1
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
4
<meta http-equiv="Content-Type"
5
content="text/html; charset=windows-1252">
6
<meta http-equiv="Content-Language" content="en-us">
7
<meta name="author" content="Olli Parviainen">
8
<meta name="description"
9
content="Readme file for SoundTouch audio processing library">
10
<meta name="GENERATOR" content="Microsoft FrontPage 4.0">
11
<meta name="ProgId" content="FrontPage.Editor.Document">
12
<title>SoundTouch library README</title>
15
.normal { font-family: Arial }
21
<h1>SoundTouch audio processing library v1.5.0
23
<p class="normal">SoundTouch library Copyright (c) Olli
24
Parviainen 2002-2009 </p>
26
<h2>1. Introduction </h2>
27
<p>SoundTouch is an open-source audio
28
processing library that allows changing the sound tempo, pitch
29
and playback rate parameters independently from each other, i.e.:</p>
31
<li>Sound tempo can be increased or decreased while
32
maintaining the original pitch</li>
33
<li>Sound pitch can be increased or decreased while
34
maintaining the original tempo </li>
35
<li>Change playback rate that affects both tempo
36
and pitch at the same time </li>
37
<li>Choose any combination of tempo/pitch/rate</li>
39
<h3>1.1 Contact information </h3>
40
<p>Author email: oparviai 'at' iki.fi </p>
41
<p>SoundTouch WWW page: <a href="http://www.surina.net/soundtouch">http://www.surina.net/soundtouch</a></p>
43
<h2>2. Compiling SoundTouch</h2>
44
<p>Before compiling, notice that you can choose the sample data format
45
if it's desirable to use floating point sample
46
data instead of 16bit integers. See section "sample data format"
47
for more information.</p>
48
<h3>2.1. Building in Microsoft Windows</h3>
49
<p>Project files for Microsoft Visual C++ 6.0 and Visual C++ .NET are
50
supplied with the source code package. </p>
51
<p> Please notice that SoundTouch
52
library uses processor-specific optimizations for Pentium III and AMD
53
processors. Visual Studio .NET and later versions supports the required
54
instructions by default, but Visual Studio 6.0 requires a processor pack upgrade
55
to be installed in order to support these optimizations. The processor pack upgrade can be downloaded from
56
Microsoft site at this URL:</p>
57
<p><a href="http://msdn.microsoft.com/en-us/vstudio/aa718349.aspx">http://msdn.microsoft.com/en-us/vstudio/aa718349.aspx</a></p>
58
<p>If the above URL is unavailable or removed, go
59
to <a href="http://msdn.microsoft.com/">http://msdn.microsoft.com</a>
60
and perform a search with keywords "processor pack". </p>
61
<p>To build the binaries with Visual C++
62
compiler, either run "make-win.bat" script, or open the
63
appropriate project files in source code directories with Visual
64
Studio. The final executable will appear under the "SoundTouch\bin"
65
directory. If using the Visual Studio IDE instead of the make-win.bat script, directories bin and
66
lib may need to be created manually to the SoundTouch
67
package root for the final executables. The make-win.bat script
68
creates these directories automatically.
70
<h3>2.2. Building in Gnu platforms</h3>
71
<p>The SoundTouch library can be compiled in
72
practically any platform supporting GNU compiler (GCC) tools.
73
SoundTouch have been tested with gcc version 3.3.4., but it
74
shouldn't be very specific about the gcc version. Assembler-level
75
performance optimizations for GNU platform are currently available in
76
x86 platforms only, they are automatically disabled and replaced with
77
standard C routines in other processor platforms.</p>
78
<p>To build and install the binaries, run the
79
following commands in the SoundTouch/ directory:</p>
80
<table border="0" cellpadding="0" cellspacing="4">
84
<pre>./configure -</pre>
87
<p>Configures the SoundTouch package for the local
96
<p>Builds the SoundTouch library &
97
SoundStretch utility.</p>
102
<pre>make install -</pre>
105
<p>Installs the SoundTouch & BPM libraries
106
to <b>/usr/local/lib</b> and SoundStretch utility to <b>/usr/local/bin</b>.
