Let's get started. I have to decode handwritten notes on the second report, but you can get going now with our first report.
I'll add the second reporter's notes as soon as they are done, and so stop back by.
Judge James Robart - Room 14106 - Wednesday, 11/14/12
Judge Robart opens with a comment that "software is not as exciting as strip clubs" -- after noticing that the crowds have thinned considerably.
Exhibits to be sealed 3417, 502, 2238, et al. Parties agree that Motorola has others to be sealed. Courtroom may need to be cleared in the afternoon. [Note: This did not happen.]
Microsoft Resumes Direct Examination of 4th Witness, Gary Sullivan:
[Microsoft employee: Principal Software Developer/Engineer. Job title is Video/Image Architect.]
Explanation of 50 percent compression ratio. Motorola made no proposals to the groups before this was achieved. Graph used to show that the VCEG draft had already got nearly the compression rate that was in the final. Proposal doc discussed. Motorola had not submitted docs prior to that. Motorola was supporting interlaced video in the standards groups.
MBAFF discussion. Similar to what was in MPEG2.
PICAFF discussion. Motorola did not create PICAFF.
Reference to email from 2005 to a Ms. Grey about policy of patent pool vs. standards formation. He told her that those activities are separate. Email refers to distinguishing between patents that are in a fundamentally important category vs. a "bozo tweak." Refers to caution in his emails saying he is not a lawyer, should contact a lawyer for legal advice. Believes Motorola did not contribute any fundamentally important patents.
Motorola Cross-examination of Gary Sullivan:
July 2001 contribution of Motorola to JPEG. Discusses lack of interlaced coding tools. In some cases performance was improved. Graphs to show performance comparison. Motorola proposed interlaced coding tools in December of 2001. In 2009 stereo high profile was updated to use interlaced coding tools. Paper published at conference, describes stereo high profile. Another extension was multi-view high profile. Very technical discussion. Others besides Motorola asked for interlaced coding tools to be included. Interlaced video was included in later versions of WMV. VC1 included interlaced video. DirectX VA to support interlaced content without first converting it to progressive--quoted from Microsoft web site. Series of papers he wrote--includes mention of high coding efficiencies for interlaced frames. Field mode vs. frame mode. More tech details. MBAFF vs. path coding. Was reported on a few "sequences." Back to prior performance graph--that graph did not include interlaced video rates. Video compression paper he coauthored. Magazine article he wrote mentions MBAFF. Another article he wrote for a conference--mentions patent licensing and patent pools and that patent holders are not forced to join pools. Some patent holders are not happy with the terms,and thought it might take up to 20 years for licensing issues to settle. Says he was being facetious--in effect saying there are no guarantees until the patents expire. Also mentions that this struggle with licensing issues will not affect standards adoption.
Mentions organizations that are supporting the H.264 standard. Specification on DirectX video acceleration, includes MBAFF support. Email to Patty, inviting her company to attend a meeting; Patty is concerned that patent pools are not equitable for companies that have more significant IP. He responded that pools and open standards are not related. Responds that he had cautioned Patty that he is not a lawyer and is not qualified to give legal advice in this regard. Email explains that the RAND requirement does not mean they have to price her company's IP cheaply. Refers back to the above "bozo tweak" remark.
Refers to PhD thesis. Not sure how available it was on UCLA library but he gave about 20 copies to friends. Talk about patents and various companies' contributions. Telinor contributed some tech that they said they had not patented and would not patent. Any way to tell if technology considered for use in the standards may have been covered by other companies' patents? Confirmed that it was not his role as chairman to determine this. But he stated that if IP was proposed for use that would not be available under RAND, then that tech would not be used in the standard.
Refers to the standard itself.
Tech discussion on SEI messages/timing. VUI-video usability info. Very technical talk.
Microsoft Redirect, Gary Sullivan:
Mentions that VCEG did not have interlaced video support until after the JVT collaboration because it was considered older tech. There were other ways to handle interlaced video. Defines a "sequence" as a video snippet of maybe 10 seconds or so (sequences were used in some of the performance graphs in exhibits). Refers to doc for Motorola stating that interlaced video was not as well supported. States that MPEG2 has interlaced tools that perform as well as H.264 for interlaced video. Tech discussion on coding types using frames vs. path. Asked why DirectX had interlaced support. Stated that they wanted to support the entire standard. VCEG draft of 1999 was essentially Telinor's proposal.
