`
`UNITED STATES DISTRICT COURT
`SOUTHERN DISTRICT OF CALIFORNIA
`BEFORE HONORABLE CATHY ANN BENCIVENGO, JUDGE PRESIDING
`________________________________
`BELL NORTHERN RESEARCH, LLC,,
`Plaintiff,
`
`))
`
`)
`
`) CASE NO. 18CV1783-CAB-BLM
`)
`
`))
`
`vs.
` SAN DIEGO, CALIFORNIA
`)
`COOLPAD TECHNOLOGIES, INC. AND
`YULONG COMPUTER COMMUNICATIONS, )
`) THURSDAY, JUNE 20, 2019
`Defendants.
`)
`--------------------------------)
`BELL NORTHERN RESEARCH, LLC,
`Plaintiff,
`
`))
`
`) CASE NO. 18CV1784-CAB-BLM
`)
`
`vs.
`HUAWEI TECHNOLOGIES Co., LTD., )
`HUAWEI DEVICE (HONG KONG) CO., )
`LTD., and HUAWEI DEVICE USA,
`)
`INC.,
`)
`Defendants. )
`--------------------------------)
`BELL NORTHERN RESEARCH, LLC.,
`) CASE NO. 18CV1785-CAB-BLM
`Plaintiff,
`)
`vs. )
`)
`KYOCERA CORPORATION and KYOCERA )
`INTERNATIONAL INC., )
`)
`Defendants. )
`--------------------------------)
`
`))
`
`))
`
`1 2 3 4 5 6 7 8 9
`
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`1
`
`LG 1029
`LG v BNR
`IPR2020-00108
`
`
`
` 2
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`--------------------------------)
`BELL NORTHERN RESEARCH, LLC., )
` )
` Plaintiff, ) CASE NO. 18CV1786-CAB-BLM
` vs. )
` )
`ZTE CORPORATION, ZTE (USA) INC. )
`ZTE (TX) INC. )
` )
` Defendants.)
`--------------------------------)
`BELL NORTHERN RESEARCH, LLC,, )
` )
` Plaintiff, ) CASE NO. 18CV2864-CAB-BLM
` )
` vs. )
` )
`LG ELECTRONICS, INC., LG )
`ELECTRONICS U.S.A. INC., and )
`LG ELECTRONICS MOBILE RESEARCH )
`U.S.A., LLC, )
` )
` Defendants. )
`________________________________)
`
`
`
`
`
`
`REPORTER'S TRANSCRIPT OF PROCEEDINGS
`CLAIMS CONSTRUCTION HEARING
`DAY TWO, VOLUME TWO, PAGES 1-122
`
`
`
`
`
`
`Proceedings reported by stenography, transcript produced by
`computer assisted software
`
`M a u r a l e e R a m i r e z , R P R , C S R N o . 1 1 6 7 4
` F e d e r a l O f f i c i a l C o u r t R e p o r t e r
` o r d e r t r a n s c r i p t @ g m a i l . c o m
`
`2
`
`
`
` 3
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`COUNSEL APPEARING:
`For The Plaintiff: Sadaf Raja Abdullah, Esq.
` Steven W. Hartsell, Esq.
` Paul J. Skiermont, Esq.
` SKIERMONT DERBY LLP
` Thanksgiving Tower
` 1601 Elm Street, Suite 4400
` Dallas, Texas 75201
`
`
`For The Defendants Thomas Nathan Millikan, Esq.
`Coolpad and Yulong: James Young Hurt, Esq.
` PERKINS COIE, LLP
` 11988 El Camino Real, Suite 350
` San Diego, California 92130
`
`for the Defendants Joanna M. Fuller, Esq.
`Huawei entities: Jason W. Wolff, Esq.
` FISH & RICHARDSON P.C.
` 12390 El Camino Real
` San Diego, California 92130
`
` Ethan J. Rubin, Esq.
` FISH & RICHARDSON, P.C.
` One Marina Park Drive
` Boston, MA 02210
`
`For The Defendants Jiaxiao Zhang, Esq.
`ZTE entities: McDERMOTT WILL & EMERY LLP
` 18565 Jamboree Road, Suite 250
` Irvine, California 92612
`
` Amol Ajay Parikh, Esq.