107
Please notice that 'root' privileges may be required to install the
108
binaries to the destination locations.</p>
113
<h4><b>2.2.1 Required GNU tools</b> </h4>
114
<p> Bash shell, GNU C++ compiler, libtool, autoconf and automake tools are required
116
the SoundTouch library. These are usually included with the GNU/Linux distribution, but if
117
not, install these packages first. For example, in Ubuntu Linux these can be acquired and
118
installed with the following command:</p>
119
<pre><b>sudo apt-get install <font SIZE="2">automake autoconf libtool build-essential</font></b></pre>
120
<h4><b>2.2.2 Problems with GCC compiler compatibility</b></h4>
121
<p>At the release time the SoundTouch package has been tested to compile in
122
GNU/Linux platform. However, in past it's happened that new gcc versions aren't
123
necessarily compatible with the assembler settings used in the optimized
124
routines. <b>If you have problems getting the
125
SoundTouch library compiled, try the workaround of disabling the optimizations</b>
126
by editing the file "include/STTypes.h" and removing the following
127
definition there:</p>
129
<pre>#define ALLOW_OPTIMIZATIONS 1</pre>
131
<h4><b>2.2.3 Problems with configure script or build process</b> </h4>
132
<p>Incompatibilities between various GNU toolchain versions may cause errors when running the "configure" script or building the source
133
codes, if your GNU tool versions are not compatible with the versions used for
134
preparing the SoundTouch kit. </p>
135
<p>To resolve the issue, regenerate the configure scripts with your local tool
137
the "<b>./bootstrap</b>" script included in the SoundTouch source code
138
kit. After that, run the <b>configure</b> script and <b>make</b> as usually.</p>
139
<h4><b>2.2.4 Compiler issues with non-x86 processors</b></h4>
140
<p>SoundTouch library works also on non-x86 processors.</p>
141
<p>However, in case that you get compiler errors when trying to compile for non-Intel processor, edit the file
142
"<b>source\SoundTouch\Makefile.am</b>" and remove the "<b>-msse2</b>"
143
flag on the <b>AM_CXXFLAGS </b>line:</p>
144
<pre><b>AM_CXXFLAGS=-O3 -fcheck-new -I../../include # Note: -msse2 flag removed!</b></pre>
145
<p>After that, run "<b>./bootstrap</b>" script, and then run <b>configure</b>
146
and <b>make</b> again.</p>
148
<h2>3. About implementation & Usage tips</h2>
149
<h3>3.1. Supported sample data formats</h3>
150
<p>The sample data format can be chosen
151
between 16bit signed integer and 32bit floating point values, the
152
default is 32bit floating point. </p>
155
In Windows environment, the sample data format is chosen
156
in file "STTypes.h" by choosing one of the following
159
<li><span style="font-weight: bold;">#define INTEGER_SAMPLES</span>
162
<li><span style="font-weight: bold;">#define FLOAT_SAMPLES</span> for
163
32bit floating point</li>
166
In GNU environment, the floating sample format is used by default, but
167
integer sample format can be chosen by giving the
168
following switch to the configure script:
170
<pre>./configure --enable-integer-samples</pre>
173
<p>The sample data can have either single (mono)
174
or double (stereo) audio channel. Stereo data is interleaved so
175
that every other data value is for left channel and every second
176
for right channel. Notice that while it'd be possible in theory
177
to process stereo sound as two separate mono channels, this isn't
178
recommended because processing the channels separately would
179
result in losing the phase coherency between the channels, which
180
consequently would ruin the stereo effect.</p>
181
<p>Sample rates between 8000-48000H are
183
<h3>3.2. Processing latency</h3>
184
<p>The processing and latency constraints of
185
the SoundTouch library are:</p>
187
<li>Input/output processing latency for the
188
SoundTouch processor is around 100 ms. This is when time-stretching is
189
used. If the rate transposing effect alone is used, the latency
191
is much shorter, see section 'About algorithms'.</li>
192
<li>Processing CD-quality sound (16bit stereo
193
sound with 44100H sample rate) in real-time or faster is possible
194
starting from processors equivalent to Intel Pentium 133Mh or better,
195
if using the "quick" processing algorithm. If not using the "quick"
197
if floating point sample data are being used, several times more CPU
198
power is typically required.</li>
200
<h3>3.3. About algorithms</h3>
201
<p>SoundTouch provides three seemingly
202
independent effects: tempo, pitch and playback rate control.
