Plaintiffs,
v. 00 Civ. 277 (LAK) SHAWN C. REIMERDES, et al,
Defendants. ------------------------------x
July 21, 2000
9:00 a.m. Before:
HON. LEWIS A. KAPLAN,
District Judge
APPEARANCES
THE COURT: A couple of preliminary things on my list before we get started. First of all, have you been able to reach agreement or otherwise come to a view on the specific findings of the Microsoft case that I asked whether you would stipulate to?
MR. SIMS: Your Honor we haven't yet. We will over the weekend.
THE COURT: Just so I am clear, if I don't hear an objection on the specific findings that I have identified, I will take judicial notice of them and the deadline for any objections is 11 o'clock on Monday. As I say, they are in the nature of definitions and I --
MR. GARBUS: I can't see why it would be a problem?
MR. SIMS: I am advised we have gotten to it, we have no objection.
THE COURT: You might consider also because I will be looking further at that opinion for other things that might be useful by way of background here, to see whether there is a description in it of Linux that you can agree on as to what it is, what its relationship to Microsoft windows is. You may be able to do that, you may be able to save time. OK.
The second item on my list is that I am going to alert you to an issue I thought I had because I am going to ask for a brief on this if by the end of the trial I think it is relevant and you might as well start thinking about it. If I am understanding at least certain of the arguments here, the defendants are arguing that there are encryption utilities other than DeCSS that would decrypt CSS protected DVDs, that there is no proof, direct or otherwise -- as to direct I think it is stipulated -- that any particular decrypted movie that may have been offered over the Internet or on a file-sharing utility or whatever other relevant modes of distribution may be, that was decrypted with DeCSS as opposed to some other decryption utility and what the defendants I think would ask me to conclude from that is that the plaintiffs therefore can't have had any injury actually or threatened by reason of these defendants' actions.
Just by way of a parenthetical digression, I am well aware that the issue of the extent to which there is sufficient proof that there are in fact decrypted movies available over the Internet has not been decided and we are going to have further discussion about that. My question presupposes, just on a hypothetical basis, if there is adequate proof to find that there are such movies out there.
The thought that occurred to me last night and on which I may want briefs is whether the burden of proving that whatever decryptions are out there came from the defendant or more broadly DeCSS is on the plaintiff as distinguished from the burden being on the defendant to prove that they did not. The case that prompted me to think about that, I know we all read this in law school a long time ago, Summers v. Tyce, decided by the California Supreme Court in 1948 which seems at least superficially to be analogous to this problem.
The case involved a group of hunters out hunting with shotguns, two hunters fired simultaneously. One shotgun pellet hit the plaintiff. Because they were shotgun pellets as opposed to bullets out of a rifled weapon, it was impossible to tell which shotgun the pellet came from. The California Supreme Court said that the burden of proof was on the defendants to show that their shotgun did not discharge the bullet that hit the plaintiff and in the absence of such proof, everybody who fired a shotgun at that time was jointly and severally liable.
I am not inviting argument on that this morning, but I wanted you to know that I am thinking about that and to give you an opportunity to address it at the appropriate time.
OK. That said -- and I should add there is a lot of development in the law of products liability, I am sure you all the know the case Summers v. Tyce as its starting point and some of you may know that. Let's continue where we are.
Mr. Corley, come on up. You are still under owes.
MR. GARBUS: We had that open question before you concerning certain aspects of Mr. Corley's testimony. I presume we are not getting into it until such time as a further discussion of the court.
THE COURT: Absolutely. Absolutely. If it is something you are going to want to go into Mr. Gold, tell me you need to talk to me privately with Mr. Garbus.
MR. GOLD: Thank you, your Honor.
Q. Good morning, Mr. Goldstein.
A. Good morning.
Q. We left off yesterday discussing a statement that you had made and I am going to give it to you again. Did you not at one point in time, after October of '99, state that DeCSS was a free DVD decoder that allows people to copy DVDs?
A. That was a statement that was made on our web site which I did not write but I take responsibility for. It was at some point after it was published on our web site I believe in November that we realized that it was a lot more complicated than just that and we -- future writings on the web site reflected this.
Q. What do you mean when you say it was a lot more complicated than that?
A. It wasn't simply a bit of code that allowed someone to copy DVDs. At that point in time, I wasn't even that familiar with the technology. I wasn't aware of how the encryption worked, I wasn't aware of the extent of existing DVD piracy and wasn't aware that simply copying a DVD was a trivial thing that had been done for years, so I misread the actual facts of the case. When I realized it was more complicated than what we had reported, the case actually became very much more interesting to me.
Q. Did you discover after you first made the statement we have just referred to that it was incorrect?
A. Did I discover -- first of all, I didn't make the statement. I discovered after looking at various public forums, discussions about the subject, that this was not a case about copying. It had nothing to do with copying. It was about something called access control which was, prior to November, I had been unfamiliar with that.
THE COURT: The question, Mr. Corley, wasn't about what this case was about. The question was about the statement that appeared on your web site.
THE WITNESS: I am sorry. Not the case but what was going on at the time, the people whose web sites were being shut down, the story itself, that apparently was not about copying DVDs as we had initially stated.
Q. Let me ask you, Mr. Goldstein, is it not true that DeCSS is a free DVD decoder that allows people to copy DVDs?
A. I believe it allows the data to be copied to a hard drive at some point during its operation as do many other utilities but I believe that it can do that if you write it in a particular way.
Q. Then the prior statement you made was correct, I guess?
A. Not entirely because it did not exist for the purpose of copying DVDs. One would not run that program to copy DVDs.
Q. But you said that it was a free DVD decoder that allows people to copy DVDs and if I understand your testimony this morning, you are saying the same thing?
A. It allows you to do that but that's not the purpose of it. Again, I am looking at this from a journalistic perspective.
Q. I am not asking you the purpose, I am asking you what it does.
MR. GARBUS: I will object to Mr. Gold cutting off the witness' answer.
THE COURT: The witness is being unresponsive.
A. I am explaining it as best I can. Basically the program exists to allow someone to bypass the CSS encryption on the DVD. Now, whether that's the purpose of that is to copy a DVD, that's something that we initially didn't understand properly and I realize, I realized in November that it was -- that was not the case at all, that it was about defeating access control, being able to have a player that worked under the Linux operating system among other things.
Q. You began posting DeCSS on your web site in November of 1999?
A. That's correct.
Q. Is it your testimony that you did that as a journalist to write a story?
A. That's correct.
Q. Could you have written the identical story without the posting, using the letters DeCSS as many times as you wanted in the story?
A. Not writing a story that would have been respected as a journalistic piece, no, because in a journalistic world, you have to pretty much put up or shut up. You have to show your evidence and in this particular case, we would be writing an evidence without showing what we were talking about and particularly in the magazine that I work for, people want to see specifically what it is that we are referring to, what bit of technology that doesn't work, what new advancement, what evidence do we have and simply saying that somebody else said something just won't cut it.
So in this particular case, we pointed to the evidence itself which was already firmly established out there in the Internet world. We just put it up on our site so we could write our perspective on it and show the world what it was all about.
Q. Getting back to the question, in November of 1999, was it possible for you to write a story about DeCSS on your web site, using the letters DeCSS next to each other as many times as you wanted, without posting DeCSS?
A. I will take another shot at it. I -- basically, the story would not hold any value to our readers if we simply printed allegations without showing evidence.
Q. I am not asking you, sir, what value your story would have.
THE COURT: Mr. Garbus.
MR. GARBUS: I object. The questions are argumentative. And bad as to form.
THE COURT: Certainly the one in the process of winding up is.
Do you have another question, Mr. Gold? Ask it, please.
Q. In order to post DeCSS, didn't you have to go to your computer and get it?
A. Go to my computer and get it?
Q. Doesn't anyone have to -- yes, go to your computer, the same one you used to write the story?
A. We had to get DeCSS from a computer, not our computer because we didn't have it. We took it from one of the other sites that had it at the time and posted it based on that.
Q. Once you did that, when you wrote DeCSS in your story, in effect now you had posted it, isn't that right?
A. That's correct.
Q. Couldn't you have written the same story using the same exact words and using DeCSS without going out and getting DeCSS and posting it?
A. No. It would not have been the same story. It is analogous to printing a story about a picture and not printing the picture. People want to see what you are talking about.
Q. Does your web site now have a slogan "Stop the MPAA"?
A. It is probably on there somewhere, yes.
Q. Doesn't it stand out in bold letters?
A. I'm not exactly sure where it would stand out, but wherever it is, I'm sure it stands out graphically.
Q. And what is the meaning of that phrase, what do you mean to convey by stopping the MPAA?
A. Well, it is basically a phrase that I didn't coin myself. It has come from a number of people that see the actions of this lawsuit and other events over the past few months as ominous and something that should be stood up to and that's what people are doing simply by vocalizing that.
Q. Have you tried to stop the MPAA by writing stories about the MPAA and describing what they do as inappropriate?
A. I don't know if they carry that much weight where I could stop the MPAA by writing a story, but I have written stories to educate people as to the facts of the case.
Q. Have you ever tried to stop the MPAA by making contributions to organizations that were espousing the repeal of the antisurf protection laws?
A. We have tried to raise funds, I am not in a position to make contributions unfortunately, but we are trying to raise funds to help pay for our defense.
Q. Have you promoted through your web site the taking or downloading of DeCSS?
A. We have promoted the linking to various sites that still have it posted because that has not been ruled illegal in any way.
Q. Is it your belief as a reporter that to write a story about trafficking in illegal drugs in Manhattan, you would have to spend time actually trafficking in illegal drugs?
A. No, I don't think it is similar at all. One is a computer file, one is an actual substance.
Q. Is it true that the number of sites posting CSS has increased since January 2000 when you started linking?
A. It is hard to say exactly how many there are at any one particular point in time. There were a lot already up when we started the story. We came in kind of late. I believe CSS was cracked, I am hearing dates of October, and we didn't post the story until nearly a month later.
At that point in time, there were already hundreds of sites that had it and that was a good part of the story that we posted. I don't know how many there are now, and I don't know how many came up in January. I do know there was an increase after the preliminary injunction against us and I'm not sure where that stands today or how many sites actually have it or how many sites that haven't been reported that have it.
Q. Mr. Goldstein, are you a leader in the hacker community?
A. Some people may say that. I don't like to refer to myself as a leader but I am certainly somebody who certainly follows that particular aspect of society and I try to speak intelligently on it.
Q. You publish the Hacker Quarterly?
A. Yes.
Q. Your web site 2600 is known to virtually every hacker throughout the United States?
A. And quite a number of nonhackers as well, yes.
Q. You have appeared on radio and television?
THE COURT: We are not going to repeat what Mr. Garbus did yesterday, are we? We went through the TV programs and magazine articles. Let's go, Mr. Gold. He has a right not to like this law and he has a right to say so.
MR. GOLD: Absolutely.
THE COURT: The question is whether he violated it and that's a different question.
MR. GOLD: My point, yes.
Q. There was a hacker convention last weekend, is that true?
A. That's correct, the third HOPE convention, call H2K.
Q. And there was a Mr. Johansen appeared at that?
A. Joe Johansen appeared on panel.
Q. And he spoke?
A. He and his father both spoke.
Q. Was a mock trial held of this case on --
A. Yes, actually we had scheduled that before the trial had been scheduled so the timing was kind of strange but yeah, that was organized by somebody else.
