Google Git
Sign in
android / platform / external / lame / c4cbe47910d2ece1617411206c6b5e4ffbe09360 / . / doc / html / history.html
blob: 34775f5b397152d6df463fd5e8ddc0a7df4b2f1f [file] [log] [blame]
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.7 [en] (X11; I; Linux 2.2.12-20 i686) [Netscape]">
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
<title>LAME Changelog</title>
</head>
<body style="color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" alink="#bbbbbb" link="#888888" vlink="#555555">
<center>
<h1>History</h1>
</center>
Starting with LAME 3.0: <br>
<font color="#ff0000">red = features and bug fixes which
affect quality</font> <br>
<font color="#3366ff">blue = features and bug fixes which
affect speed</font> <br>
black = usability, portability, other
<hr>
<hr>
<h3>LAME 3.98.4 &nbsp; &nbsp;March 22 2010</h3>
<ul>
<li>Joseph Flynn
<ul>
<li>Improvements for LAME DirectShow filter:
<ul>
<li>Added support for the DirectShow IAMStreamConfig Interface to the LAME encoder filter output pin.
</li>
<li>Modified the DirectShow filter registration section so that the LAME Encoder filter is correctly registered in the Audio Compressors filter category. This will allow third-party encoding applications using the DirectShow System Device Enumerator Interface to correctly detect the LAME encoder when querying the Audio Compressors filter group.
</li>
<li>Modified the filter registration information so that the MP3 audio subtype is correctly reported as being supported on the encoder output pin. This will allow third-party encoding applications using the DirectShow IFilterMapper2 Interface to recognize that the LAME encoder supports MP3 output.
</li>
<li>Altered the Filter Merit Value that was being used when the filter was registered so that it is now using the standard DirectShow compressor filter merit value of MERIT_DO_NOT_USE (0x200000). Previously, the filter was being registered using a value of MERIT_SW_COMPRESSOR (0x100000), which was at a lower priority (i.e. worse priority) than MERIT_DO_NOT_USE. This prevented the LAME Encoder filter from being selected for use by some third-party encoding applications.
</li>
<li>Added code to calculate the frame length of the audio frames used for the nBlockSize element of the WAVEFORMATEX output structure. Previously this value was simply hard-coded to 1.
</li>
</ul>
</li>
</ul>
</li>
<li>Robert Hegemann
<ul>
<li><font color="#ff0000">Fix for Bugtracker item <i>[ 2973877 ] A problem regarding the new drain code</i></font>
</li>
</ul>
</li>
</ul>
<h3>LAME 3.98.3 &nbsp; &nbsp;February 27 2010</h3>
<ul>
<li>Rog&eacute;rio Brito:
<ul>
<li>Update the debian packaging for the new release.
</li>
</ul>
</li>
<li>Robert Hegemann
<ul>
<li>The <b>ignore-tag-errors</b> switch had no effect when embedding album art, fixed.</li>
<li>Library API change: lame_decode functions are now obsolete but still present, please use hip_decode instead.
The reason for this change is: lame_decode functions use a single global variable within the library
to store decoder setup, hip_decode functions don't.
The encoder now uses hip_decode internally and it is now possible to use <b>clipdetect</b>
feature while reencoding mp3 to mp3.
</li>
<li>Workaround for FFMPEG bug, which uses to call lame_encode_flush more than once in a loop.
</li>
<li>Windows: program icon and version info added (when building with VC9)
</li>
<li>Fix for Bugtracker item <i>[ 2688413 ] lib name problem in Microsoft Visual Studio 6</i>
</li>
<li>Fix for Bugtracker items <i>[ 2051870, 2423650, 2928684 ] several small documentation issues</i>
</li>
<li>Fix for Bugtracker item <i>[ 2723518 ] resampling in 3.98 and 3.99alpha</i>
</li>
<li>Fix for Bugtracker item <i>[ 2891879 ] Because of Windows API change, there was a problem with the <b>prority</b> switch.</i>
</li>
<li>Fix for Bugtracker item <i>[ 2893101 ] Access Violation in BladeMP3EncDLL if UNICODE was defined.</i>
</li>
<li>Fix for Bugtracker item <i>[ 2887359 ] Wrong length in ID3v2 tag when num_samples isn't set</i>
</li>
<li>Fix for Bugtracker item <i>[ 2872590 ] LameTAG: "Music length" missmatch in LAME 3.98</i>
</li>
<li>Fix for Bugtracker item <i>[ 2824296 ] wrong enc_padding value in LAME 3.99a and 3.98.3 (from CVS)</i>
</li>
<li><font color="#ff0000">Revisiting the FhG decoder problem (FhG V1.5 build 50, ships with MS Windows):
enabling the new-drain-code seems to solve that issue better, than restricting the buffer size (see below: 3.98 beta 1, May 16 2007).</font>
</li>
<li>Patch submitted by Bernhard Doebler, tracker item <i>[ 2807676 ] Error when building Lame with NASM support</i>
</li>
<li>Patch submitted by Mancuso Raffaele, tracker item <i>[ 2406420 ] compile lame_enc.dll under cygwin</i>
</li>
</ul>
</li>
</ul>
<h3>LAME 3.98.2 &nbsp; &nbsp;September 22 2008</h3>
<ul>
<li>Robert Hegemann</li>
<ul>
<li>Fix for Bugtracker item <i>[ 2123206 ] lame 3.98.1 segfaults with -h</i>
</li>
</ul>
</ul>
<h3>LAME 3.98.1 &nbsp; &nbsp;September 21 2008</h3>
<ul>
<li>Rog&eacute;rio Brito:
<ul>
<li>More fixes for the abx tool for Unix systems:
<ul>
<li>Plugged a memory leak.</li>
<li>Fixed an endianness problem: users of big-endian machines
can now do abx tests.</li>
</ul>
</li>
<li>Fixed history's HTML doctype</li>
<li>
Fixed history so that it <em>finally</em> validates
at <a href="http://validator.w3.org/">W3's validator</a>
</li>
<li>
Fixed compilation of frontend <code>mp3rtp.c</code>. Thanks to Kris Karas.
Bugtracker item <i>[ 2015432 ] mp3rtp missing uint16_t in lame 3.98</i>
</li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li>Fix for Bugtracker item <i>[ 2031704 ] --id3v1-only didnt work in 3.98-final</i></li>
<li>Fix for Bugtracker item <i>[ 2022035 ] encoder_padding value and resampling</i></li>
<li>Fix for Bugtracker item <i>[ 2029282 ] Frequency filtering API broken in 3.98</i></li>
<li>Fix for Bugtracker item <i>[ 2039648 ] potential memory leak in parse_args() function in parse.c</i></li>
<li>Fix for some tagging issues:
<ul>
<li>Made search for ID3v1 genres more sloppy, abbrevations may match more often as some simple typos.
Examples:<ul><li>--tg "Alt. Rock" matches genre "Alternate Rock"</li>
<li>--tg "acapela" matches genre "A Cappella"</li>
</ul></li>
<li>New switch --pad-id3v2-size "n": adds ID3v2 tag with n padding bytes.</li>
</ul></li>
</ul>
</li>
</ul>
<h3>LAME 3.98 &nbsp; &nbsp;July 4 2008</h3>
<ul>
<li>Anton Sergunov:
<ul>
<li>Frontend DirectShow: enabling LAME dshow filter to connect to "File Writer Filter".
</li>
</ul>
</li>
<li>Rog&eacute;rio Brito:
<ul>
<li>Updates to the Debian Packaging</li>
<li>Fixes to the abx tool for Unix systems (so that more people
can evaluate LAME's compression against the original files)</li>
</ul>
</li>
<li>Alexander Leidinger:
<ul>
<li>explicitely link the math lib to the lame lib</li>
<li>add switch to disable the use of the compaq optimized math lib</li>
</ul>
</li>
</ul>
<h3>LAME 3.98 beta 8&nbsp; &nbsp;April 13 2008</h3>
<ul>
<li>Robert Hegemann:
<ul>
<li>LAME now accepts a floating point value in the range [0,...,10[ as VBR quality setting, like <b>-V5.678</b>
</li>
<li>Found and fixed some suspicious code in additive masking calculation for VBR-NEW
</li>
<li>bug-fix:<font color="#ff0000">experimental code was defaulted by accident for VBR-NEW</font>
</li>
<li>fix for some endianess problem on big-endian machines
</li>
</ul>
</li>
</ul>
<h3>LAME 3.98 beta 7&nbsp; &nbsp;April 6 2008</h3>
<ul>
<li>Robert Hegemann:
<ul>
<li>libmp3lame API: allow frontends to separately retrieve LAME/Xing and ID3 data, because the old library automatism
makes it impossible to make fully buffered encodes.
</li>
<li>libmp3lame API: added some experimental unicode ID3 tagging code.
</li>
<li>frontends: write itself final ID3 tags and LAME/Xing header frame
</li>
<li>lame_enc.dll: writes itself final LAME/Xing header frame
</li>
<li>Latest changes to the new VBR psymodel:
<ul>
<li>uses a different spreading function
</li>
<li><font color="#ff0000">bug-fix for out-of-bounds array access (program stack corruption possible)</font>
</li>
</ul>
</li>
</ul>
</li>
</ul>
<h3>LAME 3.98 beta 6&nbsp; &nbsp;December 16 2007</h3>
<ul>
<li>Robert Hegemann:
<ul>
<li>Feature request <i>[ 1811483 ] WAVE_FORMAT_EXTENSIBLE support (PCM)</i>
</li>
<li><font color="#ff0000">Fix for some rare scalefactor selection issue the newer vbr code had at low compression levels
</font></li>
<li>Fix for Bugtracker item<i>[ 1813496 ] AIFF parsing bug</i>
</li>
<li>Latest changes to the new VBR code:
<ul>
<li><font color="#3366ff">it now has its own psy model, a derivation from NSPSY.</font>
</li>
<li><font color="#ff0000">some more tuning has been done for this new psy model.</font>
Many thanks to Horst Albrecht and Myles Thaiss.
