Patent on a novel DSP Algorithm?

Dear all,

I have a few questions about patenting on DSP algorithms:

Suppose I have come up with a novel DSP algorithm, say, a fast
implementation scheme for real time video coding.  What is the best way to
earn profits out of it?  other than starting a company by myself?

I believe I can file a patent, and make profits when it is used by others.
 Can I still get a patent if I publish and disclose the results in a
technical conference?  Are there a lot of people out there who read
proceedings from a conference, and then file patents before the original
researchers do?  What would be the best way to protect yourself from
this?

Thanks!

CJLam
		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

See "Patent Law Essentials" by Durham for US Patents.

But... if your idea is really worth anything I would not ask around for
advice or research it on your own... see a patent lawyer.  You do it
wrong... you lose!

Dirk

dbell wrote:

> See "Patent Law Essentials" by Durham for US Patents.
> 
> But... if your idea is really worth anything I would not ask around for
> advice or research it on your own... see a patent lawyer.  You do it
> wrong... you lose!
> 
> Dirk
> 
And ask about the lawyer -- he does it wrong... you lose!

My dad lost a patent because his lawyer did it by the book -- he applied 
for a patent search, then waited to file the patent.  It turns out that 
there are firms that specialize in jumping interesting patents by 
intercepting them at the patent search then filing them without one.

Just another little way that our patent system is screwed up.

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

in article GaWdnRaNN7aZ4N_eRVn-sg@giganews.com, clam at jethrolam@gmail.com
wrote on 10/04/2005 09:00:

> Suppose I have come up with a novel DSP algorithm, say, a fast
> implementation scheme for real time video coding.  What is the best way to
> earn profits out of it?  other than starting a company by myself?

DSP engineers come up with novel DSP algorithms often.  that, among other
things, is what they get paid for.  if you're so confident that your
algorithm is both novel and very useful (that it will result in a
significant bandwidth or memory requirement reduction or improve performance
in some way that saves time and/or money) and you're not going to start a
business yourself with it, then you might want to find a venture capitalist
that you could persuade to underwrite developing and marketing of this.  you
could also try approaching an existing company that is in the same biz.


> I believe I can file a patent, and make profits when it is used by others.
> Can I still get a patent if I publish and disclose the results in a
> technical conference?

in the U.S., i think you need to file within a year of any disclosure you
make.  otherwise it becomes prior art.  dunno what it is in other countries.
of course, if someone else publishes an algorithm with the same salient
elements first, then you're scre3wed.

>  Are there a lot of people out there who read
> proceedings from a conference, and then file patents before the original
> researchers do?

if it wasn't their own paper being presented at the conference, and they try
patenting the idea of someone else, i would guess that this would be a
slam-dunk for challenging it as prior art.

>  What would be the best way to protect yourself from this?

talk to a patent attorney.  they ain't cheap but if your idea is worth
money, it might not be a bad idea.  if i recall correctly, count on spending
upward from $20K to patent anything.  if that is too costly, many companies
rely only on trade secret (so forget about publishing or presenting at a
conference) to protect their IP.

-- 

r b-j                  rbj@audioimagination.com

"Imagination is more important than knowledge."

In general, it's better to do not assume that patent will protect you.
Patenting is one thing, real commercial world is the another. Prototype
your product first and then you may see what to do next.

Kind regards,
Yuri

Hi CJ,
  I was in the same situation. You should not publish anything,
anywhere prior to getting your application in and having your "priority
date" from the patent agency. Even then you may wish to reveal it only
to a few interested parties, under a confidentiallity agreement.Later
of course the patent is published for the world to see and to contest
if anyone wishes.
 Actually, if you are a private individual with limited funds,  the
cost to protect it in all the top economies would be too great.
 Next, you might also question the knowledgeability of a patent agent
unless they are particularly specialised and expert. Some feel that
they and the patent authorities sometimes process technically complex
applications and then let the lawyers sort out any challenges.
Can you finance a legal challange ?
 Next if your invention is an algorithm, can you a) be sure it is not
based on some obscure, pre-existing technique and b) can you detail it
in such a way that someone would not be able to get the same benefit by
a different algorithm that got to the same end result, making yours
interesting but of no commecial value.

 I asked the same questions that you did a few years ago and the advice
I got from an ex-AMD exec. was "to make product" actually I have done
the patent route also, just for my own satisfaction but making an
innovative product that no-one else can is the best investment of time
and money IMHO.