Q. Is it fair to say that the hacker community in the United States is following this trial carefully?
A. I would say the hacker community, open source community, the Linux community, yes.
THE COURT: What do you mean by the open source community?
THE WITNESS: The open source community is basically people who write software, release the source code, share information, it overlaps into the Linux world. I am actually not a part of that community. I didn't know much about that community before this case.
THE COURT: OK, thank you.
Q. Is it true that the 2600.com site provides a form for people to fill out so they can add their site to the 2600.com mirror list?
A. Yes, we have two forms. We have a form where people know about a site that carries DeCSS or know of a site that carries DeCSS, they can submit that, or if they know of a site that we are listing that no longer has it, they can fill that out as well. But I should also point out that that has not been worked upon for a couple of months simply because we are too busy with other projects.
Q. But the form has been there for many months?
A. The form has been there from the beginning, yes.
Q. And it still is?
A. I believe so.
Q. You don't know?
A. I haven't looked at our web site in a while but I don't believe any changes were made, no.
Q. How long a while?
A. I don't look at every page.
Q. How long a while haven't you looked at your web site?
A. I have looked at my web site. That particular page I have not looked at probably in a couple of weeks but I believe it is still up there because we don't make changes.
Q. Do you understand that it is wrong to break through a protective device that protects a digital copyrighted work in order to take that work?
MR. GARBUS: I object to the question.
THE COURT: Sustained.
Let me ask you something about the mirror list and some of the sites to which you have linked. Are there sites to which you are linked on which by clicking on the link on your site, you reach a page at the linked site that has a variety of content on it which may or may not include DeCSS on the first page that comes up?
THE WITNESS: Yes, that's correct. I would say that's probably the case more times than not.
THE COURT: Do some of those sites then give the person who reaches it through clicking on the hyperlink on your site then give the user or visitor, whatever you want to call it, the option to go further in the -- I'll call it the transferee site if the visitor wishes to download DeCSS?
THE WITNESS: The person is given the choice as to whether or not they want to proceed further? Yes.
THE COURT: Are there sites to which you are linked or have been linked where the effect of clicking on the hyperlink on your site is to begin a download to the user of DeCSS without the user or the viewer taking any further action?
THE WITNESS: You would always have to take some bit of further action even for something as simple as verifying where you want to put the file on your hard drive. There are no instances that I know of where you can only click once on one of our hyperlinks and immediately start a download.
THE COURT: From your understanding, is it possible to construct a link of the type I have described, whether it is on your site or not?
THE WITNESS: I imagine anything is possible but I'm not familiar with that particular application.
THE COURT: Are there links listed on your mirror list in which the user, when the user clicks on the link is transferred to a point on a transferee site that has no content that comes up on the screen other than DeCSS or a dialog box that requires the user to do something like verify that the user wishes to download DeCSS? That is to say, no other content?
THE WITNESS: Yes, that's certainly possible.
THE COURT: OK, thank you. Go ahead, Mr. Gold.
Q. Mr. Goldstein, have you ever publicly advocated that it is wrong to infringe on anyone's copyright?
A. Yes, I support the copyright laws as I understand them.
Q. Have you, sir, publicly advocated that helping someone to commit a copyright infringement is wrong?
A. I believe that helping somebody to break the law is wrong, yes.
Q. What do you mean by wrong?
A. Illegal, immoral, something I wouldn't do.
Q. Now, at the end of 1999, did you know that CSS was a protected device protecting digital copyrighted movies?
A. I knew it was an encryption standard that had been applied to DVDs. I did not see it as the same thing as preventing illegal copying, no.
Q. Why not? What's the difference?
A. Again, I'm not a lawyer, so my understanding of the nuances might be a bit vague, but --
Q. Sir, I would like you to not give me any legal opinions.
THE COURT: Mr. Gold, he was trying to answer so let's hear the answer.
MR. GOLD: Thank you, your Honor.
A. My understanding is the protection, the copyright law is meant to protect the owner of the copyright from having illegal works distributed whether by copying or any other kind of infringement where the work or the copyright holder is not compensated for the work. And my analysis of CSS and what CSS accomplished was not the same thing as that. In other words, copying of DVDs was not affected by whether or not one decrypted CSS or one did not decrypt CSS.
Q. Did you know at the end of 1999 that the movies made by the major Hollywood studios were copyrighted?
A. Did I know they were copyrighted? Yes, of course.
Q. Is the quarterly published, Hacker -- forgive me --
A. Hacker Quarterly.
Q. Is that copyrighted?
A. Yes, we copyright both the magazine and material on the web page.
Q. Why?
A. We don't wish for people to be able to simply copy everything on our site and claim ownership as theirs.
Q. Did you make the copyright application or authorize that it be made?
A. Yes.
Q. Now, by providing links through the 2600.com site, are you making it possible for any member of the general public to download DeCSS if they have a computer?
MR. GARBUS: I object to the form of the question.
THE COURT: What's the objection?
MR. GARBUS: Let me hear the question again.
MR. GARBUS: I will withdraw the objection.
A. If they have a computer and a connection and they choose to go to our site, yes, they can download it that way.
Q. Is it your understanding that once someone has downloaded DeCSS through the 2600.com site, neither you nor 2600 has any control over what they do with that, what they did with DeCSS?
A. That's correct, we have no control what people do after they leave our site.
Q. Did you make any efforts to confine your providing DeCSS to people who were going to use it for any specific purpose?
A. I don't think such control is possible on the net. So the answer would be no.
Q. Is it true that anyone who downloads DeCSS can use it for any purpose that they want to put it to?
A. Any purpose they want --
Q. Any purpose at all?
A. I'm not -- I am having trouble following that. DeCSS can only do a limited number of things. It can't do anything somebody wants.
Q. Is it not true that all DeCSS can do is decrypt CSS?
A. DeCSS exists for the purpose of bypassing CSS, correct.
Q. Is it true that your posting of DeCSS and your linking to other sites that post DeCSS is being done for the sole reason that you are a journalist?
A. The reason the story appeared was because it was a journalistic piece. The reason we continued to write about it and talk about it is because we believe it continues to be that kind of a story.
Q. Writing it and being a journalist was your only purpose of posting or linking to other sites that --
A. That's the purpose for our site, that's the purpose for our magazine. It is a journalistic endeavor, yes.
Q. Did you ever write that the mirroring of DeCSS was a demonstration of electronic civil disobedience?
A. I can't say that I wrote those specific words but I believe those words may have appeared on our web site.
Q. And you're responsible for them?
A. I take responsibility for what appears on our web site, yes.
Q. What is your definition of civil disobedience?
MR. GARBUS: I object.
THE COURT: Sustained.
Q. When those words appeared on your web site, what did you understand them to mean?
A. My definition of civil disobedience, electronic civil disobedience, I suppose simply people taking a stand in a way they perceive as morally just. I really think it should be left up to the individual person to make that definition on their own. That's how I would picture it on the web site. Again, I don't know the exact context of how it appeared on our web site. I would have to look at that.
Q. Is it not true, to your knowledge, that civil disobedience means violating some law for the purpose of making a statement?
A. It many cases, it does involve sitting in front of a door, for instance, a minor violation to prove a point, yes.
Q. It is limited only to minor violations?
A. It can be bigger than that, I suppose.
Q. Now, do you write a news story on your web site or does a news story appear on your web site relating to DeCSS every single day?
A. I write a story occasionally. There are other people who write stories occasionally on the web site. They do not appear every day, no.
Q. How often in the course of a week in the last three months have stories about DeCSS appeared on your web site?
A. I would say probably an average of about one a week. And that's simply because we don't have very much in the way of staff.
Q. But I gather that you link to sites that post DeCSS every single day, every single minute, every hour of the year?
A. It is not an ongoing thing. We are not consciously doing it -- basically every story we write on our web page stays up until the end of time or until we are ordered to take it down.
Q. Is it possible for you to remove it?
A. Yes.
Q. You haven't done that?
A. No, I have not.
Q. And you don't take it down on each day when you are not writing a story about DeCSS, is that true?
A. No, we believe in keeping our stories up even when a story has a factual error in it. Our philosophy is it would be wrong to rewrite history and pretend we didn't say something that was wrong.
Q. Have you made any money by posting DeCSS or linking it to other sites that post DeCSS?
A. I would strongly suspect not.
Q. Why would you only suspect? Do you know?
A. I'm certain I haven't made any money from posting DeCSS. I have probably lost quite a bit of money because I have not been able to devote my time to the things I am supposed to be doing.
Q. There has been a great increase in people coming to your web site since you first started to post DeCSS, is that not true?
A. I have no way of knowing that since we haven't kept a counter since last summer.
Q. Wasn't that because so many people were coming to your web site?
A. That was last summer before all this started. I have no way of knowing how many people are hitting our site. We have no ads on our site so we have no economic need for more people to come there. I do note a lot of people are talking about it. I can suspect that more people are coming to our site but I have no real way of proving that.
Q. I gather al of your income comes from the Hacker Quarterly?
A. That's correct.
THE COURT: Mr. Gold, isn't that whole line irrelevant, in line of Greeging and a whole line of cases all the way back to Greeging?
Q. Have you talked with anyone at all with about advertisements in the Hacker Quarterly in the last several months, the possibility of advertising?
A. No, our policy is never to accept advertising.
Q. And you have had a discussion with no one about that?
A. No.
Q. I gather, Mr. Goldstein, that you have never been involved personally in reverse engineering?
A. No, I'm not an engineer.
Q. And I gather you have never been involved in cryptographic research?
A. No, I am not a cryptologist.
MR. GOLD: Your Honor, I believe it would be appropriate to have a sidebar at this point before I get into another subject.
THE COURT: Come to the sidebar.
MR. GARBUS: I would ask that this part of the record be deemed confidential.
THE COURT: The transcript of the sidebar will be separately bound and filed under seal. It will be available to counsel but not otherwise.
Implicit in that of course is there will be no discussion by counsel or revelation to anybody about what the sidebar was all about.
THE COURT: Anything else, Mr. Gold?
MR. GOLD: Your Honor, I would like to offer pages 58 and 59 of Mr. Goldstein's deposition transcript as an exhibit in the record at this point.
THE COURT: You have your objection, Mr. Garbus. I take it subject to the objection. If I don't refer to it in my decision, you should assume that I sustained the objection.
Q. Mr. Corley, I have noticed that a picture of a telephone and a telephone booth appears to an appear on the back cover of every issue of the Hacker Quarterly?
A. That's correct. We print foreign pay phone pictures.
Q. For what purpose?
A. People seem to be interested in what pay phones in other countries look like so they send us their photos when they go on vacation or if they live in that particular country and we compile them together and print four pictures on the back of every issue.
Q. And the 2600, that title of your quarterly, is that -- does that relate to a situation occurring with frequency in the '80s and '90s relating to phone service?
A. No, actually as I said yesterday, that dates back to the 1960s. It is something known as in-band signaling where a particular frequency sent over a long distance telephone line cause certain conditions to occur. And committing a 2600 hertz tone would drop a subscriber into what was known as operator mode where they could route phone calls, route themselves to internal operator and explore the entire system. That has not worked for quite a while and it is a symbolic type of name.