</li>
<li><font color="#ff0000">the "out-of-bits" strategy is reworked</font>
</li>
</ul>
<li><font color="#ff0000">It was possible, that the "interchannel masking effects feature"
was used by the dual-channel-mode for bi-lingual encodings too. It was meant to work
on stereo L/R channels only.</font>
</li>
</ul>
</li>
</ul>
<h3>LAME 3.98 beta 5&nbsp; &nbsp;August 12 2007</h3>
<ul>
<li>Jonathan Stott:
<ul>
<li>Bug tracker items: <i>[ 1590693 ] ID3v2 tag not writing, [ 1636267 ] ID3v2 tags overwritten</i><br>
If the output file is opened 'write-only', then LAME can't update the LAME tag.
In this case LAME silently overwrote the first bytes of the file and an
optional ID3v2 tag disappeared. Now an error message will be printed and
no data is written in this case.
</li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li>Fix for Bugtracker item <i>[ 1719593 ] Track numbers > 255 not allowed even with --id3v2-only</i>
<li>Fix for Bugtracker item <i>[ 1742623 ] fail(lame --mp3input -m m -b 128 --resample 8 *.mp3 **.mp3)</i><br>
The problem here was, the input files are MPEG-1 Layer2 files named as MP3s. Even if you leave out
the --mp3input switch LAME tried to decode the input files as Layer3 files because of the file name
extension and because it found some valid looking Layer3 synchronization header.
The fixed LAME version does not assume the file name extension is always correct and treats the files
depending on the first found MPEG sync word. The files in question are now correctly detected as
Layer2 files and transcoding does succeed.
</li>
<li>Fix for Bugtracker item <i>[ 1445175 ] Input being stdin fails in Windows on WAV files</i><br>
The problem here was, seeking on pipes shows some different behaviour depending on C-Library
implementations. The workaround tries to detect it's working on a pipe and doing some
reading instead of seeking in that case.
</li>
<li>Fixing some memory leak in the 'lame_enc.dll'.
</li>
<li>Fix for Bugtracker items <i>[ 1160757, 1160741 ] --little-endian / --big-endian not working</i><br>
These switches where originally intended to be used together with Libsndfile only.
</li>
<li>Fix for Bugtracker item <i>[ 1746336 ] Incorrect Bitrate with ABR und --resample, LAME 3.98b4</i><br>
Some earlier bug-fix had some typo. As a result, when adding a '--resample 123' switch,
the average bitrate rised upto maximum bitrate.
</li>
</ul>
</li>
</ul>
<h3>LAME 3.98 beta 4&nbsp; &nbsp;June 23 2007</h3>
<ul>
<li>Dennis Lambe Jr: Added support for total track count (id3v2) in the frontend
</li>
<li>Nyaochi:
<ul>
<li>Ability to set user-defined ID3v2.3 frame</li>
<li>Ability to include albumArt in ID3v2.3 tag</li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li>Bugfix: the "play length in ms", which is stored in the ID3v2 tag TLEN, was not correctly computed.
Some hardware and software players were confused by this garbage data.</li>
<li><font color="#ff0000">Out of bits strategy for the newer VBR code overhauled</font></li>
<li>LAME API: the ID3 tag functions do not store the pointers passed anymore,
they do make deep copies of strings passed as parameters.
</li>
<li>Changes in LAME frontend switches regarding ID3 tags:
<br><tt>--tg "MyGenre"</tt> will route unknown ID3v1 genres to "Other" for ID3v1 tags
and will be stored as plain text "MyGenre" for ID3v2 tags. Genres given by known
ID3v1 numbers will be stored as its corresponding text in ID3v2 tags.
<br><tt>--tn "02/02"</tt> will store the track number specified as plain text as-is
for ID3v2 tags.
</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.98 beta 3&nbsp; &nbsp;May 22 2007</h3>
<ul>
<li>Robert Hegemann:
<ul>
<li>Fixes regarding max number of bits limitation</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.98 beta 2&nbsp; &nbsp;May 20 2007</h3>
<ul>
<li>Robert Hegemann:
<ul>
<li>Bug tracker item: <i>[ 1693461 ];</i>
Fixed memory leaks in ACM codec</li>
<li>Fixed encoding of non-standard sampling rates in CBR</li>
<li><font color="#ff0000">Improved VBR strategy when running out of bits</font></li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.98 beta 1&nbsp; &nbsp;May 16 2007</h3>
<ul>
<li>Alexander Leidinger:
<ul>
<li>Add TLEN (ID3v2) support (Submitted by: Linus Walleij).</li>
<li>Add number of total tracks per album (ID3v2) support
(Submitted by: Kyle VanderBeek).</li>
<li>Some seatbelts for overflowing arrays in the ID3v2
support.</li>
<li>Update the RPM spec (Submitted by: Kyle VanderBeek).</li>
<li>Fix some mem-leaks in the error case.</li>
<li>Update to newer autotools versions.</li>
<li>Update to use a recent libsndfile (submitted by
libsndfile author).</li>
<li><font color="#3366ff">Intrinsics support
enabled for gcc</font></li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li><font color="#3366ff">The newer VBR code is now LAME's default
VBR routine</font></li>
<li><font color="#ff0000">Fixed: in case of
not enough bits the new vbr code incorrectly used old vbr routine</font></li>
<li><font color="#ff0000">Improved ATH
adjustment in low volume cases</font></li>
<li><font color="#ff0000">Fixed (PSY model): mapping from convolution bands
to partition bands was broken since we replaced tables by own calculation
several years ago</font></li>
<li><font color="#ff0000">Fixed (PSY model): loss of fraction in equal loudness weighting</font></li>
<li><font color="#ff0000">Fixed (PSY model): in NSPSY highpass filter, out of bounds access in fircoef</font></li>
<li><font color="#ff0000">Known problem samples for the new VBR code:
many of them are at an acceptable quality level now;</font>
with a big <b>'Thank You'</b> to Francis Niechcial</li>
<li><font color="#ff0000">Modified VBR strategy to handle out of bits cases</font></li>
<li>Restricted bitreservoir size for 320 kbps frames to
the size used for sideinfo, because of decoding problems
with FhG decoders installed on almost every Windows system</li>
<li>LAME aborts on unsupported input files
or unrecognized parameter options passed more often now </li>
<li>Bug tracker item: <i>[ 1596306 ] "fatal error during initialization";</i>
an invalid MPEG samplerate was returned by optimum_samplefreq function</li>
<li>Bug tracker item: <i>[ 1585942 ] lame not --silent when TERM not set;</i>
in case LAME was build with TERMCAP defined and no TERM
environment is defined, now we do not issue an error message and
silently fallback to the default behaviour as if LAME was
compiled without TERMCAP defined.</li>
<li>Bug tracker item: <i>[ 1711980 ] LAME writes invalid Xing header when ID3 tags exist;</i>
LAME was sometimes writing an invalid Xing/Info header</li>
<li>Feature request: <i>[ 1588283 ] Flushing output stream in lame.exe;</i>
'flush' option added</li>
<li><font color="#3366ff">Added FFTSSE and FFT3DNOW assembler code from
Lame4 branch</font></li>
<li>Changes in lame frontend switches:
-k removed, add lowpass and highpass switches if you need to change
them; --short/noshort/allshort - degraded into DEVELOPER ONLY switches
normal users shouldn't use them; -X -Z degraded to
DEVELOPER ONLY switches, -X is too tough to communicate to
end users and -Z isn't used actualy</li>
<li>Fixed some console printing problems</li>
<li>Windows: ACM code now uses LAME library API only, all
references to private include files are removed</li>
<li>Windows: DirectShow code now uses LAME library API only, all
references to private include files are removed</li>
<li>Windows: disabled code that resets processor affinity,
because this doesn't belong to LAME, but seems to work around
some problems the parent process has (in most cases EAC)</li>
</ul>
</li>
<li>John33:
<ul>
<li>Fixed mp2 and mp3 decoding: For mp3 and mp2 decoding,
this now yields the same output as foobar2000 but the error checking
remains unchanged</li>
</ul>
</li>
<li>Gabriel Bouvigne:
<ul>
<li>VC8 project files</li>
<li>Added support for x64 under VC8</li>
<li>Restricted MPEG 2.5 to 64kbps frames</li>
</ul>
</li>
<li>Takehiro Tominaga:
<ul>
<li><font color="#3366ff">SSE version of FFT</font></li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.97&nbsp; &nbsp;September 24 2006</h3>
<ul>
<li>3.97 beta 3 becomes 3.97 </li>
</ul>
<br>
<h3>LAME 3.97 beta 3&nbsp; &nbsp;August 19 2006</h3>
<ul>
<li>Gabriel Bouvigne:
<ul>
<li><font color="#ff0000">Workaround against a