Motorola Recross, Gary Sullivan:
Back to his email to Patty--"not a lawyer" reference--says these are personal opinions. But that email does say he strongly believes what he wrote and that most people would share his opinions. Refers to other doc with H.264 vs. MPEG2 performance graphs. Agreed he did not personally test those using interlaced video.
What was changed in the two designs to achieve the desired bit rate savings?
Answer: Many things involved in that design. Discussion of importance of supporting interlaced video. Microsoft wanted to comply with the standards.
Did Microsoft seek to have their patents be declared as standards-essential?
No follow up.
Microsoft's 5th Witness, Jennifer Ochs:
Works for Marvell who makes semiconductor chips. Director of IP litigation. There for 3 years now. Has a masters in EE. Microsoft and Motorola are customers. Sells Wi-Fi chips for xBox. Chips cost $3 to $4. Used by many other products including Sony Playstation and even in cars. Chips are commodity products.
Marvell engineers are involved in the 802.11 process and has key patents. Patents they have are very important for newer versions of 802.11.
Letter was sent to Motorola asking for RAND license. Letter was sent at Microsoft's request for indemnification. She understood that Motorola would offer FRAND terms for licensing. Emails and letters back and forth. License was for products that use their chips, but Microsoft and Apple were excluded in Motorola's proposed license. She stated that was discriminatory.
Royalties would end up being more than the cost of the chip if costs of the end products are used. Compares to use in an Audi A8--royalties would be over $2000 for that $3 to $4 chip. Not practical for them because Marvell may not know in advance the chips' use when sold. She has not heard of this end-product rate used for chips' licenses before. States that one percent on chip cost would be considered a high ceiling for a patent licensing. She does not consider 2.25 percent to be a reasonable starting point.
Motorola Cross-examination of Jennifer Ochs:
She has negotiated RAND licenses for chips before. Details about what is part of such license discussions. "Defensive suspension" terms in RAND licenses discussed. Mentions it is possible to have a RAND license with no royalties paid if there is a cross-licensing agreement.
Agreed that Motorola told Marvell that it would not assert its patents against Marvell because Motorola had a program to license end products.
Asserted that Marvell was contractually obligated to try to indemnify Microsoft.
Review of letters on the licensing agreements and pending litigation with Microsoft. Letter states Marvell believes that it does not owe Microsoft an indemnification--that the legal issues are between Motorola and Microsoft.
In the past Microsoft had not asked Marvell to get licenses for patents.
Review of emails from a Mr. Kowalski of Motorola about licenses. Months later, Marvell had not responded to Motorola's proposals. There was a meeting set up, but canceled. This meeting was to have Motorola review the value of Marvell's patents. Then later Motorola offered a counter-proposal. Marvell is working on a counter-offer in reply. Discussion of time-line for the proposals and meetings to discuss. Appears that Marvell was not moving quickly to get these licenses. [PJ: Is it possible that this unusual move by Marvell was just to get a negative and some information for Microsoft to use in litigation, as opposed to a sincere effort to get a licenes? That seems to be what Motorola is implying.]
Microsoft Redirect of Jennifer Ochs:
Does not agree that Motorola and Marvell's patents are equally important--that Marvell's patents were more important, but proposed a royalty-free license simply to help Microsoft get the indemnification they wanted.
What else do their chips do?
A: These chips only do Wi-Fi.
Background on profiles--what types of uses need more or fewer licenses? Who are Marvell's competitors?
A: Intel, Qualcomm, etc.
Microsoft's 6th Witness, Prof. Michael Orchard:
Before he started, attorneys for both sides wanted to tell the judge that the concept of profiles only applies to H.264 licenses, not to 802.11. (Referred previous testimony and question from judge.)
EE professor at Rice U. Other background. Research on image and video compression. IEEE Fellow. Followed the development of the standard, talked to Sullivan. Says Motorola patents are not critical; some are about to expire. Not critical to Microsoft products. Reading Motorola and Microsoft patents into the record. Much longer list for Microsoft.
Judge says his criminal calendar may allow convening on Monday and finish on Tuesday.
Microsoft Continues with Prof. Orchard:
Definition of video compressions and why it is needed. Encoding and decoding, coding tools. Exhibit on difference between progressive and interlaced video. Time of capture determines the format. Interlaced is sometimes converted to progressive, but the reverse is not true.
Coding tools increase the efficiency of coding. Background on interlaced. Virtually all modern displays on TV and PCs are progressive.
