Use correct spelling for 'dB'.
[kugel-rb.git] / manual / configure_rockbox / sound_settings.tex
blob61b450e761844ce5c163b96d3974b014a1415f08
1 % $Id$ %
2 \screenshot{configure_rockbox/images/ss-sound-settings}{The sound settings screen}{}
4 The sound settings menu offers a selection of sound settings you may
5 change to customise your listening experience.
7 \section{\label{ref:volume}Volume}
8 This setting adjusts the volume of your music. Like most professional
9 audio gear and many consumer audio products, Rockbox uses a decibel scale
10 where 0~dB is a reference that indicates the maximum volume that the \dap{}
11 can produce without possible distortion (clipping). All values lower than
12 this reference will be negative and yield a progressively softer volume.
13 \opt{player,recorder,recorderv2fm,ondio,x5,m5,ipodnano,ipodnano2g,ipodvideo,
14 ipodcolor,ipod1g2g,h10,h10_5gb,sansa,sansaAMS,gigabeats,gigabeatf,vibe500}{
15 Values higher than 0~dB are available and can be used to raise the
16 volume more than would otherwise be possible. These volume levels will
17 ordinarily lead to distorted sound, but might work nicely for music that has
18 an otherwise low volume level.}
19 The volume can be adjusted from a
20 \opt{player}{minimum of -78~dB to a maximum of +18~dB.}%
21 \opt{recorder,recorderv2fm,ondio}{minimum of -100~dB to a maximum of +12~dB.}%
22 \opt{h100,h300}{minimum of -84~dB to a maximum of 0~dB.}%
23 \opt{x5,m5,ipod3g,ipod4g,gigabeatf,mrobe100}{minimum of -73~dB to a maximum of +6~dB.}%
24 \opt{ipodnano}{minimum of -72~dB to a maximum of +6~dB.}%
25 \opt{ipodvideo}{minimum of -89~dB to a maximum of +6~dB.}%
26 \opt{ipodnano2g,ipodcolor,ipod1g2g,h10,h10_5gb,sansa,sansaAMS}{minimum of
27 -74~dB to a maximum of +6~dB.}%
28 \opt{gigabeats}{minimum of -90~dB to a maximum of +6~dB.}%
29 \opt{gigabeatf,vibe500}{minimum of -74~dB to a maximum of +6~dB.}%
30 \opt{ipodvideo}{\\Remark: Lowering the volume below -57~dB will also affect the line-out
31 and the recording gain.}
33 \section{Bass}
34 This setting emphasises
35 \nopt{h100,h300}{or suppresses}
36 the lower (bass) frequencies in the sound. A value of 0~dB means that bass
37 sounds are unaltered (flat response).
38 \opt{masd}{The minimum setting is -15~dB and the maximum is 15~dB.}%
39 \opt{masf}{The minimum setting is -12~dB and the maximum is 12~dB.}%
40 \opt{h100,h300}{The minimum setting is 0~dB and the maximum is 24~dB.}%
41 \opt{ipodnano,ipodnano2g,ipodcolor}{The minimum setting is -6~dB and the maximum is 9~dB.}%
42 \opt{ipodvideo}{The minimum setting is -12~dB and the maximum is 12~dB.}%
43 \opt{x5,sansa,sansaAMS,h10,h10_5gb,vibe500}{The minimum setting is -24~dB and the maximum is 24~dB.}
45 \opt{ipodvideo}{
46 \section{Bass Cutoff}
47 This setting controls the frequency below which the bass adjustment applies.
48 The setting has a range from 1 to 4, where a bigger number affects a bigger
49 range of bass frequencies. The actual cutoff frequency used for each setting
50 value will vary with sample rate.
53 \section{Treble}
54 This setting emphasises
55 \nopt{h100,h300}{or suppresses}
56 the higher (treble) frequencies in the sound. A value of 0~dB means that
57 treble sounds are unaltered (flat response).
58 \opt{masd}{The minimum setting is -15~dB and the maximum is 15~dB.}%
59 \opt{masf}{The minimum setting is -12~dB and the maximum is 12~dB.}%
60 \opt{h100,h300}{The minimum setting is 0~dB and the maximum is 6~dB.}%
61 \opt{ipodnano,ipodnano2g,ipodcolor}{The minimum setting is -6~dB and the maximum is 9~dB.}%
62 \opt{ipodvideo}{The minimum setting is -12~dB and the maximum is 12~dB.}%
63 \opt{x5,sansa,sansaAMS,h10,h10_5gb,vibe500}{The minimum setting is -24~dB and the maximum is 24~dB.}
65 \opt{ipodvideo}{
66 \section{Treble Cutoff}
67 This setting controls the frequency above which the treble adjustment applies.
68 The setting has a range from 1 to 4, where a bigger number affects a smaller
69 range of treble frequencies. The actual cutoff frequency used for each setting
70 value will vary with sample rate.