short blocks detection issue</font></li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.97 beta 2&nbsp; &nbsp;November 26 2005</h3>
<ul>
<li>Gabriel Bouvigne:
<ul>
<li>Fixed an initialization error when input is not using a
standard sampling frequency</li>
<li>Fixed a possible assertion failure in very low bitrate
encoding</li>
<li><font color="#ff0000">Slight change
regarding ATH adjustment with V5</font></li>
<li><font color="#ff0000">Reinstated bit
reservoir for 320kbps CBR</font></li>
<li><font color="#3366ff">ReplayGain analysis
should now be faster when encountering silent parts</font></li>
</ul>
</li>
<li>Takehiro Tominaga:
<ul>
<li>Fixed a possible link problem of assembly code</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.97 beta 1&nbsp; &nbsp;September 12 2005</h3>
<ul>
<li>Takehiro Tominaga:
<ul>
<li>Fixed an out of array access in mp3rtp </li>
<li><font color="#ff0000">Fixed a quality
setting in DLL</font></li>
<li>Fixed display when using --silent</li>
</ul>
</li>
<li>Vitaly Ivanov:
<ul>
<li>Updated DirectShow interface</li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li><font color="#ff0000">Fixed an out of
array access</font></li>
<li><font color="#ff0000">Fixed some small
rounding problem in vbr-new quantization routines</font></li>
<li><font color="#ff0000">Fixed a bug in
vbr-new regarding high frequencies (sfb21) when using -Y</font></li>
<li><font color="#ff0000">Fixed a few bugs in
vbr-new when using -Y</font></li>
<li><font color="#ff0000">Updated scalefactors
allocation scheme in vbr-new</font></li>
<li>Fixed mingw32 configure problems</li>
<li>Resolved some compiler warnings</li>
<li>Updated command-line visualisation</li>
</ul>
</li>
<li>Gabriel Bouvigne:
<ul>
<li>Changed some FLOAT8 to FLOAT</li>
<li>Added project files for VC7</li>
<li><font color="#ff0000">Reworked -q1 and -q0</font></li>
<li><font color="#ff0000">Updated presets</font></li>
<li><font color="#ff0000">Fixed an error in
ISO quantization on systems not using the IEEE754 hack</font></li>
<li><font color="#3366ff">Faster quantization</font></li>
<li><font color="#3366ff">SSE version of
init_xrpow</font></li>
</ul>
</li>
<li>Rog&eacute;rio Brito:
<ul>
<li>Updated Debian packaging</li>
<li>Documentation work</li>
</ul>
</li>
<li>Chris Miller:
<ul>
<li>Support for x64 platform SDK in makefile.msvc</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.96.1 &nbsp; &nbsp;July 25 2004</h3>
<ul>
<li>Robert Hegemann:
<ul>
<li><font color="#ff0000">Fixed a rare bug in
vbr-new (could lead to crashes or data corruption)</font></li>
</ul>
</li>
<li>Gabriel Bouvigne:
<ul>
<li>some fixes in ACM codec</li>
<li>fixed padding when encoding to 320kbps</li>
<li><font color="#ff0000">fixed block size
selection for mid and side channels</font></li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.96 &nbsp; &nbsp;April 11 2004</h3>
<ul>
<li>Gabriel Bouvigne:
<ul>
<li><font color="#ff0000">new quantization
selection mode (used in ABR/CBR)</font></li>
<li><font color="#ff0000">set sfscale for
ABR/CBR up to 160kbps</font></li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.96 beta 2 &nbsp; &nbsp;March 28 2004</h3>
<ul>
<li>Takehiro Tominaga:
<ul>
<li><font color="#3366ff">removed unnecessary
integer convertion in resampling</font></li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li><font color="#ff0000">reworked scalefactor
allocation in vbr-new</font></li>
<li>fixed a freeformat decoding problem</li>
</ul>
</li>
<li>Gabriel Bouvigne:
<ul>
<li><font color="#ff0000">updated minimal
bitrate for V1 and V2</font></li>
</ul>
</li>
<li>Aleksander Korzynski:
<ul>
<li>added ability to disable ReplayGain analysis</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.96 beta &nbsp; &nbsp;March 7 2004</h3>
<ul>
<li>Takehiro Tominaga:
<ul>
<li>fixed decoding issue</li>
</ul>
</li>
<li>Aleksander Korzynski:
<ul>
<li>changed internal ReplayGain handling</li>
<li>fixed some issues when ReplayGain is used with
resampling</li>
</ul>
</li>
<li>Robert Hegemann:
<ul>
<li>added standard ISO quantization for vbr-new, used at
lower quality settings</li>
<li><font color="#3366ff">faster count_bits
for vbr-new</font></li>
<li><font color="#3366ff">faster
find_scalefac_ave function for vbr-new</font></li>
<li><font color="#ff0000">fixed an out of
array access in psychoacoustic models; this bug could make some psy
calculations worthless and sometimes let lame crash</font></li>
<li><font color="#ff0000">fixed an error on
silent scalefactor bands; this bug resulted in huffman data overrun
problems while decoding, resulting in audible glitches</font></li>
<li>fixed a freeformat decoding bug</li>
</ul>
</li>
<li>Gabriel Bouvigne:
<ul>
<li><font color="#ff0000">adjusted short block
thresholds</font></li>
<li>fixed some array addressing bugs</li>
<li>made ReplayGain analysis reentrant</li>
</ul>
</li>
<li>David Chandler: fixed a crash in quantize_xrpow </li>
<li>Michal Bacik: fixed a crash when using 8kHz </li>
<li>Goran Markovic: fixed some decoding bugs </li>
<li>John Edwards: fixed a too small buffer in ReplayGain code</li>
</ul>
<br>
<h3>LAME 3.95.1 &nbsp; &nbsp;January 12 2004</h3>
<ul>
<li>Gabriel Bouvigne:
<ul>
<li>fixed a crash when using vbr-new</li>
<li>changed ReplayGain reference level to 89dB</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.95 &nbsp; &nbsp;January 11 2004</h3>
<ul>
<li>Gabriel Bouvigne:
<ul>
<li><font color="#ff0000">fixed lowpass values
when using vbr with mono files</font></li>
<li><font color="#3366ff">faster quantization
loops</font></li>
<li><font color="#3366ff">faster count_bits</font></li>
<li>fixed a buffer requirement error in ACM codec</li>
</ul>
</li>
<li>Takehiro TOMINAGA:
<ul>
<li>fixed mpglib and other decoding support code to prevent
the crash when invalid mp3 input</li>
</ul>
</li>
<li>removed Layer I decoding support</li>
<li><font color="#3366ff">use FastLog and IEEE 754
hack on PowerPC too (approx. 10 percent faster)</font></li>
</ul>
<br>
<h3>LAME 3.94 beta December 15 2003</h3>
<ul>
<li>Takehiro Tominaga:
<ul>
<li><font color="#ff0000">fixed block
switching of nspsytune</font></li>
<li><font color="#ff0000">best huffman divide
in the inner loop.</font> This should improve the quality, but
PAINFULLY slow. So it is not enabled by default. Use -q0 to use it.</li>
<li>Changed -q option mapping. "-q2" until version 3.93 is
now "-q3".</li>
<li><font color="#ff0000">saving bits by
better scalefactor storing</font></li>
<li>removed Vorbis support</li>
<li><font color="#ff0000">substep quantization.</font>This
should help breaking the SFB21 bloating problem</li>
<li><font color="#ff0000">made psychoacoustic
model aware of ATH adjustements</font></li>
<li><font color="#ff0000">use ATH value as
short block masking lower limit</font></li>
<li><font color="#ff0000">several fixes in
psychoacoustic model</font></li>
<li>more robust decoding</li>
</ul>
</li>
<li>Mark Taylor / Gabriel Bouvigne: fixed issues in VBR header</li>
<li>Mark Taylor: workaround against some hardware decoder
defficiencies</li>
<li>Aleksander Korzynski: ability to compute the "Radio"
ReplayGain and detect clipping on the fly. The ReplayGain value is
stored in the Lame tag.</li>
<li>Gabriel Bouvigne:
<ul>
<li><font color="#ff0000">work on presets</font></li>
<li><font color="#ff0000">use presets by
default for cbr/abr</font></li>
<li><font color="#ff0000">use presets by
default for vbr</font></li>
<li><font color="#ff0000">analog silence
detection in partitionned sfb21</font></li>
<li><font color="#3366ff">do not compute noise
in upper 0 part of the spectrum</font></li>
<li><font color="#3366ff">only compute noise
in modified scalefactor bands</font></li>
</ul>
</li>
<li>Guillaume Lessard:
<ul>
<li>nogap related changes</li>
</ul>
</li>
<li>Alexander Leidinger:
<ul>
<li>prevent closing the input fd prematurely if the input
is a named pipe</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.93.1 &nbsp; &nbsp;December 1 2002</h3>
<ul>
<li>Gabriel Bouvigne:
<ul>
<li>preset medium added to the dll interface</li>
<li><font color="#ff0000">fix for abr/cbr
presets</font></li>
<li><font color="#ff0000">fix -q0 switch</font></li>
</ul>
</li>
<li>Alexander Leidinger: fix link problem on systems where
socket() resides in libsocket</li>
</ul>
<br>
<h3>LAME 3.93 &nbsp; &nbsp;November 16 2002</h3>
<ul>
<li>Takehiro Tominaga:
<ul>
