Mercurial > hg > sv-dependency-builds

annotate src/libsamplerate-0.1.8/doc/quality.html @ 168:ceec0dd9ec9c

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Replace these with versions built using an older toolset (so as to avoid ABI compatibilities when linking on Ubuntu 14.04 for packaging purposes)
author Chris Cannam <cannam@all-day-breakfast.com>
date Fri, 07 Feb 2020 11:51:13 +0000
parents 545efbb81310
children
rev   line source
cannam@85 1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
cannam@85 2 <HTML>
cannam@85 3
cannam@85 4 <HEAD>
cannam@85 5 <TITLE>
cannam@85 6 Secret Rabbit Code (aka libsamplerate)
cannam@85 7 </TITLE>
cannam@85 8 <META NAME="Author" CONTENT="Erik de Castro Lopo (erikd AT mega-nerd DOT com)">
cannam@85 9 <META NAME="Version" CONTENT="libsamplerate-0.1.8">
cannam@85 10 <META NAME="Description" CONTENT="The Secret Rabbit Code Home Page">
cannam@85 11 <META NAME="Keywords" CONTENT="libsamplerate sound resample audio dsp Linux">
cannam@85 12 <LINK REL=StyleSheet HREF="SRC.css" TYPE="text/css" MEDIA="all">
cannam@85 13 </HEAD>
cannam@85 14
cannam@85 15 <BODY TEXT="#FFFFFF" BGCOLOR="#000000" LINK="#FB1465" VLINK="#FB1465" ALINK="#FB1465">
cannam@85 16 <!-- pepper -->
cannam@85 17 <CENTER>
cannam@85 18 <IMG SRC="SRC.png" HEIGHT=100 WIDTH=760 ALT="SRC.png">
cannam@85 19 </CENTER>
cannam@85 20 <!-- pepper -->
cannam@85 21 <BR>
cannam@85 22 <!-- pepper -->
cannam@85 23 <TABLE ALIGN="center" WIDTH="98%">
cannam@85 24 <TR>
cannam@85 25 <TD VALIGN="top">
cannam@85 26 <BR>
cannam@85 27 <DIV CLASS="nav">
cannam@85 28 <BR>
cannam@85 29 <A HREF="index.html">Home</A><BR>
cannam@85 30 <A HREF="license.html">License</A><BR>
cannam@85 31 <A HREF="history.html">History</A><BR>
cannam@85 32 <A HREF="download.html">Download</A><BR>
cannam@85 33 <A HREF="quality.html">Quality</A><BR>
cannam@85 34 <A HREF="api.html">API</A><BR>
cannam@85 35 <A HREF="bugs.html">Bug Reporting</A><BR>
cannam@85 36 <A HREF="win32.html">On Win32</A><BR>
cannam@85 37 <A HREF="faq.html">FAQ</A><BR>
cannam@85 38 <A HREF="lists.html">Mailing Lists</A><BR>
cannam@85 39 <A HREF="ChangeLog">ChangeLog</A><BR>
cannam@85 40 <BR>
cannam@85 41 <DIV CLASS="block">
cannam@85 42 Author :<BR>Erik de Castro Lopo
cannam@85 43 <!-- pepper -->
cannam@85 44 <BR><BR>
cannam@85 45 <!-- pepper -->
cannam@85 46
cannam@85 47 </DIV>
cannam@85 48 <IMG SRC=
cannam@85 49 "/cgi-bin/Count.cgi?ft=6|frgb=55;55;55|tr=0|md=6|dd=B|st=1|sh=1|df=src_qual.dat"
cannam@85 50 HEIGHT=30 WIDTH=100 ALT="counter.gif">
cannam@85 51 </DIV>
cannam@85 52
cannam@85 53 </TD>
cannam@85 54 <!-- pepper -->
cannam@85 55 <!-- ######################################################################## -->
cannam@85 56 <!-- pepper -->
cannam@85 57 <TD VALIGN="top">
cannam@85 58 <DIV CLASS="block">
cannam@85 59
cannam@85 60 <H1><B>SRC Quality</B></H1>
cannam@85 61 <CENTER><P>
cannam@85 62 <B>This document not yet complete.</B><BR>
cannam@85 63 </P></CENTER>
cannam@85 64 <P>
cannam@85 65 When measuring the performance of a Sample Rate Converter, there are three
cannam@85 66 factors to consider:
cannam@85 67 </P>
cannam@85 68 <UL>
cannam@85 69 <LI><B>Signal-to-Noise Ratio</B> - a measure of how much noise the sample
cannam@85 70 rate conversion process adds to the signal.
cannam@85 71 This is measured in decibels (dB) and the higher this value the
cannam@85 72 better.
cannam@85 73 For most sample rate converters, the SNR will vary depending on
cannam@85 74 the input signal and the ratio between input and output sample
cannam@85 75 rates.
cannam@85 76 The only valid comparison of SNR is between the worst case for
cannam@85 77 for each converter.
cannam@85 78 <LI><B>Bandwidth</B> - most sample rate converters attenuate high
cannam@85 79 frequencies as part of their operation.
cannam@85 80 Bandwidth can be measured by finding the frequency where the
cannam@85 81 attenuation is 3dB and expressing that as a percentage of the full
cannam@85 82 bandwidth at that sampling rate.
cannam@85 83 <LI><B>Speed</B> - the faster the better <B>:-)</B>.
