[Speech-reco] Re: Can I copy this code directly?

[Bill Cox]

I talked to one of the authors of the patent.  Actually, it was pretty
cool.  Here's more or less what he had to say:

The technology was never really tried for speech recognition, but he
thinks it has promise.  Compute power was one of the major drawbacks
back then, but that was in the 1970s and early 80s, and compute power
is no longer a problem.  He doesn't know if I can copy the code, but
thinks any good patent lawyer should be able to answer that.  He hopes
that we can copy the code, and that we can find a good use for it.  In
any case, the ideas in the patent are now free for use.

Bill

> http://www.freepatentsonline.com/4625286.html
>
> The other one is not very interesting to me.
>
> Bill
>
> On Thu, Oct 14, 2010 at 8:51 AM, Bill Cox <address@hidden> wrote:
>> I think the technology in this patent rocks:
>>
>> http://www.freepatentsonline.com/5012518.html
>>
>> In 1978, this guy compressed speech to the most important 4 formants
>> (poles), plus a second order correction polynomial.  It's based on the
>> same technology I've been playing with, LPC.  He even does what I
>> wanted to do, making traces of formant movement over time.  The patent
>> has expired, which means I can duplicate as much of this work as I
>> like for free.  Even better, the patent includes an Apendix A, which
>> is the entire Fortran source code for the algorithm.  I see no
>> copyright notice on the code.
>>
>> So, am I actually allowed to copy the Fortran code directly?  Doesn't
>> putting it in a patent application with no copyright notice place it
>> into the public domain?  Too bad we don't have blind lawyers on the
>> list!
>>
>> Bill
>>
>