73 \section{Balance}
74 This setting controls the balance between the left and right channels. The
75 default, 0, means that the left and right outputs are equal in volume.
76 Negative numbers increase the volume of the left channel relative to the
77 right, positive numbers increase the volume of the right channel relative
78 to the left.
80 \section{Channels}
81 A stereo audio signal consists of two channels, left and right. The
82 \setting{Channels} setting determines if these channels are to be combined in
83 any way, and if so, in what manner they will be combined.
84 Available options are:
86 \begin{table}
87 \begin{rbtabular}{\textwidth}{lX}%
88 {\textbf{Setting} & \textbf{Description}}{}{}
89 Stereo
90 & Leave the audio signal unmodified. \\
92 Mono
93 & Combine both channels and send the resulting signal to both stereo
94 channels, resulting in a monophonic output. \\
96 Custom
97 & Allows you to manually specify a stereo width with the
98 \setting{Stereo Width} setting described later in this chapter. \\
100 Mono Left
101 & Plays the left channel in both stereo channels. \\
103 Mono Right
104 & Plays the right channel in both stereo channels. \\
106 Karaoke
107 & Removes all sound that is common to both channels. Since most
108 music is recorded with vocals being equally present in both channels
109 to make the singer sound centrally placed, this often (but not
110 always) has the effect of removing the voice track from a song. This
111 setting also very often has other undesirable effects on the sound. \\
112 \end{rbtabular}
113 \end{table}
115 \section{Stereo Width}
116 Stereo width allows you to manually specify the effect that is applied
117 when the \setting{Channels} setting is set to ``custom''.
118 All values below 100\% will progressively mix the contents of one channel
119 into the other. This has the effect of gradually centering the stereo image,
120 until you have monophonic sound at 0\%. Values above 100\% will progressively
121 remove components in one channel that is also present in the other. This has
122 the effect of widening the stereo field. A value of 100\% will leave the
123 stereo field unaltered.
125 \opt{masf}{
126 \section{Loudness}
127 When listening at low volumes, the ear will tend to make bass and treble
128 frequencies sound quieter than they really are. To compensate for this,
129 \setting{Loudness} is an effect which emphasises bass and treble in a fashion
130 suited to the human ear. Frequencies in the vocal range are unaffected, since
131 the human ear picks these up very easily at any sound level.
132 It is of course also possible to use this effect at higher volumes for
133 enhanced bass and treble.
136 \opt{masf}{
137 \section{Auto Volume}
138 Auto volume is a feature that automatically lowers the volume on loud parts,
139 and then slowly restores the volume to the previous level over a time
140 interval. This setting allows this time interval to be configured. Short
141 values like 20~ms are useful for ensuring a constant volume for in-car use and
142 other applications where background noise makes a constant loudness desirable.
143 A longer timeout means that the change in volume back to the previous level
144 will be smoother, so there will be fewer sharp changes in volume level.
147 \opt{masf}{
148 \section{Super Bass}
149 This setting changes the threshold at which bass frequencies are affected by
150 the \setting{Loudness} setting, making the sound of drums and bass guitar
151 louder in comparison to the rest of the sound. This setting only has an
152 effect if \setting{Loudness} is set to a value larger than 0~dB.
155 \opt{masf}{
156 \section{MDB {}-- Micronas Dynamic Bass}
157 The rest of the parameters in this menu relate to the Micronas Dynamic
158 Bass (MDB) function. MDB is designed to enable the user to hear bass
159 notes that the headphones and/or speakers are not capable of reproducing.
160 Every tone has a fundamental frequency (the ``main tone'') and also several
161 harmonics, which are related to that tone. The human brain has a mechanism
162 whereby it can actually infer the presence of bass notes from the higher
163 harmonics that they would generate.
165 The practical upshot of this is that MDB produces a more authentic sounding
166 bass by tricking the brain into believing it is hearing tones that the
167 headphones or speakers are not capable of reproducing.
