Mercurial > hg > sv-dependency-builds

comparison src/libvorbis-1.3.3/doc/vorbis-fidelity.html @ 1:05aa0afa9217

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Bring in flac, ogg, vorbis
author Chris Cannam
date Tue, 19 Mar 2013 17:37:49 +0000
parents
children
comparison
equal deleted inserted replaced
0:c7265573341e 1:05aa0afa9217
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 (&trade;) of Xiph.Org.<br/>
175
176 These pages &copy; 1994 - 2005 Xiph.Org. All rights reserved.
177 </div>
178
179 </body>
180 </html>