EXHIBIT 2013
`
`EXHIBIT 201 3
`
`

`
`Martin Colloms
`
`sixth edition
`
`@?)WILEY
`
`THX Ltd. Exhibit 2013 Page 1
`IPR2014-00235
`
`

`
`High Performance
`Loudspeakers
`
`Sixth Edition
`
`Martin Colloms and Paul Darlington
`
`Col/oms Electmtlcoustics, UK
`Apple Dynamics, UK
`
`John Wiley & Sons, Ltd
`
`THX Ltd. Exhibit 2013 Page 2
`IPR2014-00235
`
`

`
`Copyright © 1005
`
`John Wrlcy & Son' L!tl, The Atri um . Solllhe rn Gate. Chic hester.
`We~t Sussex POI <J RSQ, England
`
`Tele phone ( + -W ) 1243 779777
`
`Emai l (fo r orde rs ancl c usto me r wrvtcc c nqui ric>): C•·books@ wi ley.co.uk
`Visit o ur ll omc Page o n www.wi ley.co m
`
`All Riglm Reserved. No pa n o f this pu blication rnay be reproduced, sto red in a retrieval system or tran~miued in
`any fomt or by a ny means, electronic. mechan ical. ph01ocopyi ng. record ing, scanning or o therwise, e xcept under the
`te nns of the Copy right. Desig ns nnd Pa te nts Ac t 1988 or unde r the tenns of a licence issued by the Copyrig ht
`Licensing Agency Ltd, 90 Touc nham Court Rond, London WIT 4 LP, UK , without the pe rmi ssion in wri ting o f the
`Puhlishcr. Requests to the Publi sher should be :uldn:.<secl to the Penniss io ns Dc partrnem. John Wi ley & Sons l, td,
`Titc Atrium . Southern Gmc. Chichc., tcr. West S u,scx POI 9 SSQ. Eng land, o r e mailcd to pcmtreq@wiley.co.uk, o r
`faxed to ( +44) 12-13 770620.
`
`This publi catiou is des ig ned to (li'Ovid c accunuc ami a utho ritlll ive infonnati cm in regard to the subject maller
`cove red. It is so ld o n the understa nding lhat the Publishe r is not e ngaged in rende ring profession~! services. If
`professional ad vice o r o ther ex pe n a.~sistance is requi red. the services of a compete nt professional shou ld be sought.
`
`Other ll'iley Editorial O.Oices
`
`Jo hn Wiley & Son~ Inc .. Ill Ri ve r S tree t. Hoboken. NJ 0 7030, USA
`
`Jossey- Bnss. 989 Mnrke t Street. Sa n Francisco, Ci\ 94 103- 174 1. USA
`
`Wi lcy-VCII Verlag Gmb ll. Boschw . 12, D-69,169 Wcinheim, Gc nnany
`
`Jo hn Wiley & So ns 1\nstra.lia Ltd, 42 1\lcDougn ll Street. Mi lto n, Queensland 4064, Australi n
`
`John Wi ley & Sons (As ia) Pte Ltd, 2 C lc memi Loop 1102-0 1, Jin Xing Distripark . Singapore 129809
`
`Jo hn Wiley & So ns Canada Ltd, 22 Worcester Road. Etobicokc. Ontario. Canada 1\19\V ILl
`
`Wiley :llso publishes its books in a va riety o f electronic fo nn ats. Some conte nt that a ppears
`in print may not be available in electronic boo ks.
`
`Library of Congress Cntaloging-in-l'ublicnrion /Jnta:
`
`Collo ms, Martin.
`Hi gh perfomt ancc lo udspeakers I Mani n Co ll oms and Paul Dnrlington.
`-6th cd.
`p. em.
`Includes bibliog rnphical refere nces a nd index.
`IS (l N 0-470-09430-3 (a lk. paper)
`I. Darlingto n, Pa ul.
`I. Lo udspeakers.
`TK5983.C6-l 2005
`62 1.382'8-l - dc22
`
`II. Title.
`
`2005024427
`
`British Ubrary Cntalo~:uing in Publica/inn Data
`
`A catalogue record for this book is availa ble fro m the British Lihrury
`
`IS BN- 13: 9 78-0-4700 94-30 -3
`ISBN- 10: 0-470094-30-3
`
`Typeset in I0/12pt 1i mcs by La,erwords Pri vate l. imi tcd. Ch~nnai, Ind ia
`Pri nted ;md bound in Great llrilili n by Amony Rowe Ltd, Chippcnh:un. Wiltshire
`This hook is prin ted on ~cid-fn:c p:•pc r respo nsibly manu factured from sustainable forestry
`in which a t least two trees arc plant ed for each one '"eel for p:•per production.
