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