` Thomas DaMario, Esq.
` McDERMOTT WILL & EMERY LLP
` 444 West Lake Street, Suite 4000
` Chicago, Illinois 60606
`
`ALSO PRESENT:
`
`
`
`
`
`
`
`3
`
`
`
` 4
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`San Diego, California; Thursday, June 20, 2019; 9:00 a.m.
`(Cases called)
`MS. ABDULLAH: Sadaf Abdullah on behalf of plaintiff,
`Bell Northern Research.
`MR. HARTSELL: Steven Hartsell on behalf of Bell
`Northern Research.
`THE COURT: Thank you.
`MR. SKIERMONT: Good morning your Honor. Paul
`Skiermont on behalf of Bell Northern Research.
`MS. ZHANG: Good morning, your Honor. Jiaxiao Zhang
`from McDermott Will & Emery on behalf of ZTE. With me is Amol
`Parikh and Thomas DaMario.
`MS. FULLER: Good morning. Joanna Fuller on behalf of
`Huawei with Fish & Richardson, and with me is Jason Wolff and
`Ethan Rubin.
`MR. MILLIKEN: Good morning, your Honor. Tom Milliken
`from Perkins Coie on behalf of Coolpad and Yulong. With me is
`James Hurt.
`THE COURT: Thank you. All right. We're back. So
`let's get started on the '842.
`MR. HARTSELL: Your Honor, may I approach?
`THE COURT: Yes. Go ahead.
`MR. HARTSELL: Good morning, your Honor. Again this
`is Steven Hartsell on behalf of Bell Northern Research. The
`'842 patent was developed by engineers at Broadcom and filed in
`
`4
`
`
`
` 5
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`January of 2010. The '842 patent is a continuation of U.S.
`Patent Number 7,646,703 which claims priority to at least
`July 2004. The '842 patent is directed to long training
`sequences with minimum peak-to-average power ratios, and today
`I would like to provide some background and a few common steps
`that I hope the Court would find useful in today's discussion.
`The '842 patent is taught against the backdrop of the
`802.11 WiFi standard which is promulgated by IEEE, which is the
`Institute for Electrical and Electronic Engineers. This
`standard governs how different wireless devices are designed
`and how they communicate with one another. Now as technology
`evolves, the 802.11 standard has been amended periodically to
`add additional capabilities, usually resulting in faster speeds
`and better coverage.
`As you can see on our slide, in 1999, the 802.11
`standard was amended to implement OFDM, which stands for
`orthogonal frequency-division multiplexing, to increase data
`throughput. I'm going to show you what that means on slide 5.
`At the top, you can see this is how data was transmitted OFDM.
`Basically we have single carriers that are separated. When
`OFDM is implemented, the carrier waves are essentially smushed
`together allowing you to send more data found within the given
`bandwidth. As you can see on the OFDM, there's an overlap in
`the subcarriers which is necessary to achieve high data rates.
`In slide 6, each colored peak is a subcarrier which
`
`5
`
`
`
` 6
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`carries data essentially, for example, the data you might need
`to load your website. The carriers are designed to be
`orthogonal which allows them to occupy the same bandwidth
`without interfering with which other.
`Now as with many things while OFDM provides throughput
`improvements and other advantages, it also brings certain
`disadvantages. And one of the disadvantages to using OFDM
`systems is they are known to have high peak-to0average power
`ratio, in other words, PAPR, when compared to single carrier
`systems. PAPR is the ratio of peak power to the average power
`signal.
`Now due to the presence of large numbers of
`independently modulated subcarriers in an OFDM system, the peak
`value of a system can be very high as compared to the average
`of the system as a whole. This is a problem -- PAPR is a
`problem because it reduces the power efficiency of radio
`frequency amplifiers, and this results essentially in high
`power consumption battery drain.
`Therefore, the RF amplifiers are operated usually with
`a certain safety margin called a power back-off. Increasing
`the power back-off can result in lower amplifier efficiency and
`higher overall power consumption.