<li><font color="#ff0000">bit allocation for
pre-echo control improved for single channel encodings</font></li>
<li><font color="#ff0000">substep noise shaping</font></li>
<li><font color="#3366ff">optimizations by
changing data structure</font></li>
<li><font color="#ff0000">noise shaping model
2 fix</font></li>
<li><font color="#3366ff">nspsytune FIR filter
clean up</font></li>
<li><font color="#ff0000">fix small psymodel
bugs(DC current estimation, preecho detection of non-VBR mode, and
nspsymode initialization)</font></li>
<li>portability fixes for Tru64 UNIX</li>
</ul>
</li>
<li>Albert Faber: some fixes in the DLL</li>
<li>Simon Blandford: fixes for channel scaling in mono mode</li>
<li><font color="#3366ff">Dominique Duvivier: some
optimizations and a faster log10 function</font></li>
<li>Mark Taylor:
<ul>
<li>some tag related fixes in the direct show filter and in
the ACM codec</li>
<li><font color="#3366ff">fixed a mono
encoding bug found by Justin Schoeman</font></li>
<li>calc_noise bug fix</li>
<li>other fixes</li>
</ul>
</li>
<li>Alexander Leidinger:
<ul>
<li>update to autoconf 2.53, rewrite some configure tests</li>
<li>Akos Maroy: determine gcc version even with gcc 3.1</li>
<li>Andrew Bachmann: compile shared libs on BeOS (and
perhaps other arches)</li>
<li>ultrasparc switches for gcc 3.1</li>
<li>fixes for SunOS 4.x</li>
<li>fixes for 64bit arches</li>
<li>CFLAGS fix for IRIX</li>
<li>don't override CFLAGS if exptopt isn't requested</li>
</ul>
</li>
<li>Robert Hegeman:
<ul>
<li><font color="#3366ff">some fixes</font></li>
<li><font color="#ff0000">some fixes for VBR</font></li>
</ul>
</li>
<li>Gabriel Bouvigne:
<ul>
<li>--noasm switch. Might help Cyrix/Via users</li>
<li><font color="#ff0000">presets and
alt-presets merged</font></li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.92 &nbsp; &nbsp;April 14 2002</h3>
<ul>
<li><font color="#ff0000">Alexander
Leidinger:&nbsp; add non linear psymodel (compile time option,
disabled by default)</font>, workaround a bug in gcc 3.0.3
(compiler options, based upon suggestions from various people, see
archives and changelog for more)</li>
<li>Steve Lhomme:&nbsp; ACM wrapper (MS-Windows codec)</li>
<li><font color="#3366ff">Steve Lhomme:&nbsp;
less memory copying on stereo (interleaved) input</font></li>
<li> <font color="#ff0000">Takehiro Tominaga:
Inter-channel masking, enables with --interch x option</font></li>
<li> For buggy versions of gcc compiler (2.96*), back off on
some of the advanced compiler options<br>
</li>
</ul>
<br>
<h3>LAME 3.91 &nbsp; &nbsp;December 29 2001</h3>
<ul>
<li><font color="#ff0000">Darin
Morrison:&nbsp; Bugfix for --alt-preset (for content with low
volume, clean vocals), only important for the "fast standard" preset</font>
</li>
<li>Alexander Leidinger:
<ul>
<li>add some missing files to the distribution</li>
<li>add --alt-preset to the man page</li>
</ul>
</li>
</ul>
<br>
<h3>LAME 3.90 &nbsp; &nbsp;December 21 2001</h3>
<ul>
<li><font color="#ff0000">Many small improvements
and bug fixes not added to history</font></li>
<li><font color="#ff0000">John
Dahlstrom:&nbsp; more fine tuning on the auto adjustment of the ATH</font></li>
<li><font color="#3366ff">Robert
Hegemann:&nbsp; small speed and quality improvements for the old
VBR code (--vbr-old).</font> </li>
<li><font color="#ff0000">Robert
Hegemann:&nbsp; some short block bug fixes</font> </li>
<li><font color="#ff0000">Robert
Hegemann:&nbsp; Big improvements to --vbr-mtrh, now encodes much
more frequencies over 16khz</font> </li>
<li><font color="#ff0000">Robert
Hegemann:&nbsp; --vbr-new code disabled (outdated and lower
quality) and replaced with --vbr-mtrh (Both --vbr-new and --vbr-mtrh
now default to mtrh)</font> </li>
<li>Robert Hegemann:&nbsp; reordering of --longhelp to give
more information, --extrahelp dropped </li>
<li>Darin Morrison:&nbsp; Totally revamped and extremely
high quality unified preset system and other general quality
improvements now available with --alt-presets:
<ul>
<li> <font color="#ff0000">some improvements
to psychoacoustics (vast improvements over default L.A.M.E. modes) when
--alt-preset is used including:</font>
<ul>
<li> <font color="#ff0000">Improved tuning
of short block usage.</font></li>
<li> <font color="#ff0000">Improved
quantization selection usage (the -X modes), now adapts between
appropriate modes on the fly. Also helps on "dropout" problems and with
pre-echo cases.</font></li>
<li> <font color="#ff0000">Improved joint
stereo usage. Thresholds are better tuned now and fix some "dropout"
problems L.A.M.E. suffers from on clips like serioustrouble.</font></li>
<li> <font color="#ff0000">Improved noise
shaping usage. Now switches between noise shaping modes on the fly
(toggles -Z on and off when appropriate) which allows lower bitrates
but without the quality compromise.</font></li>
<li> <font color="#ff0000">Clips vastly
improved over default L.A.M.E. modes (vbr/cbr/abr, including --r3mix):
castanets, florida_seq, death2, fatboy, spahm, gbtinc, ravebase, short,
florida_seq, hihat, bassdrum, 2nd_vent_clip, serioustrouble, bloodline,
and others. No degraded clips known.</font></li>
<li> VBR bitrates are now more "stable" with less
fluctuation -- not dipping too low on some music and not increasing too
high unnecessarily on other music. "--alt-preset standard" provides
bitrates roughly within the range of 180-220kbps, often averaging close
to 192kbps.</li>
</ul></li>
<li> --alt-presets replace the --dm-presets and "metal"
preset is removed and replaced with generic abr and cbr presets.</li>
<li> --alt-preset extreme (note the 'e') replaces xtreme to
help eliminate some confusion</li>
<li> --alt-preset vbr modes now have a fast option which
offers almost no compromise in speed.</li>
<li> --alt-preset standard (and "fast standard") are now
much lower in bitrate, matching --r3mix with an overall average, though
offering higher quality especially on difficult test samples.</li>
<li> --alt-presets are no longer just "presets" as in a
collection of switches, instead they are now quality "modes" because of
special code level tunings (those mentioned above).</li>
<li> Use --alt-preset help for more information.</li>
</ul>
</li>
<li>Roel VdB: more tuning on the --r3mix preset </li>
<li>Jon Dee, Roel VdB:&nbsp; INFO tag</li>
<li>Alexander Leidinger, mp3gain@hotmail.com:&nbsp; added
--scale-l and --scale-r to scale stereo channels independantly </li>
<li>Takehiro Tominaga:&nbsp; <font color="#ff0000">new
noise shaping mode, offering more "cutting edge" shaping according to
masking, enabled via -q0</font> </li>
<li>Mark Taylor:&nbsp; More work on --nogap </li>
<li>Gabriel Bouvigne:&nbsp; Small changes to abr code for
more accurate final bitrate </li>
<li>Gabriel Bouvigne, mp3gain@hotmail.com:&nbsp;
Preliminary <a href="http://www.replaygain.org">
ReplayGain</a> analysis code added (not functional yet) </li>
<li>Gabriel Bouvigne, Alexander Leidinger:&nbsp;
Documentation updates </li>
<li>John Dahlstrom, DSPguru@math.com:&nbsp; floating point
interface function in the Windows DLL</li>
</ul>
<br>
<h3>LAME 3.89beta&nbsp; &nbsp;July 5 2001</h3>
<ul>
<li> John Stewart:&nbsp; long filename support for Win9x/NT.</li>
<li> Takehiro Tominaga:&nbsp; LAME can calculate the CRC of
VBR header, so now "lame -pv" works fine.</li>
<li><font color="#ff0000">Robert
Hegemann:&nbsp; Improvements of the new VBR code (--vbr-mtrh).</font></li>
<li><font color="#3366ff">Robert Hegemann: New VBR
code (--vbr-mtrh) is now defaulted to get more feedback. The VBR speed
is now on par with CBR. We will use the old VBR code in the release.</font></li>
<li><font color="#ff0000">Gabriel Bouvigne: Change
of the maximum frame size limit. LAME should now be more friendly with
hardware players.</font></li>
<li>Gabriel Bouvigne: Size of VBR is now more balanced
according to the -V value.</li>
<li>Alexander Leidinger: Finished the implementation of the
set/get functions.</li>
<li>John Dahlstrom: LAME now handles 24bits input</li>
<li>Mark Taylor: bugs in lame --decode causing truncation of
mp3 file fixed</li>
<li>Mark Taylor: preliminary --nogap support</li>
<li>"Final" API completed: shared library safe! &nbsp;This
API is frozen and should be backwords compatiable with future versions
of libmp3lame.so, but we will continue to add new functionality.