`
`THX Ltd. Exhibit 2013 Page 3
`IPR2014-00235
`
`

`
`184
`
`High Performance L oudspeakers
`
`aluminium foil to aid hem conduction. The dome itself. depending on the metal or alloy
`chosen, may be drawn from foil of 30 to 100 ttm thickness. and wi ll require careful
`handling.
`
`Tolemncing
`The control of sensiti vity is vital for prcctston loudspeakers. Without this control a
`designer' s good intentions count for liulc as quite large changes in sound quaJity occur
`with relatively smaJI changes in relative driver level. In one test example. a stereo pair
`was found to have poor focus, where the image was pulled laterally, depending on the
`frequency reproduced, and one channel sounded louder than the other. When individually
`measured the two speakers fell within the tolerance defined for the overall specification
`namely ±3 dB, 100 Tlz to 15 k ll z. Close examination showed errors of as liule as I <lEi
`for each driver relative to a design centre defined for the system. llowevcr a careful com(cid:173)
`parison of the systems showed that a worse case error scenario pertained. For one channel
`the tweeter was I dB high, the woofer I dB low; f(>r the other the tweeter was l dB low,
`the woofer 1 dB high. Using the criterion that for a consistent sound 0.5 dB tolerancing
`is worthwhile, for the treble level relative to the mid-range, this pair of speakers was
`delivering an unacceptable 2 dB error- one 2 dB too 'bri ght', the other 2 dB too 'dim'.
`Production procedures arc advisable which can trnp such errors and provide data on the
`causes of these variations, thus facilitating tighter quality control.
`In the above example. swapping over the dri ver combinations would result in two
`tonally balanced speakers, though with one measuring a noticeable. but not so damaging,
`2 dB greater than the other.
`Where variations occur, pre-checking driver sensitivity in a simple test fix ture is really
`useful. For critical applications they may be sorted into colour-coded groups and matched
`with other drivers of equivalent sensitivity. Good sample and pair matching docs result
`in a significant performance gain, which may not always be specifically detectable by an
`observer. Cont rolled subjective testing docs show n link between close tolerance matching
`and bcuer sound quality overall.
`Virtually every component of a driver can affect performance and sensitivity. Key
`factors arc the cone or diaphragm mass, the number of turns and the exact gauge of wire
`in the voice coil, coil centration, the gap tolerance and the build quality of the magnet
`system. Small errors c;m accumulate, resulting in serious I or 2 dB changes in loudness,
`and these must be controlled. For driver quality control. tbere is also no substitute for
`a flux probe which can read directly the field strength auained by sample production
`assemblies and methods.
`
`Effer.Js of Pam111e1er \!arimio11. l'articlllarly Cli111a/e
`ln a number of the references. sensitivity analysis has been applied to l!xaminc how the
`target responses are affected by variations in the parameters. It would appear that more
`complex the alignment order, the more sensiti ve it becomes to such variations.
`Consider a system designed for a worldwide market where air cond itioning and tem(cid:173)
`perature control cannot be assumed in the place of usc. llumidi ty can vary from almost
`0 to I 00%. 1\ significant change in mov ing mass ca n result with pulp cones clue to the
`hygroscopic nature of nwny diaphragms including. though to a lesser ex tent. the synthetic
`
`THX Ltd. Exhibit 2013 Page 4
`IPR2014-00235
`
`

`
`185
`. S ,, tern Analysis: Room Environment\ and 2rr Theory
`. frctJtlcncY ·.:.:>~~-__:..-----------------------
`~
`·~d aterials such as polyamide: Ambient temperature can vary from 15 °C to
`11yton-bi~St; ?~1!c as a general mean, but a speaker system with its back to a sunlit window
`.10 °C wrtl~ - - e an overa ll temperawre rise of 10 oc or so above ambient, making a total
`c~n cxr~r!~~\hese arc serious eff~cts since the ~nr:round of many modem lo.w colo~rnllion
`nsc to 4.
`11, 10y synthetic matcnals such as n11nlc rubber whose mcchantcal resistance
`11111ts en
`.
`.