MPEG2 was the first compression standard for interlaced. 14 Motorola patents in the standard are for interlaced video.
Refers to VCEG contribution docs. He has reviewed them all. Some are key, some trivial. 170 companies contributed about 3000 docs. Motorola contributed 25 docs. All of Motorola docs that were accepted related to interlaced video. Microsoft contributed docs that significantly improved the standard--many related to improved predictions. None related to interlaced.
Discussion on what is essential to the standard. MPEG LA doc relating to essential patents reviewed. Patents are broad and rich in what topics are covered. Summary doc he prepared was reviewed. Includes non-MPEG LA patents. Considers the Motorola patents miniscule. Motorola claims 16 of their patents are essential and 14 relate to interlaced. Doc shows categories (field coding, motion vector prediction, etc.) that each patent applies to. Says that there were acceptable alternatives to Motorola patents were available and use of alternatives would not have degraded the standard.
Motorola objects to exhibit prepared by him. Overruled. Entered as demonstrative.
Discussion of adaptive frame field coding. This is an interlaced coding tool. PICAFF vs MBAFF exhibit (a picture). Exhibits of pictures showing single macroblock vs. paired macroblock MBAFF. No difference that he can see. Motorola patents were for the paired macroblock. Motorola did not compare single vs. paired--they compared them to things that were not good alternatives.
Contribution doc from Video Telly discussed. PICAFF use predates Motorola's patents. Alternatives were available, e.g., MBAFF. Tests showed that PICAFF did not perform as well as MBAFF. Motorola did not invent the use of scanned paths. Sony had submitted scanned path proposals too. Compares Sony to Motorola scanned path proposals, but no direct comparison tests were done. Says they are roughly equivalent, but thinks the Sony proposal was slightly better. Motorola did not submit a comparison to the standards group. Docs shown that others had compared Sony vs. zigzag scan paths and Motorola vs. zigzag, and Sony's were better.
Doc showing prior art for Motorola's patent. Opines that this patent is trivial. Other patent ('419) claim is for motion compensation, which expired over a year ago. Discussion on related German patent. The 1991 patent should relate to hardware, not a general algorithm.
Exhibits 462 and 432 -- Motorola objection due to lack of foundation. Discussion of foundation by Prof. Orchard. Argument that prior art is not established, but overruled.
968 patent. Motion compensation. 1993, will expire in March. Another related German patent. Like the earlier one, this should relate to hardware, not software algorithms. Other prior art exhibit entered.
Identified 40 Microsoft patents as essential, reviewed them. Considers them fundamental advances. Patent numbers read [zzzz...]. Other patents--contributions that add features but may not be used by everyone. More patents--ones that add advanced functionality.
Refers to Gary Sullivan's testimony. Relating Motorola's patents to Microsoft's products like xBox, but he considers them unimportant. On to Win/7 video use, shows an exhibit on most common video usage by PC users. DVD playing on PCs uses MPEG2 for interlaced, not H.264. Motorola's examples are not commercially relevant, one was from a Pirate Bay site. YouTube does not use interlaced. Shows exhibit saying that YouTube will not support interlaced--videos must be un-interlaced prior to uploading. Windows does not use the interlaced coding tools. Windows phone and other Microsoft products do not support interlaced., including Skype, Lync, Silverlight, Zune. Again says that Motorola patents are very minor, unimportant.
16 vs. 17 patents in exhibit vs. the number he stated?
A: One was removed.
Motorola Cross-examination of Prof. Orchard:
TV has provided interlaced content for decades. Newer TVs use H.264. So do satellite TV providers. Compares DVR use and Microsoft encouraging PC users to record broadcast TV.
Exhibit from Microsoft web site about Windows Media Center, saying you can watch and record TV including HDTV. He does not know the details. Also did not know that it allows TV shows to be sent to xBox or a TV. Did not know that xBox can be used as a Windows Media extender so that all recorded TV shows can be shown on any TV in your house. Mentions other entertainment providers, including Netflix, Uverse, many others worldwide. Accepts that such interlaced content would be decoded by H.264. Would also include formats used by older camcorders. May want to upload such content or burn it to DVD. Back to YouTube, makes a point that users must have interlaced content to upload because YouTube provides instructions on how to handle it.
Did not analyze other MPEG LA pool patents other than Microsoft or Motorola's or assess their validity. Did not do an infringement analysis of all Motorola's patents. Did not conduct any tests of validity of Motorola's patents.
