow down LAME.</li>
 
  </ul>
 <b>May 18 1999</b> 
  <ul>
 
    <li> Version 3.03 released.</li>
  
    <li> <font color="#3366ff">Faster (20%) &amp; cleaner FFT (Thanks to
Chris Matrakidis&nbsp; http://www.geocities.com/ResearchTriangle/8869/fft_summary.html)</font></li>
  
    <li> mods so it works with VC++ (Thanks to Gabriel Bouvigne, www.mp3tech.org)</li>
  
    <li> MP3s marked "original" by default&nbsp; (Thanks to Gabriel Bouvigne,
www.mp3tech.org)</li>
  
    <li> Can now be compiled into a BladeEnc compatible .DLL&nbsp;&nbsp;
(Thanks to&nbsp; Albert Faber, CDex author)</li>
  
    <li> Patches for "silent mode" and stdin/stdout&nbsp; (Thanks to Lars
Magne Ingebrigtsen)</li>
  
    <li> <font color="#ff0000">Fixed rare bug: if a long_block is sandwiched
between two short_blocks, it must be changed to a short_block, but the short_block 
ratios have not been computed in l3psy.c.&nbsp; Now always compute short_block 
ratios just in case.</font></li>
  
    <li> <font color="#ff0000">Fixed bug with initial quantize step size
when many coefficients are zero.&nbsp; (Thanks to Martin Weghofer).</font></li>
  
    <li> Bug fixed in MP3x display of audible distortion.</li>
  
    <li> improved status display (Thanks to Lauri Ahonen).</li>
 
  </ul>
  
  <h4> May 12 1999</h4>
  
  <ul>
 
    <li> Version 3.02 released.</li>
  
    <li> <font color="#ff0000">encoder could use ms_stereo even if channel
0 and 1 block types were different.&nbsp; (Thanks to Jan Rafaj)</font></li>
  
    <li> <font color="#ff0000">added -k option to disable the 16 kHz cutoff
at 128kbs.&nbsp; This cutoff is never used at higher bitrates. (Thanks to
Jan Rafaj)</font></li>
  
    <li> <font color="#ff0000">modified pe bit allocation formula to make
sense at bit rates other than 128kbs.</font></li>
  
    <li> fixed l3_xmin initialization problem which showed up under FreeBSD.&nbsp; 
(Thanks to Warren Toomey)</li>
 
  </ul>
 <b>May 11 1999</b> 
  <ul>
 
    <li> Version 3.01 released</li>
  
    <li> max_name_size increased to 300&nbsp; (Thanks to Mike Oliphant)</li>
  
    <li> patch to allow seeks on input file (Thanks to Scott Manley)</li>
  
    <li> fixes for mono modes (Thanks to everyone who pointed this out)</li>
  
    <li> overflow in calc_noise2 fixed</li>
  
    <li> bit reservoir overflow when encoding lots of frames with all zeros&nbsp; 
(Thanks to Jani Frilander)</li>
 
  </ul>
  
  <p><br>
  <b>May 10 1999</b> </p>
  <ul>
 
    <li> Version 3.0 released</li>
  
    <li> <font color="#ff0000">added GPSYCHO (developed by Mark Taylor)</font></li>
  
    <li> <font color="#000000">added MP3x (developed by Mark Taylor)</font></li>
  
    <li> LAME now maintained by Mark Taylor</li>
 
  </ul>
 <b>November 8 1998</b> 
  <ul>
 
    <li> Version 2.1f released</li>
  
    <li> 50% faster filter_subband() routine in encode.c contributed by James
Droppo</li>
 
  </ul>
 <b>November 2 1998</b> 
  <ul>
 
    <li> Version 2.1e released.</li>
  
    <li> New command line switch <b>-a</b> auto-resamples a stereo input
file to mono.</li>
  
    <li> New command line switch <b>-r</b> resamples from 44.1 kHz to 32 kHz
[this switch doesn't work really well. Very tinny sounding output files.
Has to do with the way I do the resampling probably]</li>
  
    <li> Both of these were put into the ISO code in the encode.c file, and
are simply different ways of filling the input buffers from a file.</li>
 