Q. Going back to the description you just gave us, was it possible to use what you described to make phone calls without paying for them?
A. That's one use, yes, it could have been used in that way as well.
Q. Yesterday we were reviewing certain articles or your counsel was reviewing with you articles that appeared in the Hacker Quarterly. I am not going to go through them line by line at this point. However, do you recall whether that warning about -- that goes something like if I remember it, we don't advocate the breaking of the law so we are not advocating that you perform the acts described above, did that appear in every column in the last four or five years?
A. It is something that has recurred. We don't print it on every page in every issue. So for instance, if somebody writes a letter to the editor and I think you could point to virtually every expression of this, somebody expresses an intent to commit a crime, we will without fail advise them not to do that or try to convince them that's a really bad idea.
Q. Such descriptive articles about acts which were in fact illegal did not contain that warning?
A. The articles are written from the author's perspective. If the author wrote the article from a different perspective as ours, there would be no such correction in there.
Q. By the way, how do we know that those authors weren't you?
A. Well, I can tell you they weren't me but other than that, I don't know how I can prove it.
Q. So --
A. Without violating --
Q. So Bull Finch and Crypton and whoever people sign, are you sure those people aren't you?
A. I'm very sure, very sure. For one thing, I don't have the technical expertise to write many of those articles. I write the editorial, I write the responses to the letters and occasionally news updates.
Q. Do you recall one such article where there was a description of how one could climb up a tree on his street where there would be a black box that related to telephone service in the community --
A. I believe you made reference to that yesterday. It is not really a black box.
Q. What color is it?
A. It is different colors, but beige, white, silver.
Q. Those are the boxes?
A. Basically that particular story entailed people who could climb poles or go into basements or closets in buildings and make phone calls off subscriber lines.
Q. Do you remember in one paragraph, there was a description of how you could climb the tree or pole and get to the black box and whatever color it was and rip it apart so there would be no telephone service to the community?
A. I don't recall specifically that paragraph but that could have been something that the author wrote.
Q. And you did not write that article?
A. No, I did not.
Q. Does that article, if you remember, contain a warning about don't break the law and don't do the things that are set forth above?
A. No, as I said, we warned people through letters and our editorial policy. We don't insert our comments into other people's articles.
Q. I am asking if you remember that that particular article did not contain such a warning?
A. I don't remember the particular article but our policy states we don't insert our editorial policy into articles.
THE COURT: Mr. Gold, I think you made your point.
MR. GOLD: Thank you, Mr. Goldstein. Thank you I have no further questions.
Oh, yes, if I may, there is one area that I haven't gotten into. I thank my colleague.
Q. Mr. Goldstein, is it true that unless you are enjoined, you intend to keep providing links on 2600.com to other sites where DeCSS is available, do you not?
A. It is my understanding that that's legal so until we are told otherwise, yes.
Q. The answer is yes?
A. Yes.
MR. GOLD: Thank you.
THE COURT: Thank you, Mr. Gold.
MR. GARBUS: I have one question.
Q. The judge asked you about linking. If www.garb.com comes to you and asks you to link, do you know what material I will have on my site when you give me permission to link to you?
A. Well, first of all, I believe you mean permission for us to link to you.
Q. Yes?
A. What we had been doing while we were keeping the list updated was basically going to the site, seeing if the file names were there, were present, and if so, and this is because we were getting so many submissions, we would just say yes, OK, we will post this one.
What we later found out, there were some dummy files out there. Some sites had the file names but they weren't really the files. They were gibberish or even empty files. It is simply impossible for us to go out there and verify that each file is exactly what it says it is. That's one of the reasons we stopped doing it, because it was kind of pointless after a while.
Q. Turning to the judge's question, do you know if you are going to link to www.garb.com who allegedly has DeCSS there, whether it is on the first page, second page, fifth page or what other information is contained on that site?
A. I am sorry, could you rephrase that a little?
Q. Let's, let's assume on your site 2600.com, you had ww.garb.com as a linking site. Do you know -- and that linking site allegedly has something about DeCSS, do you know whether -- it has comments on DeCSS, do you know whether it has the actual mirror and do you know where that appears on the first page, the fifth page of the site?
MR. GOLD: Objection, your Honor.
THE COURT: Overruled.
A. We would know when we initially verified the existence of the site if it was on the fifth page or first page or wherever but we wouldn't retain that knowledge and that could change, too. A person could take down DeCSS the day after we link to them and until somebody told us that, we wouldn't know to take it off the list. As I mentioned over the past few months, we haven't maintained it at all.
Q. Do you know how many clicks it takes once you get to that site to get to the actual mirror of DeCSS?
A. It differs with every site. Some may have a diatribe of have a various political ideology that they want you to read before you get to the particular mirror of the file where you click on it. Some might have nothing but the file name when you get to their site that you click on. So it varies with every site. We have no way of controlling that.
THE COURT: Ever heard of the term "deep linking"?
THE WITNESS: I have heard the term. I am a little confused by it because I know of one type of linking. I might have an idea of what that is. I think it involves going into a site beyond the main page, I think Ticket Master was doing something like that or somebody was doing something like that to Ticket Master.
THE COURT: Now, your web site has a home page, right?
THE WITNESS: Correct.
THE COURT: That's the page you get if you enter the universal resource locator www.2600.com into your browser, true?
THE WITNESS: That is right.
THE COURT: You have other pages that are, so to speak, behind the home page, right?
THE WITNESS: That's correct.
THE COURT: Every one of those other pages has its own universal resource locator or URL, correct?
THE WITNESS: That's correct, yes.
THE COURT: So someone who wished to link to your site and not to your home page but rather to some specific page beyond your home page could enter in his own web page a hyperlink not to your home page, but rather to the page behind your home page in which the person was interested in linking, right?
THE WITNESS: That's true, yes.
THE COURT: So if, for example, let's just make an assumption, the New York Times, which is www.NYTimes.com, has a web site, the home page has the URL that I just indicated, right?
THE WITNESS: That's correct.
THE COURT: And let's just suppose that as somebody suggested somewhere in the bowels of the New York Times web site is the DeCSS code, all right? Are you with me?
THE WITNESS: Yes.
THE COURT: That code would appear on a page or out there in hyperspace with its own unique URL, true?
THE WITNESS: Correct.
THE COURT: Now, if the New York Times wanted you to carry a link to its mirror, you might link to that mirror in more than one way, isn't that true?
THE WITNESS: There are several ways you can do it, yes.
THE COURT: You could link by inserting the hyperlink to www.NYTimes.com and then leave the user to navigate through the New York Times web site, true?
THE WITNESS: Correct.
THE COURT: Or you could link by inserting the appropriate URL to the specific page on the New York Times web site that had the DeCSS code, am I correct?
THE WITNESS: That's correct.
THE COURT: And it was your practice to verify the existence of proposed mirrors before you linked to them, when you put them on your mirror list, right?
THE WITNESS: That's correct.
THE COURT: So the means of knowing whether the link that you posted on your mirror list took someone who clicked on that hyperlink to a page containing a whole bunch of content, some of it relating to DeCSS and others relating to God only knows what, as compared to taking you directly to the source code for DeCSS was in your hands, true?
THE WITNESS: If we had spent more time on it, we could have refined it so it took you directly to the actual files, yes.
THE COURT: Or not?
THE WITNESS: Yes. What we did was we took what they gave us and looked at that and said does this go to DeCSS, either on the first page, second page or whatever and if it did, we just put it in the way they gave it to us.
THE COURT: OK. Anybody want any further examination?
MR. GARBUS: One thing.
If the court were to issue a declaratory judgment saying that the posting of DeCSS were illegal or inappropriate, would you stop posting the linking?
THE WITNESS: Yes, of course.
THE COURT: Mr. Gold?
Q. One question, do you know what a declaratory judgment is?
A. I am sorry.
Q. Do you know what a declaratory judgment is?
A. It is similar to an injunction, preliminary injunction.
MR. GOLD: Thank you.
THE COURT: OK, anything else? Mr. Corley, you are excused. Thank you.
MR. GARBUS: Your Honor, can I approach the bench for a moment.
THE COURT: OK, you and Mr. Cooper or Mr. Sims, whoever wants to come.
MR. GARBUS: I think the Summer v. Tyce issue that you raised is an issue that we have been conscious of since the beginning and just to use the analogy, if one of the bullets is defective --
THE COURT: I don't want to hear argument about it now.
MR. GARBUS: I don't want to also, but I just want to go into the question and relate it and I don't want to get contumacious, the whole question of discovery and the question of robustness is directly related to that issue and it has always been our argument, which we have not had a chance to document through discovery, there is the Merden report, the Macrovision report which indicated that DeCSS was not a bullet in the same way that the other utilities were bullets and it has been, that's been our position from the beginning.
THE COURT: I am glad to hear that, Mr. Garbus, because this is the very first time you have said it.
MR. GARBUS: That robustness was an issue and the efficacy of DeCSS --
THE COURT: You just told me the Summer v. Tyce issue has been in your mind from the beginning and the first person to raise it in this case is me on the 5th day of trial. You are trying to make a record for another purpose. If you want to make a motion at an appropriate point for whatever relief you think is appropriate, you make that motion, but we are going to hear evidence right now.
THE COURT: Mr. Hernstadt, your next witness?
MR. HERNSTADT: Very briefly, before you go to the next witness, yesterday we sent a letter in asking the Court's permission to use a computer to do a demonstration. We thought it was for next week but we have managed to bring the witnesses in for today, so we have another letter asking permission to do it this afternoon.
THE COURT: Tell me what it is about. Tell me what the demonstration is supposed to be about, Mr. Hernstadt.
MR. HERNSTADT: This afternoon we would like to do a demonstration of the LiVid Linux DVD player. The software is still in an alpha -- in other words, even before it is released for general testing, but it is functioning software so we would like to show you how it works.
THE COURT: All right, I have signed the letter. I didn't see the letter you sent yesterday.
MR. HERNSTADT: The other possible demonstration is similar to the one Dr. Shamos did.
THE COURT: Mr. Cooper?
MR. COOPER: Your Honor, we have had no notice of either of these demonstrations to my knowledge. It is difficult for me to understand the relevancy of the first described demonstration to any of the issues in this lawsuit and we would like an opportunity to get more detail from opposing counsel regarding what the demonstrations are.
THE COURT: How long is it going to take, Mr. Hernstadt?
MR. HERNSTADT: The demonstration?
THE COURT: Yes.
MR. HERNSTADT: A few minutes.
THE COURT: Relax, Mr. Cooper.
Next witness?
MR. HERNSTADT: The defendants call Larry Peterson.
Q. Good morning, Professor Peterson. Can you tell us about your educational background please?
A. I have a Ph.D. in computer science from Purdue University.
Q. What year did you graduate?
A. 1985.
Q. What was the subject of your dissertation?
A. It had to do with several issues relating to computer networks and focused primarily on e-mail and use of e-mail.
Q. After you received your degree, where were you employed?
A. I then took an assistant professorship position at University of Arizona. I was there for 13 years, last two of which I was department head.