End-of-day Administrative Matters Handled by the Judge:
Talked about the number of hours allocated to each side; the judge is keeping track.
Depositions need to be "moved to be published." (Similar to how exhibits are accepted into evidence. Judge had earlier cautioned both sides about forgetting to formally request that exhibits be entered. He did not want them to back-track to do that.
There is a bit more to the notes, but probably the best thing to do is to post the
Microsoft v. Motorola - Courtroom 14106 - 9:00 AM (Day 2)
Arrived at 8:40 AM. Current display is of the H.264 development timeline, along with a video display for documents.
On the right, there are seven people at a table. On the left, 6 (though people are moving about a bit). Four more on the right or a bench inside the gate. No bench on the left. Five people are in a public area, in pews, right behind the sticky note saying: "Sullivan, Ochs, Orchard, Delcosteo, Gibson." One per line.
Behind that, there are two people in black, then behind three more, one with documents and a red pen, then me. One guy with no jacket, in jeans, with a MacBook. Then a guy with a laptop and in tweed sportcoat. On the left, one guy on the first pew, then a second with white hair in a tan suit, with a MacBook. A woman standing. One guy has an iPad. I think on the left is Microsoft.
More people arrive. On the left, the lawyers are using mostly Dell and Lenovo laptops. There is one woman at each lawyers' table. 24 people outside gate. On the right, most at the table seem to be using Firefox (saw error page)
Judge: "Software still doesn't hold a candle to strip clubs." Judge and lawyers discuss confidential exhibits and various housekeeping issues. Lists provisional sealings, and there is a request to have room cleared when future products are referenced.
Then Dr. Sullivan takes the stand again.
Dr. Sullivan's testimony was
primarily on the details of what contribution Motorola made to the H264
standard and when it was made. The H264 standard began with the VCEG process
which produced the "TML" encoders. The last of these was TML-9. The VCEG
process then gave way to a joint video standard which had the "JMT" (or
maybe JM) encoders. Of these JMT-6.1 basically reflects the final encoder in
the standard. A series of graphs were shown that were taken from papers
showing the bit rate vs PSNR of various encoders. The takeaway from this was
that the vast majority of the improvement from MPEG-4 and MPEG-2 had already
occurred with the work done by VCEG. However, the VCEG encoders were worse
on interlaced video than MPEG-4 according to the report Motorola provided.
This was to be expected as interlaced video had been entirely left out of
the standard thus far so the interlaced specific features of MPEG-4 weren't
Dr. Sullivan was never involved in any licensing negotiations, but an email
he wrote (exhibit 2345) to Patty from E-Prepit (?) was concerned about her
company contributing patents and if they'd end up in a patent pool and how
standards bodies and patent pools were related. Sullivan wrote 'not a
lawyer, but ... "
The majority of the testimony was the lawyers going 'stuff stuff stuff, is
this right?' or 'stuff stuff stuff' 'oops forgot to phrase that as a
question' or occasionally 'stuff stuff, right?' 'not sure about that' 'let
me refresh your memory based on this sworn statement you made at this other
time' So mostly I paraphrased as the lawyer spoke and then modified it when
the witnesses occasionally dissented. I tended to abbreviate Motorola as MT
and Microsoft as MS.
Sullivan: By time VCEG had finished proposal "50%" compression improvement.
"What is that?".- Overall reduction of 50% varies
by video sequence. Unclear cut in half from what. While achieving 1/2 the
number of bits.
How did Summer 2K1 compare. Very similar to VCEG final standard.
Motorola made no proposals that were part of VCEG standard.
424a -> (chart of bit rate charts. TML - 9.
JMG.1 - approximate final standard. Joint video standard.
Bit rate vs. PSNR. By time it was joint project. Vast majority of
improvements already made. Motorola did not contribute to JML-9.
2271 - Submission to MPEG by Motorola - contribution doc. contains no
proposals has report on performance on existing work.
Special features for interlaced coding features. Motorola made JDT
proposals for interlaced video, MPAFF was in MPEG 2 as well and MPEG4.
Some MT contributions [side mark] MPAFF not originated by MT. Best
alternative very similar to
that contributed in MPEG2, copying frame switching. "Paired
macroblocks". PICAFF in prior MPEG2. MT did not orig PCAFF no significant
Email in 2005. 2345 exhibit unpublished quoted in MT brief.