` nee vary strongl y With temperature 111 the range concerned. As an example, a
`bass
`d cornp "1
`·u a 110111inal 25 Hz resonance may have a 17 liz resonance at 27 C. but alter-
`a
`'
`'
`.111
`1 ·v·r wr 1 •
`< n .'", 35 Hi'. at 10 °C. With the surround frequently the dominant driver compliance.
`nauvcll~ ·be worth considaing LF alignments exhibiting reduced sensitivity to such vari(cid:173)
`~~ ."'~.uovcr the specific range cited, .a rcOex dcsi~n migh.t misbehave .in such a mar~ncr
`'11:
`reflex loading which may be a n sk that a destg,ner nught not cons1der worth tak1ng.
`th
`regards other factors, almost any driver parameter can and will vary due to pro-
`A~on tolernnce. probably the greatest of these bt.:ing tlte IJI factor. Appearing ' squared·
`clucll
`·
`I .
`. .
`. I
`I
`fi
`I
`.
`I
`I
`I • LF analysis relat1ons ups II IS ccrtam y t 1c 1rst c 101ce to to erancc c osely. both
`111 tiC
`.
`.
`.
`. producrjon !>pread and 1ts effect on the result1ng system ahgnment (sec Chapter 3.
`~~.:tion 3.9). The use or smurated magnetic parts and good coil cent ration arc key factors.
`
`1.
`
`4.7 Transmission-line E nclosures
`In theory a transmission line or labyrinth is capable of being extended to infinity, providing
`,1 perfectly resistive termination to the driver by absorbing all the rear directed energy.
`llcncc in tl1is respect it may be regarded as a special case of a very large scaled box.
`which is itself the practical realization of an ' infinite barne·.
`Physically the line resembles an acoustic pipe which, in its folded form, is also known
`,,, a labyrinth. The cross-section must be sufficiently large in order that the rear directed
`~ncrgy is not impeded. and yet it must also be long enough to tcrrninate the energy
`dm•m to the lowest desired frequencies. Si nce the path length must be comparable to the
`''""\!length absorbed, if genuine low-frequency operation is the aim, ideal transmission
`lines quickly become over-large, even when folded. For a given closed line, the absorption
`reduces as the cut-off frequency is approached. This menns that the energy is beginning
`to be rcnected back to the driver and the system will then approximate to a sealed box
`of the same totHI volume. For example, a filled, non-resonnnt absorptive line, intended
`10 oper~te effecti vely clown to 35 Hz, must be at least 8 m long. I f a cross-section of
`900 em
`is adopted (roughly equivalent to that of a 36 em diameter bass driver) then the
`~nclosurc volume will be inordinately large at 720 litres.
`Pra~tical lines nre of necessity smaller in cross-section 10 reduce the total volume.
`and Pipe resonances and reflections at higher frcquencie may become a problem unless
`a~s,orption.is very carefully arranged. It was discovered by Bailey that a filling composed
`~ ong-h:urcd sheep's wool at a density of 8 kg m 3 possessed better acoustic properties
`~~~arcd _with other materials, such as glass fibre. More specifically, at low frequencies
`1m 60 I I~) the speed or sound in the line appears to be reduced by 50%. which has
`por~ant tmplicntions [4R-5 1] .
`,, ~~hllc, in theory, a damped line is a perfect absorber and the termi nation at the far end
`lnvarin~l~onsequcnce: that is, it may be closed or open. in practice, commercial designs
`_a ) ) leave the lrne open. Energy propagated hdow the line's cut-off frequency will
`
`.
`.
`• E\cn \Ill. lwc ·r. d
`c ct nme, 111:1y vary wuh hmmdlly.
`
`II igh Performance 1
`~dspea~cr.
`~
`:pending on the 111 , • et,tl or 1
`a to,
`ness, and will req .
`tllrc ca f ·
`rc Ul
`
`akers. Without this
`.
`· control .
`I
`a
`langes lfl sound qualit
`•ne test example a st y OCcur
`crco n •
`,.lur
`ed laterally, dependinn
`o on the
`I
`n t tc other. When indiv'd
`d for the overall spcc'tfit u:lll}
`·
`C.1hon
`wed errors of as little as 1 dU
`
`'
`
`I
`
`stem. However a careful
`com.
`.
`.
`trro pertamcd. For one channel
`her the tweeter was 1 dB 10,,
`:tent s~und .0.5 dB tolerancin~
`tge. thts patr of speakers \\ '"
`dll'. the other 2 dB too 'dim'
`!rrors and provide data 011 the
`:ontrol.