  </ul>
 <b>October 31 1998</b> 
  <ul>
 
    <li> Version 2.1d released</li>
  
    <li> Fixed memory alloc in musicin.c (for l3_sb_sample)</li>
  
    <li> Added new command line switch (-x) to force swapping of byte order</li>
  
    <li> Cleaned up memory routines in l3psy.c. All the mem_alloc() and free()
routines where changed so that it was only done <i>once</i> and not every
single time the routine was called.</li>
  
    <li> Added a compile time switch -DTIMER that includes all timing info.
It's a switch for the time being until some other people have tested on their 
system. Timing code has a tendency to do different things on different platforms.</li>
 
  </ul>
 <b>October 18 1998</b> 
  <ul>
 
    <li> Version 2.1b released.</li>
  
    <li> Fixed up bug: all PCM files were being read as WAV.</li>
  
    <li> Played with the mem_alloc routine to fix crash under amigaos (just
allocating twice as much memory as needed). Might see if we can totally do
without this routine. Individual malloc()s where they are needed instead</li>
  
    <li> Put Jan Peman's quality switch back in. This reduces quality via
the '-q&nbsp;<int>' switch. Fun speedup which is mostly harmless if you're
not concerned with quality.</int></li>
  
    <li> Compiling with amiga-gcc works fine</li>
 
  </ul>
 <b>October 16 1998</b> 
  <ul>
 
    <li> Version 2.1a released. User input/output has been cleaned up a bit.
WAV file reading is there in a very rudimentary sense ie the program will
recognize the header and skip it, but not read it. The WAV file is assumed
to be 16bit stereo 44.1 kHz.</li>
 
  </ul>
 <b>October 6 1998</b> 
  <ul>
 
    <li> Version 2.1 released with all tables now incorporated into the exe.
Thanks to <b>Lars Magne Ingebrigtseni</b>([e-mail])</li>
 
  </ul>
 <b>October 4 1998</b>In response to some concerns about the quality of the
encoder, I have rebuilt the encoder from scratch and carefully compared output
at all stages with the output of the unmodified ISO encoder. <a href="http://www.uq.net.au/%7Ezzmcheng/lame/download.html">
Version2.0</a> of LAME is built from the ISO source code (dist10), and incorporates
modifications from myself and the 8hz effort. The output file from LAME v2.0
is <i>identical</i> to the output of the ISO encoder for my test file.Since
I do not have heaps of time, I left the ISO AIFF file reader in the code,
and did not incorporate a WAV file reader.Added section on <a href="http://www.uq.net.au/%7Ezzmcheng/lame/quality.html">
quality</a><b>October 1 1998</b> 
  <ul>
 
    <li> Updated web page and released LAME v1.0</li>
 
  </ul>
 <b>Up to September 1998</b> 
  <ul>
Working on the 8hz source code... 
    <ul>
 
      <li> Patched the <a href="http://www.8hz.com/">8hz</a> source code</li>
  
      <li> 45% faster than original source (on my freebsd p166).</li>
  
      <ul>
 
        <li> m1 - sped up the mdct.c and quantize() functions [MDCTD, MDCTD2,
LOOPD]</li>
  
        <li> m2 - sped up the filter_subband routine using <b>Stephane Tavenard</b>
 's work from musicin [FILTST]</li>
  
        <li> m2 - minor cleanup of window_subband [WINDST2]</li>
  
        <li> m2 - Cleaned up a few bits in l3psy.c. Replaced a sparse matrix
multiply with a hand configured unrolling [PSYD]</li>
  
        <li> m3 - (amiga only) Added in the asm FFT for m68k (based on sources
from <b>Henryk Richter</b> and <b>Stephane Tavenard</b>)</li>
  
        <li> m4 - raw pcm support back in</li>
  
        <li> m5 - put in a byte-ordering switch for raw PCM reading (just
in case)</li>
  
        <li> m6 - reworked the whole fft.c file. fft now 10-15% faster.</li>
  
        <li> m7 - totally new fft routine. exploits fact that this is a real-&gt;complex 
fft. About twice as fast as previous fastest fft (in m6). (C fft routine is
faster than the asm one on an m68k!)</li>
  
        <li> m8</li>
  
        <ul>
 
          <li> - Now encodes from stdin. Use '-' as the input filename. Thanks
to <b>Brad Threatt</b></li>
  
          <li> - Worked out that the 1024point FFT only ever