`Another concept that may come up today is BPSK. BPSK
`stands for binary phase shift keying which is a digital
`modulation process by changing or modulating a phase of a
`
`6
`
`
`
` 7
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`constant frequency reference signal. The patent explains at
`column 2, lines 29 to 34 that in the 802.11a and 802.11g,
`versions of the standard when data packets are inserted, they
`include a preamble, and that preamble contains a short training
`sequence followed by a long training sequence which are used to
`synchronize -- which are used for synchronization between the
`sender and receiver devices.
`Now the long training sequence uses BPSK and,
`therefore, each subcarrier in the training sequence consists of
`either a +1 or a -1. That's just an artifact of using BPSK.
`So there are very few symbols that are actually available
`behind using BPSK coding, making it very important to be able
`fine tune the timing so that data in the packet is accurately
`read and and interpreted.
`In slide 10, this is a three-dimensional
`representation of an OFDM channel. At the top left in the kind
`greenish-gray area, you can see these are the short training
`fields. To the right, the blue squares represent the long
`training fields, and the gray blocks further to the right
`represent the data that is actually being transmitted. And as
`you can see in OFDM, there a lot of overlapping data occurring
`at the same time.
`Now with higher data throughputs, the patentees
`recognized the need to create longer training sequences to
`ensure proper synchronization between sending and receiving
`
`7
`
`
`
`8
`
`wireless devices especially since we were going to start
`compacting more data than we were before. The solution that
`the inventors devised built upon the existing training
`sequences by adding subcarriers which are selected in a manner
`to minimize PAPR. You can see in the last slide essentially
`they took the existing long training sequences and they add
`subcarriers to either side. And there's a couple of examples
`in the patent. And they select these subcarriers such that the
`PAPR is minimal.
`And as we saw on the previous slide, these preambles
`are sent with every data packet so they're constantly being
`sent, so it's desirable to minimize the PAPR as much as
`possible.
`And unless the Court has any questions, I would hand
`it over to defendants' counsel.
`THE COURT: I'm sure I will, but go ahead.
`MR. HURT: Good morning, your Honor. James Hurt from
`Perkins Coie on behalf of the defendants.
`THE COURT: Thank you.
`MR. HURT: So today for you, I am going to present a
`tutorial. The roadmap, I have four basic modules. Those four
`modules are going to be wireless basics, then switching to
`frequency and time domain, then talk a little bit about
`orthogonal frequency-division multiplexing or OFDM.
`(Court reporter interruption)
`
`1 2 3 4 5 6 7 8 9
`
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`8
`
`
`
`9
`
`MR. HURT: Oh, I'm sorry.
`And then we'll talk a little about the 802.11
`standards themselves. So what is wireless digital
`communications? Fundamentally this is getting bits from the
`transmitting apparatus to the receiving apparatus. It involves
`the movement of information from the transmitter to the
`receiver. All it is moving, information from point A to point
`B.
`
`But that information needs to go to something called
`"the channel." What is the channel? I like to think of the
`channel like a hose. It's just a pipe that connects the
`transmitting device to the receiving device. The more
`bandwidth you use, the fatter the hose is going to be. So in
`802.11n, we're using a 20 megahertz channel. There are other
`technologies out there such as like CDMA that only use the 1.25
`megahertz channel.
`(Court reporter interruption)
`MR. HURT: I'm sorry. The channel, the wireless
`channel bandwidth affected the more data you can get through.
`But to get that information through, you must pass through that
`channel and that channel impairs and degrades the signal.
`So let's look at a typical WiFi environment. Here
`assuming your home office, you have a transmitter device called
`an AP going to your client. The signals are going to travel
`through that space. You might have a direct line path that
`
`1 2 3 4 5 6 7 8 9
`
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`9
`
`
`
` 10
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`goes from the transmitting device to the receiving device. You
`may have a path that bounces off the wall or you may have a
`path that bounces off your couch. Those three paths combine at
`the receiver. This is known as a multipath environment. It is
`the multipath environment that is one source of channel
`degradation. The signals bounce around the environment, they
`arrive at the receiver with different replicas at different
`times.