 I did also approach a number of well-known companies, once the patent
was filed and the result was   ....... Silence. Either the "Not
Invented Here" syndrome or maybe they wait and then use superior "fire
power" to challenge should the idea have some merit.

 I wish you all the best in your adventure, Steve

Go to uspto dot gov
then go to the patent section
and do an advanced search to get
an idea of what the US patent office
is looking for.

Good luck,
Dave

Well, are you really sure that your idea is that good?
Most of the DSP patents in USPTO database are complete and utter junk -
they are totally useless for a small company to have and can only be
used by big multi-nationals and their legal departments to prevent
other companies from entering the field and for cross-licensing.
This is not a game a small start-up company can play with just one or
few patents.
For the same reason you should not probably try to make and sell your
own product (it really depends on a type of product), while trying to
enforce your patent rights against other companies: the big guys will
gang together to fight your patent and will enforce their own combined
patent portfolio against your product. Whatever your product is, they
will find something that infringes on their patents and, at the very
best, will force you to cross-license your patent -essentially give
it to them for free.

What else to say?  Patenting is not for the faint of the pocketbook.
Unless you have some cash reserves, stocks, investments, equity etc.,
or better yet, a rich uncle to back you up in case you run our of
money, you might as well just forget about patenting.
Most probably you will not see any money from patent licensing for
5-10-15 years, that is, if you are lucky to see any money at the end...
On the contrary, you will be spending considerable sums on patent
layers, PTO fees...
Patent is a long-term commitment to make ...

And don't expect big companies to respect you patent: as soon as your
idea is made public they will just implement it in their products,
patent or no patent, they don't care...
As somebody said it before, when dealing with big companies,  "there
is no such thing as a free lunch, unless you're the lunch".

The best advice I can give you right now is to buy a book by David
Pressman "Patent It Yourself", study it for a while and, after you
are comfortable with terminology, file your Provisional Patent
application with USPTO by yourself without paying anything to patent
lawyers yet (it's cheap - 80 bucks or something to file). This will
lock your priority date and will give you a year to decide if you
really want to continue along this route.
After all, it is your decision and your luck...

If I "wasn't quite sure" and had limited funds. I would spend some time on 
line searching through prior art http://www.uspto.gov/patft/index.html . 
Then I would read lots patents to try to understand the lingo and to 
understand why they are written the way that they are. I would look for a 
provisional patent example and use that as a template to write my own 
provisional patent. File it for $100. Wait for receipt. I am now safe (for a 
year) to promote my idea to whom ever I please and to see if it has merit. 
If it gets interest, I would pay for a formal patent before the year is up.

Thomas

"clam" <jethrolam@gmail.com> wrote in message 
news:GaWdnRaNN7aZ4N_eRVn-sg@giganews.com...
> Dear all,
>
> I have a few questions about patenting on DSP algorithms:
>
> Suppose I have come up with a novel DSP algorithm, say, a fast
> implementation scheme for real time video coding.  What is the best way to
> earn profits out of it?  other than starting a company by myself?
>
> I believe I can file a patent, and make profits when it is used by others.
> Can I still get a patent if I publish and disclose the results in a
> technical conference?  Are there a lot of people out there who read
> proceedings from a conference, and then file patents before the original
> researchers do?  What would be the best way to protect yourself from
> this?
>
> Thanks!
>
> CJLam
>
> This message was sent using the Comp.DSP web interface on
> www.DSPRelated.com 

Thomas Magma wrote:
> ... I would read lots patents to try to understand the lingo and to
> understand why they are written the way that they are. I would look for a
> provisional patent example and use that as a template to write my own
> provisional patent. File it for $100. ... If it gets interest, I would pay
> for a formal patent before the year is up.
>
> Thomas

You can write an informal application and file it as a provisional
application for a patent.  The object of the provisional application is
to establish an early priority date for a subsequent regular
application.  HOWEVER, if your provisional application is informal and
does not meet the requirements of the patent law (specifically 35 USC
112) the provisional app. may prove to be of little value if you try to
use it as the basis of a subsequent regular application for a patent.
In other words an unprofessional, hastily written provisional
application may not even be worth your time, effort, and the $100
filing fee.

Richard Tanzer
Patent Agent