Extensively. (not on screen) Patty Gray for E-Prepit (?). She was
concerned about IP standards and wanted to know how pools worked.
Never personally in a licensing negotiation or involved in a
patent pool. Should not mix up patent pools and standards. Pools
happen outside of standards. Colorful language. You decide if your IP
is in the fundamental category or the "bozo tweak" category. Patent
holders decide. Chaired JDT. Viewed MT proposals. Did not consider
MT contributions fundamentally important.
Cross-examination: Pinstripe suit glasses --> Mark Roland met during
deposition. July 2001 Contribution to MPEG by Motorola. 2271
comparing MPEG-4 to H.26L Attributed to lack of interlaced coding tools.
[Exhibit 2271 is a paper and figure 16 is a graph]
Comparing UVLC configuration of H26L and MPEG4 with and without
interlaced video. Don't know what that is, UVLC. Sullivan mentions
UVLC each time on particular test sequence/configuration bit rates.
Figure fifteen. Another similar chart. "relatively low range of
fidelity". MT proposed other interlaced coding tools in Dec. 2001.
As first released had multiple profiles. Baseline main, extend.
Interlaced content tools in main profiles. May 2K4 included a "high"
profile also included interlace content.
2K9/2K10, "stereo high profile" also has interlaced coding tools.
3398 --> paper by Sullivan (MS person) at a conference. "Standards
based approach" to stuff. Section 2.4 has discussion of stereo high
"Multiview high profile". Did not support interlaced video. Might qualify
"could not be coded as individual fields" meant PCAFF
(picture level). Industry requests received for interlaced and joined
MS added interlaced to WMV or VC9 as a standard. Added in VC-1 WMV-9 [a
specific version of WMV]
may have gotten VC-1 later. V-A had interlaced. And some version with
interlaced in one profile has VC-1.
3381 - in reference to main goal of VC-1. MT lawyer said it was
encoding interlaced without first converting to progressive. 3381
agrees with MT lawyer. From MS website.
Has a series of papers related to H.264. Page 566 of exhibit 4224 - 742
[legalese page number on bottom right ends in]
Adaptive frame field coding stuff. Explaining that H.264 lots
choose different way of encoding video. Frame mode and field mode.
Encoder could put a progressive video in field mode (but you probably
wouldn't want to). Pairs of macroblocks. PAFF coding reduced bit
rates by 16% on some sample sequences. "Reducing bit rates is a good
thing." On some sequences MBAFF reported better than PAFF by 14%-16%.
Back to Fig 18 424a does not show interlaced video performance.
These are not "high resolution". This is about 1/4 of standard
resolution. At least some in improvement after MT started some sort
Exhibit 575 - another paper by Sullivan. 2005. P. 27 of article
"several lawyers numbers" Adaptive frame field coding operation.
Interlaced has different properties than progressive (statistical).
574 - report magazine article, 2006, "core coding technology of H264.
Main innovative features. Discussion of MPAFF.
3399 -- mentions licensing. H264 and deployment status - conference
paper (does MT counsel not know conf. vs. industry paper
he keeps doing this [refers to an exhibit as an industry paper and getting
corrected]) Section on patent licensing and patent pools.
Holders don't have to use such pools. Licensing issues may take 20
years. Usage fees. Being facetious. Licensing issues do not appear
to be a widespread hindrance to deployment efforts. Short list of
people/companies that have deployed it Bskyb DTV-3 in
broadcast TV, bluRay.
DirectX video acceleration.
591 - DX video acceleration spec for H264 by Sullivan.
Software decoder talks to hardware. Some ophy Dx some by SW MPAFF. [I
think this is some decoding done by hardware, some by software]
Back to email 2345 - email chain. Invited Patty to talk at
workshop. "She was concerned about patent pools not being
equitable" MT lawyer - and in email response.
"Standards body have NO OPINION whatsoever on specific licensing pools"
Pools/standards have nothing to do with each other. "I'm not a lawyer
and she shouldn't rely on me for legal advice" 2005 -- involved with
Discussion of what RAND means. Proper licenses will be available
though it's not guaranteed to be cheap. Can charge more for
fundamental IPR than another for bozo tweak. You are in control.
Has not been in negotiations. "Bozo tweak has changed my entire"
perspective/view on software engineers". Laughter. Judge Robart.