cannam@85 84 </UL>
cannam@85 85
cannam@85 86 <P>
cannam@85 87 There are a number of sample rate converters available for downloading
cannam@85 88 but I will limit the comparison ot Secret Rabbit Code to the following:
cannam@85 89 </P>
cannam@85 90 <UL>
cannam@85 91 <LI><A HREF="http://www.mega-nerd.com/SRC/download.html">sndfile-resample</A>
cannam@85 92 which is a program (which uses libsamplerate) from the <B>examples/</B>
cannam@85 93 directory of the Secret Rabbit Code source code distribution.
cannam@85 94 <LI><A HREF="http://www-ccrma.stanford.edu/~jos/resample/Available_Software.html">
cannam@85 95 Resample</A>
cannam@85 96 by Julius O Smiths which seems to have been the first high quality converter
cannam@85 97 available as source code.
cannam@85 98 <LI><A HREF="http://www.tsp.ece.mcgill.ca/MMSP/Documents/Software/AFsp/ResampAudio.html">ResampAudio</A>
cannam@85 99 which is part of
cannam@85 100 <A HREF="http://www.tsp.ece.mcgill.ca/MMSP/Documents/Software/AFsp/AFsp.html">
cannam@85 101 Audio File Programs and Routines</A>
cannam@85 102 by Peter Kabal.
cannam@85 103 <LI><A HREF="http://home.sprynet.com/~cbagwell/sox.html">SoX</A> which is maintained
cannam@85 104 by Chris Bagwell.
cannam@85 105 SoX is also able to perform some low quality sample rate conversions but these
cannam@85 106 will not be investigated.
cannam@85 107 <LI><A HREF="http://shibatch.sourceforge.net/">Shibatch</A> which seems to be a
cannam@85 108 frequency domain sample rate converter.
cannam@85 109 Unfortunately, this converter does not handle arbitrary conversion ratios and
cannam@85 110 hence could not be properly compared to the other converters.
cannam@85 111 <LI><A HREF="http://sr-convert.sourceforge.net/">sr-convert</A> is another
cannam@85 112 converter which does not handle arbitrary conversion ratios.
cannam@85 113 </UL>
cannam@85 114
cannam@85 115 <P>
cannam@85 116 It should be noted that the first three converters above are based on the algorithm
cannam@85 117 by <A HREF="http://www-ccrma.stanford.edu/~jos/resample/"> Julius O. Smith</A>
cannam@85 118 which emulates the conversion of the digital signal to an analogue one and then
cannam@85 119 sampling the analogue signal at the new sample rate.
cannam@85 120 </P>
cannam@85 121
cannam@85 122 <!--+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-->
cannam@85 123
cannam@85 124 <H3><B>Methodology</B></H3>
cannam@85 125 <P>
cannam@85 126 Measuring the SNR of a converter is relatively straight forward.
cannam@85 127 Generate an input signal consisting of a windowed sine wave, sample rate
cannam@85 128 convert it and measure the signal-to-noise ratio of the output signal.
cannam@85 129 A typical length for the original file is 30000 samples.
cannam@85 130 </P>
cannam@85 131 <P>
cannam@85 132 The bandwidth of a sample rate converter is a little more difficult to measure.
cannam@85 133 Currently this is done by generating two short files containing a windowed
cannam@85 134 sine wave.
cannam@85 135 The frequencies of the sine waves are 0.35 and 0.495 of the sample rate.
cannam@85 136 These file are then upsampled by a factor of 2 using the converter under test.
cannam@85 137 If the attenutaion of the lower frquency is less than 3dB and higher frequency is
cannam@85 138 more than 3dB, it is then possible to iteratively increase the lower frequency
cannam@85 139 and decrease the upper frequency keeping the -3dB point bracketed.
cannam@85 140 When the distance between the upper and lower frequency is sufficiently small,
cannam@85 141 it is possible to obtain a very accurate estimate of the -3dB frequency.
cannam@85 142 </P>
cannam@85 143 <P>
cannam@85 144 The speed of a sample rate converter is easy to measure; simply perform a
cannam@85 145 conversion on a large file or a number of smaller files and time the conversion
cannam@85 146 process.
cannam@85 147 </P>
cannam@85 148
cannam@85 149 <P>
cannam@85 150 The above measurement techniques are built into a test program which is delivered
cannam@85 151 with the Secret Rabbit Code source code distibution.
cannam@85 152 This program is able to test the first four of the above converters.
cannam@85 153 </P>
cannam@85 154 <!--+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-->
cannam@85 155
cannam@85 156 <H3><B>SoX</B></H3>
cannam@85 157 <P>
cannam@85 158 SoX provides three methods of resampling; a linear interpolator, a polyphase
cannam@85 159 resampler and the Julius O. Smith simulated analogue filter method.
cannam@85 160 </P>
cannam@85 161
cannam@85 162 <H3><B>Shibatch</B></H3>
cannam@85 163 <P>
cannam@85 164 Shibach
cannam@85 165 </P>
cannam@85 166
cannam@85 167 <P>
cannam@85 168 <B>More Coming Soon.</B>
cannam@85 169 </P>
cannam@85 170
cannam@85 171 <!-- <A HREF="mailto:aldel@mega-nerd.com">For the spam bots</A> -->
cannam@85 172
cannam@85 173 </DIV>
cannam@85 174 </TD></TR>
cannam@85 175 </TABLE>
cannam@85 176
cannam@85 177 </BODY>
cannam@85 178 </HTML>
cannam@85 179