`
`Another impairment is what's known as signal fading or
`variation in received signal power. You can see, as you might
`expect, the further away you move from the transmitting device,
`your received signal gets lower. Here we have an example of
`the actual received signal. You see that the signal is moving
`up and down and doesn't follow that straight line path. Where
`does that come from? That comes from what's called small scale
`interference. This possible small scale interference is a
`result of the multipath environment, the signals bouncing
`across the different objects in the environment and then
`combining at the receiver either constructively or
`destructively.
`Channel estimation. This is an important concept
`particularly to the '842 patent. For a receiver to actually
`receive the information from the transmitter, it needs to know
`what the channel did to the signal. To do so, the receiver
`needs to know in advance what the transmitter is actually going
`
`10
`
`
`
` 11
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`to be transmitting. The '842 patent is about training
`sequences. As an example, we take the transmitter. It sends
`the signal through the channel. We see that the channel
`degrades the signal. It does something to it. The receiver
`gets that signal and it needs to look at it. It says hey, what
`did I receive from the channel? Oh. I see something that's
`distorted from the known signal that I'm expecting to receive.
`Once it sees that signal, it can correct for it. It makes that
`correction and says okay. Now I know what the channel is going
`to do to my signal.
`So moving on to the second part of the tutorial,
`frequency and time domain. I like to use an analogy for
`frequency and time domain. Here on the left, you see music
`notes on a scale. To the right, you see a speaker. The notes
`on a staff represent the frequency domain. These are the
`frequencies that you want to hear. But you don't actually hear
`those. What you hear is the time domain sequence or the sound.
`Something in between was transformed, the frequency into time.
`What does that? In this case, it's the piano. It's the
`transformer. It's the device that converts frequency, notes,
`to sound, time.
`Here's a visual demonstration. You can see as I take
`the frequency to the left, the period of the wave form
`increases. This corresponds to the low note on the scale. As
`we increase the frequency, the period of the wave form
`
`11
`
`
`
` 12
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`decreases. That will correspond to the high note on the scale.
`A signal can be described in both time or the frequency domain.
`They're effectively equivalent representations of the same
`signal, but they're described differently. One is saying here
`is what you look like in time, the other is saying here's what
`you look like in frequency.
`So here's an example. Here's a cosign of 128 hertz.
`This is saying I'm a cosign and the one sample is 128, that's
`what I want to transmit. You take the Inverse Fourier
`Transformer, this signal, you end up with an actual cosign wave
`in the time domain at 128 hertz. Similarly you take a 256
`hertz cosign wave. You have a single sample saying, I want
`256. Take the Inverse Fourier Transform of that, you end up
`with a cosign 256 hertz.
`So you might ask yourself, what happens if I combine
`them? What is this going to look like? So we put on the left
`both 128 and 256, take that Inverse Fourier Transformer. What
`do we have? We have something that doesn't look like a cosign
`wave anymore because the signals have combined and now we have
`the combined representation of both 128 and 256. We know that
`the time domain signal on the right was synthesized or created
`from the frequency domain signal on the left.
`Moving briefly into OFDM or Orghogonal Frequency
`Division Multiplexing. I want to explain exactly what OFDM is
`compared to some other techniques and talk a little bit more
`
`12
`
`
`
` 13
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`about subcarriers. So before we get specifically into OFDM, I
`want to talk a little bit more about wireless spectrum.
`I'm sure, as your Honor knows, back in the 80's, back
`in the 90's, we had radio bands. Oftentimes we would have to
`scan our FM radios to figure out what music channel we wanted
`to listen to. Here you can see as we scan the FM radio band,
`the frequency or peak of what channel we want to tune to
`increases. Once we see that specific channel, we go ahead and
`tune back to there. And we see, boom, here's the signal that
`we want, here's the frequency at which it was present.
`The point being here is, a signal may be transmitted
`at different frequencies, as if using different channels of a
`FM radio without changing the information content. What this
`means is that you can have the same song playing on 88.3 as
`91.1. They're on two different frequency channels, but it's
`the same information content. It's the ability to send that
`information on separate frequencies at the same time. That's
`the basis of OFDM.
`So going back, what is OFDM? Here is an analogy I
`like to think of. Going back to the hose or that fat pipe, you
`have a single fat pipe of water. That's your bandwidth in a
`single carrier system. We're going to take that pipe of water
`and we're going to divide into multiple independent parallel
`streams, like from a showerhead. That's the picture to your
`right.
