Mercurial > hg > sv-dependency-builds

annotate src/libvorbis-1.3.3/doc/vorbis-fidelity.html @ 88:fe7c3a0b0259

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add some MinGW builds
author Chris Cannam <cannam@all-day-breakfast.com>
date Wed, 20 Mar 2013 13:49:36 +0000
parents 98c1576536ae
children
rev   line source
cannam@86 1 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
cannam@86 2 <html>
cannam@86 3 <head>
cannam@86 4
cannam@86 5 <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-15"/>
cannam@86 6 <title>Ogg Vorbis Documentation</title>
cannam@86 7
cannam@86 8 <style type="text/css">
cannam@86 9 body {
cannam@86 10 margin: 0 18px 0 18px;
cannam@86 11 padding-bottom: 30px;
cannam@86 12 font-family: Verdana, Arial, Helvetica, sans-serif;
cannam@86 13 color: #333333;
cannam@86 14 font-size: .8em;
cannam@86 15 }
cannam@86 16
cannam@86 17 a {
cannam@86 18 color: #3366cc;
cannam@86 19 }
cannam@86 20
cannam@86 21 img {
cannam@86 22 border: 0;
cannam@86 23 }
cannam@86 24
cannam@86 25 #xiphlogo {
cannam@86 26 margin: 30px 0 16px 0;
cannam@86 27 }
cannam@86 28
cannam@86 29 #content p {
cannam@86 30 line-height: 1.4;
cannam@86 31 }
cannam@86 32
cannam@86 33 h1, h1 a, h2, h2 a, h3, h3 a {
cannam@86 34 font-weight: bold;
cannam@86 35 color: #ff9900;
cannam@86 36 margin: 1.3em 0 8px 0;
cannam@86 37 }
cannam@86 38
cannam@86 39 h1 {
cannam@86 40 font-size: 1.3em;
cannam@86 41 }
cannam@86 42
cannam@86 43 h2 {
cannam@86 44 font-size: 1.2em;
cannam@86 45 }
cannam@86 46
cannam@86 47 h3 {
cannam@86 48 font-size: 1.1em;
cannam@86 49 }
cannam@86 50
cannam@86 51 li {
cannam@86 52 line-height: 1.4;
cannam@86 53 }
cannam@86 54
cannam@86 55 #copyright {
cannam@86 56 margin-top: 30px;
cannam@86 57 line-height: 1.5em;
cannam@86 58 text-align: center;
cannam@86 59 font-size: .8em;
cannam@86 60 color: #888888;
cannam@86 61 clear: both;
cannam@86 62 }
cannam@86 63 </style>
cannam@86 64
cannam@86 65 </head>
cannam@86 66
cannam@86 67 <body>
cannam@86 68
cannam@86 69 <div id="xiphlogo">
cannam@86 70 <a href="http://www.xiph.org/"><img src="fish_xiph_org.png" alt="Fish Logo and Xiph.Org"/></a>
cannam@86 71 </div>
cannam@86 72
cannam@86 73 <h1>Ogg Vorbis: Fidelity measurement and terminology discussion</h1>
cannam@86 74
cannam@86 75 <p>Terminology discussed in this document is based on common terminology
cannam@86 76 associated with contemporary codecs such as MPEG I audio layer 3
cannam@86 77 (mp3). However, some differences in terminology are useful in the
cannam@86 78 context of Vorbis as Vorbis functions somewhat differently than most
cannam@86 79 current formats. For clarity, then, we describe a common terminology
cannam@86 80 for discussion of Vorbis's and other formats' audio quality.</p>
cannam@86 81
cannam@86 82 <h2>Subjective and Objective</h2>
cannam@86 83
cannam@86 84 <p><em>Objective</em> fidelity is a measure, based on a computable,
cannam@86 85 mechanical metric, of how carefully an output matches an input. For
cannam@86 86 example, a stereo amplifier may claim to introduce less that .01%
cannam@86 87 total harmonic distortion when amplifying an input signal; this claim
cannam@86 88 is easy to verify given proper equipment, and any number of testers are
cannam@86 89 likely to arrive at the same, exact results. One need not listen to
cannam@86 90 the equipment to make this measurement.</p>
cannam@86 91
cannam@86 92 <p>However, given two amplifiers with identical, verifiable objective
cannam@86 93 specifications, listeners may strongly prefer the sound quality of one
cannam@86 94 over the other. This is actually the case in the decades old debate
cannam@86 95 [some would say jihad] among audiophiles involving vacuum tube versus
cannam@86 96 solid state amplifiers. There are people who can tell the difference,
cannam@86 97 and strongly prefer one over the other despite seemingly identical,
cannam@86 98 measurable quality. This preference is <em>subjective</em> and
cannam@86 99 difficult to measure but nonetheless real.</p>
cannam@86 100
cannam@86 101 <p>Individual elements of subjective differences often can be qualified,
cannam@86 102 but overall subjective quality generally is not measurable. Different
cannam@86 103 observers are likely to disagree on the exact results of a subjective
cannam@86 104 test as each observer's perspective differs. When measuring
cannam@86 105 subjective qualities, the best one can hope for is average, empirical
cannam@86 106 results that show statistical significance across a group.</p>
cannam@86 107
cannam@86 108 <p>Perceptual codecs are most concerned with subjective, not objective,