169 The MDB parameters are as follows:
171 \begin{description}
172 \item[MDB enable.]
173 This turns the MDB feature on or off. For many users this will be the
174 only setting they need, since Rockbox picks sensible defaults for the
175 other parameters. MDB is turned off by default.
176 \item[MDB strength.]
177 How loud the harmonics generated by MDB will be.
178 \item[MDB Harmonics.]
179 The percentage of the low notes that is converted into harmonics.
180 If low notes are causing speaker distortion, this can be set to 100\%
181 to eliminate the fundamental completely and only produce harmonics in the
182 signal. If set to 0\% this is the same as turning the MDB feature off.
183 \item[MDB Centre Frequency.]
184 The cutoff frequency of your headphones or speakers. This is usually
185 given in the specification for the headphones/speakers.
186 \item[MDB shape.]
187 It is recommended that this parameter be set to 1.5 times the centre frequency.
189 This is the frequency up to which harmonics are generated. Some of the
190 lower fundamentals near the cut{}-off range will have their lower
191 harmonics cut, since they will be below the range of the speakers.
192 Fundamentals between the cut{}-off frequency and the lower frequency
193 will have their harmonics proportionally boosted to compensate and restore
194 the `loudness' of these notes.
196 For most users, the defaults should provide an improvement in sound
197 quality and can be safely left as they are. For reference, the defaults
198 Rockbox uses are:
200 \begin{table}[h!]
201 \begin{rbtabular}{0.5\textwidth}{Xc}{Setting & Value}{}{}
202 MDB Strength & 50~dB \\
203 MDB Harmonics & 48\% \\
204 MDB Centre Frequency & 60~Hz \\
205 MDB Shape & 90~Hz \\
206 \end{rbtabular}
207 \end{table}
209 \end{description}
212 \opt{swcodec}{
213 \section{Crossfeed}
214 Crossfeed attempts to make the experience of listening to music on
215 headphones more similar to listening to music with stereo speakers. When you
216 listen to music through speakers, each ear will hear sound originating from
217 both speakers. However, the sound from the left speaker reaches your right
218 ear slightly later than it does your left ear, and vice versa.\\
220 The human ear and brain together are very good at interpreting the time
221 differences between direct sounds and reflected sounds and using that
222 information to identify the direction that the sound is coming from. On the
223 other hand, when listening to headphones, each ear hears only the stereo
224 channel corresponding to it. The left ear hears only the left channel and
225 the right ear hears only the right channel. The result is that sound from
226 headphones does not provide the same spatial cues to your ear and brain as
227 speakers, and might for that reason sound unnatural to some listeners.\\
229 The crossfeed function uses an algorithm to feed a delayed and filtered
230 portion of the signal from the right channel into the left channel and vice
231 versa in order to simulate the spatial cues that the ear and brain receive
232 when listening to a set of loudspeakers placed in front of the listener. The
233 result is a more natural stereo image that can be especially appreciated in
234 older rock and jazz records, where one instrument is often hard-panned to
235 just one of the speakers. Many people will find such records tiring to listen
236 to using earphones and no crossfeed effect.\\
238 Crossfeed has the following settings:
239 \begin{description}
240 \item[Crossfeed.]
241 Selects whether the crossfeed effect is to be enabled or not.
242 \item[Direct Gain.]
243 How much the level of the audio that travels the direct path from a speaker
244 to the corresponding ear is supposed to be decreased.
245 \item[Cross Gain.]
246 How much the level of the audio that travels the cross path from a speaker
247 to the opposite ear is to be decreased.
248 \item[High-Frequency Attenuation.]
249 How much the upper frequencies of the cross path audio will be dampened.
250 Note that the total level of the higher frequencies will be a combination
251 of both this setting and the \setting{Cross Gain} setting.
252 \item[High-Frequency Cutoff.]
253 Decides at which frequency the cross path audio will start to be cut
254 by the amount described by the \setting{High-Frequency Attenuation}
255 setting.
256 \end{description}
258 Most users will find the default settings to yield satisfactory results, but
259 for the more adventurous user the settings can be fine-tuned to provide a
260 virtual speaker placement suited to ones preference.
261 % TODO: adapt the guidelines for crossfeed settings found here?
262 % http://www.ohl.to/interests-in-audio/crossfeed-and-eq-for-headphones/
263 Beware that the crossfeed function is capable of making the audio distort
264 if you choose settings which result in a too high output level.
