Vorbis dev - Dec 2000 - patents and separate entropy coding

Hi everyone,

first of all, I searched through the archives for any posts resembling
this, I didn't read all posts about the patents so if this has been
suggested before I apologise.

I read in an article on C|Net (I think, it was linked from Slashdot
anyway) that Thompson are threatening to sue you if Ogg Vorbis becomes a
success. Which is evil, and I'm also mad at them because they never
answered my mail (which was about how to get an mp3 patent license, of
all things). But I digress.

Thing is, I understand that they have a patent on compressing audio by
transforming to the frequency domain, quantising, and then entropy
encoding, which is what Ogg Vorbis does, so there's trouble. Now, I've
been working on a Haar wavelet image compressor, and since I haven't
written the entropy coder yet I just dump my quantised output to a new
file. This file is as large as the original file. Then I simply use
gzip, which compresses well because of the preprocessing.

So, should trouble arise, you could make Ogg Vorbis from a compressor
into a transformer. Since you're not doing any entropy encoding, in fact
you're not even compressing anything, they won't have anything against
you.

There may be a problem with this, which is that it may in fact then be
illegal to create .ogg.gz files without a license from them, but it
would be pretty hard for them to hold that since they don't have any
patents on transforming audio data or on entropy encoding random data,
just the combination.

Well, IANAL, but I thought that it might be useful.

Cheers,

Lourens

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

On Fri, 15 Dec 2000, Lourens Veen wrote:
> Hi everyone,
> 
> first of all, I searched through the archives for any posts resembling
> this, I didn't read all posts about the patents so if this has been
> suggested before I apologise.
> 
> I read in an article on C|Net (I think, it was linked from Slashdot
> anyway) that Thompson are threatening to sue you if Ogg Vorbis becomes a
> success. Which is evil, and I'm also mad at them because they never
> answered my mail (which was about how to get an mp3 patent license, of
> all things). But I digress.
> 
> Thing is, I understand that they have a patent on compressing audio by
> transforming to the frequency domain, quantising, and then entropy
> encoding, which is what Ogg Vorbis does, so there's trouble.
IANAL, either, but that's awfully broad a patent.  I don't see how it
could possibly hold up.  A patent on a specific method for doing that,
maybe, but not on the general steps....  It's hard to claim that you've
come up with a substantial contribution to the patent if the only thing
between it and prior technologies is what you feed it.  When you look at
ogg-vorbis from such a general perspective, you'd have to consider things
like JPEG as prior art, and at such a general level, the difference is so
subtle that I doubt the patent would have any chance of holding up under
careful scrutiny, if such a patent even exists.

That having been said, it would be nice if a patent lawyer were willing to
look over the patents and offer an opinion.

Later,
David

---------------------------------------------------------------------
                    Check out my weekly web comic:
                     http://www.techmagazine.org

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

> Thing is, I understand that they have a patent on compressing audio by
> transforming to the frequency domain, quantising, and then entropy
> encoding, which is what Ogg Vorbis does, so there's trouble. 
No.  The patent (well, patents; there are several that read
approximately like this) is not anywhere as broad as you're saying.
Go read them all; there's a list linked from the LAME pages. Remember
that patent claims are like concentric circles, each further narrowing
the claims made before.  You have to read *all* the claims.

Our IP attourneys say we're totally clear.  Relax.  Vorbis does not
infringe any patent we've looked at.

Monty

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

You've probably already attacked this claim, but here's my $0.02:

        Claim # 10
        10. A coding process according to claim 9, further comprising the
steps of: 
        utilizing different code tables for different spectral ranges; and 
        setting borders between the different spectral ranges dependent upon
the signal. 