Q. Of which department?
A. The computer science department.
Q. For the last couple of years, where have you been?
A. Princeton University.
Q. What courses do you teach at Princeton?
A. At Princeton, I have been teaching the computer networks class and introductory programming class.
Q. What is involved in the computer networks class?
A. Primarily I walk students through the motivation behind the Internet, the problems that cropped up over a number of years as the Internet evolved, and how researchers had solved them to evolve the Internet to where it is today.
Q. How long have you been studying the Internet?
A. Since I was a graduate student in 1980.
Q. How long has the Internet been around?
A. For a few more years than that. In 1980 is when it took on the form that we are familiar with today roughly. Before that, it was a packet switch network called the Arpanet. Because of the advent of ethernet and wanting to bring more networks into the fold, the Internet technology was deployed in the very early '80s.
Q. And what are your research subjects?
A. My is research crosses between networks and operating systems. Both of them are broadly the study of computer systems. We build systems, we measure them, and evaluate how they work, so it primarily has been focusing on how we can build operating systems so that networks can be more effective for users. We take a very, what we call, end-to-end perspective so we are concerned not just with one particular thing and how fast you can transmit over it or even the fact that the links are catenated together to form the network or that there is a computer attached to either end and application programming or processes right on those computers, so we have been focused on how to get data from an application on this side of the network to an application on the other side of the network.
Q. Have you ever researched a high speed connections from a network to a computer?
A. Back during the early '90s, we were a part of what was called the gigabit test, it was a national initiative to push the edges of gigabit network technology and at that time, what we in particular were doing was trying to move bits between a processor on one computer to a processor on another computer at the same speeds that the link connecting those two computers operated which at the time was 622 megabits a second, technology called OC12, and we were in fact able to move the data from one machine to another at that speed. There are a number of obstacles we had to overcome to do that.
Q. Is OC12, could you explain what OC12 is?
A. It's stands for Optical Carrier 12. There is a base unit of transmission -- let me go back. There is basically the technology having to do with fiber optics and there is a base rate at which people can transmit which is roughly 55 megabits per second and you keep multiplying that by 2 or 3 or 6 or 12, so 6 times that base rate gives you 622. something megabits per second. It is the way the bandwidth is measured on optical links.
Q. Is OC12 used in the Internet today?
A. Yes, it is.
Q. Where do you find it?
A. You would find it in the backbone of the Internet. So let me just give you a broad picture of the Internet and I can tell you where different technology applies.
You can characterize the very core of the Internet which is a set of long hold networks running across the country from Los Angeles to New York and so on and cities in between as the backbone of the Internet. A number of carriers have provided backbones. In the original days of the Internet, they were typically provided by government agencies so there was the NSF net in the early '90s and eventually commercial companies replaced that. So today you go to Sprint or AT&T and they have a backbone. The links that make up the backbone running between Chicago and Houston or whatnot might be at 622 megabits per second.
Out near the edges of that backbone then, you connect in various ISPs, Internet Service Providers, so your local cable company might in one direction provide you a service at home, but they have to turn around and connect into the Internet at the high end, they would possibly use a 622 megabit link from their site into SprintNet, for example.
MR. HERNSTADT: Your Honor, may I approach?
THE COURT: Yes.
Q. I have handed you Exhibit BDC. By way of explanation, could you tell us what this is?
A. This is just the layout of Sprint's Internet backbone network. I found it on the web a few days ago. It shows links connecting various cities. I know Sprint has a significant presence in Kansas City, for example, so it will show you the link running from Kansas City to Fort Worth and collectively this would make up the Sprint backbone.
Q. Can you explain what the different links are?
A. This is color coded and it is hard to tell for sure which are which, but these are the various links technologies being used in the Sprint backbone, DS3, an older technology, 45 megabits per second. OC3 is 155; OC12, 622; 0C48, that's 6.4 gigabits per second.
MR. COOPER: From a foundational standpoint, can we get some information from the witness about where this document comes from?
THE COURT: He said he found it on the Internet a couple of days ago.
MR. COOPER: We have not seen it before and I am looking for some indication about how he went about locating it and whether this is some area of his expertise or whether he did some research and pulled it down.
THE COURT: Is that about the size of it?
THE WITNESS: Yeah, I went to Sprint.com as a good guess of where to find Sprint's backbone, some links off there, I couldn't give you the exact URL, and I found this page.
THE COURT: What do you want me to do with it if anything?
MR. COOPER: I am trying to explore the foundation of documents of this type so we have an even playing field as far as cross-examination is concerned, your Honor.
MR. HERNSTADT: Your Honor, we would move this into evidence insofar as it assists your Honor in understanding what this --
THE COURT: Now you are not worried that things that appear on the Internet that are hearsay. But there is no objection so I will receive it.
MR. HERNSTADT: Thank you.
Q. Professor Peterson, have you received any honors or awards in connection with your work?
A. Several of my papers have been Outstanding Papers at different conferences, two in particular were Student Paper awards where I was co-author with one of my students at SigComm which is premiere conference for the networking community.
Q. Are you an editor now of one of the --
A. I am editor and chief of ECM which is the computer sciences professional organizations tracking on computer systems. That's the premiere journal on computer systems.
Q. Have you authored any publications?
A. Several. I think I counted this morning over 50. Technical papers.
THE COURT: Do you have a CV?
MR. HERNSTADT: I was about to ask if I could approach.
THE COURT: Yes.
Q. Professor Peterson, is this your curriculum vitae?
A. Yes, it is.
Q. Does this accurately reflect your publications?
A. Yeah, I believe it does.
MR. HERNSTADT: Your Honor, we offer this into evidence.
THE COURT: Received.
Q. Professor Peterson, what is Internet 2?
A. Internet 2 is a consortium of 150-odd universities. In the old days, the Internet was a play thing of the researchers at universities, but of course it has become a very commercial entity today and that has impacted the way research is conducted. So primarily research universities have gotten together to build an another piece, another corner of the Internet that they can use for research purposes and done in collaboration with some of the carriers, equipment companies and the like, Cisco and Nortel, I am not precisely sure who all the players are, so that they can have high speed connectivity so their astrophysicists can send huge files to each other in the national labs.
Q. Will it replace the Internet?
A. No, it is supplemental purely for the use of education. In fact, to join the consortium, you have to agree you will only use the connectivity and bandwidth for research and teaching purposes.
Q. Are you being paid for your appearance today?
A. No, I am not.
Q. Or for any of the work you have done in connection with this matter?
A. No.
Q. Can you describe how the Internet works by, for example, if you were to send a file from your office to a colleague's office, what are the various steps it would take?
A. As I have already briefly described that there is a backbone and off the edge -- there are multiple backbones. Off the edge of those backbones will be individual service providers. They may be cable companies, local phone companies. There are a variety of people who have gotten in the business of providing Internet service. They will then connect at some number of links into the various backbones They might connect for one or might connect to more. Collectively, you can think of that as the majority of the Internet in terms of its capacity.
What happens at the edges then is that individual users or consumers will buy some link into one of those service providers. So maybe it is just they have a 56-kilobit modem and they dial into the service provider periodically or maybe they have released a line which is sometimes referred to as the last mile, last mile link into the Internet.
So now we have the backbone surrounded by service providers surrounded by the last mile links. That constitutes then all the bandwidth that makes up the Internet. One point I should make, if you go into the Internet, it is not just links. There are nodes, computers, in essence, that connect those links together sometimes called packet switchers, sometimes called routers. So at every one of the connection points, there is a router of some sort. Of course at the edges, individual PCs and laptops and whatnot that the end user has is connected to that last mile link.
Q. OK, could you tell us what file transfer speed is?
A. File transfer speed would be the number of bytes in the file divided by how long it took to transfer the file. That would give you a bit rate or byte rate.
Q. Are connections to the Internet or pieces of the Internet rated by how fast you are capable of --
A. Any one of these links, you can talk about the bandwidth it could potentially deliver. Whether or not you get the same transfer speed as you have individual links depends on a number of factors. It certainly depends on the fastest link between the sender and receiver. So I couldn't go -- the transfer time couldn't be any faster than the slowest link between the source and the destination. There are other factors that come into play though.
Q. In practical terms, does one ever get the maximum theoretical speed of any piece of the Internet?
MR. COOPER: Your Honor, it is ambiguous and lacks foundation.
THE COURT: Overruled.
A. We get it in the line. We can get conditions set up just right that we can transfer at the theoretical maximum rate that the link would provide. On any given day at any given time, you might see something approaching that speed, might see half of it, a tenth of it, it is hard to predict. It depends on a number of factors, not the least of which is how many other people are attempting to use the network at the same time and so -- I could expand upon that, but there is clearly going to be a point at which, as I have described it, we have got millions and millions of end users connecting to their ISPs, connecting to backbone. There is some point at which there is a link that goes from Kansas City to Fort Worth that is trying to carry all of that traffic and that link could potentially be a bottleneck, a pinch point in the network.
Q. Is that, what you have just described, is that what is known as congestion?
A. Let me be more precise, congestion is a state where you have multiple incoming flows of packets attempting to go out on a shared link and the input rate exceeds the output capability. So what happens is the packets get skewed as much as they can be, but there is a limited amount of memory in any one of those routers. As soon as that queue is full, packets are dropped and they simply are not delivered. The state at which you drop packets is the state of congestion.
Q. What is the impact of the congestion on the rates of speed?
A. Because congestion means packets are being dropped and not being delivered, if you go to the software on the in-points, they are charged with the responsibility of making sure every packet gets through exactly as it was presented, so as the packets are coming out is exactly the same as the packets that go in. If you ever detect a packet didn't get through, then the sending software is responsible for retransmitting that packet. But the bottom line, you will not get the same rate that the theoretical capacity of the links might have suggested you would.
THE COURT: When you say they are dropped, does that mean they are never transmitted or they are delayed?
THE WITNESS: They are dropped. They will be delayed as long as they are queued in any one of the routers, but they are literally, if I have no room to hold them, I reject the packet as it is coming in.
THE COURT: The question whether a packet is dropped or simply delayed depends on the capacity of the router?
THE WITNESS: That is right.
THE COURT: Go ahead.
Q. What are some of the other pinch points that you mentioned?
A. Well, the software that I have been talking about is TCP, transmission control protocol. It is the key protocol of the Internet technology. TCP is a very delicate protocol in that it has to be tuned just right to get the transfer rates that the underlying bandwidth might suggest. So, for example, TCP will allow some number of packets to be in transit before getting an acknowledgement which is a packet that says I got packet 42, before getting an acknowledgement back.
You would not want to be in a situation where I sent you one packet and waited for you to respond I got it and then send a second packet and then respond because then I have only one packet in flight and cross-country legacies being what they are, you have very low bandwidth, you get very far from the capability of underlying links.
So what TCP does, it will send multiple packets and multiple packets in flight, ideally as many as the Internet capacity can hold. As I said, it is a simple matter of tuning. If your TCP has not been tuned to keep enough packets in flight, you will get far less bandwidth than you might have expected.
Q. What are some of the solutions if any for congestion?
A. Well, there is no solution for congestion aside from putting in more capacity. All we can do is program nodes so they detect when congestion is happening so the sources stop sending so fast because if they continue to send fast, the network will eventually collapse, which is get no useful work done, I am spending all my time sending packets that will eventually be dropped so nothing gets through. So the solution is what TCP does today which is slow down whenever it detects that congestion is happening on the network.