`
`13
`
`
`
` 14
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`So why would we use OFDM? OFDM is more efficient.
`Here is a spectrum comparison for the same data rate
`transmission, if we use multi-carrier, multiple faucets or like
`an FM radio we have to have guard bands in between each
`station, but we're able to go ahead and use every channel on
`that FM radio band to transmit data, or we can decide to try to
`glop all that data together and do something called "single
`carrier."
`Now single carrier when you spread the data rate, it
`causes bandwidth to expand. That's the basis for a technology
`called CDMA, which was actually invented here in San Diego by a
`company called Qualcomm. A similar technology called frequency
`hopping spread spectrum was actually invented in the '40s by
`Austrian-born actress Hedy Lamarr. She, during the 1940s,
`worked with our allies to help the Allies defeat the Germans by
`coming up with a system that would hop frequencies to overcome
`the German jamming of the Allied torpedoes.
`Similarly though when you take away from a single
`carrier, we can crunch even more. We can get down to OFDM
`because we're able to overlap these subcarriers and these
`signals in a very special way. This is a very similar slide to
`what co-counsel has shown you before.
`I want to point out a couple key things about this
`one. When you look at the peak of the red signal, you'll see
`that all of the other colors go to zero. That's what it means
`
`14
`
`
`
` 15
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`to be orthogonal in the context of the '842. The '842 is
`saying use all these different signals, use them in a
`non-interfering way to bring the data across all subcarriers.
`The subcarriers spacing is an important feature in
`OFDM. They need to be spaced at certain regular spacing so
`they maintain orthogonal. In this case, we call that Delta F.
`And the K or the index value is just a number how far away from
`the center.
`So in 802.11a, there are 52 subcarriers. They range
`from -26 to +26. In 802.11n, the technology used today, we go
`from -28 to +28. You've added four subcarriers that we're
`using. But to be clear, those subcarriers were already there.
`There are 64 defined subcarriers in the system. The question
`is not were they added: Were they used. That's the primary
`difference between 802.11a and 802.11n.
`The patentee did not invent subcarriers. They were
`present. They were simply not used before. In fact, none of
`the stuff I discussed today so far was invented by the
`patentee. All this was known technology, known techniques.
`So moving quickly into the 802.11 family of standards.
`I know this is a busy slide. I just want to point out a couple
`of things. In 1999, 802.11a was introduced, using 20 megahertz
`of bandwidth channel. It was based on OFDM. It's max data
`rate was 54 megabits per second. Ten years later in 2009,
`802.11n was introduced. It also has an option or capability to
`
`15
`
`
`
` 16
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`use 20 megahertz channels, also based on OFDM technology, but
`its max data rate goes to 600 megabits per seconds. WiFi has
`evolved both in the technology used and the max data rate that
`it supports.
`The key thing about 802.11 though was is it was
`designed to be backwards compatible. That meant the older
`devices and newer devices need to be able interoperate
`together, but more fundamentally, it put constraints on newer
`standards. The standards cannot go and change things that the
`older devices are expecting to see. So during training
`sequences, the values that the receiver is going to use to
`determine what the channel did to its signal already defined
`the value. The BPSK value for that subcarrier, it cannot be
`changed.
`The '842 patent was about determining those four
`values that they're going to use on the two extra subcarriers
`on the left and the two extra subcarriers on the right. That's
`the invention. That's what they're claiming, this inventive
`sequence that's four defined values for subcarriers on the left
`and the right.
`Here it is. This is the actual 802.11a training
`sequence. Again, 64 subcarriers already there, existed the
`entire time. Only 52 were active and it has a -26 to a 26 with
`dc with a zero index not being used. This training sequence
`was already defined in 802.11.
`
`16
`
`
`
` 17
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`The sequence itself is on the bottom. 53 subcarriers
`is OFDM training symbol, modulated by a sequence of L. Those L
`sequence values are all +1 or -1s BPSK. You can think of a
`training sequence just like the notes on a scale. The receiver
`knows what the transmitter is going to be sending during this
`training sequence. It's used so the receiver can figure out
`what did the channel do to my signal.