Since Vorbis does not separate the frequency ranges via a filter bank, it
cannot use "different code tables for different spectral ranges". 
Right?
> -----Original Message-----
> From: owner-vorbis-dev@xiph.org [mailto:owner-vorbis-dev@xiph.org]On
> Behalf Of Jack Moffitt
> Sent: Friday, December 15, 2000 3:09 PM
> To: vorbis-dev@xiph.org
> Subject: Re: [vorbis-dev] patents and separate entropy coding
> 
> 
> > Roger, sorry 'bout that. I am talking about patent claim
#US5579430
> > which is entitled Digital encoding process and can be found at
> > http://www.delphion.com/details?pn=US05579430__
> 
> You are talking about patent #US5579430
> 
> What you quoted from the patent is NOT a claim, even though it is
> similar.  It's just the abstract.  Abstracts tend to be broad.
> 
> Clicking on "expand details" and hunting down the page for
"Claims"
> Gives us:
> 
> Under Claim #1, it says "transforming the samples of said acoustical
> signal using a transform/filter bank"
> 
> Note "transform/filter bank".
> 
> Vorbis does not use a filterbank.  This is a specific reference to the
> polyphase filterbank stage that is BEFORE teh mdct in Mpeg audio
> encoder.  Vorbis does not use this stage, and therefore our audio
> encoding process is different.
> 
> This patent is broad though.  You could claim it covered many things,
> including things like jpeg.  I don't think claim #1 would stand up.
> Some of the more specific claims (delphion shows a total of 
> 32 claims).
> 
> So now that we all understand how to reference patents and 
> claims, this
> discussion should get much more to the meat of any problems :)
> 
> jack.
> 
> --- >8 ----
> List archives:  http://www.xiph.org/archives/
> Ogg project homepage: http://www.xiph.org/ogg/
> To unsubscribe from this list, send a message to 
> 'vorbis-dev-request@xiph.org'
> containing only the word 'unsubscribe' in the body.  No 
> subject is needed.
> Unsubscribe messages sent to the list will be ignored/filtered.
> 
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

I have an idea....    Break the encoder into 2 'separate' projects.....

One project (named 'O')  takes PCM audio and transforms it into the
frequecy
domain....  Great...

The other project (name 'gg') takes frequecy domain *data* and performs
entropy
encoding on it... Great again....

Noone will actually want to use each project separately (well, they might)....
But the normal user will always use the two separate projects in succession.....
And no one here is at fault for that....

Problem solved: ;)
    --Davy

Then another totally separate project

Lourens Veen wrote:
> Hi everyone,
>
> first of all, I searched through the archives for any posts resembling
> this, I didn't read all posts about the patents so if this has been
> suggested before I apologise.
>
> I read in an article on C|Net (I think, it was linked from Slashdot
> anyway) that Thompson are threatening to sue you if Ogg Vorbis becomes a
> success. Which is evil, and I'm also mad at them because they never
> answered my mail (which was about how to get an mp3 patent license, of
> all things). But I digress.
>
> Thing is, I understand that they have a patent on compressing audio by
> transforming to the frequency domain, quantising, and then entropy
> encoding, which is what Ogg Vorbis does, so there's trouble. Now,
I've
> been working on a Haar wavelet image compressor, and since I haven't
> written the entropy coder yet I just dump my quantised output to a new
> file. This file is as large as the original file. Then I simply use
> gzip, which compresses well because of the preprocessing.
>
> So, should trouble arise, you could make Ogg Vorbis from a compressor
> into a transformer. Since you're not doing any entropy encoding, in
fact
> you're not even compressing anything, they won't have anything
against
> you.
>
> There may be a problem with this, which is that it may in fact then be
> illegal to create .ogg.gz files without a license from them, but it
> would be pretty hard for them to hold that since they don't have any
> patents on transforming audio data or on entropy encoding random data,
> just the combination.
>
> Well, IANAL, but I thought that it might be useful.
>
> Cheers,
>
> Lourens
>
> --- >8 ----
> List archives:  http://www.xiph.org/archives/
> Ogg project homepage: http://www.xiph.org/ogg/
> To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
> containing only the word 'unsubscribe' in the body.  No subject is
needed.
> Unsubscribe messages sent to the list will be ignored/filtered.
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

I have some similar thoughts.  I'm offering them up with some hesitation,
mainly because I do not have as much invested personally in Vorbis as many
of you.  

Monty, I see your point about not wanting to tip the legal defense hand too
much, but ... 
a highly public, professional, well thought-out refutation of all questioned
patents might make the whole thing go away much faster. 

Let me explain,

It may be that Thomson genuinely believes that Vorbis is an encroachment.
If so, we should attempt to convince them otherwise.

It seems like the purpose of the suit, and especially Linde's comments, is
to increase FUD about Vorbis.  If we fight the FUD, the legal fight would be
fruitless for them to pursue.

Thomson hopes to quell Vorbis development, but the more publicity this case
gets, the worse the backlash for them if they fail.  If they fail in a court
to prove that Vorbis infringes on MP3 patents, they will not only have
exposed themselves to the world as the royalty-extracting monopoly that they
chose to be, but they will also have introduced the world to an alternative,
Vorbis.