Q. Have you ever heard of technology called Napster?
A. Yes, I have.
Q. Is Napster or use of Napster clogging the Internet?
MR. COOPER: Foundation.
THE COURT: Sustained.
Q. Are you aware of any impact on the Internet of Napster?
A. I am not aware of any studies that say Napster is accounting for "X" percentage of the packets being exchanged. It has not come up in any circles I have been in where people point to Napster as detectable in the traffic.
Q. Are you aware of any other file-sharing technology similar to Napster?
A. Gnutella is the only one I can think of.
Q. Are you aware of any impact on the Internet of Gnutella?
A. No.
Q. Is there, at Princeton University, is there what is reported in the press as a Napster problem? Do you understand what I mean by that?
A. Princeton has not taken any action to limit the use of Napster is all that I know.
Q. What is the network topology?
A. Well, roughly speaking it is like this, and I'm not the Princeton system administrator so I am sure there are details that I don't know. Each dorm at Princeton currently has a 10- megabit shared network for all the students within that dorm that would then be connected by 100-megabit ethernet into the central facilities of the university. So it is one of a -- like a tree, like the central facility, the core, 100 megabits going down the dorms and various departments as well and in the dorms, you would have 10 megabits shared. Within individual apartments, you have richer connectivity because there are other things they are trying to do.
Q. Are you aware of use of Napster in uploading, downloading MP3 files at Princeton?
A. Not that I am aware of, no.
Q. Could the Internet sustain file transfers of a size of 650 megabytes in any kind of significant volume?
MR. COOPER: Object to the form, and foundation.
THE COURT: Yes, sustained.
Q. Are you aware of the size of a Napster file?
A. Napster files are in the neighborhood of 3 or 4 megabytes.
Q. Are you aware of a general volume of Napster files that are being transferred over the Internet?
A. I'm not aware of the volume of transfers, no.
Q. Could the Internet as it is now configured handle a significant number of transfers of 650 megabit files?
MR. COOPER: Object to the form and foundation.
THE COURT: Sustained as to both.
Q. Based on your experience and your expertise with networking the Internet, is it your opinion that the Internet as it is presently configured could handle thousands of 650 megabyte files transferred?
MR. COOPER: Same objection.
THE COURT: Sustained.
Q. I would like you to assume the following scenario, that tens of thousands of 650-megabit files are being transferred each day and could you tell us what your opinion --
A. A minute ago you asked me if I knew what the volume of Napster was and I don't know in any detail what the volume of Napster is. But I can say, I can make a comparison with 650 megabyte video files and say for whatever fault that volume is, I can talk about how that would strain the Internet.
Q. Please do so.
A. It doesn't matter what that volume is. If you take Napster to be a problem and 10,000, 100,000 or a million and you suddenly just translated 3 megabytes into 650 megabytes, then I could talk about that.
MR. COOPER: I would still like a foundation if I could, your Honor.
THE COURT: I think you need it.
MR. HERNSTADT: I am asking him to assume -- he stated he knows the size of a Napster file. I am asking him to assume there are tens of thousands --
THE COURT: Look, this is a little bit along these lines, you have the Tappan Zee bridge going across the Hudson River up there, and we don't know how many cars are going across it, and so far we don't know how many lanes it is and you are saying if instead of whatever numbers of cars are going across it, we had a lot more very big trucks, would there be a problem, and I think you can see that there are some problems with that question.
MR. GARBUS: Your Honor, can we take a five-minute break?
THE COURT: Yes.
Q. Professor Peterson, have you ever heard of DivX?
A. I heard of DivX, the intake for compression tool, yes.
Q. Do you know what -- do you understand a part of this lawsuit is about the threat that plaintiff has alleged will be posed by movies being on DVDs being decrypted by DeCSS compressed to a 150-megabyte file and then sent on the Internet?
A. Yes, I understand that's the basic idea.
Q. Does that as a new technology and new threat -- in other words, a year ago, DeCSS has been around since approximately October of 1999.
THE COURT: Is there a question?
MR. HERNSTADT: Yes.
Q. Let me ask you to assume that as of July 2000, there is additional traffic on the Internet consisting of these 650-megabyte DivX files of DVD movies?
A. If people started to transmit 650-megabyte DivX --
MR. COOPER: Your Honor, is there a question?
THE COURT: There is not yet.
Patience, Mr. Peterson.
Q. My question to you, Professor Peterson, based on your experience, do you have an opinion as to what the effect on the Internet would be posed by a new flow of 650-megabyte files being added to it?
MR. COOPER: Objection to the form. It is an incomplete hypothetical.
THE COURT: It sure is. I am going to hear it sooner or later, so I may as well hear it now and then I will ask the next question and then we will be past this I think.
Go ahead, answer.
A. I believe the question being asked is if the transfer of 650-megabyte files in the Internet -- I'm not sure if that's what he asked -- I am sorry. I am having a little trouble knowing exactly what it is I should answer.
THE COURT: I think I know where Mr. Hernstadt is going and I will help him get there.
The net has a finite capacity today, right?
THE WITNESS: Correct.
THE COURT: It has whatever level of congestion or lack of congestion that it has, right?
THE WITNESS: Right.
THE COURT: If there is a big new load placed on it, depending on how big the load is and what the current capacity is, it may or may not put a strain on the capacity of the system, true?
THE WITNESS: True.
THE COURT: When networks such as the Internet in the past have encountered limitations due to the volume of traffic pressing their capacity, the tendency has been to expand the capacity, right?
THE WITNESS: Right.
THE COURT: Kind of like the interstate highway system?
THE WITNESS: That's an accurate analogy.
THE COURT: All right, let's go, Mr. Hernstadt.
MR. HERNSTADT: Thank you. Actually, I thought you were doing very well, your Honor.
Q. Professor Peterson, in such an instance, how much of an expansion would there have to be?
A. Well --
MR. COOPER: Your Honor.
THE COURT: Obviously the answer is it depends. How many messages, what the capacity is, whether there is overcapacity or undercapacity, unless you have got some specifics here, you are nowhere, and even if you have the specifics, you have to address the expansion question.
Q. If we are talking about the addition of 650-megabyte files being traded or being sent via the Internet by let's say -- and let us assume for the purposes of this at an additional rate of 10,000, 10,000 650-megabyte files per day, will that pose a strain on the Internet?
A. I can't say whether 10,000 transfers exactly would pose a strain. What I can say, if you look at the Internet today as to what is a common-sized file that's transferred and make the assumption that the Internet is engineered for what is happening today, that you see files in the neighborhood of like the -- like the music example, MP3, 3, 4, 5, 6, 7 megabytes, that's a typical transfer file size at the high end. There are certainly smaller transfers, but that is considered to be a sizable transfer in today's Internet, 650 megabytes is two orders of magnitude or a hundred times larger than that.
Q. And based upon your experience, do you have an opinion as to how long it would take for the backbone of the Internet to be increased by two orders of magnitude?
MR. COOPER: No foundation for that question, your Honor.
THE COURT: Sustained.
Q. Professor Peterson, do you have any experience with the growth of the backbone of the Internet?
A. Yes, I do.
Q. Do you know how fast the Internet, the backbone of the Internet, has developed, how much speed on the Internet has increased over the last ten year?
A. Yes, I could walk through that if you'd like.
Q. Please.
MR. COOPER: Your Honor, I would like to have some foundation for the information the professor is about to give.
THE COURT: I'm satisfied. Overruled.
A. Sir, there are different ways of measuring exactly what we are talking about here, but let me give you a couple of example ways of looking at this.
If you start in 1990, approximately ten years ago, the Internet backbone at that time run by the National Science Foundation consisted of 1.5 megabits per second links. In 1993 that was upgraded to 45 megabit per second links.
THE COURT: What was the second one?
THE WITNESS: 45 megabits. If you go to 1995, we are now seeing some new technology introduced, we start to see the commercial carriers become bigger and bigger players. A cutting age network that was a follow onto the Nation Science Foundation's network was running at 155 megabits per second. That was the cutting edge technology in 1995 for the long haul lengths that make up the Internet.
If you keep following that history, and I don't have an exact date here, but it's in the '98 neighborhood, I don't know exactly when this particular turnover took place, we started to go see the OC12 that we were talking about a little earlier in the backbone, 622 megabits per second.
Today you can find OC48 links, they are at 2.4 gigabits. So if you look at the peak technology available over a ten year period, you will see that it's increased two orders of magnitude over that ten year period, if I've got my numbers right.
If you focus just on the last five years, which I think is probably a little bit more appropriate because that's the technology that's carrying us today, OC3, OC12, OC48, we've gone from bandwidths measured in the hundreds of megabits per second to gigabits per second. That's one order of magnitude improvement in the backbone's capacity in a five year period. It could be a little bit more precise, but for now I'll stay with orders of magnitude.
We are looking at if you take the assumption that you need to improve the capacity of Internet by two orders of magnitude, so that 650 megabyte file transfers, because as common place as today, 3 megabyte or 4 or 5 megabyte transfers, that's going to take two orders of magnitude improvement in capacity. As I just said, it was a five year period to see the last order of magnitude improvement.
Q. Could you define what is an order of magnitude?
A. Factor of ten. Two orders of magnitude is a factor of 100.
Q. What is the connection between the increased capacity of the backbone and the increased connection speed for the average consumer, connection to the home?
MR. COOPER: Can we have a foundation, your Honor?
THE COURT: I don't even understand the question. Try again.
Q. Professor Peterson, is there a connection between how fast a connection is available to the home and how much capacity is available on the backbone?
A. Yes, because more consumers in the home start sending in and receiving data, that data is traveling over the shared capacity of the backbone, the backbone capacity has to keep pace as you add more users or you add larger files that those users are transferring.
Q. Professor Peterson, in your opinion, what would be the effect on the effective transfer schemes of files if the average file size being transferred increased but the Internet backbone did not?
MR. COOPER: I object to the form of the question, because it's an incomplete hypothetical, and I think it still lacks foundation. I think it's the same question.
THE COURT: Look, I am going to take it, but let me say this: Obviously I have a lot of respect for Professor Peterson's expertise, but this just has so many limitations that it's close to being of no value, because there are so many factors here that you are not controlling for and so many things that we don't know about, that it's just not of much help. But go ahead.
Q. Professor Peterson, what are the factors that determine the transfer rate that a home user would obtain when sending a file from his or her home to somebody else?
A. The factors that would come into play would be the capacity of the links between the sender and receiver, and we have already been talking about the backbone, but it might most likely be limited by those two edge links which would be the uplink speed of the sender and the download speed of the receiver, because these links are typically not symmetric into the home. And while I may be the only one using my end link, the links that are in the middle of the network are being shared with thousands and millions of other users, and so congestion is going to happen on those links, and of course it depends on the level of congestion exactly what transfer fee I'm going to get.
Q. Is the Internet capacity today sufficient so that congestion is only an occasional problem in terms of slowing transfer rate down?
A. Congestion happens all the time on the Internet today.
Q. If everything else remained the same but the size of the average file being sent increased, is it predictable what the impact would be on the transfer speeds?
MR. COOPER: Same objections, your Honor.
THE COURT: Overruled.
