|
|
Posted by PTravel on 07/13/06 16:39
"Toby" <kymarto123@ybb.ne.jpp> wrote in message
news:44b5b22a$0$65966$bb4e3ad8@newscene.com...
>> Sorry, but that's not what I'm asking. The contention here is that there
>> is generational loss in straight copies of digital video data, and I've
>> never heard that anywhere else. For generational loss to occur, there
>> have to be gross drop out errors that can be fixed by conventional ECC
>> and, per you and Martin, are fixed by extrapolating based on prior or
>> subsequent video data, i.e. pixels are filled in. I'll take your word
>> that happens, but I don't believe it happens often enough to be a
>> concern. I'd like to see some data that suggests that, notwithstanding
>> every source I've ever seen that says digital video dupes are lossless,
>> this isn't so (see the quote from Adobe in a post I made to Martin).
>
> For most practical purposes they are loseless, and Adobe would probably
> win in court with their fradulent claim.
Either they are lossless or they are not. If they are not, Adobe is wrong.
As for whether they'd win in court, let's just leave at this: I'm a lawyer,
this is the kind of law that I practice, and you're wrong.
> That doesn't mean that it is truly loseless, for in that case *every*
> single instance of error correction would have to be bit-perfect
> reconstruction based on redundancy, and it is not.
What is the uncorrectable error rate? Until that is established the
discussion that this thread has veered into is moot.
>
> I think you'd better define what you mean by "gross drop out errors". At
> what point do they become gross, counselor?
Errors that the ECC can't restore to the original data and, instead, must
extrapolate or simply guess -- what Martin has been calling "mitigated"
errors (a pretty accurate term, actually).
>
> Toby
>
Navigation:
[Reply to this message]
|