Sullivan's thesis 618 - indexed etc. Not aware of any availability of
thesis. Turned it in August/September. Gave out about 20 copies to
friends. Oct. 1991. From 2002. Paper has references to publications
XIV - Sullivan and R. L. Baker (?) submitted to conf. for May 1992.
(would be publicly available later).
Sullivan looks slightly disturbed by questions about his thesis. [His
thesis listed something that was in the future -- it was written down as
Telinore doesn't have any tech covered by patents in standard; but Sullivan
doesn't know if contributions were covered by other patents.
Participants were supposed to make declarations.
Only evaluated things for patents unless indication it wouldn't be
there on RAND
No indication of this. Didn't analyze any particular patents.
Some sections of standard are informative. Annex B.3 Sullivan doesn't
Handed copy of something about 4 inches thick.
March 2010 B.3 is indeed marked as informative Ex 421.
Supplemental enhancement information. Decoder not necessarily
required to use if formally speaking. Picture timing SEI messages. Has
special status. In Annex C there's some remarks about that data and
timing conformance to the standard. MT lawyer - picture timing info
not required on page 308. Picture timing SEI messages must be given to
decoder but not necessity in the bitstream (other means not
VUI - video usability Information -- color description. Syntax
allows it to be present or not present.
Back to MS counsel
2271. Up until time of formation of JDT, no interlaced encoding tools
included. WEren't particularly interested in interlaced as it was an
old tech. Someone could do interlace or chop it up into fields, or
all pictures as fields without anything in std itself.
MPEG 2 and 4 had interlaced. Summer 2007 submitted doc criticizing
interlaced video perf, not surprising. They emphasized getting
Could have used tools from MPEG2 - no sig difference in H264 and MPEG2
424 page 567. PAFF reduced bit rates over frames only coding.
PICAFF/PAFF not comparison with MPEG 2 coding tools.
MBAFF 14%-16% better than PAFF. Not a comparison between MPEG-2 and H264.
Why did you include interlaced support in DX? Because it was in the
VCEG Aug1999 -- essentially was Telinore proposal in essence.
Young guy on MT bench seems to have found some patent point system
somewhere, but it's hard to hear as they are quiet (a parabolic
mic would be nice).
MT woman in green (and entire MT table) drinks FIJI water. Less of it
is visible on MS table (one old guy has a bottle that might be it
ReCross. 345 - email. Not a lawyer but don't understand how
different impression can persist. Receive cont where MPEG2 and H264 as
a whole were compared. 574 (magazine article?) Figure 7. H264 and MPEG2.
H264 is better (a shock? I think not) Did he run tests himself of
coding tools? No.
Judge Robart --> why do we have H264? replacing MPEG2. Main reason
compress better. Fewer bits/quality.
What changes were made? Methods of chopping, precision of info
increased. Filtering techniques, entropy coding more sophistication.
Are you familiar with what MT patents have been declared essential?
How frame field adaptive coding for interlaced video? How important is
this? In Sullivan's view areas are quite similar. Not aware of any
Sullivan would not categorize this as fundamental. But didn't say
Bozo tweak. Why is his chairing Reptor important? MS does video things.
Interoperability knowing what the standard is. [Reptor is the standards
body thing that did this whatever it's actually called]
Thinks streaming video was recognized as increasingly important.
Did MS seek to have any of its patents declared H.264 patent
essential? Yes, I think so.
50% improvement in bit rate for progressive scan. Would that have
been there for interlaced without MBAFF MBAFF in MPEG2 is
When talking to ITU do they [patent holders] ever indicate that they are
not ready to license
under RAND terms? There's a form, but we never had a submission that
used that form.
Do you need a license to each patent to practice? Well different profiles?
What are profiles? Profiles are packages for different subset. Generally
a decoder does entire profiles. You would get license to each patent
for the profile.
Jennifer Ochs next witness. (Some MT people leave to go get food/work)
[they discussed wandering off for this witness during the recess]
Marvell employee. Subpoenaed. Director IP litigation. Oversee patent
litigation, licenses, indemnity. Has familiarity with Marvell's
products. Years, then 97 - something as partner. MS in EE
Motorola and MS customers of Marvell. Sells WiFi chip to MS for xBox.
IEEE 802.11 they contain almost all functionality needed for WiFI.
$3-$4 per chip, and sells similar chips. Ranges from gaming systems.
PS3 to automobile. Audi A8 will have one. Gaming consoles from a few
$100 to $80K for Audi. Their chips are commodity products.