&nbsp;<br>
</li>
</ul>
<h3> LAME 3.88beta&nbsp;&nbsp; March 25 2001</h3>
<ul>
<li> <font color="#ff0000">A lot of work that was
never added to the History!</font></li>
<li> <font color="#ff0000">Frank Klemm and
Gabriel Bouvigne:&nbsp; New ATH formula.&nbsp; Big improvement
for high bitrate encodings.</font></li>
<li> <font color="#ff0000">Takehiro Tominaga:
Temporal masking</font></li>
<li> <font color="#ff0000">Gabriel Bouvigne/Mark
Taylor: auto adjustment of ATH</font></li>
<li> <font color="#ff0000">Robert
Hegemann:&nbsp; Better outer_loop stopping criterion.&nbsp;
Enabled with -q2 or better.</font></li>
<li> <font color="#ff0000">Robert Hegemann/Naoki
Shibata:&nbsp; slow/carefull noise shaping.&nbsp;
-q3..9:&nbsp; amplify all distorted bands.&nbsp; -q2: amplify
distorted bands within 50%.&nbsp; -q1-0:&nbsp; amplify only
most distorted band at each iteration.</font></li>
<li> <font color="#ff0000">Takehiro Tominaga:
Interframe, shortblock temporal masking.</font></li>
<li> Takehiro Tominaga:&nbsp; LAME restructured into a
shared library and front end application.&nbsp; Slight changes to
the API. More changes are coming to turn LAME into a true shared
library (right now you have to recompile if you upgrade the library :-(</li>
<li> <font color="#000000">Naoki Shibata:</font>
<ul>
<li> <font color="#ff0000">improvements to
psychoacoustics</font><font color="#000000">
(--nspsytune)</font>
<li> <font color="#ff0000">BUG in long block
pre echo control fixed </font><font color="#000000">
(some out of range array access in M/S psychoacoustics)</font></li>
</ul>
</li>
<li> <font color="#000000">Ralf
Kempkens:&nbsp;&nbsp;&nbsp; Visual Basic Script for lame,
suggested to put it on your Windows Desktop and you can drag'n'drop
Waves to encode on it.</font></li>
<li> <font color="#000000">Alexander
Stumpf:&nbsp;&nbsp;&nbsp; improved lame.bat for 4Dos users</font></li>
<li> <font color="#000000">Mark Taylor: Several
bugs fixed in the resampling code.</font></li>
<li> <font color="#000000">Frank Klemm, Robert
Hegemann:&nbsp;&nbsp;&nbsp; added assembler code for CPU
feature detection on runtime (MMX, 3DNow, SIMD)</font></li>
<li> <font color="#3366ff">Takehiro Tominaga:
3DNow FFT code.</font></li>
<li> <font color="#000000">Florian Bome,
Alexander Leidinger:&nbsp;&nbsp;&nbsp; more work on
configure stuff</font></li>
<li> <font color="#000000">Alexander
Leidinger:&nbsp;&nbsp; automake/libtool generated Makefiles and
TONS of other work.</font></li>
<li> <font color="#000000">Alexander
Leidinger:&nbsp;&nbsp; Much work towards shared library style
API.</font></li>
<li> <font color="#000000">Anonymous: New more
efficient RTP code.</font></li>
<li> <font color="#ff0000">Mark Taylor:
psycho-acoustic data now computed for all scalefactor bands (up to 24
kHz)</font></li>
<li> <font color="#ff0000">Mark Taylor, Takehiro
Tominaga: All ISO table data replaced by formulas - should improve
MPEG2.5 results for which we never had correct table data.</font></li>
</ul>
<h3> LAME 3.87alpha&nbsp; September 25 2000</h3>
<ul>
<li> Mark Taylor:&nbsp; Bug fixed in LAME/mpglib error
recovery when encountering a corrupt&nbsp; MP3 frame during
*decoding*.</li>
<li> Albert Faber:&nbsp; added LayerI+II decoding support</li>
<li> <font color="#000000">Frank Klemm:&nbsp;
added improved CRC calculation</font></li>
<li> <font color="#000000">Frank Klemm:
substantial code cleanup/improvements</font></li>
<li> Robert Hegemann:&nbsp; Bug fixes
<ul>
<li> <font color="#ff0000">in huffman_init</font>,
could lead to segmentation faults (only in rare cases, most likely at
lower sample rates)</li>
<li> <font color="#ff0000">M/S switching at
lower sample rates</font> (the fact there is no 2nd granule was
ignored)</li>
</ul>
</li>
<li> <font color="#3366ff">Robert
Hegemann:&nbsp; speed up in&nbsp; VBR</font></li>
<li> Jarmo Laakkonen:&nbsp; Amiga/GCC settings for
Makefile.unix.</li>
<li> Magnus Holmgren:&nbsp; README and Makefile for (free)
Borland C++&nbsp; compiler. Will also compile lame_enc.dll, but
this is untested.</li>
<li> Florian Bome:&nbsp;&nbsp; LAME finally has
a&nbsp; ./configure script!!</li>
</ul>
<h3> LAME 3.86beta&nbsp; August 6 2000</h3>
<ul>
<li> Christopher Wise:&nbsp; A makefile for DJGPP, the DOS
version of gcc.&nbsp; Now most windows users should be able to
compile LAME with minimal effort.</li>
<li> <font color="#ff0000">Robert
Hegemann:&nbsp; old VBR:&nbsp;&nbsp; fixed some bugs and
Takehiro's scalefac_scale feature (not yet on by&nbsp;
default.)&nbsp; older LAME versions did not allow to spent more
than 2500 bits of 4095 possible bits to a granule per channel, now
fixed.</font></li>
<li> Robert Hegemann:&nbsp; new VBR:&nbsp;&nbsp;
analog silence treatment like in old VBR</li>
<li> William Welch:&nbsp; Improved options for Linux/Alpha
gcc and ccc compilers in Makefile.</li>
<li> Mathew Hendry:&nbsp; setting appropriate CRC bit for
additional Xing-VBR tagging frame</li>
<li> Don Melton:&nbsp; added ID3 version 2 TAG support</li>
<li> <font color="#000000">John Dahlstrom: fixed
bug allowing timing information (for status in command line encoder) to
overflow.</font></li>
<li> <font color="#000000">Tamito KAJIYAMA, Fixed
several bugs in the LAME/Vorbis interface.</font></li>
<li> <font color="#000000">Mark Taylor:&nbsp;
lame --decode will recognize <a href="http://albumid.cjb.net">Album
ID tags</a></font></li>
<li> <font color="#ff0000">Naoki
Shibata:&nbsp; Additive masking and other improvements to psycho
acoustics.&nbsp; (not yet on by default)</font></li>
</ul>
<h3> LAME 3.85beta&nbsp;&nbsp; July 3 2000</h3>
<ul>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; mid/side stereo demasking thresholds updated.</font></li>
<li> Takehiro Tominaga: New short block MDCT coefficient data
structure.&nbsp; Should allow for future speed improvements.</li>
<li> Robert Hegemann:&nbsp; fixed bug in old VBR routine,
the --noath mode messed up the VBR routine resulting in very large files</li>
<li> Robert Hegemann: found bugs in some sections when using 32
bit floating point.&nbsp; Default is now back to 64bit floating
point.</li>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; Modified PE formula to use ATH.</font></li>
<li> <font color="#000000">S.T.L.:&nbsp;
README.DJGPP - instructions for compiling LAME with DJGPP, the dos
version of gcc.</font></li>
</ul>
<h3> LAME 3.84beta&nbsp; June 30&nbsp; 2000</h3>
<ul>
<li> Mark Weinstein:&nbsp; .wav file output (with --decode
option) was writing the wrong filesize in the .wav file.&nbsp; Now
fixed.</li>
<li> Mark Taylor:&nbsp; (optional) Vorbis support, both
encoding and decoding.&nbsp; LAME can now produce .ogg files, or
even re-encode your entire .ogg collection into
mp3.&nbsp;&nbsp; (Just kidding: it is always a bad idea to
convert from one lossy format to another)</li>
<li> ?: Bug fixed causing VBR to crash under
windows.&nbsp;&nbsp; (pretab[] array overflow)</li>
<li> Sergey Sapelin: Another bug found in the mpg123 MPEG2
tables.&nbsp; Now fixed for the mpg123 based decoder in LAME.</li>
<li> Marco Remondini:&nbsp; VBR histogram works in
win32.&nbsp; compile with -DBRHIST -DNOTERMCAP</li>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; LAME CBR will now use scalefac_scale to expand the
dynamic range of the scalefactors.</font></li>
<li> <font color="#000000">Iwasa Kazmi:&nbsp;
Library improvements:&nbsp; exit()'s, printf, fprintf's are being
replaced by interceptable macros.</font></li>
</ul>
<h3> LAME 3.83beta&nbsp; May 19&nbsp; 2000</h3>
<ul>
<li> <font color="#ff0000">Mark Taylor:&nbsp;
Bug in buffering routines:&nbsp; in some cases, could cause
MDCT&nbsp; to read past end of buffer.&nbsp; Rare in MPEG2,
even more rare for MPEG1, but potentially serious!</font></li>
<li> Mark Taylor:&nbsp; MDCT/polyphase filterbank was not
being "primed" properly.&nbsp; Does not effect output unless you
set the encoder delay lower than the default of 576 samples.</li>
<li> <font color="#ff0000">Mark Taylor:&nbsp;
"vdbj" and "Caster"&nbsp; found several VBR bugs (now
fixed):&nbsp;&nbsp; 1.&nbsp; Analog silence detection only
checked frequencies up to 16 kHz.&nbsp; 2.&nbsp; VBR mode could
still somehow avoid -F mode.&nbsp; 3.&nbsp; VBR mode would
ignore noise above 16 kHz (scalefactor band 22), Now calc_noise1 will
compute the noise in this band when in VBR mode.&nbsp; Not
calculated in CBR&nbsp; mode since CBR algorithm has no way of
using this information.</font></li>
<li> Mark Taylor:&nbsp; scalefactor band 22 info
(masking(=ATH),&nbsp; noise and energy) now displayed in frame
analyzer.</li>
<li> <font color="#ff0000">VBR code ATH tuning
was disabled by accident in 3.81, now fixed.</font></li>