267 \opt{swcodec}{
268 \section{\label{ref:EQ}Equalizer}
269 \screenshot{configure_rockbox/images/ss-equalizer}{The graphical equalizer}{}
270 Rockbox features a parametric equalizer (EQ). As the name suggests, a
271 parametric EQ lets you control several different parameters for each
272 band of the EQ. In some ways the EQ is similar to the \setting{Bass}
273 and \setting{Treble} settings described earlier, but the EQ allows you to
274 control the sound much more carefully.\\
276 Rockbox's parametric EQ is composed of five different bands:
277 \begin{description}
278 \item[Band 0: Low shelf filter.]
279 The low shelf filter boosts or lowers all frequencies below a certain
280 frequency limit, much like what a ``bass'' control found on ordinary
281 stereo systems does.
282 Adjust the ``cutoff'' frequency parameter to decide where the shelving
283 starts to take effect. For example, a cutoff frequency of 50~Hz will
284 adjust only very low frequencies. A cutoff frequency of 200~Hz, on the
285 other hand, will adjust a much wider range of bass frequencies.
286 The ``gain'' parameter controls how much the loudness of the band is
287 adjusted. Positive numbers make the EQ band louder, while negative
288 numbers make that EQ band quieter.
289 The ``Q'' parameter should always be set to 0.7 for the shelving
290 filters. Higher values will add a small boost around the cutoff
291 frequency that is almost always undesirable.
292 \item[Bands 1-3: Peaking filters.]
293 Peaking EQ filters boost or lower a frequency range centered at the
294 centre frequency chosen.
295 Graphic equalizers in home stereos are usually peaking
296 filters. The peaking filters in Rockbox's EQ lets you adjust three
297 different parameters for EQ bands 1 through 3. The ``centre'' parameter
298 controls the centre frequency of the frequency range that is affected
299 as described above.
300 The ``gain'' parameter controls how much each band is adjusted, and
301 works as for the low shelf filter.
302 Finally, the ``Q'' parameter controls how wide or narrow the affected
303 frequency range is. Higher Q values will affect a narrower band of
304 frequencies, while lower Q values will affect a wider band of
305 frequencies.
306 \item[Band 4: High shelf filter.]
307 A high shelf filter boosts or lowers all frequencies above a certain
308 frequency limit, much like what a ``treble'' control found on ordinary
309 stereo systems does.
310 The high shelf filter is adjusted the same way as the low shelf filter,
311 except that it works on the high end of the frequency spectrum rather
312 than the low end.
313 \end{description}
314 As a general guide, EQ band 0 should be used for low frequencies, EQ bands 1
315 through 3 should be used for mids, and EQ band 4 should be used for highs.
317 \begin {description}
318 \item[Enable EQ.]
319 This option controls whether the EQ is on or off.
321 \item[Graphical EQ.]
322 This option brings up a graphic EQ screen, which allows adjustment of each of
323 the three parameters described above (gain, centre frequency, and Q) for each
324 of the five EQ bands.
326 \begin{table}
327 \begin{btnmap}{}{}
328 \opt{IPOD_4G_PAD,IPOD_3G_PAD,IPOD_VIDEO_PAD}{\ButtonScrollFwd}
329 \opt{IRIVER_H100_PAD,IRIVER_H300_PAD,IAUDIO_X5_PAD,PBELL_VIBE500_PAD}{\ButtonUp}