`So 802.11n came along. What do we want? What do we
`always want? We want better, faster, cheaper. 802.11n
`increased the data rate from a 54 megabits per second to 600
`megabits per second. Many different ways for the system
`designers to achieve that goal. One of the ways they achieved
`that goal was to increase the used subcarriers.
`So again, only have 64. Using 52 in 802.11a.
`802.11n, all right, let's use four more. What enabled that was
`improved digital filtering technology. Technology not invented
`by the patentee here.
`So now instead of having six subcarriers on the left
`and five on the right, we can decrease it, use those extra
`subcarriers to carry more data. To do that, you have to define
`values for those subcarriers during the training sequence, so
`you can determine what did the channel do to that specific
`subcarrier.
`So the actual patent itself was a patent application
`filed by Broadcom during the 802.11n standardization process.
`
`17
`
`
`
` 18
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`The specification disclosed the exact same training sequence as
`specified in the eventual standard. Again, you were required
`to start with 802.11a. They didn't invent the entire sequence.
`It was already there for them. Here it is again, the 802.11a
`training sequence.
`Now I want to talk a little bit about peak-to-average
`power ratios as counsel discussed as well. Here's the sequence
`on the left. You take the Inverse Fourier Transform with this
`the sequence, you end up with this sequence on the right. The
`sequence to the right is the power sequence that you actually
`will get out when you take the Inverse Fourier Transform.
`I have shown the solid red line, the average value.
`And the dotted green line is the peak value. As counsel
`indicated, depending upon the ratio of the peak to the average,
`it's going to matter how much variability you have going into
`your power amplifier. The more variability, the more back-off
`you need. So he's right, minimizing peak-to-average power
`ratio is an important aspect of OFDM's system.
`But one thing I want to know, if you look at the left
`sequence, the one in frequency domain, it consists of only +1s
`and -1s. If you take the power of that sequence, its peak
`power and its average power are identical. They are both 1.
`Because when you take a 1 or a -1 and you square it or multiply
`it by itself, 1 x 1, 1, -1 x -1, 1. So it's the peak in the
`average in the frequency domain where a sequence is defined by
`
`18
`
`
`
`19
`
`BPSK is 1. It has no peak-to-average because they're exactly
`the same thing which means in the context of the '842 patent,
`peak-to-average power ratio is a time domain property. There
`is no peak-to-average power ratio for a frequency domain
`signal.
`Here is the '842 patent and the 802.11n training sequence.
`The four red dots, that's the supposedly inventive sequence of
`the '842 patent. Again, those subcarriers already existed.
`They were already there. What the patentee had to figure out
`was what do I want to put on these four subcarriers? Do I want
`to put a +1 or do I want to put a -1 because there were only
`four additional subcarriers and we were restricted to +1s and
`-1s, there are only 16 possibilities the patentee could have
`chosen from. It turns out that this selection, 1 out of 16,
`this is the one that gives you the minimal peak-to-average
`power ratio when converted to time domain.
`That property of the peak-to-average power ratio in the
`time domain is an inherent characteristic of the frequency
`domain sequence that you selected. Had you changed any one of
`these red dots from a +1 to a -1 or take a -1 to a +1, the
`corresponding peak-to-average power ratio will go up.
`Let's go ahead and do that. I take the Inverse Fourier
`Transformer, the extended long training sequence defined in the
`'842 patent and, again, we get to the right a power domain
`sequence. And you'll notice the peak-to-average power ratio
`
`1 2 3 4 5 6 7 8 9
`
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`19
`
`
`
` 20
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`from 8011a to 802.11n went up just a little bit. It went from
`3.2 dB to 3.6 dB. dB is a relative scale that engineers like to
`use. Approximately 3 dB is a factor of 2. From .2 to .6 is
`just a smidgen more. Not a big deal. But the patentee and BNR
`are correct, you do want to try to minimize this. But you only
`had four values to mess with to figure out how you wanted to do
`this. Thank you, your Honor.
`THE COURT: Okay. All right. Do you want to start
`with the first term that's at issue here?
`MR. HARTSELL: I understood that the defendants would
`be presenting first since they're the ones who put this term up
`for construction.
`MR. HURT: I'm happy to present first, your Honor.