`>inmions would result in two
`ticeable, but not so damaging.
`
`n a simple test fixture is rcull)
`Jur-coded groups and matched
`and pair matching does result
`e spccificnlly detectable b)' .111
`ween close tolerance matching
`
`orrnance and sensitivity. Kt.:)
`IS and the exact gauge of wrr~
`e build quality of the magnet
`or 2 dB changes in loudncs'l.
`there is also no substitute for
`ttained by sample production
`
`n applied to examine hoW the
`·
`that mor~
`rs. It would appe.Jr
`s 10 such variations.
`d ICIO"
`.
`· ·
`·
`ere air condruo111ng an
`1 ·1hl10SI
`I.
`Jmidiry can vary ron '
`•
`1
`. d e to t it.:
`11
`It with (1trlp cones
`•
`nthCllC
`0 a lesser ex tent, the sy
`
`THX Ltd. Exhibit 2013 Page 5
`IPR2014-00235
`
`

`
`186
`
`High Pcrfonnancc Loudspeakers
`
`emerge from the aperture and, depending on its phase, will augment or cancel the main
`contribu tion from the front of the driver. Over a limi ted range of frequencies such an
`enclosure will behave as a reflex, the propagation delay down the line providing the
`required phase inversion. This is where the wool filling comes into its own since, for a
`given length of line, the delay is effectively doubled, thus lowering the cut-off frequency
`of the system as a whole (Fig. 4.42)
`It has been suggested that the mechanism of this increased delay is at least partially due
`to the air velocity in the line reducing al low frequencies, due to the movement of the wool
`mass. As such, the lallcr represents a non-linear inertial component in the energy path
`down the line, and the impulse response may be irregular and poor subjecti ve rhythm.
`A funher effect is produced at low frequencies by the air mass in the line moving with
`the driver, added to the driver's own moving mass. Whereas in conventional boxes the
`
`Figure -t..t2 A four-way, Boor standing ~ystcm with separate transmission-line loading at mid and
`low frequencies. The larger line is folded and a thick foam lining helps absorb highcr-order pipe
`modes (courtesy TDL)
`
`THX Ltd. Exhibit 2013 Page 6
`IPR2014-00235
`
`

`
`mancc Loudspeakers
`
`or cancel the main
`cq ucncies such an
`line providing the
`s own since. for a
`: cut-off frequency
`
`least partia lly due
`'emcnt of the wool
`n the energy path
`bjective rhythm.
`: line moving witJ1
`~111 ionaJ boxes IJ1c
`
`loading at mid and
`higher-order pipe
`
`(.,ow-frequency System Analysis: Room Environments and 2rr Theory
`
`187
`
`nir mass is quite tow, often less thnn 20% of the driver, in an open transmission line, the
`mass addition may be equal to that of the driver itself, thus reducing the Iauer's loaded
`resonant frequency by a factor of J2.
`Resonant modes are common in this conve111ional type of open pipe, occurring at odd
`(as opposed 10 even) mulliples of the quarter wavelength . Taking the example of a 2 m
`fi lled pipe with moderate absorption (packing density below 5 kg/m 3) the quarter-wave
`mis-termination occurs at about 30 Hz. If the pipe is over-damped, too much energy
`will be absorbed, and the sound output will possess a gradual roll-off from as high as
`70 liz. Tn commercia ll y viable systems. i.e. those of reasonable size, it may be difficult to
`find a satisfactory compromise between a uniformly extended LF response and an upper
`bass/lower mid-performance with minimal, colouration~ inducing line resonances.
`
`Quarter-wave 'Line' Loading
`II is worth keepi ng in mind that for a useful I 00 dB s.p. l. al a mid-bass frequency of
`50 Hz c even a 2 10 mm frame (8 in) bass unit needs to move ±7.5 mm under scaled-box
`loading.
`Considering the drive for a column of sufficient length, e.g. 1.5 m, intending 'quarter(cid:173)
`wave' operation, the system behaves as an infinite baffle above I 00 Hz, but by 50 Hz
`the open-ended column is imparting a mass load on the driver, the larger radiating area
`maintaining the overall acoustic output, while diaphragm excursion in this example is
`reduced to just ±2.5 mm at 50 Hz (as of course does reflex loading). The total moving
`mass is typically increased four-fold at low frequencies, reducing the f s of a 50 Hz driver
`1025 Hz.