203
These three controls are implemented as combination of two primary
204
effects, <em>sample rate transposing</em> and <em>time-stretching</em>.</p>
205
<p><em>Sample rate transposing</em> affects
206
both the audio stream duration and pitch. It's implemented simply
207
by converting the original audio sample stream to the desired
208
duration by interpolating from the original audio samples. In SoundTouch, linear interpolation with anti-alias filtering is
209
used. Theoretically a higher-order interpolation provide better
210
result than 1st order linear interpolation, but in audio
211
application linear interpolation together with anti-alias
212
filtering performs subjectively about as well as higher-order
214
<p><em>Time-stretching </em>means changing
215
the audio stream duration without affecting it's pitch. SoundTouch
216
uses WSOLA-like time-stretching routines that operate in the time
217
domain. Compared to sample rate transposing, time-stretching is a
218
much heavier operation and also requires a longer processing
219
"window" of sound samples used by the
220
processing algorithm, thus increasing the algorithm input/output
221
latency. Typical i/o latency for the SoundTouch
222
time-stretch algorithm is around 100 ms.</p>
223
<p>Sample rate transposing and time-stretching
224
are then used together to produce the tempo, pitch and rate
227
<li><strong>'Tempo'</strong> control is
228
implemented purely by time-stretching.</li>
229
<li><strong>'Rate</strong>' control is implemented
230
purely by sample rate transposing.</li>
231
<li><strong>'Pitch</strong>' control is
232
implemented as a combination of time-stretching and sample rate
233
transposing. For example, to increase pitch the audio stream is first
234
time-stretched to longer duration (without affecting pitch) and then
235
transposed back to original duration by sample rate transposing, which
236
simultaneously reduces duration and increases pitch. The result is
237
original duration but increased pitch.</li>
239
<h3>3.4 Tuning the algorithm parameters</h3>
240
<p>The time-stretch algorithm has few
241
parameters that can be tuned to optimize sound quality for
242
certain application. The current default parameters have been
243
chosen by iterative if-then analysis (read: "trial and error")
244
to obtain best subjective sound quality in pop/rock music
245
processing, but in applications processing different kind of
246
sound the default parameter set may result into a sub-optimal
248
<p>The time-stretch algorithm default
249
parameter values are set by the following #defines in file "TDStretch.h":</p>
251
<pre>#define DEFAULT_SEQUENCE_MS AUTOMATIC
252
#define DEFAULT_SEEKWINDOW_MS AUTOMATIC
253
#define DEFAULT_OVERLAP_MS 8</pre>
255
<p>These parameters affect to the time-stretch
256
algorithm as follows:</p>
258
<li><strong>DEFAULT_SEQUENCE_MS</strong>: This is
259
the default length of a single processing sequence in milliseconds
260
which determines the how the original sound is chopped in
261
the time-stretch algorithm. Larger values mean fewer sequences
262
are used in processing. In principle a larger value sounds better when
263
slowing down the tempo, but worse when increasing the tempo and vice
266
By default, this setting value is calculated automatically according to
269
<li><strong>DEFAULT_SEEKWINDOW_MS</strong>: The seeking window
270
default length in milliseconds is for the algorithm that seeks the best
271
possible overlapping location. This determines from how
272
wide a sample "window" the algorithm can use to find an optimal mixing
273
location when the sound sequences are to be linked back together. <br>
275
The bigger this window setting is, the higher the possibility to find a
276
