Bits are bits

[March Audio]

I'm not obsessed I'm just saying nothing is perfect,otoh I'm an R&D guy I have to
Lets imagine that the majority of the drives in your democratic database are low cost ones with poor performance .
100% certainty is just like bit perfect.( I'm a fan of Heisenberg)

Quite, nothing is perfect, but it is more than good enough to be a total non issue.

Yeah sure, imagine. We would see a wide range of different crc values. In the real world this isnt happening.

 

[RayDunzl]

The Phone Companies were satisfied with 9,998 call completions per 10,000 attempts during acceptance testing.

9,997 and we had to pretend we fixed something and do it again. For 24 hours. Number of total calls scaled to the size and engineered capacity of the office.

Lots of load boxes running sometimes.

 

[Thomas savage]

Iv deleted a few insults, this is not the place for that kind of thing.

 

[Jinjuku]

Iv deleted a few insults, this is not the place for that kind of thing.

Honest to god. I want to know what equipment has had his hands on it so I can avoid either recommending it or purchasing it!

Serious as a heart attack.

 

[j_j]

The OP, if it's of either a SPDIFF or AES signal, either in optical (with receiver) or voltage, is meaningless.

It ignores the veriest basics of information theory. What the question remains is "what happens when the bits are demodulated", no more, no less.

Remember, unlike analog, for which error must grow every copy, without fail, due to basic physics, error in digital does not grow at all until you fail completely to detect anything meaningful, and furthermore, in a real digital transmission system, that failure is evident at the receiver. With an intelligent receiver, you can even tell you have failure without error correction, by looking (mechanically, not with the human eye!) at the "eye pattern".

Please see "Shannon Information Theorem" for more detail.

 

[j_j]

As to CRC's and other verification methods. With sufficient length of CRC (and the math of how to do these is hardly in question) one can push the chance of error down to ridiculously low levels when comparing two files.

Consider a 128 bit CRC/Hash of a CD. Please. Figure out how many things have to be different, even using the first 128 prime numbers (yes, that's weak, even) residuals. That, thar, is a (*&(* big number.