cannam@86 109 quality. This is why evaluating a perceptual codec via distortion
cannam@86 110 measures and sonograms alone is useless; these objective measures may
cannam@86 111 provide insight into the quality or functioning of a codec, but cannot
cannam@86 112 answer the much squishier subjective question, "Does it sound
cannam@86 113 good?". The tube amplifier example is perhaps not the best as very few
cannam@86 114 people can hear, or care to hear, the minute differences between tubes
cannam@86 115 and transistors, whereas the subjective differences in perceptual
cannam@86 116 codecs tend to be quite large even when objective differences are
cannam@86 117 not.</p>
cannam@86 118
cannam@86 119 <h2>Fidelity, Artifacts and Differences</h2>
cannam@86 120
cannam@86 121 <p>Audio <em>artifacts</em> and loss of fidelity or more simply
cannam@86 122 put, audio <em>differences</em> are not the same thing.</p>
cannam@86 123
cannam@86 124 <p>A loss of fidelity implies differences between the perceived input and
cannam@86 125 output signal; it does not necessarily imply that the differences in
cannam@86 126 output are displeasing or that the output sounds poor (although this
cannam@86 127 is often the case). Tube amplifiers are <em>not</em> higher fidelity
cannam@86 128 than modern solid state and digital systems. They simply produce a
cannam@86 129 form of distortion and coloring that is either unnoticeable or actually
cannam@86 130 pleasing to many ears.</p>
cannam@86 131
cannam@86 132 <p>As compared to an original signal using hard metrics, all perceptual
cannam@86 133 codecs [ASPEC, ATRAC, MP3, WMA, AAC, TwinVQ, AC3 and Vorbis included]
cannam@86 134 lose objective fidelity in order to reduce bitrate. This is fact. The
cannam@86 135 idea is to lose fidelity in ways that cannot be perceived. However,
cannam@86 136 most current streaming applications demand bitrates lower than what
cannam@86 137 can be achieved by sacrificing only objective fidelity; this is also
cannam@86 138 fact, despite whatever various company press releases might claim.
cannam@86 139 Subjective fidelity eventually must suffer in one way or another.</p>
cannam@86 140
cannam@86 141 <p>The goal is to choose the best possible tradeoff such that the
cannam@86 142 fidelity loss is graceful and not obviously noticeable. Most listeners
cannam@86 143 of FM radio do not realize how much lower fidelity that medium is as
cannam@86 144 compared to compact discs or DAT. However, when compared directly to
cannam@86 145 source material, the difference is obvious. A cassette tape is lower
cannam@86 146 fidelity still, and yet the degradation, relatively speaking, is
cannam@86 147 graceful and generally easy not to notice. Compare this graceful loss
cannam@86 148 of quality to an average 44.1kHz stereo mp3 encoded at 80 or 96kbps.
cannam@86 149 The mp3 might actually be higher objective fidelity but subjectively
cannam@86 150 sounds much worse.</p>
cannam@86 151
cannam@86 152 <p>Thus, when a CODEC <em>must</em> sacrifice subjective quality in order
cannam@86 153 to satisfy a user's requirements, the result should be a
cannam@86 154 <em>difference</em> that is generally either difficult to notice
cannam@86 155 without comparison, or easy to ignore. An <em>artifact</em>, on the
cannam@86 156 other hand, is an element introduced into the output that is
cannam@86 157 immediately noticeable, obviously foreign, and undesired. The famous
cannam@86 158 'underwater' or 'twinkling' effect synonymous with low bitrate (or
cannam@86 159 poorly encoded) mp3 is an example of an <em>artifact</em>. This
cannam@86 160 working definition differs slightly from common usage, but the coined
cannam@86 161 distinction between differences and artifacts is useful for our
cannam@86 162 discussion.</p>
cannam@86 163
cannam@86 164 <p>The goal, when it is absolutely necessary to sacrifice subjective
cannam@86 165 fidelity, is obviously to strive for differences and not artifacts.
cannam@86 166 The vast majority of codecs today fail at this task miserably,
cannam@86 167 predictably, and regularly in one way or another. Avoiding such
cannam@86 168 failures when it is necessary to sacrifice subjective quality is a
cannam@86 169 fundamental design objective of Vorbis and that objective is reflected
cannam@86 170 in Vorbis's design and tuning.</p>
cannam@86 171
cannam@86 172 <div id="copyright">
cannam@86 173 The Xiph Fish Logo is a
cannam@86 174 trademark (&trade;) of Xiph.Org.<br/>
cannam@86 175
cannam@86 176 These pages &copy; 1994 - 2005 Xiph.Org. All rights reserved.
cannam@86 177 </div>
cannam@86 178
cannam@86 179 </body>
cannam@86 180 </html>