A. There would be increased congestion. I would expect there to be increased congestion.
Q. And what is the connection between increased congestion and transfer speeds?
A. Congestion is probably the predominant factor after the actual link speed that affect the transfer rate.
Q. And to get a sense of what two orders of magnitude means, could you compare that to an increase, for example, in microprocessor speed of two orders of magnitude?
A. Just as an example a lot of people today would consider having a 500 megahertz PC as being pretty good. Two orders of magnitude back in time, that was 5 megahertz, which has been quite a while ago in a lot of people's memory.
MR. HERNSTADT: Thank you very much, Professor Peterson.
THE COURT: Thank you. Mr. Cooper.
Q. Good morning, professor.
A. Good morning.
Q. I believe you testified regarding some information regarding the Internet backbone and made reference to OC12. Those are routers, is that correct?
A. OC12 is a link speed.
Q. Okay. And the speed is 622.08 megabytes, is that correct?
A. I forget if the .08 is correct, but it's 622. Something megabits per second.
Q. And the speed at which the switchers and transfers are I'll to operate now have increased significantly from the OC12, have they not?
A. Today there are OC48 links in the backbone.
Q. Are there not OC96?
A. There are probably some OC96. I'm not aware specifically but that's possible.
Q. Are there not OC92 that have an effective rate of 9.93 megabytes current any use?
A. I'm not aware of their use in the Internet backbone.
Q. Are you aware of their use in connection with work being done at the high speed connectivity consortium?
A. I'm not specifically aware of that consortium.
Q. Have you not heard of the consortium, a consortium that involves the Carnegie Mellon University, Cisco Systems Inc., the Corporation for National Research Initiatives among others?
A. There are lots of consortiums like that. I'm not familiar with that particular one.
Q. Your writing partner in your book is Mr. David, is that correct?
A. Um-hum.
Q. He is a fellow with Cisco Systems, is he not?
A. That's correct.
Q. The only experiment you conducted in connection with your testimony today, as I understand it, was to check the effective download speed of your home DSL, is that correct?
A. That's correct.
Q. And the effective download speed that you experienced was 2 meg, correct?
A. 2 megabits per second, yes.
MR. COOPER: I have no further questions, your Honor.
THE COURT: Mr. Hernstadt.
Q. One question. Professor Peterson, what is your Internet connection from your home?
A. It's a DSL.
Q. Where does it go?
A. It goes directly into the department, so I have a leased line directly into the department, so the department is in essence my ISP.
Q. What is the connection of your department?
A. The department into the Internet?
Q. Yes.
A. We are connected by 100 megabit into the campus. The campus is then connected to the Internet at I believe 50 megabits per second.
MR. HERNSTADT: Thank you very much.
THE COURT: Professor Peterson. Thank you very much.
MR. ATLAS: We are going to be calling Professor Peter Ramadage. I wonder if we could have a short break so he can set up a laptop computer.
THE COURT: Sure. We will take 15 minutes.
THE COURT: Mr. Atlas.
MR. ATLAS: Good morning, your Honor. Defense calls professor Peter Ramadage.
DEPUTY COURT CLERK: State your name, spelling your last name.
THE WITNESS: Peter Ramadage, R-A-M-A-D-G-E.
THE COURT: Proceed, Mr. Atlas, please.
Q. Professor Ramadage, where are you presently employed?
A. Princeton University.
Q. What do you do at Princeton?
A. A professor in the department of electrical engineering at Princeton University, so I am engaged in teaching both graduate and undergraduate students, and also research.
Q. Do you teach in any specialized area?
A. Yes, I do teaching and research in the area of digital video libraries, digital signal processing for video, search techniques for digital video, video compression and transcoding of digital video.
Q. Could you briefly go through each one of those areas in which you teach and just describe them for the Court, please.
A. My primary focus at the moment is in digital video libraries. With the ever increasing amount of digital video that's available, it becomes important to have an indexing, search and browsing mechanism for retrieving video content once it has been archived. We are developing tools which will go into producing systems to achieve that goal. This is a relatively new area, so we are only at the very beginning of the research. We are developing very elementary building blocks for different types of searches, different types of browsing mechanisms to aid in the formation of digital libraries.
Q. Could you describe the area of signal processing for us, please.
A. Okay. Signal processing generally is the area in which you input a given signal, let's suppose it's a video signal, for example, and then you pass those bits that you are waiting through an algorithm, whose purpose is to come up with either a different version of what has been input or to answer some question that the user has input based on the video content that you are searching or inputting into the algorithm.
Q. I think you also mentioned compression technology.
A. Yes.
Q. What do you do in compression technology?
A. Let's restrict our attention to video files. Because video is a very, very intensive signal to convert to digital format, it requires very many bits to convert adequately into the right format. It results in a very large file. Now to store those files in an efficient way, or to transmit them to another person in an efficient way one normally uses a compression technology. That compression technology comes in two forms. It's a form called loss less compression, and in loss less compression the file is transformed into a format which is smaller and requires less space but which can be decompressed and you get back the original content completely without any change. So it's called loss less, because no information was lost.
Another type of compression technology is lossy compression, and lossy compression you are willing to trade off accuracy of the decompressed files, so the decompressed file would be an approximation to the original, but you will benefit by being able to get much, much greater compression. That's typically used for consumer video, video on the Internet, for example, video phones, wireless multimedia, they all use lossy compression.
Q. Are you familiar with any other types of compression technology?
A. Those are the main two. There is also a general purpose compression technology used on computer files. Typically it's under the name of zip files or LZW double compression techniques. That's a loss less compression technology.
Q. You also mentioned transcoding. If you could briefly describe that for us.
A. Okay. There are various applications when after a video has been produced and it has been stored in a compressed format that one needs to actually compress it further or to transform it inn some other way to a different format.
To give you an example, if I have a video stored and I wanted to transmit it to someone over a wireless communication channel, there simply may not be enough bandwidth on the wireless communication channel to transmit the video in a reasonable amount of time in its current format, so you would take that current format and you would transcode it or basically you would take an existing compressed file, process it, and to achieve greater compression or change other attributes of the video such as the frame rate or the frame size to make it better suited to transmit over the wireless communication channel. That would be one application of transcoding.
Q. How long have you taught at Princeton?
A. 16 years.
Q. What degrees do you hold and from what schools?
A. I have two undergraduate degrees, one in physics, one in electrical engineering from the University of New Castle in Australia. I have a masters degree in electrical engineer from the University of New Castle. I have a Ph.D. from the University of Toronto in Canada.
Q. Have you received any professional honors?
A. Yes. I have been awarded a complication medal from my undergraduate institution. I have been awarded a best paper award by the IEEE. I have been awarded several teaching awards both from inside Princeton University and agencies outside Princeton University; NSF research initiation grants; IBM young investigator awards, I think.
Q. What is the IEEE?
A. The Institute of Electrical and Electronic Engineers.
Q. Have you received any teaching awards separate and award from what you described?
A. Those are the main ones.
Q. Have you been published?
A. Yes, I have published over 80 referee journals and conference articles.
MR. ATLAS: I am going to show the witness what we have marked as Defendants' Exhibit BDE. It's a copy. I will let the witness describe what it is.
THE COURT: Any objection?
MR. MERVIS: No objection, your Honor.
THE COURT: Your name is?
MR. MERVIS: Michael Mervis.
THE COURT: BDE is received.
Q. Can you tell us what BDE is?
A. This is my curriculum vitae. I would say it's the latest one that I just produced about two or three weeks ago.
Q. Does this reflect your published articles and conference papers?
A. Yes. Yes, it's got the most recent articles we will be presenting this September at the IEEE conference on image conferencing. Those have been added.
Q. Have you ever testified before as an expert?
A. Not in court, no. I have been employed as an expert witness on other court cases, but I haven't actually testified.
Q. Have you been deposed before as an expert?
A. Yes, I have.
Q. In which case?
A. There is a case I. Omega v. Cyquest concerning the patent infringement on removable media disk drives.
Q. Were you being compensated for your time in that case?
A. Yes, I was. I was being paid an hourly rate of -- it started at $200 an hour and it went up to $250 an hour by the end of the case.
Q. Are you being compensated for your time in this case?
A. No, I am not.
Q. Do I take that to understand that you are offering your service as an expert for no compensation?
A. That's correct. I believed it is important to present my point of view and perhaps the point of view of other researchers in this area, that it's important for academics and researchers in industry to have fair access use of digital content in my own area of research. That happens to be digital video. And I think it's very important for researchers in digital video today to have access to the digital video that's in the marketplace for fair use, use in research. And down the road when this technology is mature I think it would be important for other researchers in disciplines not necessarily related to technology also to have access to this digital information.
Q. Can you describe in a little bit more detail the specific areas of research you are currently engaged in now at Princeton?
A. Okay. Sir, we have I think several projects running currently. One project has to do with searching video by example, so we have stored in a data base a large amount of video. This is a hypothetical example. What we would like to do is someone comes in and says I would like to see what you've got in your data base, clips that look like this example, and they bring with them an example of what they would like to see, what they would like to retrieve from the data base, and we have been working on algorithms for quickly searching through the data base to try and match the example of what is in the data base and pull those thing out of the data base.
Now because the video in the data base is stored in compressed form to save space, it is important that these algorithms actually operate on the compressed video. So we have been developing algorithms that do that. That's one project.
Another project has to do with multicamera video. We have video from two cameras taking a video of a scene simultaneously, and we are creating a synthetic video as seen from a virtual camera on the baseline between the two real cameras. That might have application in replays of sporting events, it might have application in surveillance. The Navy is very interested in this because they are putting cameras on the decks of aircraft carriers, so we will also be talking to the Navy about these possible applications.
Another project is using the information which is embedded in the compressed video to quickly search through that based on camera motion, try to estimate how the camera was moving by using the information embedded in the compressed video. And once you extracted what the camera was doing, that can often help you say what is happening in the video.
To give you a specific example, in a basketball game the camera often follows the ball very closely, so by determining the camera motion you can quite easily pick out things like fast breaks, jump shots, lay-ups and things like this.
Q. Any other areas of research you are currently engaged in?
A. We also are developing some novel compression algorithms. Based on some other work we think we might have some ideas that will lead to novel compression algorithms. These are not general purpose algorithms. These are rather specialized compression algorithms that might be applicable in computer graphics, compressing computer graphics or compressing video game video.
Q. What is the general purpose of your research in this area?
A. We are working towards providing tools for the browsing, manipulation and searching and indexing of digital video. That's our general objective. This is a very new discipline, so we are not working on everything right now. We are working taking the very first steps and working on the very elementary building blocks that would go into building such a system.
Q. When you say this is a new discipline, you are referring specifically to your research or the general research in this area?
A. The general research in this area is relatively new, really has come to the forefront in the past five years.
Q. Do you have any expectation of how this research will develop in the future?
A. The expectation is that after maybe in ten years time, after we have all the elementary pieces worked out, people will start to put these together into commercial systems or even public domain systems, and then people from the humanities, people from the social sciences, as well as people from technology and the sciences will be able to use these tools as part of their research. It will become a research tool for people outside of the immediate technological area where they were developed.
Q. Now in your research, and specifically in the development of the algorithms you testified about, do you use digital content?