330 \opt{SANSA_E200_PAD}{\ButtonScrollBack}
331 \opt{GIGABEAT_PAD,GIGABEAT_S_PAD}{\ButtonRight}
332 \opt{COWON_D2_PAD}{\TouchMidRight}
333 \opt{HAVEREMOTEKEYMAP}{&
334 \opt{IRIVER_RC_H100_PAD,GIGABEAT_RC_PAD}{\ButtonRCFF}
336 & Raises the highlighted parameter.\\
338 \opt{IPOD_4G_PAD,IPOD_3G_PAD,IPOD_VIDEO_PAD}{\ButtonScrollBack}
339 \opt{IRIVER_H100_PAD,IRIVER_H300_PAD,IAUDIO_X5_PAD,PBELL_VIBE500_PAD}{\ButtonDown}
340 \opt{SANSA_E200_PAD}{\ButtonScrollFwd}
341 \opt{GIGABEAT_PAD,GIGABEAT_S_PAD}{\ButtonLeft}
342 \opt{touchscreen}{\TouchMidLeft}
343 \opt{HAVEREMOTEKEYMAP}{&
344 \opt{IRIVER_RC_H100_PAD,GIGABEAT_RC_PAD}{\ButtonRCRew}
346 & Lowers the highlighted parameter.\\
348 \opt{IRIVER_H100_PAD,IRIVER_H300_PAD,IAUDIO_X5_PAD}{\ButtonLeft}
349 \opt{IPOD_4G_PAD,IPOD_3G_PAD,IPOD_VIDEO_PAD,PBELL_VIBE500_PAD}{\ButtonLeft}
350 \opt{SANSA_E200_PAD,GIGABEAT_PAD,GIGABEAT_S_PAD}{\ButtonUp}
351 \opt{COWON_D2_PAD}{\ActionStdPrev}
352 \opt{HAVEREMOTEKEYMAP}{& }
353 \opt{GIGABEAT_RC_PAD}{\ButtonRCVolUp}
354 & Moves to the previous EQ band. \\
356 \opt{IRIVER_H100_PAD,IRIVER_H300_PAD,IAUDIO_X5_PAD}{\ButtonRight}
357 \opt{IPOD_4G_PAD,IPOD_3G_PAD,IPOD_VIDEO_PAD,PBELL_VIBE500_PAD}{\ButtonRight}
358 \opt{SANSA_E200_PAD,GIGABEAT_PAD,GIGABEAT_S_PAD}{\ButtonDown}
359 \opt{COWON_D2_PAD}{\ActionStdNext}
360 \opt{HAVEREMOTEKEYMAP}{& }
361 \opt{GIGABEAT_RC_PAD}{\ButtonRCVolDown}
362 & Moves to the next EQ band. \\
364 \opt{IRIVER_H100_PAD,IRIVER_H300_PAD,GIGABEAT_PAD,GIGABEAT_S_PAD}{\ButtonSelect}
365 \opt{IAUDIO_X5_PAD,IPOD_4G_PAD,IPOD_3G_PAD,IPOD_VIDEO_PAD,SANSA_E200_PAD}{\ButtonSelect}
366 \opt{PBELL_VIBE500_PAD}{\ButtonOK}
367 \opt{touchscreen}{\ButtonMenu{} or \TouchCenter}
368 \opt{HAVEREMOTEKEYMAP}{&
369 \opt{IRIVER_RC_H100_PAD}{\ButtonRCSelect}
370 \opt{GIGABEAT_RC_PAD}{\ButtonRCPlay}
372 & Toggles the cursor among the three parameters (gain, centre frequency,
373 Q) for the selected EQ band.\\
375 \opt{IRIVER_H100_PAD,IRIVER_H300_PAD}{\ButtonMode}
376 \opt{IPOD_4G_PAD,IPOD_3G_PAD,IPOD_VIDEO_PAD}{\ButtonMenu}
377 \opt{IAUDIO_X5_PAD}{\ButtonPower/\ButtonRec}
378 \opt{SANSA_E200_PAD}{\ButtonPower/\ButtonRight}
379 \opt{GIGABEAT_PAD,GIGABEAT_S_PAD}{\ButtonPower}
380 \opt{PBELL_VIBE500_PAD}{\ButtonCancel}
381 \opt{touchscreen}{\ButtonPower{} or \TouchTopLeft}
382 \opt{HAVEREMOTEKEYMAP}{&
383 \opt{IRIVER_RC_H100_PAD}{\ButtonRCStop}
384 \opt{GIGABEAT_RC_PAD}{\ButtonRCDsp}
386 & Exits the graphic EQ screen.\\
387 \end{btnmap}
388 \end{table}
390 \item[Pre-cut.]
391 If too much gain is added through the graphical EQ, your music may distort.
392 The \setting{Precut} setting allows you to adjust the overall gain of the EQ.
394 If your music distorts when using the EQ, trying changing this setting to a
395 negative value.
397 \item[Simple EQ.]
398 This option provides an easier alternative for those who are daunted by all of
399 the parameters that can be adjusted using the graphical EQ. With the
400 \setting{Simple EQ}, the only parameter that can be adjusted is the gain.
402 \item[Advanced EQ.]
403 This sub menu provides options for adjusting the same parameters as the
404 \setting{Graphical EQ}. The only difference is that the parameters are
405 adjusted through textual menus rather than through a graphic interface.
407 \item[Save EQ Preset.]