`THE COURT: Okay. Go ahead.
`MR. HURT: Do you mind if I do a have a quick swig of
`
`water?
`
`THE COURT: No, go right ahead.
`MR. HURT: All right. So we're here back to talk
`about the proper construction of the Inverse Fourier
`Transformer. Here is the claim language:
`Wherein the Inverse Fourier Transformer processes the
`extended long training sequence, which we've discussed quite a
`bit before, from the signal generator and provides what? An
`optimal extended long training sequence with a minimal
`peak-to-average ratio.
`
`20
`
`
`
` 21
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`On the bottom left, I've shown again, here is the
`extended long training sequence of the '842. Take the Inverse
`Fourier Transform, you end up with this. It has
`peak-to-average power ratio of 3.6 dB.
`Let's look at the proposed constructions. Defendants
`propose: A circuit and/or software that performs a defined
`mathematical function that transforms a series of values from
`the frequency domain into the time domain.
`BNR proposes: Plain and ordinary meaning, or
`alternatively, circuit and/or software that at least performs
`Inverse Fourier Transform.
`Let's talk about the first, plain and ordinary
`
`meaning.
`
`THE COURT: This is a fundamentally, perhaps, stupid
`question, but why does it bounce back and forth from Inverse
`Fourier Transform to Inverse Fourier Transformer?
`MR. HURT: So the transform is the actual defined
`mathematical formula or the function. The transformer is just
`something that implements that function.
`THE COURT: And this is clearly something that people
`who do this stuff would recognize. I mean, it's written in the
`patent in initial caps. So while I'm not exactly sure what it
`is, I would suspect you certainly would know, people who would
`practice this sort of technology are going to recognize this,
`and I think in both the briefing, it was recognized that this
`
`21
`
`
`
` 22
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`is a mathematical function that an electrical engineer is going
`to recognize. They're going to know what this is.
`MR. HURT: Yes. Absolutely agree with your Honor.
`THE COURT: So why do I have to do anything more with
`it when it says that it's there, it's operatively coupled with
`the signal generator, and it's going to process this extended
`long training sequence from the signal generator and provide an
`optimal extended long training sequence with a minimal
`peak-to-average ratio? Isn't it just doing what the formula
`does?
`
`MR. HURT: Yes, your Honor. But two reasons why you
`should construe this term. One is to provide -- resolve the
`dispute between the parties as to the exact scope of this claim
`term. Second, provide clarity and guidance to the finder of
`fact. Going back to the first, we have a fundamental dispute
`with respect to what defendants believe the Inverse Fourier
`Transformer of the '842 is doing relative to what BNR proposes
`that an Inverse Fourier Transformer in the abstract can do.
`BNR has already proposed and argued that the Inverse
`Fourier Transformer can be multi-dimensional, can operate
`between multiple domains. Defendants do not dispute that
`mathematical concept in the abstract. What defendants -- our
`concerns are is that even when we get to expert reports, if we
`have a fundamental dispute, the arguments are not going to be
`joined. We're going to be talking about the '842 Inverse
`
`22
`
`
`
` 23
`
` 1
` 2
` 3
` 4
` 5
` 6
` 7
` 8
` 9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`Fourier Transformer, the one that takes frequency domain
`signals into time domain. BNR will be talking about this
`amorphous transform that can -- according to them, can do
`anything. It can take any number of dimensions, go anywhere to
`any space to any other space. Yet the '842 patent never talks
`about anything else other than frequency in time.
`THE COURT: So fundamentally, it's not this
`mathematical functionality but rather that this claim is
`directed as a wireless communication device that comprises this
`transformer?
`MR. HURT: Absolutely, your Honor.
`THE COURT: In the context of the claim language
`itself that says this is a transformer that is comprised in a
`wireless communications device, your argument is how it
`operates that mathematical principle is limited?
`MR. HURT: Absolutely.
`THE COURT: And limited to this frequency into time
`
`domain?
`
`MR. HURT: Yes, your Honor. That's exactly correct.
`THE COURT: Okay. And now help me find out other than
`beyond the fact that it says that it's a wireless
`communications device, why would someone recognize that?
`MR. HURT: Absolutel