A. Almost exclusively everything is based on digital content.
Q. How do you use the digital content?
A. One of the advantages of having digital video is that you can search through it in a very quick way. Digital video enables many things which are not possible with analog video. You can search through the video, you can jump into it in a random access type of way rather than sequentially have to start from the beginning. You can store it very conveniently, and you can create data bases and libraries of it very conveniently. The idea is you can create a library of video content, both video, sound as well as text.
Q. What are the sources of digital video content that you use in your research?
A. We have a variety of sources, and I also should tell you that at different stages of the research we use different types of video, so we have our own camera, it's a digital camera and it stores its video on a high quality digital tape. From there we can transcode it into different formats, whatever format we find most convenient. We find MPEG1 the most convenient to work with, so we often transcode into MPEG 1 format.
Q. Explain what MPEG1 is.
A. MPEG 1 -- there are various standard compression technologies. One of the very first is called MPEG1. MPEG stands for Motion Picture Expert Group. It was a consortium of people interested, industries interested in digital video, as well as representatives of the standards organizations, and they produced over a period of three to four years an ISO standard called MPEG1.
We also download pieces of video that other researchers have, and they make small clips of video available on their web pages, either displaying the result of their algorithms, or often in conjunction with that they will put up the original video, unprocessed video, to allow other people to try to replicated their results on videos. Those are very short videos, less than ten seconds.
Occasionally my students are able to download some digital video from the web that a movie studio has put up as a publicity piece for a movie or something like this. So we also have something like that.
In addition, on two occasions we have been able to negotiate through an industrial partner permission to use an extensive piece of video, copyrighted video from the content producer. That is subject to much more restricted use.
Q. Those are the four primary sources of digital content you have available to you now?
A. There is another source, and that is we can take analog tapes and convert them into digital form. Several years ago that was how we relied for getting digital video, but it wasn't totally satisfactory. It's subject to the noise of the analog recording process, then the noise of the analog playback process, and then the peculiarity of the particular digitizer that you use and compression hardware that you use to actually do the digitization and compression. So we prefer not to have to rely on that too much.
As I said, algorithms rely on processing the compressed domain video, and we don't want to get tied to a particular compressor, a particular piece of hardware that does the compression, because then maybe our algorithms will only work with that particular piece of hardware, so it's important for us not to get too dependent on that mechanism.
Q. In terms of the variety of digital content that you have just gone through, the five categories, do you find the variety available in those categories best suits the type of work you're doing?
A. There are problems with each of those video sources that I have mentioned. Let me elaborate on that a little bit.
First we have our own camera, but it's quite a steep learning curve to learn how to use the digital camera, so our graduate students spend some time learning it if they need to produce video, but the resultant video is not particularly high quality. They are not professional cameramen, and they are faced with various hurdles that they have to come up with, and they are quite innovative in trying to overcome those hurdles, but the resulting video is not commercial quality video. For example, we can't hold the camera very steady. It's a hand-held camera. We can't take multicamera video. We only have one camera. And we can't control the lighting very well.
The video that we download from the web from other researchers is typically very, very short and sometimes has a lower frame rate than we would like to work with, and sometimes a very small frame size, small number of pixels in it.
Video that we digitize from analog tape I think I said already there is a couple of problems with that. I won't go over those again.
By far the best quality digital video that we can obtain is from DVDs.
Q. Have you heard of DeCSS?
A. Yes, I have.
Q. Have you used DeCSS?
A. Yes, I have. I have used it. I went to the --
Q. Wait. What do you understand DeCSS to be?
A. It's a program which will read the contents, the table of contents of a DVD disk, and then you can ask it to descramble the scrambled VOB files, the video object files on that disk, and store them to your hard drive.
Q. You testified a moment ago that you used DeCSS. How did you obtain DeCSS?
A. I went to a search engine, Google.com and I just typed in DeCSS and did a search. It came up with about over 7,000 hits, so I did a bit of a search through those hits until I found a site that had the software, and I downloaded the source and executable program.
Q. Do you recall the site you downloaded it from?
A. No, because it involved a search through all of the various hits I don't actually remember the site I eventually found it on. Sometimes when you go to these sites you don't actually end up on the front page of the site, you end up on some lower page that has the feel you want, but it's not immediately clear what site this is ultimately connected to.
Q. Do you know whether it's 2600.com?
A. I don't believe it was 2600.com, no.
Q. What form was DeCSS in when you downloaded it?
A. It was a zip file which is a loss less compression, that's a general type of loss less compression used to compress general computer files. When I decompressed that, I believe it gave me the source, the executable and maybe another document, I can't remember, a "read me" file or something like that.
Q. Do you have an opinion on whether DeCSS would be useful to you in your area of research?
MR. MERVIS: Objection, your Honor. There is no foundation.
THE COURT: Overruled.
A. My initial experiments have indicated that's by far the best source of high quality digital video available to us today.
Q. What is the best?
A. The DVD. In terms of the breadth of video that's available to us, the wide variety of video available to us from various sources, and the high quality of the content, DVDs are the best source for us.
Q. What I'm asking you is do you have an opinion on whether the DeCSS utility would be useful to you in your area of research?
A. Oh, absolutely, because the video content on DVDs is scrambled, and so to get access to it we have to unscramble it and that's exactly what DeCSS does, it unscrambles that video content.
Q. If you were to have access to a wide variety of high quality digital content like on DVDs, why would that be helpful to your research?
A. Okay. Initially when we first start developing an algorithm, we usually use very short pieces of video, because videos are a very time consuming object to work with and it takes a lot of space. But after we've got the prototype working, it's very important to test the algorithm on a wide variety of different video. It's important for two reasons. First, you want to make sure your algorithm isn't somehow dependent on the type of video you took or the type of encoder you used. That's step number one.
Step two is you would like to search for video for which it doesn't work, and to do that you need to get out there and search through a whole range of different types of video. You are specifically looking for video where your algorithm fails to help you improve your algorithm or maybe start up a different research direction.
Q. In order to use the video, the high quality video that's available on DVDs, do you have an understanding of whether you need to decrypt them first?
A. Yes. There are several types of files in the DVDs. There are files with the extension IFO, which I believe is an abbreviation for information, and those contain like a table of content type information about what is on the DVD. There are files with the extension BUP, which I believe stands for back up. Those I think are back-ups for the IFO files. Then there are files with the extension VOB, which I believe is an abbreviation for video object. And those actually contain the video and those are scrambled, and those are the things that need to be unscrambled before you can actually use the video.
Now I should also add that that video is already in a compressed form. It's in a compressed form called MPEG2, which was the second extension of the ISO standard from MPEG1 to MPEG2, and that dealt with a higher quality video. The initial base standard for MPEG2 was intended to produce compressed video of comparable quality to broadcast video. But it also includes higher level video standards as well.
Q. In terms of the digital video content that's commercially available on DVDs, is that preferable to the digital content that you described before, the digital content that's currently available here?
A. Yes, it's very high quality, clean, no noise. It is already digitized, which is excellent, which means we avoid having to digitize it ourselves and then pick out any peculiarities our own digitizer exhibits. So, it's a very sound and preferable source of digital content. Also because of the wide variety of DVDs available, it satisfies our need for a large source of different varieties of video.
And in addition, I want to add one other thing. When we use our own camera we try to avoid introducing any bias. We don't want to sort of take video which is too favorable to what we are trying to do, because the scientific method demands that if you want to thoroughly test your algorithm you have to use the video which has been taken independently by somebody else. That's a basic fundamental premise of the scientific method.
Q. Is there a relationship between the amount or variety of digital video that you would use in connection with your research and the confidence you would have in the results of your research?
A. Absolutely. One of the criticisms -- I have been to conferences. I have also given talks in industry, and one of the criticisms that industry often has mentioned is that you really need to test your algorithms on a wider variety of video. I often hear that from people in the industry.
Q. Are there any other applications of your research that would benefit from having access to the high quality digital content of DVDs?
A. I think I have covered the main points.
Q. In terms of compression technology, do you use compression technology in your research?
A. Yes, as I have said, our algorithms are designed to work with compressed files, and because of that we need to have a working knowledge of the compression technologies that are employed. We also do our own transcoding of video from different compression standards to a second compression standard in order to test our algorithms on a variety of compression standards.
Q. I believe earlier you testified that there was lossy compression and loss less compression?
A. Yes.
Q. Do you know what type of compression is used on consumer-oriented video like the DVDs you buy in a store?
A. Yes, both the MPEG1 standard and the MPEG2 standard are lossy compression technologies. That means when the video is compressed, information is thrown away. If it's a good encoder, a good compressing algorithm, it will first try to throw away information which is least perceptually significant, but the more compression you ask for, the more bits it has to discard or the more information it has to discard, and eventually it is discarding important information, and that shows up as artifacts in the resultant video when you uncompress it.
Q. As a general matter, why would you use lossy compression if you can use loss less compression?
A. Okay. Loss less compression that can achieve compression factors of 2 is quite typical. If you are very lucky you might get higher than 2. So, a 2 gigabyte file would be reduced to 1 gigabyte with a compression factor of 2. By applying loss compression, you can -- I'm quoting here from the MPEG standard -- you can achieve compression factors of around 20 to 30 without any visible perceptual difference between the uncompressed video and the compressed video. So, that's a factor of ten better at least than the loss less compression technology. And in practice you often find compression factors as high as 40 used.
Q. Now, could you just describe for us a little bit in greater detail as you use more and more compression on content what happens to that content.
A. Okay. Almost all of these compression technologies are block based. What I mean by that is that each frame of the video is divided into small blocks, nonoverlapping blocks in the simplest case. These blocks are about 16 by 16, 16 pixels horizontally by 16 pixels vertically.
The compression is based on those elementary blocks, so when you start to throw away too many pieces of information, those blocks don't get represented correctly in the uncompressed video, so you start to actually visibly see these blocks in the uncompressed video, and you start to see miscoloring of the blocks.
One of the things that underlies most of the compression technologies is that high frequency information is discarded first. Now, high frequency information encodes things like edges, where there are sharp transitions, sudden changes in the image, so when you have discarded enough of this high frequency information or too much of this high frequency information, you start to see artifacts at the edge boundaries, a phenomenon called ringing, where there seems to be a time varying fluctuation around the edge, and you can visibly see this in highly compressed video.
Q. Is that what you referred to before as an artifact?
A. That's an artifact. The fact that you can see the blocks is an artifact. The fact that the blocks are miscolored is an artifact, and this ringing is an artifact. There are other artifacts as well, but those are the main ones.
Q. Do you have an understanding of an average size of a DVD film?
A. Well, I have looked at the DVD for the movie Contact, and the total amount of information on that DVD disk was about seven and a half gigabytes. Most DVDs contain the original movie as well as some additional add-ons such as director's comments, or things that were cut from the movie might be added on there, or special effects might be added on.
In the particular case of the DVD Contact, the actual video file and the audio file took up about 6 gigabytes, and the remaining gigabyte and a half were the extras and the add-ons.
Q. Do you have an understanding of how much an average CDR can hold in terms of megabytes or gigabytes?
A. The standard is 650 megabytes.
Q. So, if I wanted to copy a DVD that I went out and purchased and copied it onto a CD, I would have to compress the content to go from 6 or 7 gigabytes down to 650 megabytes?