In the end, this could be a good thing for Vorbis.  If it is proven, either
in a court of law or the court of public opinion, that Vorbis is not an MP3
ripoff and has some advantages over MP3, both legal and technical; then this
could actually act as a springboard for Vorbis.
> -----Original Message-----
> From: owner-vorbis-dev@xiph.org [mailto:owner-vorbis-dev@xiph.org]On
> Behalf Of Bruno Barreto
> Sent: Friday, December 15, 2000 3:47 PM
> To: vorbis-dev@xiph.org
> Subject: Re: [vorbis-dev] patents and separate entropy coding
> 
> 
> > Jack, do you think it's a good idea to post my detailed 
> patent reading
> > we did as a first round for the original IP work?  It's 
> good meat, but
> > would potentially be giving the meat of our argument away to Thomson
> > if this ever does come to litigation
> 
> I agree you could be giving the meat of your arguments away...
> But i think at this differently... I know one thing... IF 
> Vorbis had the
> specs
> out now open to the public (i for one don't know where i can 
> get it, let me
> know
> if it's out already) and IF vorbis developers had made clear 
> their points of
> why
> they are absolutely sure vorbis does not infringe any patent, 
> then i'm sure
> that,
> those people that today post their concerns about vorbis 
> future against
> possible
> sues, would be completely convinced and relaxed about vorbis 
> being clearly
> not
> infringing any patents. Fhg and Thompson would not sue xiph, 
> if they knew
> that
> it is clear to most people on the internet who deal with 
> audio, that vorbis
> does not
> infringe any patents indeed. Now you see... As it is not 
> clear for many of
> us, if that
> is acctually true, how can we expect that this would also be 
> clear to "Henri
> Linde " Thompson's
> vice president, who said you were infringing thomson's 
> patents? If it had
> reached his ears
> that it's widely avaiable in vorbis site, how vorbis 
> acctually does not
> infringe the patents,
> he would not have said that coz he'd knew that there was a 
> proof that what
> he was saying was wrong.
> 
> In other words, Monty, i'd guess that posting your detailed 
> patent reading,
> and your arguments,
> would calm things down, and would seal Fhg and thompson's 
> mouths, in my
> opinion.
> 
> I fear that if this is not done, you could be giving them 
> space to sue you
> guys, and thus, putting away
> investors possibly, and maybe putting vorbis project at risk, 
> if xiph runs
> out of batteries...
> 
> Regards,
> Machado, Bruno.
> 
> Keeping our fingers crossed ;) - "Aleksandar Dovnikovic"
> 
> 
> --- >8 ----
> List archives:  http://www.xiph.org/archives/
> Ogg project homepage: http://www.xiph.org/ogg/
> To unsubscribe from this list, send a message to 
> 'vorbis-dev-request@xiph.org'
> containing only the word 'unsubscribe' in the body.  No 
> subject is needed.
> Unsubscribe messages sent to the list will be ignored/filtered.
> 
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

> Thing is, I understand that they have a patent on compressing audio by
> transforming to the frequency domain, quantising, and then entropy
> encoding, which is what Ogg Vorbis does, so there's trouble. Now,
I've
Vorbis uses a vector quantizer. A vector quantizer is _not_ the same
as first quantizing, and then entropy encoding. Part of the entropy
encoding is already included in the quantizer, and they are not
separable. So this patent won't work.
> been working on a Haar wavelet image compressor, and since I haven't
Patented.
> written the entropy coder yet I just dump my quantised output to a new
> file. This file is as large as the original file. Then I simply use
> gzip, which compresses well because of the preprocessing.
> 
> So, should trouble arise, you could make Ogg Vorbis from a compressor
> into a transformer. Since you're not doing any entropy encoding, in
fact
> you're not even compressing anything, they won't have anything
against
> you.
Problem is, any application that uses Vorbis, _will_ have to license the
patent. Otherwise, Vorbis would be clean trivially, as most patents
talk about "storing a file" or "transmitting a signal",
which libvorbis
does not. But Vorbis has problems with patents that prevent applications
from using Vorbis as well as with patents which threaten Vorbis itself,
as Vorbis is meant to provide license/patent/claim-free natural audio
compression.
> There may be a problem with this, which is that it may in fact then be
> illegal to create .ogg.gz files without a license from them, but it
Read the patents (very carefully) please; this doesn't apply. Oh,
and if you first create an .ogg file, and than an .ogg.gz file, it
won't apply anyway, as some user action is required (patent says: a
device which does... etc.)
> would be pretty hard for them to hold that since they don't have any
> patents on transforming audio data or on entropy encoding random data,
> just the combination.
> 
> Well, IANAL, but I thought that it might be useful.
At least you were thinking; most people complaining about patent
stuff ain't. Most of those people are just reiterating (with
some mistakes)  what they read on some news-site (which had obvious
errors to begin with). But no hard feelings to anyone, of course.
just getting tired of reading false statements over and over again.

Dagdag,

Segher

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.