better mixing position becomes, but at the same time large values may
277
cause a "drifting" sound artifact because neighboring sequences can be
278
chosen at more uneven intervals. If there's a disturbing artifact that
279
sounds as if a constant frequency was drifting around, try reducing
282
By default, this setting value is calculated automatically according to
285
<li><strong>DEFAULT_OVERLAP_MS</strong>: Overlap
286
length in milliseconds. When the sound sequences are mixed back
287
together to form again a continuous sound stream, this parameter
288
defines how much the ends of the consecutive sequences will overlap with each other.<br>
290
This shouldn't be that critical parameter. If you reduce the
291
DEFAULT_SEQUENCE_MS setting by a large amount, you might wish to try a
292
smaller value on this.</li>
294
<p>Notice that these parameters can also be
295
set during execution time with functions "<strong>TDStretch::setParameters()</strong>"
296
and "<strong>SoundTouch::setSetting()</strong>".</p>
297
<p>The table below summaries how the
298
parameters can be adjusted for different applications:</p>
302
<td valign="top"><strong>Parameter name</strong></td>
303
<td valign="top"><strong>Default value
304
magnitude</strong></td>
305
<td valign="top"><strong>Larger value
306
affects...</strong></td>
307
<td valign="top"><strong>Smaller value
308
affects...</strong></td>
309
<td valign="top"><strong>Effect to CPU burden</strong></td>
313
<pre>SEQUENCE_MS</pre>
315
<td valign="top">Default value is relatively
316
large, chosen for slowing down music tempo</td>
317
<td valign="top">Larger value is usually
318
better for slowing down tempo. Growing the value decelerates the
319
"echoing" artifact when slowing down the tempo.</td>
320
<td valign="top">Smaller value might be better
321
for speeding up tempo. Reducing the value accelerates the "echoing"
322
artifact when slowing down the tempo </td>
323
<td valign="top">Increasing the parameter
324
value reduces computation burden</td>
328
<pre>SEEKWINDOW_MS</pre>
330
<td valign="top">Default value is relatively
331
large, chosen for slowing down music tempo</td>
332
<td valign="top">Larger value eases finding a
333
good mixing position, but may cause a "drifting" artifact</td>
334
<td valign="top">Smaller reduce possibility to
335
find a good mixing position, but reduce the "drifting" artifact.</td>
336
<td valign="top">Increasing the parameter
337
value increases computation burden</td>
341
<pre>OVERLAP_MS</pre>
343
<td valign="top">Default value is relatively
344
large, chosen to suit with above parameters.</td>
345
<td valign="top"> </td>
346
<td valign="top">If you reduce the "sequence
347
ms" setting, you might wish to try a smaller value.</td>
348
<td valign="top">Increasing the parameter
349
value increases computation burden</td>
353
<h3>3.5 Performance Optimizations </h3>
354
<p><strong>General optimizations:</strong></p>
355
<p>The time-stretch routine has a 'quick' mode
356
that substantially speeds up the algorithm but may degrade the
357
sound quality by a small amount. This mode is activated by
358
calling SoundTouch::setSetting() function with parameter id
359
of SETTING_USE_QUICKSEEK and value "1", i.e. </p>
361
<p>setSetting(SETTING_USE_QUICKSEEK, 1);</p>
363
<p><strong>CPU-specific optimizations:</strong></p>
365
<li>Intel MMX optimized routines are used with
366
compatible CPUs when 16bit integer sample type is used. MMX optimizations are available both in Win32 and Gnu/x86 platforms.
367
Compatible processors are Intel PentiumMMX and later; AMD K6-2, Athlon
369
<li>Intel SSE optimized routines are used with
370
compatible CPUs when floating point sample type is used. SSE optimizations are currently implemented for Win32 platform only.