<li> <font color="#000000">Mark Taylor:&nbsp;
lame --decode will produce .wav files.&nbsp; (oops - size is off by
a factor of 4)</font></li>
</ul>
<h3> LAME 3.82beta&nbsp;&nbsp; May 11 2000</h3>
<ul>
<li> Robert Hegemann:&nbsp; Fixed bug in high bitrate joint
stereo encodings.</li>
<li> <font color="#3366ff">Naoki
Shibata:&nbsp; new long block MDCT routine</font></li>
</ul>
<h3> LAME 3.81beta&nbsp; May 8 2000</h3>
<ul>
<li> all ISO code removed!</li>
<li> <font color="#3366ff">Takehiro Tominaga and
Naoki Shibata:&nbsp; new window subband routines.</font></li>
<li> <font color="#000000">Naoki
Shibata:&nbsp; Bug fix in mpglib (decoding) lib:&nbsp; in some
cases, MDCT coefficients from previous granule was incorrectly used for
the next granule.</font></li>
<li> <font color="#ff0000">ISO 7680 bit buffer
limitation removed.&nbsp; It can be reactivated with
"--strictly-enforce-ISO"&nbsp; Please report any trouble with high
bitrates.</font></li>
</ul>
<h3> LAME 3.80beta&nbsp; May 6 2000</h3>
<ul>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; more efficient and faster huffman encoding!</font></li>
<li> <font color="#ff0000">Takehiro Tominaga and
Mark Taylor:&nbsp; much improved short block compression!</font></li>
<li> <font color="#000000">Tomasz Motylewski and
Mark Taylor:&nbsp; MPEG2.5 now supported!</font></li>
<li> <font color="#000000">Mark Taylor:
incorporated Takehiro's bitstream.c!&nbsp; bitstream.c used by
default, but old ISO bitstream code can also be used.</font></li>
<li> <font color="#ff0000">Scott&nbsp; Manley
and Mark Taylor:&nbsp; good resampling routine finaly in
LAME.&nbsp; uses a 19 point FIR filter with Blackman
window.&nbsp; Very slow for non integer resampling ratios.</font></li>
<li> <font color="#000000">Iwasa Kazmi:&nbsp;
fixed SIGBUS error:&nbsp; VBR and id3 tags were using data after it
was free()'d.</font></li>
<li> <font color="#ff0000">Robert
Hegemann:&nbsp; Improved VBR tuning.&nbsp; #define
RH_QUALITY_CONTROL and #RH_SIDE_VBR now the defaults.</font></li>
<li> <font color="#000000">Robert
Hegemann:&nbsp;&nbsp; LAME version string now added to
ancillary data.</font></li>
<li> Kimmo Mustonen:&nbsp; VBR histogram support for Amiga.</li>
<li> Casper Gripenberg:&nbsp; VBR stats (but not histogram)
for DOS verson.</li>
<li> Robert Hegemann:&nbsp; rare VBR overflow bug fixed.</li>
<li> Zack:&nbsp; -F option strictly enforces the VBR min
bitrate.&nbsp; Without -F, LAME will ignore the minimum bitrate
when encoding analog silence.</li>
<li> Shawn Riley:&nbsp; User can now specify a compression
ratio (--comp &lt;arg&gt;) instead of a bit rate.&nbsp;
Default settings based on a compression ratio of 11.0</li>
<li> Mark Taylor:&nbsp; free format bitstreams can be
created with --freeformat, and specify any integer bitrate from 8 to
320kbs with -b.</li>
<li> Mark Taylor: lame be used as a decoder (output raw pcm
only):&nbsp; lame --decode input.mp3 output.pcm</li>
</ul>
<h3> LAME 3.70&nbsp;&nbsp; April 6 2000</h3>
<ul>
<li> "LAME 3.69beta" becomes LAME 3.70 "stable"</li>
</ul>
<h3> LAME 3.69beta&nbsp;&nbsp; April 6 2000</h3>
<ul>
<li> "spahm":&nbsp; default mode selection bug
fixed.&nbsp; In some cases, lame was defaulting to regular stereo
instead of jstereo when the user did not specify a mode.</li>
</ul>
<h3> LAME 3.68beta&nbsp; April 4 2000</h3>
<ul>
<li> Mark Taylor: mono encoding bug in DLL fixed.</li>
<li> Ingo Saitz: bug in --cwlimit argument parsing fixed.</li>
<li> <font color="#ff0000">Scott Manly: bug in
4-point resample code fixed.</font></li>
</ul>
<h3> LAME 3.67beta&nbsp; March 27 2000</h3>
<ul>
<li> <font color="#ff0000">Robert
Hegemann:&nbsp; jstereo now enabled for MPEG2 encodings</font></li>
<li> Mark Taylor: old M/S stereo mode which used L/R maskings
has been removed.</li>
<li> Mark Taylor: Xing MPEG2 VBR headers now working.</li>
<li> <font color="#ff0000">Mark Taylor:&nbsp;
When quantized coefficients are all 0 in a band, set scalefactors to 0
also to save a few bits.</font></li>
<li> <font color="#000000">Ingo Saitz:&nbsp;
Problems with framesize calculation when using -f fast-math option
fixed.</font></li>
</ul>
<h3> LAME 3.66beta March 21 2000</h3>
<ul>
<li> Bug fixes in BladeEnc DLL, possible click in last mp3
frame, VBR historgram display, byteswapping option, ASM quantize
routines work for both float and double.</li>
</ul>
<h3> LAME 3.65beta&nbsp;&nbsp; March 17 2000</h3>
<ul>
<li> Enabled ASM version of quantize_xrpow() - accidently
disabled in lame3.64.</li>
</ul>
<h3> LAME 3.64beta&nbsp; March 16 2000</h3>
<ul>
<li> Don Melton:&nbsp; id3v1.1 tags &amp; id3 bugfixes</li>
<li> <font color="#ff0000">Gabriel
Bouvigne:&nbsp; L/R matching block type fix</font></li>
<li> <font color="#ff0000">Bug fixed which was
allowing quantized values to exceed the maximum when not using -h</font></li>
<li> <font color="#3366ff">Mark Taylor: Fitlers
based on polyphase filterbank.&nbsp; should be slightly better
since the responce is independent of the blocktype, and they are
slightly faster.</font></li>
<li> Mark Taylor: API:&nbsp; the API changed slightly - and
this should be the final version. There is a new routine:
lame_encode_buffer() which takes an arbritray sized input buffer,
resamples &amp; filters if necessary, encodes, and returns the
mp3buffer.&nbsp; There are also several new #defines, so it is
possible to compile a simple encoding library with no decoding or file
I/O or command line parsing.&nbsp; see the file API for details.</li>
<li> Mark Taylor: MSVC stuff:&nbsp; lame.exe (with and
without the frame analyzer) and the CDex lame_enc.dll
should compile under MSVC.&nbsp; The MSVC5 project files may need
some tweaking.&nbsp; In particular,
you need to make sure LAMEPARSE, LAMESNDFILE and HAVEMPGLIB
are defined.&nbsp; (and HAVEGTK for the GTK stuff).</li>
</ul>
<h3> LAME 3.63beta&nbsp; February 20&nbsp; 2000</h3>
<ul>
<li> Robert Hegemann:&nbsp; FPE with -h fixed?</li>
<li> Mathey Hendry:&nbsp; FPE error catching for Cygwin,
FPE fix for vbr mode and output to /dev/null</li>
<li> Jeremy Hall:&nbsp; Fixed problems with input files
where the number of samples is not known.</li>
<li> <font color="#3366ff">Mathew
Hendry:&nbsp; ASM quantize_xrpow() for GNU i386</font></li>
<li> <font color="#3366ff">Wilfried
Behne&nbsp; quantize_xrpow ()for PowerPC and non-ASM</font></li>
<li> <font color="#3366ff">Takehiro
Tominaga:&nbsp; GOGO FFTs&nbsp; (not yet used?)</font></li>
</ul>
<h3> LAME 3.62beta&nbsp;&nbsp; February 9 2000</h3>
<ul>
<li> <font color="#000000">Iwasa Kazmi:&nbsp;
frame analyzer short block display of single subblocks (press
1,2&nbsp; or 3)</font></li>
<li> <font color="#000000">Ingo Saitz:&nbsp;
--help option added, with output to stdout</font></li>
<li> <font color="#ff0000">Alfred Weyers: short
block AAC spreading function bug fixed</font></li>
<li> <font color="#3366ff">Takehiro
Tominaga:&nbsp; new scalefac data structure - improves performance!</font></li>
<li> <font color="#ff0000">Lionel
Bonnet:&nbsp; Bug fixed in MPEG2 scalefactor routine: scalefactors
were being severly limited.</font></li>
<li> <font color="#3366ff">Takehiro
Tominaga:&nbsp; faster FFT routines from.&nbsp; These routines
are also compatible with the GOGO routines, in case someone is
interested in porting them back to LAME.</font></li>
<li> <font color="#3366ff">Sigbj&oslash;rn
Skj&aelig;ret, Takehiro Tominaga:&nbsp; faster pow() code.</font></li>
<li> <font color="#ff0000">Joachim
Kuebart:&nbsp; Found some unitialized variables that were effecting
quality for encodings which did not use the -h option (now fixed).</font></li>
<li> Mark Taylor: More modularization work.&nbsp;&nbsp;
It is now possible to use LAME as a library where you can set the
encoding parameters directly and do your own file
i/o.&nbsp;&nbsp; The calling program is now it's own mp3
output.&nbsp; For an example of the LAME API, see main.c, or
mp3rtp.c or mp3x.c.&nbsp; These can all be compiled as stand alone
programs which link with libmp3lame.a.</li>
<li> Felix vos Leitner:&nbsp; mp3rtp fixes.&nbsp;
mp3rtp is a standalone program which will encode and stream with RTP.</li>
<li> Robert Hegemann:&nbsp; Information written to stderr
displaying exactly which type of lowpass filter (if any) is being used.</li>
<li> Iwasa Kazmi:&nbsp; mpglib (the mpg123 decoder) scsfi
decoding fixes.</li>
<li> Takehiro Tominaga:&nbsp; More mpglib scsfi decoding
fixes.</li>
</ul>
<h3> LAME 3.61beta&nbsp; January 14 2000</h3>
<ul>
<li> <font color="#ff0000">Mark Taylor: Fixed bug
with lowpass filters when using VBR with a 64kbs or lower min bitrate
setting.</font></li>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; more efficient huffman encoding splitting.</font></li>