Several Marvell engineers are involved in standards. Marvell has a
few 100 issued patents in 802.11 in US. Their patents relate to
latest standards. N and AC, patents very important to AC.
Personally involved in discussions.
608 - letter to Motorola drafted requesting they provide a RAND license
signed by her subordinate. She drafted.
MS made request to Marvell to get one relating to [Microsoft's use of
Understood "committed to licence on" FRAND terms.
Our intent to get a license that would protect MS (and told MT this).
16 - email and draft license - Kowalski. Nov 5 2011
Did it include a license for Marvell's chip. Would cover products that
use our chips. Section on repudiating parties which are excluded
specifically MS and Apple. Considered license discriminatory. [when asked
by lawyer about ones that excluded some Marvell customers]
2.25% of net selling price. Assumed resulting price would change with
net selling price $4.50 for console > chip. $22.50 for a $1K
A8 $100K --> $2250. [actual answer was 'more than $2000']
Marvell did not consider them commercially reasonable, and they don't
know the NSP. Other people also have patents.
Profit margin on chips is small. Patent stacking is a concern.
Has not heard of a chip maker paying % NSP. [a percentage of net selling
Arm Holdings does 1% of chip price as benchmark. 1% of chip is high
License much more from more from Arm than 802.11.
Do you think 2.25 average selling price of customer end product, or
chip reasonable starting point for negotiation? No.
MT counsel. Palo Alto. Accepted email service. Not sure she's
compelled to serve. Negotiation can be very complex, time consuming.
Warrants, (missed something).
Geographic, products covered defensive suspension does not agree is
Testimony before ITC on Jan 18 page 1937 line 21-25 swore to tell truth
1979, line 3
Says typically there is a defensive suspension in that testimony.
Hadn't previously included them. [in Marvell's agreements ...becoming more
1980 - license you had previously offered
Cross licensing - affects royalty rate typically. Ochs "opening offer
should be same order of magnitude"
Marvell made counter proposal though redline without a number. Marvel's
important patents and Motorola's would cancel out to 0 royalties
MT never told Marvell they needed a license for 802.11, and had no
intention of asserting 802.11 against Marvell and has a history of not
requesting them against chip sets.
Marvell has never conceded to an obligation under contract to seek Rand
license from MS, but tried due to request.
1610 - not sure if original drafter. Letter Marvell to MS Leonard
Smith at ITC "I believe I drafted this one". March 2011
Nov 30 2004 agreement has exclusions to indemnity. Basically says unclear
Marvell needs to get MS a license, and MT seems to be retaliatory for MS's
start of legal proceedings.
In essence saying we don't owe indemnity and it's MS's fight, not ours.
1608 - another letter, drafted. Marvell has not requested a RAND
license for Motorola. 7 years after Marvell started selling chips to
July 2011 7 years after selling started.
They have on several occasions asked for indemnity.
Typically MS asks Marvell to pay attorney's fees.
MS doesn't say give us a license.
Nov 24 2010 - first request for RAND. Marvell waits 8 months to send MT
Oh also someone's phone did go off on vibrate on the MT side of the pews, I
think it was a suited person in the second row so a MT attorney type. He
let it go to voicemail rather than pressing a button to make it be quiet.
Dr. Michael Orchard. MS witness (was guy in a tan suit)
Lawyers explaining that profiles only a pply to H.264 by agreements.
Has C.V. as an exhibit. Professor E.E. at Rice. PhD Princeton and 2 MS in EE. Worked in DSP prePhD and since done consulting. Research. IEEE fellow. Here for Sullivan's testimony. [PJ: What??? Does this judge allow witnesses to listen to other witnesses?] Followed VCEG work closely. Made proposals himself.
Summary not conclusive. MT patents are a small sliver. Limited to interlaced. Marginal to H.264. Two other patents, one expired, another set to expire. MT's patents are mostly irrelevent to MS.
Back from lunch: Video compression. Processing to shrink video. Explanation of encoding/decoding. Coding tools are particular ways of shrinking.
Difference between progressive and interlaced. Progressive frame by frame. Interlaced is 2 images off in time. Progressive is never converted to interlaced. (Really? Isn't it used fro some broadcasts? I think he's flat out wrong on that.)
Interlaced video is from cathode ray tubes. Interlaced video not that common. Solution to a problem that no longer exists. Today's displays are progressive.