371
Processors compatible with SSE extension are Intel processors starting
372
from Pentium-III, and AMD processors starting from Athlon XP. </li>
373
<li>AMD 3DNow! optimized routines are used with
374
compatible CPUs when floating point sample type is used, but SSE
375
extension isn't supported . 3DNow! optimizations are currently
376
implemented for Win32 platform only. These optimizations are used in
377
AMD K6-2 and Athlon (classic) CPU's; better performing SSE routines are
378
used with AMD processor starting from Athlon XP. </li>
381
<h2><a name="SoundStretch"></a>4. SoundStretch audio processing utility
383
<p>SoundStretch audio processing utility<br>
384
Copyright (c) Olli Parviainen 2002-2009</p>
385
<p>SoundStretch is a simple command-line
386
application that can change tempo, pitch and playback rates of
387
WAV sound files. This program is intended primarily to
388
demonstrate how the "SoundTouch" library can be used to
389
process sound in your own program, but it can as well be used for
390
processing sound files.</p>
391
<h3>4.1. SoundStretch Usage Instructions</h3>
392
<p>SoundStretch Usage syntax:</p>
394
<pre>soundstretch infilename outfilename [switches]</pre>
397
<table border="0" cellpadding="2" width="100%">
401
<pre>"infilename"</pre>
403
<td valign="top">Name of the input sound
404
data file (in .WAV audio file format). Give "stdin" as filename to use
405
standard input pipe. </td>
409
<pre>"outfilename"</pre>
411
<td valign="top">Name of the output sound
412
file where the resulting sound is saved (in .WAV audio file format).
413
This parameter may be omitted if you don't want to save the
415
(e.g. when only calculating BPM rate with '-bpm' switch). Give "stdout"
416
as filename to use standard output pipe.</td>
420
<pre> [switches]</pre>
422
<td valign="top">Are one or more control
427
<p>Available control switches are:</p>
428
<table border="0" cellpadding="2" width="100%">
434
<td valign="top">Change the sound tempo by n
435
percents (n = -95.0 .. +5000.0 %) </td>
441
<td valign="top">Change the sound pitch by n
442
semitones (n = -60.0 .. + 60.0 semitones) </td>
448
<td valign="top">Change the sound playback rate by
449
n percents (n = -95.0 .. +5000.0 %) </td>
455
<td valign="top">Detect the Beats-Per-Minute (BPM) rate of the sound and adjust the tempo to meet 'n'
456
BPMs. When this switch is
457
applied, the "-tempo" switch is ignored. If "=n" is
458
omitted, i.e. switch "-bpm" is used alone, then the BPM rate is
459
estimated and displayed, but tempo not adjusted according to the BPM
466
<td valign="top">Use quicker tempo change
467
algorithm. Gains speed but loses sound quality. </td>
473
<td valign="top">Don't use anti-alias
474
filtering in sample rate transposing. Gains speed but loses sound
481
<td valign="top">Displays the program license
488
<li>To use standard input/output pipes for processing, give "stdin"
489
and "stdout" as input/output filenames correspondingly. The
490
standard input/output pipes will still carry the audio data in .wav audio
492
<li>The numerical switches allow both integer (e.g. "-tempo=123") and decimal (e.g.
493
"-tempo=123.45") numbers.</li>
494
<li>The "-naa" and/or "-quick" switches can be
495
used to reduce CPU usage while compromising some sound quality </li>
496
<li>The BPM detection algorithm works by detecting
497
repeating bass or drum patterns at low frequencies of <250Hz. A
498
lower-than-expected BPM figure may be reported for music with uneven or
499
complex bass patterns. </li>
501
<h3>4.2. SoundStretch usage examples </h3>
502
<p><strong>Example 1</strong></p>
503
<p>The following command increases tempo of
504
the sound file "originalfile.wav" by 12.5% and stores result to file "destinationfile.wav":</p>
506
<pre>soundstretch originalfile.wav destinationfile.wav -tempo=12.5</pre>
508
<p><strong>Example 2</strong></p>
509
<p>The following command decreases the sound
510
pitch (key) of the sound file "orig.wav" by two
511
semitones and stores the result to file "dest.wav":</p>
513