</ul>
<h3> LAME 3.60beta&nbsp;&nbsp; January 9 2000</h3>
<ul>
<li> Mark Taylor: Distribution now comes with self
test.&nbsp; Needs work to be automated, see 'make test' in Makefile.</li>
<li> <font color="#ff0000">Mark Taylor: AAC
spreading function now the default</font></li>
<li> Gabriel Bouvigne: updated HTML docs</li>
<li> Felix von Leitner: compute correct file length from Xing
header (if present) when input file is a mp3 file</li>
<li> Felix von Leitner: mp3rtp (standalone) program now
included.&nbsp; Not yet tested.&nbsp; mp3rtp
ip:port:ttl&nbsp; &lt;infile&gt;&nbsp;&nbsp;
/dev/null will stream directly to ip:port using&nbsp; RTP.</li>
</ul>
<h3> LAME 3.59beta&nbsp; January 4 2000</h3>
<ul>
<li> Takehiro Tominaga:&nbsp; --noath option.&nbsp;
Disables ATH maskings.</li>
<li> Gabriel Bouvigne:&nbsp; updated HTML docs.</li>
<li> Iwasa Kazmi:&nbsp; makefile fixes</li>
<li> Mark Taylor:&nbsp; Fixed bug where first frame of data
was always overwritten with 0's.&nbsp; Thanks to 'gol'</li>
<li> <font color="#ff0000">Mark Taylor:&nbsp;
bug fixes in mid/side masking ratios (thanks to Menno Bakker)</font></li>
<li> Mark Taylor:&nbsp; replaced norm_l, norm_s table data
with formulas.</li>
</ul>
<h3> LAME 3.58beta&nbsp; December 13 1999</h3>
<ul>
<li> <font color="#ff0000">Segher
Boessenkool:&nbsp; More accurate quantization procedure!&nbsp;
Enabled with -h.</font></li>
<li> <font color="#3366ff">Mathew Hendry, Acy
Stapp and Takehiro Tominaga: ASM optimizations for quantize_xrpow and
quantize_xrpow_ISO.</font></li>
<li> Chuck Zenkus:&nbsp; "encoder inside" logo on web page</li>
<li> Mark Taylor:&nbsp;&nbsp; a couple people have
asked for this.&nbsp;&nbsp; Allow LAME to overide
VBR_min_bitrate if analog_silence detected.&nbsp;&nbsp;
Analog_silence defined a la Robert:&nbsp;&nbsp;
energy&nbsp; &lt; ATH.</li>
<li> An Van Lam: Valid bitrates were being printed for layer 2,
not layer 3!</li>
<li> Ethan Yeo:&nbsp; Makefile.MSVC updated</li>
<li> Mark Stephens:&nbsp; updated all MSVC project files</li>
<li> Robert Hegemann:&nbsp; lowpass and highpass filters
can be enabled with --lowpass, --highpass</li>
<li> <font color="#ff0000">Mark Taylor:&nbsp;
MS switching is now smoother: ms_ratio average over 4 granules</font></li>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; Scalefactor pre-emphasis fixed (and now turned back
on)</font></li>
<li> <font color="#ff0000">Takehiro
Tominaga:&nbsp; Bug in M/S maskings:&nbsp; switch to turn on
stereo demasking code was buggy.</font></li>
</ul>
<h3> LAME 3.57beta&nbsp; November 22 1999</h3>
<ul>
<li> Sigbj&oslash;rn Skj&aelig;ret, patch to allow
encoding from 8bit input files when using LIBSNDFILE</li>
<li> Mark Taylor: Automatic downsampling to nearest valid
samplerate.</li>
<li> Mark Taylor: Scalefactor bands demarked on MDCT plot in
frameanalyzer</li>
<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>
</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>
</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>
</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>
</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>
</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 one channel of one granule.&nbsp; A couple of people
reported problems that might be caused by this, especially at higher
bitrates.</li>
<li> Nils Faerber updated the man page and fixed many of the
compiler warnings.</li>
</ul>
<h3> LAME 3.24beta&nbsp;&nbsp; August 15 1999</h3>
<ul>
<li> This release contains the following new code (for
developers) which is disabled by default:</li>
<li> Robert Hegemann:&nbsp; Completely overhauled VBR
code.&nbsp; Now computes exact number of bits required for the
given qualty and then quantized with the appropriate bitrate.</li>
<li> Several new quantization quality measures.</li>
</ul>
<h3> LAME 3.23beta&nbsp; August 8 1999</h3>
<ul>
<li> Very nice continuously updated VBR histogram display from
Iwasa Kazmi.&nbsp; (disabled with --nohist).</li>
<li> More modulerization work.&nbsp; The encoding engine
can now be compiled into libmp3lame, but the interface is awkward.</li>
<li> <font color="#ff0000">Bug fixed in FFT Hann
window formula (Leonid A. Kulakov).</font></li>
<li> New LAME logo on the download page.&nbsp; Created by
Chris Michalisles.</li>
<li> <font color="#ff0000">Several VBR algorithm
improvements from Robert Hegemann.&nbsp; New quantization noise
metrics and VBR quality measure takes into account mid/side
encoding.&nbsp; Should produce smaller files with the same quality,
especially when using jstereo.</font></li>
</ul>
<h3> LAME 3.22beta&nbsp; July 27 1999</h3>
<ul>
<li> Downsampling (stereo to mono) bug with MPEG2
fixed.&nbsp; (Mike Oliphant)</li>
<li> Downsampling now merges L &amp; R channels - before it
only took the L channel.</li>
<li> More modularization and code cleanup from Albert Faber and
myself.</li>
<li> Input filesize limit removed for raw pcm input
files.&nbsp; For other file types, LAME will still only read the
first 2^32 samples, (27 hours of playing time at 44.1 kHz).</li>
</ul>
<h3> LAME 3.21beta&nbsp; July 26 1999</h3>
<ul>
<li> <font color="#ff0000">Correct Mid/Side
masking thresholds for JSTEREO mode!&nbsp; This is enabled with
-h.&nbsp; It&nbsp; makes LAME about 20% slower since it
computes psycho-acoustics for L,R Mid and Side channels.</font></li>
<li> <font color="#ff0000">"Analog silence"
threshold added.&nbsp; Keeps VBR from upping the bitrate during
very quite passages.&nbsp; (Robert.Hegemann)</font></li>
<li> <font color="#ff0000">New VBR quality
setting from Robert Hegemann.&nbsp; It is based on the idea that
distortion at lower bit rates sounds worse than at higher bitrates, and
so the allowed distortion (VBR quality setting) is proportional to the
bitrate.&nbsp; Because of this, default minimum bitrate is now
32kbs.</font></li>
<li> <font color="#ff0000">Expermental subblock
gain code enabled with -Z.</font></li>
<li> New "-r" option for raw pcm input files.&nbsp; With
-r, LAME will not do any fseek()'s or look for wav and aiff headers on
the input file.</li>
<li> Bug fixes in mp3x (frame analyzer) for viewing frames near
end of the file.</li>
<li> Bug fixed to allow setting the sampling rate of raw pcm
input files.</li>
</ul>
<h3> LAME 3.20beta&nbsp; July 19 1999</h3>
<ul>
<li> Bug in get_audio.c fixed.&nbsp; Libsndfile wrappers
would not compile (Miguel Revilla&nbsp; Rodriguez)</li>
<li> Nils Faerber found some unitialized variables and some
wierd extranous computations in filter_subband, now fixed.&nbsp;
This was causing seg faults on some machines.</li>
</ul>
<h3> LAME 3.19beta&nbsp; July 18 1999</h3>
<ul>
<li> <font color="#ff0000">Oops!&nbsp; Robert
Hegemann immediatly found a bug in the new&nbsp; (old -Z option)
quantization code.&nbsp; calc_noise1 was not returning tot_noise,
so non ms-stereo frames were buggy.</font></li>
</ul>
<h3> LAME 3.18beta&nbsp; July 17 1999</h3>
<ul>
<li> <font color="#ff0000">Many psycho-acoustic
bug fixes.&nbsp; Dan Nelson discovered a bug in MPEG2: For short
blocks, the code assumes 42 partition bands.&nbsp; MPEG1 sometimes
has less, MPEG2 can have more.&nbsp; In MPEG1, this bug would not
have effected the output if your compiler initializes static variables
to 0 on creation.&nbsp; In MPEG2 it leads to array out-of-bounds
access errors. Finally, there was a related bug in MPEG1/MPEG2, short
&amp; long blocks where the energy above 16 kHz was all added to
partition band 0. (the lowest frequeny partition band!)</font></li>
<li> <font color="#ff0000">The -Z option (Gabriel
Bouvigne's idea of using total quantization noise to choose between two
quantizations with the same value of "over") is now the
default.&nbsp; I believe this helps remove the trilling sound in
Jan's testsignal4.wav.&nbsp; The quality of testsignal2.wav and
testsignal4.wav are now better than Xing and getting closer to FhG.</font></li>
<li> Bug fixes in frame &amp; sample count for downsampling
mode. (ben "jacobs")</li>
<li> Patches to improve modulization.&nbsp; (ben "jacobs")</li>
</ul>
<h3> LAME 3.17beta&nbsp; July 11 1999</h3>
<ul>
<li> substantial code cleanup towards goal of making LAME more
modular.</li>
</ul>
<h3> LAME 3.16beta&nbsp; July 11 1999</h3>
<ul>
<li> <font color="#ff0000">New tunings of window
switching, and better bit allocation based on pe.&nbsp; (Jan
Rafaj.&nbsp; improves both testsignal2.wav and testsignal4.wav).</font></li>
<li> <font color="#ff0000">Bug in mid/side
quantization when side channel was zero fixed.&nbsp; (Albert Faber)</font></li>
<li> Removed some extranous computations in l3psy.c (Robert
Hegemann)</li>
<li> More detailed timing status info, including hours display.