A. That's correct.
Q. How would that be done?
A. Well, you would use a transcoder. Your video that you have from the DVD after you've descrambled it, you must descramble it first or else it won't work at all. Let's say you descrambled the VOB files, now they are in MPEG2 format. That has a bit rate -- and I will explain what bit rate is in a second -- it has a bit rate of about 6 to 10 megabits per second. Now the bit rate is how many bits are coming out of the player per second in order to display the video on the screen on average. That's an average rate. Sometimes it's higher, sometimes it's lower. But on average it's about 6.7 megabytes per second. You have to now transcode that down to a much, much lower rate.
You have available on the CD a space of 650 megabytes. Let's take the movie Contact. That's a two and a half hour movie, so I have to get two and a half hours at 6.7 megabits per second down into a 650 megabyte disk. If you do the conversion of units to make all the units appropriate, and then work out the math, the bit rate after you have done the transcoding for both the audio and the video needs to be around 590 kilobits per second. So you need to go from 6.7 megabits per second down to 490 kilobits per second.
That's an enormous compression.
Q. Would that type of compression result in artifacts being present on the ultimate product you end up on the CDR?
A. Yes, most definitely. The artifacts would be most visible when there are scene changes, camera motion or when there is movement of objects in the scene.
Just as a point of comparison, MPEG1, the rate for MPEG1 which is generally to be believed below broadcast quality video is 1.5 megabits per second.
Q. Based on your experience, if someone were going to compress a film file to send over the Internet, is it more likely that such a person would use lossy or loss less compression?
A. If they used loss less compression you would take a 6 gigabyte file down to about 3 gigabytes. That is way too big to transmit over the Internet. So that forces you, you must use lossy compression if you are going to get it down to any reasonable size. And even at 650 megabytes that's a very large file.
Q. In terms of the use of lossy compression, are there trade-offs to the user in connection with this type of compression?
A. The user, if you want to do a transcoding from say MPEG2, which is the DVD format, down to a much more highly compressed format, there are various parameters you get to choose. The most important parameter is the bit rate. Because you only have a limited amount of space, you have 650 megabytes, and you have a two and a half hour movie, that determines the bit rate. That's fixed.
Now you can play with other factors. You can change the size of the image. You can make the image smaller. You can change the frame rate so the frame rate of the DVD is 29.97 frames per second. You could attempt to do better by decreasing the frame rate.
Q. What effect would decreasing the frame rate have in terms of viewing the ultimate product?
A. You would see fewer frames per second. If the video was interlaced, which MPEG2 can be an interlaced video, I think the interlacing would be much more prominent at the lower frame rate.
Q. Could you just explain to us what the term "interlacing" refers to.
A. Yes. Broadcast video doesn't present each image in the video all at once on the screen. It actually does it in two passes. In the first pass it draws every second line on the screen. Then 1/60th of a second later it comes back and draws the lines it missed. That's called interlacing. These two scans are interlaced and they are 1/60th of a second apart.
Now if there is motion in the scene, then during that 1/60th per second things will have moved, so one of the problems with interlacing is that you begin to see slight jagged edges around objects that are moving because of the interlacing.
Computer screens don't use interlacing. They use what is called progressive scan, and progressive scan you start at the top and you draw each line sequentially down the screen.
Q. In your opinion, what is the current state of compression technology?
A. Okay. Sir, compression technology has been a very hot and active area of research for about the past 12 years. The MPEG1 standard started around the end of the 1980s, started to meet and form MPEG1 standard. The MPEG2 standard was even started before the MPEG1 standard was finalized in the early 1990s. The JPEG standard was also in the first half of the 1990s. We are now just past, at the end of 1999 the second generation of the JPEG standard, JPEG 2000. At this point compression technology is maturing and we are seeing fewer and fewer gains in the compression for each new generation of technology that comes along.
In the past there have been some interesting suggestions about potential ways of increasing the compression rates dramatically. None of those have panned out yet. In fact some of them have even been put aside now as areas of research that are not likely to be productive.
Q. Looking out for the next five years or so, do you see any breakthroughs in compression technology that would allow a person to compress a file to a greater extent than is currently available now?
MR. MERVIS: Your Honor, I object. There is no foundation for this testimony.
THE COURT: I will take the answer. Overruled.
A. In talking to my colleagues, we often have lunch together, some of them are very very actively involved in the compression community and actually have graduate students working together putting the pieces into the next ISO standards.
THE COURT: Excuse me. I had understood from your earlier testimony that you were offering yourself as an expert on compression, is that right?
THE WITNESS: I am. I also want to let you know that this issue is a hot topic of discussion in the department about where is compression going.
THE COURT: Go ahead.
THE WITNESS: And it seems right now that the only technology which looks promising is the wavelet technology, and that has been the basis of the JPEG 2000 standard which has just recently been issued.
So, it looks like from this point on there is only going to be incremental improvement. There is nothing that is known to be on the horizon which promises drastic gains. Everything right now looks like it's going to be incremental for the next five years. I can't see any further ahead than the next five years, so I don't want to speculate beyond that.
Q. Fair enough. Have you heard the term fractal compression?
A. Yes.
Q. Do you have an understanding of what that is?
A. Around 1988 a researcher by the name of Barnsley proposed this compression technology called fractal compression. I think he has two patents on that technology issued around 1990 and 1991. He formed a company I think which is called Iterated Systems, but since then nothing has really panned out from that technology. The latest I heard in the journals and other reading that I have done is that his latest version produced by his company is not quite as good as JPEG.
THE COURT: Now, isn't there software on the market in the digital photographic area known as genuine fractals?
THE WITNESS: I haven't heard of that software.
THE COURT: Isn't it in commercial use and supported by service bureaus for high compression of digital photographic files?
THE WITNESS: I have no information on that.
THE COURT: All right. Go ahead, Mr. Atlas.
Q. Have you heard the term DivX before?
A. Yes, I have.
Q. What is DivX, as you understand it?
MR. MERVIS: Your Honor, can I be heard on this? I have an objection. This witness was deposed a week ago on July 14, and as of that time the testimony was that he heard of DivX, that he downloaded some things from the Internet that he thought might be DivX, but he had done no further work on the subject. So, whatever we are going to hear from the witness in response to that question is something we have had no discovery on, no notice of, and we will not have any ability to effectively cross-examine.
THE COURT: Mr. Atlas, what about it?
MR. ATLAS: Actually at his deposition he testified about what he had done with DiVX and what he intended to do with DivX between the time of his deposition and the time he would be testifying, and I suggest that if plaintiffs wanted any further deposition before trial, they had an opportunity to seek it. But I think the witness clearly put them on notice of what he intended to do and what he looked to get out of it.
THE COURT: Let me see the deposition and tell me what you're referring to.
MR. MERVIS: Yes, your Honor. Can I approach?
THE COURT: I'm sure you will tell me the numbers.
MR. MERVIS: The subject of the DivX was covered on the following pages: 15 through 17, 21 through 22, 25 to --
THE COURT: I can't remember them all that fast.
MR. MERVIS: I'm sorry. Perhaps you let me know when you're ready.
THE COURT: Yes. What's after page 17?
MR. MERVIS: Pages 21 to 22, your Honor.
THE COURT: Next?
MR. MERVIS: 25 through 28, Judge.
THE COURT: Next.
MR. MERVIS: Page 70, your Honor.
THE COURT: Was there any representation made to you about whether or not he would testify in this area?
MR. MERVIS: The only representation that I know of, your Honor, was that his studies were continuing. And I would simply point out in response to my adversary's comment, this is expert testimony. Of course obviously we have been proceeding somewhat informally, but the notion that we would be given no notice of any further developments in this expert's experiments I think is inconsistent with the general standards of expert discovery. All of a sudden we are here today hearing about DiVX, and we have never heard a peep out of defense counsel.
MR. ATLAS: Your Honor, at page 16 he was asked whether he had made any evaluation of DivX. He said he was in the process of doing it.
THE COURT: I read it.
MR. ATLAS: You read that. I don't think this comes as any surprise, and we have been talking to each other every day -- not Mr. Mervis and I -- but they certainly were aware, and I think if they wanted further examination of this witness before trial, with eight letters a day, I think they could have included that in a letter.
THE COURT: Is that all you get?
MR. ATLAS: It's not a surprise. I think it's clear, given the time constraint, he put them on notice that he would be doing this before trial.
THE COURT: I will hear the evidence.
MR. MERVIS: Your Honor, may I make just one further application? I am prepared to cross-examine Professor Ramadage on everything so far. I don't know what he is going to say right now, but I do think that at the very least we should have the opportunity -- and I understand the Court wishes to hear the evidence now, but I do think we should have the opportunity to take the witness's deposition on these points and perhaps bring him back or perhaps just submit that deposition testimony, because I don't think it's appropriate, and I don't think it's fair to the Court, the witness or the defense counsel for me to be doing cross-examination on that topic. I think it's a better use of time to have a short deposition and we can submit the appropriate exerts if there are not --
THE COURT: Why not do that?
MR. ATLAS: I have no objection to that.
THE COURT: It's easy. You will finish your direct, Mr. Mervis will cross-examine him to wherever he gets to where he wants to insert DivX. You will get a deposition done in the next three days sometime, indeed before the trial is over. I certainly don't necessarily want to interfere with the professor's weekend if I don't have to. Lawyers are different.
MR. MERVIS: My son might disagree, your Honor, but I understand.
THE COURT: I understand, but it's an occupational hazard for sons of lawyers. And then we will proceed that way.
MR. MERVIS: Thank you.
THE COURT: All right.
MR. ATLAS: Could you repeat back the last question and answer, please.
THE COURT: I might observe this, and I do so because I want to be corrected if I misunderstand. Don't I have in evidence here Sleepless in Seattle and the Matrix, both in the form of the original DVD and in the form of a DeCSS decoded DivX'd compressed CD-ROM copy? Don't I have that?
MR. MERVIS: You do, your Honor, that's correct. I'm not sure we have moved the -- yes, we have. Your Honor, you do have both of them.
THE COURT: Okay. So I mean I suppose the purpose of all that is to ask me to look at the two movies and see to what extent I think that the copy is or isn't a commercial threat in the marketplace.
MR. ATLAS: That's part of it, your Honor. I think your Honor permitting, the witness has done his own test along the lines of what Mr. Shamos did. It will not take very long.
THE COURT: That's fine. I said I would hear it, but I'm just, you know, I don't know what form the testimony is coming in in, but I think movie fans probably don't think in terms of artifacts per movie. I think they look at the movie and they say is this an acceptable quality to watch.
MR. ATLAS: I agree. That's exactly where we are going in terms of whether it is acceptable quality. Can I hear the Last question and answer.
A. DivX is what is called a codec, and a codec is an abbreviation of two words, coder, decoder. So when you compress a video, you also have to at some later time uncompress it or decompress it. Each coder comes together with its appropriate decoder. That pair is called a codec. And DivX is an example of a codec, a coder/decoder pair.
Q. Have you used a DivX before?
A. Yes, I downloaded it from the web, installed it on my laptop computer and on the work station in my office. Then I employed a program called flask MPEG.
Q. Is this in connection with your work for this lawsuit?
A. Yes, it was with this lawsuit. I used a program called flask MPEG to select the DivX codec. Now, let me add here that there