408 This option saves the current EQ configuration in a \fname{.cfg} file.
410 \item[Browse EQ Presets.]
411 This menu displays a list of EQ presets, as well as any EQ configurations saved
412 using the \setting{Save EQ Preset} option. Users unfamiliar with the
413 operation of a parametric EQ may wish to use the presets instead of trying to
414 configure the EQ, or use the presets for designing their own custom EQ
415 settings.
417 \end{description}
420 \opt{swcodec}{
421 \section{Dithering}
422 This setting controls the dithering and noise shaping functionality of Rockbox.
424 Most of Rockbox' audio file decoders work at a higher bit depth than the 16 bits
425 used for output on the \daps{} audio connectors. The simplest way to
426 convert from one bit depth to another is simply discarding all the surplus bits.
427 This is the default behaviour, and adds distortion to the signal that will
428 vary in character along with the desired sound.
430 Dithering adds low-level noise to the signal prior to throwing away the surplus
431 bits, which gives the resulting signal a uniform noise floor which is
432 independent of the signal. Most people find this noise preferable to the
433 time-varying noise heard when not performing dithering.
435 After dithering, noise shaping is performed. This basically just pushes the
436 dithering noise to the parts of the frequency spectrum humans cannot hear so
437 easily. In Rockbox' case, some of the noise is pushed up to above 10~kHz.
439 This setting will be put to its best use when listening to dynamic music with
440 frequently occuring quiet parts, classical music being a typical example. It is
441 worth noting that the effects of dithering and noise shaping are very subtle,
442 and not easily noticable.
444 Rockbox uses highpass triangular distribution noise as the dithering noise
445 source, and a third order noise shaper.
448 \opt{swcodec}{
449 \section{Timestretch}
450 Enabling \setting{Timestretch} allows you to change the playback speed without
451 it affecting the pitch of the recording. After enabling this feature and
452 rebooting, you can access this via the \setting{Pitch Screen}. This function is
453 intended for speech playback and may significantly dilute your listening
454 experience with more complex audio.
457 \opt{swcodec}{
458 \section{Compressor}
459 The \setting{Compressor} reduces, or compresses, the dynamic range of the audio
460 signal. This makes the quieter and louder sections closer to the same volume
461 level by progressively reducing the gain of louder signals. When subsequently
462 amplified, this has the effect of making the quieter sections louder while
463 keeping the louder sections from clipping. This allows listening to the quiet
464 sections of dynamic material in noisy environments while preventing sudden loud
465 sections from being overbearing.
467 There are several settings associated with the compressor. The first, and most
468 important, is the \setting{Threshold}. The threshold is the audio input level
469 at which the compressor begins to act. Any level louder than the threshold
470 will be compressed to some extent. The maximum amount of compression, or the
471 quietest level at which the compressor will operate, is -24~dB. The default of
472 Off disables the compressor.
474 The \setting{Makeup Gain} setting has two options: Off and Auto. Off means
475 that the compressed audio will not be amplified after compression. The default
476 of Auto will amplify the signal so that the loudest possible signal after
477 compression will be just under the clipping limit. This is desirable because
478 the compressed signal without makeup gain is quieter than the input signal.
479 Makeup Gain in Auto restores the signal to the maximum possible level and
480 brings the quieter audio up with it. This is what makes it possible to hear
481 the quieter audio in noisy environments.
483 The \setting{Ratio} setting determines how aggressively the compressor reduces
484 gain above the threshold. For example, the 2:1 setting means that for each
485 two decibels of input signal above the threshold, the compressor will only
486 allow the output to appear as one decibel. The higher the ratio, the harder
487 the signal is compressed. The ratio setting of Limit means essentially a ratio
488 of infinity to one. In this case, the output signal is not allowed to exceed
489 the threshold at all.
491 The \setting{Knee} setting determines how abrupt the transition is from a
492 non-compressed signal to a compressed signal. Hard Knee means that the
493 transition occurs precisely at the threshold. The Soft Knee setting smoothes
494 the transition from plus or minus three decibels around the threshold.
496 The \setting{Release Time} setting sets the recovery time after the signal is
497 compressed. Once the compressor determines that compression is necessary,
498 the input signal is reduced appropriately, but the gain isn't allowed to
499 immediately return to normal levels. This is necessary to reduce artifacts
500 such as "pumping." Instead, the gain is allowed to return to normal at the
501 chosen rate. Release Time is the time for the gain to recover by 10~dB.