(Sakari Ailus) and percentage indicator (Conrad Sanderson).</li>
<li> <font color="#3366ff">Window_subband and
calc_noise1,calc_noise2 speedups.&nbsp; Quantize_xrpow speedup
should be significant on non GNU/intel systems. (Mike Cheng)</font></li>
<li> <font color="#3366ff">Better initial guess
for VBR bitrate.&nbsp; Should speed up VBR encoding.&nbsp;
(Gabriel Bouvigne)</font></li>
<li> More advanced .wav header parsing.&nbsp; fixes bugs
involving click in first frame. (Robert.Hegemann)</li>
<li> Correct filesize and total frame computation when using
LIBSNDFILE (ben "jacobs")</li>
<li> Click in last frame (buffering problem) when using
libsndfile fixed.</li>
<li> Audio I/O code overhauled.&nbsp; There is now a
uniform audio i/o interface to libsndfile or the LAME built in wav/aiff
routines.&nbsp; All audio i/o code localized to get_audio.c.</li>
</ul>
<h3> LAME 3.15beta</h3>
<ul>
<li> times()/clock() problem fixed for non-unix OS.&nbsp;
(Ben "Jacobs")</li>
<li> Fixed uninitialized pe[] when using fast mode.&nbsp;
(Ben "Jacobs")</li>
</ul>
<h3> LAME 3.13&nbsp; June 24 1999</h3>
<ul>
<li> Patches for BeOS from Gertjan van Ratingen.</li>
<li> Makefile info for OS/2 Warp 4.0&nbsp; (from dink.org).</li>
<li> Status display now based on wall clock time, not cpu time.</li>
<li> mem_alloc no longer allocates twice as much memory as
needed (Jan Peman).</li>
</ul>
<h3> 3.12pre9</h3>
<ul>
<li> Updated BLADEDLL code to handle recent changes (Albert
Faber).</li>
<li> Bug fixed in parsing options when not using GTK (Albert
Faber).</li>
<li> <font color="#ff0000">MPEG2 Layer III psycho
acoustics now working.</font></li>
<li> <font color="#3366ff">Improved huffman
encoding Chris Matrakidis. (10% faster).&nbsp; I dont know how he
finds these improvements!&nbsp; LAME with full quality now encodes
faster than real time on my PII 266.</font></li>
<li> Fixed time display when encoding takes more than 60
minutes.</li>
</ul>
<h3> 3.12pre8</h3>
<ul>
<li> <font color="#ff0000">New <a href="gpsycho/ms_stereo.html">mid/side stereo</a>
criterion.&nbsp; LAME will use mid/side stereo only when the
difference between L &amp; R masking thresholds (averaged over all
scalefactors) is less then 5db.&nbsp; In several test samples it
does a very good job mimicking the FhG encoder.</font></li>
<li> <font color="#ff0000">Bug in mid/side stereo
fixed:&nbsp; independent variation of mid &amp; side channel
scalefactors disabled.&nbsp; Because of the way outer_loop is
currently coded, when encoding mid/side coefficietns using left/right
thresholds, you have to vary the scalefactors simultaneously.</font></li>
<li> <font color="#ff0000">Bug in side/mid energy
ratio calculation fixed. (Thanks to Robert Hegemann)</font></li>
<li> Default mode is stereo (not jstereo) if bitrate is chosen
as 192kbs or higher.&nbsp; Tero Auvinen first pointed out that FhG
seems to think at 160kbs, their encoder is so good it doesn't need
jstereo tricks. Since LAME is not as good as FhG, I am going to claim
that 192kbs LAME is so good it doens't need jstereo tricks, and thus it
is disabled by default.</li>
<li> WAV header parsing for big-endian machines, and automatic
detection of big-endian machines.&nbsp; (Thanks to&nbsp;
Sigbj&oslash;rn Skj&aelig;ret).</li>
<li> added 56 sample delay to sync LAME with FhG.</li>
<li> MP3x (frame analyzer) can now handle MPEG2 streams.</li>
</ul>
<h3> 3.12pre7</h3>
<ul>
<li> MPEG2 layer III now working!&nbsp; lower bit rates
(down to 8kbs) and 3 more sampling frequencies:&nbsp; 16000, 22050,
24000Hz. Quality is poor - the psy-model does not yet work with these
sampling frequencies.</li>
<li> Fixed "ERROR: outer_loop(): huff_bits &lt; 0." bug
when using VBR.</li>
<li> bash and sh scripts to run LAME on multiple files now
included.&nbsp; (from Robert Hegemann and Gerhard Wesp respectively)</li>
<li> bug fix in encoding times for longer files from&nbsp;
(Alvaro Martinez Echevarria)</li>
<li> yet another segfault in the frame analyzer fixed.</li>
<li> ISO psy-model/bit allocation routines removed.&nbsp;
This allowed makeframe() to be made much simpler, and most of the
complicated buffering is now gone. Eventually I would like the encoding
engine to be a stand alone library.</li>
</ul>
<h3> 3.12pre6</h3>
<ul>
<li> <font color="#ff0000">Better VBR
tuning.&nbsp; Find minimum bitrate with distortion less than the
allows maximum.&nbsp; A minimum bit rate is imposed on frames with
short blocks (where the measured distortion can not be
trusted).&nbsp;&nbsp; A minimum frame bitrate can be specified
with -b, default=64kbs.</font></li>
<li> <a href="http://www.zip.com.au/%7Eerikd/libsndfile">LIBSNDFILE</a>
support.&nbsp; With libsndfile, LAME can encode almost all sound
formats.&nbsp; 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>
<ul>
<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>
</ul>
<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>
<h3> LAME 3.04 May 22 1999</h3>
<ul>
<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>
<h3>LAME 3.03 May 18 1999 </h3>
<ul>
<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>
<h3> LAME 3.02 May 12 1999</h3>
<ul>
<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>
<h3><b>LAME 3.01 May 11 1999</b> </h3>
<ul>
<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>
<hr>
<h3>LAME 3.0 May 10 1999</h3>
<ul>
<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>
<h3>November 8 1998</h3>
<ul>
<li> Version 2.1f released</li>
<li> 50% faster filter_subband() routine in encode.c
contributed by James Droppo</li>
</ul>
<h3>November 2 1998</h3>
<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>
<h3>October 31 1998</h3>
<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>
<h3>October 18 1998</h3>
<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;&lt;int&gt;' switch. Fun speedup which is
mostly harmless if you're not concerned with quality.</li>
<li> Compiling with amiga-gcc works fine</li>
</ul>
<h3>October 16 1998</h3>
<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>
<h3>October 6 1998</h3>
<ul>
<li> Version 2.1 released with all tables now incorporated into
the exe. Thanks to <b>Lars Magne Ingebrigtseni</b></li>
</ul>
<h3>October 4 1998</h3>
<ul>
<li>
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.
</li>
<li>
<a href="http://www.uq.net.au/%7Ezzmcheng/lame/download.html">
Version 2.0</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 <em>identical</em> 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.
</li>
<li>
Added section
on <a href="http://www.uq.net.au/%7Ezzmcheng/lame/quality.html">
quality</a>.
</li>
</ul>
<h3> October 1 1998</h3>
<ul>
<li>Updated web page and released LAME v1.0</li>
</ul>
<hr>
<h3>Up to September 1998</h3>
<p>Working on the 8hz source code...</p>
<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).
<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
<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 uses the first 6
phi values, and the first 465 energy values. Saves a bunch of
calculations.
</li>
<li>
Added a speed-up/quality switch. Speed is increased but
quality is decreased <i>slightly</i>. My ears are bad enough
not to be able to notice the difference in quality at low
settings :). Setting '-q 1' improves speed by about 10%. '-q
100' improves speed by about 26%. Enoding of my test track
goes from 111s (at default '-q 0') to 82s (at -q 100). Thanks
to <b>Jan Peman</b> for this tip.
</li>
</ul>
</li>
<li>
m9 - fixed an error in l3psy.c. numlines[] is overwritten with
incorrect data. Added a new variable numlines_s[] to fix
this. Thanks again to <b>Jan Peman</b>.
</li>
<li>
m10 - Down to 106 seconds by selecting a few more compiler
options. Also added a pow20() function in l3loop.c to speed up
(ever so slightly) calls to pow(2.0, x)
</li>
<li>m11
<ul>
<li>
No speedups. Just cleaned up some bits of the code.
</li>
<li>
Changed K&amp;R prototyping to 'normal' format. Thanks
to <b>Steffan Haeuser</b> for his help here.
</li>
<li>
Changed some C++ style comments to normal C comments in
huffman.c
</li>
<li>
Removed the #warning from psy_data.h (it was getting
annoying!)
</li>
<li>
Removed reference in bitstream.c to malloc.h. Is there a
system left where malloc.h hasn't been superceded by
stdlib.h?
</li>
</ul>
</li>
</ul>
<li>In Progess:
<ul>
<li>
my PSYD hack for the spreading functions is only valid for
44.1 kHz - Should really put in a "if freq = 44.1 kHz"
switch for it. Someone might want to extend the speedup for
48 and 32 kHz.
</li>
<li>
Putting in Jan Peman's quantanf_init speedup.
</li>
</ul>
</li>
</ul>
<hr>
<center>
<p>
<a href="http://validator.w3.org/check?uri=referer">
<img src="http://www.w3.org/Icons/valid-html401"
alt="Valid HTML 4.01 Transitional" height="31" width="88">
</a>
</p>
</center>
</body>
</html>