<pre>soundstretch orig.wav dest.wav -pitch=-2</pre>
515
<p><strong>Example 3</strong></p>
516
<p>The following command processes the file "orig.wav" by decreasing the sound tempo by 25.3% and
517
increasing the sound pitch (key) by 1.5 semitones. Resulting .wav audio data is
518
directed to standard output pipe:</p>
520
<pre>soundstretch orig.wav stdout -tempo=-25.3 -pitch=1.5</pre>
522
<p><strong>Example 4</strong></p>
523
<p>The following command detects the BPM rate
524
of the file "orig.wav" and adjusts the tempo to match
525
100 beats per minute. Result is stored to file "dest.wav":</p>
527
<pre>soundstretch orig.wav dest.wav -bpm=100</pre>
529
<p><strong>Example 5</strong></p>
530
<p>The following command reads .wav sound data from standard input pipe and
531
estimates the BPM rate:</p>
533
<pre>soundstretch stdin -bpm</pre>
536
<h2>5. Change History</h2>
537
<h3>5.1. SoundTouch library Change History </h3>
539
<p><strong>1.5.0:</strong></p>
541
<li>Added normalization to correlation calculation and improvement automatic seek/sequence parameter calculation to improve sound quality</li>
545
<li>Fixed negative array indexing in quick seek algorithm</li>
546
<li>FIR autoalias filter running too far in processing buffer</li>
547
<li>Check against zero sample count in rate transposing</li>
548
<li>Fix for x86-64 support: Removed pop/push instructions from the cpu detection algorithm. </li>
549
<li>Check against empty buffers in FIFOSampleBuffer</li>
550
<li>Other minor fixes & code cleanup</li>
554
<li>Fixes in compilation scripts for non-Intel platforms</li>
555
<li>Added Dynamic-Link-Library (DLL) version of SoundTouch library build,
556
provided with Delphi/Pascal wrapper for calling the dll routines</li>
557
<li>Added #define PREVENT_CLICK_AT_RATE_CROSSOVER that prevents a click artifact
558
when crossing the nominal pitch from either positive to negative side or vice
563
<p><strong>1.4.1:</strong></p>
565
<li>Fixed a buffer overflow bug in BPM detect algorithm routines if processing
566
more than 2048 samples at one call </li>
570
<p><strong>1.4.0:</strong></p>
572
<li>Improved sound quality by automatic calculation of time stretch algorithm
573
processing parameters according to tempo setting</li>
574
<li>Moved BPM detection routines from SoundStretch application into SoundTouch
576
<li>Bugfixes: Usage of uninitialied variables, GNU build scripts, compiler errors
577
due to 'const' keyword mismatch.</li>
578
<li>Source code cleanup</li>
585
<li>Changed static class declaration to GCC 4.x compiler compatible syntax.</li>
586
<li>Enabled MMX/SSE-optimized routines also for GCC compilers. Earlier
587
the MMX/SSE-optimized routines were written in compiler-specific inline
588
assembler, now these routines are migrated to use compiler intrinsic
589
syntax which allows compiling the same MMX/SSE-optimized source code with
590
both Visual C++ and GCC compilers. </li>
591
<li>Set floating point as the default sample format and added switch to
592
the GNU configure script for selecting the other sample format.</li>
599
<li>Fixed tempo routine output duration inaccuracy due to rounding
601
<li>Implemented separate processing routines for integer and
602
floating arithmetic to allow improvements to floating point routines
603
(earlier used algorithms mostly optimized for integer arithmetic also
604
for floating point samples) </li>
605
<li>Fixed a bug that distorts sound if sample rate changes during the
607
<li>Fixed a memory leak that appeared in MMX/SSE/3DNow! optimized
609
<li>Reduced redundant code pieces in MMX/SSE/3DNow! optimized
610
routines vs. the standard C routines.</li>
611
<li>MMX routine incompatibility with new gcc compiler versions </li>
612
<li>Other miscellaneous bug fixes </li>
614
<p><strong>v1.2.1: </strong></p>
616
<li>Added automake/autoconf scripts for GNU
617
platforms (in courtesy of David Durham)</li>
618
<li>Fixed SCALE overflow bug in rate transposer
620
<li>Fixed 64bit address space bugs.</li>
621
<li>Created a 'soundtouch' namespace for
622