`The one-quarter wavelength constitutes the low-frequency limit; 30 Hz has a 12 m
`wavelength, so the required column will be 3 n1.
`While in theory columns differ from transmission lines, they also suffer from harmonic
`modes, and since damping is discouraged on the grounds of maintaining efficiency, the
`unwanted higher modes are often troublesome. The most severe is the third harmonic, and
`this may be controlled by mounting the drive unit nt the upper third of the column length,
`on the node. In practice, further steps such as tapering the duct (which also is usuaJly
`folded within the enclosure) and judicious placement of sections of fibre absorbent arc
`required to control colouration due to these secondary resonances.
`In one example of a quarter-wave system the ' mouth ' of the column was allerccl (suc(cid:173)
`cessfully) 10 a quasi -reflex type by adding n slotted port c which tuned 111e volume to
`30 Hz-the system cut-off. In this instance a mix of the two types resulted wh ich was
`subjectively clean sounding.
`
`Pipes and Lines Oven,iew
`
`1\ugsperger succinctly rev iewed transmiSSion lines under the title ' Damped Pipes'.
`1\ugsperger, a specialist in reponing acoustic patents, reminds us of the early US patent
`of 1936 by Olney, for an 'Acoustic Labyrimh' fitted with an absorptive lining.
`One primary question concerns the fibre stuffing, whether it moves and adds to mass
`and clamping, or whether it is largely stationary, thus amenable to fluid flow consideration ,
`for example using an analysis method based on Locanthi · s horn analog.
`
`THX Ltd. Exhibit 2013 Page 7
`IPR2014-00235
`
`

`
`188
`
`I ligh Performance Loudspeaker!i
`
`Low
`
`A plain pipe resonates at odd multiples of its fundamental quarter-wave resonance and
`these arc strongly coupled to both the cone, and thus its output, and to the pipe exit.
`Even worse, the cone and pipe outputs move in and out of phase at even multiples of the
`fundamental modes, and n most uneven nnd coloured output results.
`It is shown that quite high levels of stuffing are required for a satisfactory ±I dB of
`pass-band ripple, this averaging I 0 grams per litre for system Q values of around 0.5.
`The weight of stuffing varies greatly with the grade; for example, a 1.22 m line with
`7 1 Hz natural fp required 17.5 g/1 of polyester wadding, 9.5 g/1 of fibreglass and just
`7.5 g when using 'micro fibre' {vadding.
`Specified for a greater but perhaps acceptable ripple, the analysis indicates that a well(cid:173)
`designed line can provide. the simjlar pass-band efficiency of a dosed box of lhe same
`volume, but with 2-3 dB more output at the low-frequency limil. This output could
`be traded for reduced cone excursion and would allow for a little hi gher overall power
`handling and acoustic output for the same investment. (Sec Figure 4.43(a-c).)
`Augspergcr also considers some variations on the line theme, for example, tapered lines,
`offset driver placerncnL and use of an upper coupling chamber. The possible directivity
`effccls of separated low-frequency sources arc noted while the investigalion indicates that
`usc of sufficient stuffing is essential to suppress tllC strong resonant behaviour of what
`is acouslically still a pipe. When so satisfaclorily clamped (not shown), a simple line or
`pipe then has typically half the efficiency of an equi valently sized closed box and the
`pipe output is actually so low as not to figure significantly in the output.
`
`Figw
`(af1e1
`
`a
`1:
`
`' •
`
`Figm
`drivel
`
`ID
`-o
`a. (/)
`
`110.00
`
`105.00
`
`100.00
`
`95.00
`
`90.00
`
`85.00
`
`80.00
`
`75.00
`
`70.00
`
`65.00
`
`60.00
`10
`
`100
`Frequency (Hz)
`
`1000
`
`(a) Total (~olid). driver (doued), pipe (dashed). for a medium damped, driven line/pipe
`Figure 4.43
`(after Augspergcr)
`
`~ v .... .... .....
`.
`~
`:,/ .. .. ·······
`~ .. ·
`--
`-- ~
`-....
`v.
`T,. ~··
`. •
`.... ~--~---
`.....
`~·
`Cone ··
`h:>Jpe
`
`r--..
`
`/
`
`....
`
`....
`
`'t--
`
`I;
`
`....
`
`THX Ltd. Exhibit 2013 Page 8
`IPR2014-00235