SAMPLETYPE definitions.</li>
624
<p><strong>v1.2.0: </strong></p>
626
<li>Added support for 32bit floating point sample
627
data type with SSE/3DNow! optimizations for Win32 platform (SSE/3DNow! optimizations currently not supported in GCC environment)</li>
628
<li>Replaced 'make-gcc' script for GNU environment
629
by master Makefile</li>
630
<li>Added time-stretch routine configurability to
631
SoundTouch main class</li>
634
<p><strong>v1.1.1: </strong></p>
636
<li>Moved SoundTouch under lesser GPL license (LGPL). This allows using SoundTouch library in programs that aren't
637
released under GPL license. </li>
638
<li>Changed MMX routine organiation so that MMX optimized routines are now implemented in classes that are derived from
639
the basic classes having the standard non-mmx routines. </li>
640
<li>MMX routines to support gcc version 3. </li>
641
<li>Replaced windows makefiles by script using the .dsw files </li>
643
<p><strong>v1.01: </strong></p>
645
<li>"mmx_gcc.cpp": Added "using namespace std" and
646
removed "return 0" from a function with void return value to fix
647
compiler errors when compiling the library in Solaris environment. </li>
648
<li>Moved file "FIFOSampleBuffer.h" to "include"
649
directory to allow accessing the FIFOSampleBuffer class from external
652
<p><strong>v1.0: </strong></p>
654
<li>Initial release </li>
657
<h3>5.2. SoundStretch application Change
660
<p><strong>1.4.0:</strong></p>
662
<li>Moved BPM detection routines from SoundStretch application into SoundTouch
664
<li>Allow using standard input/output pipes as audio processing input/output
669
<p><strong>v1.3.0:</strong></p>
671
<li>Simplified accessing WAV files with floating
675
<p><strong>v1.2.1: </strong></p>
677
<li>Fixed 64bit address space bugs.</li>
679
<p><strong>v1.2.0: </strong></p>
681
<li>Added support for 32bit floating point sample
683
<li>Restructured the BPM routines into separate
685
<li>Fixed big-endian conversion bugs in WAV file
686
routines (hopefully :)</li>
688
<p><strong>v1.1.1: </strong></p>
690
<li>Fixed bugs in WAV file reading & added
691
byte-order conversion for big-endian processors. </li>
692
<li>Moved SoundStretch source code under 'example'
693
directory to highlight difference from SoundTouch stuff. </li>
694
<li>Replaced windows makefiles by script using the .dsw files </li>
695
<li>Output file name isn't required if output
696
isn't desired (e.g. if using the switch '-bpm' in plain format only) </li>
698
<p><strong>v1.1:</strong></p>
700
<li>Fixed "Release" settings in Microsoft Visual
701
C++ project file (.dsp) </li>
702
<li>Added beats-per-minute (BPM) detection routine
703
and command-line switch "-bpm" </li>
705
<p><strong>v1.01: </strong></p>
707
<li>Initial release </li>
710
<h2 >6. Acknowledgements </h2>
711
<p >Kudos for these people who have contributed to development or submitted
713
SoundTouch v1.3.1: </p>
717
<li>Stanislav Brabec</li>
718
<li>Christian Budde</li>
719
<li>Brian Cameron</li>
720
<li>Jason Champion</li>
721
<li>Patrick Colis</li>
722
<li>Justin Frankel</li>
723
<li>Jason Garland</li>
724
<li>Takashi Iwai</li>
725
<li>Paulo Pizarro</li>
729
<p >Moral greetings to all other contributors and users also!</p>
731
<h2 >7. LICENSE </h2>
732
<p>SoundTouch audio processing library<br>
733
Copyright (c) Olli Parviainen</p>
734
<p>This library is free software; you can
735
redistribute it and/or modify it under the terms of the GNU
736
Lesser General Public License version 2.1 as published by the Free Software
738
<p>This library is distributed in the hope
739
that it will be useful, but WITHOUT ANY WARRANTY; without even
740
the implied warranty of MERCHANTABILITY or FITNESS FOR A
741
PARTICULAR PURPOSE. See the GNU Lesser General Public License for
743
<p>You should have received a copy of the GNU
744
Lesser General Public License along with this library; if not,
745
write to the Free Software Foundation, Inc., 59 Temple Place,
746
Suite 330, Boston, MA 02111-1307 USA</p>
749
$Id: README.html 81 2009-12-28 20:51:18Z oparviai $