Posted by Martin Heffels on 01/12/64 11:52

On Mon, 10 Jul 2006 13:22:17 -0700, "PTravel" <ptravel@ruyitang.com> wrote:

[...]
>That's still apples and oranges. I didn't say that this kind of
>"correction" isn't done. I said that it's extremely rare and the kind of
>common "drop out" that you've been discussing is fixed by ECC in such a way
>that no data is lost.

Google a bit on "bit error rate", and hop into the hits at the patent
office. You'll find some interesting documents about designs of tapes, and
how reducing it's size increased the BER above accepteable levels. It says
to me that they occur much more than we think.
It is very hard though to find any sensible figures on it. I have read an
article about it in an SMPTE-magazine, but their magazines are not online
:-(

Here's some links with some info/tables/test-results:
http://www.pmdmagnetics.com/displayfile.asp?id=48140
http://www.urbanfox.tv/articles/formats/fm3test.htm#generation

This one quotes some figures:
http://www.creativevideo.co.uk/public/pdf/sony_dvcam_2005-6_inc_xdcam.pdf
DV: 100 counts/min, DVCAM: 50 counts/min

>You see above. You understand how ECC works, and you also understand that
>ECC is capable of correcting small errors without data loss.

Yes.

>
>>
>>>We're talking apples and oranges here. On the one hand, there's
>>>everyone's
>>>DV deck that can invent data if it's missing,
>>
>> There is no deck which "invents" missing data. They interpolate, based on
>> what is there. Inventing means coming-up with something completely new
>> which didn't exist before.
>
>Now who's arguing semantics? Interpolated data is data that is different
>than what was actually recorded, hence it's invented.

No it is not. It is based on something which exists, and most likely to be
correct. If you see nasty blocks, that is data which is invented!

[...]
>Doesn't matter. The kind of gross data correction that results in
>"interpolation" of data is extremely rare.

You just assume this, because you have never seen it.

[...]
>No, I didn't. I meant that analogue waveforms get smoothed in an analogue
>copy, that noise is introduced, etc. Impedence mismatch is one way that an
>analogue signal can be degraded, but it's hardly the only way.

Understood. Indeed, not the only way.

[...]
>That's right. How often do you think miniDV experiences drop out so severe
>that it results in an uncorrectable error?

Well, watching tv, I see digital drop-outs quite a lot, from stuff recorded
on tape. On a personal level, I once had this bad batch of DVCAM-tapes,
which had a drop-out every 5 seconds. On my personal material maybe once
every 100 tapes (discounting the older stuff shot on LP).

[...]
>If it's been error-corrected, it doesn't matter, because a bit-for-bit copy
>results and there is no generational loss. If it has been concealed, it
>does matter but I don't believe that happens often enough (with proper
>technique -- I'm sure you can induce significant drop out if you put your
>mind to it) to be a consideration when duping digital data.

Well I'm glad you believe us now that there is a difference in
error-correction methods.

>>>I don't think it's pristine. I do think that the ECC for D-25 is
>>>sufficiently robust so that, by using quality tapes only once, and
>>>maintaining my camcorder, I can produce digital transfers to and from the
>>>computer without any dataloss.
>>
>> Yes and no.
>
>In my case, yes and yes, as well as in the case of others who have posted
>about drop out to this newsgroup, e.g. drop out reported due to a bad tape
>batch, etc.

Agreed, Above mentioned will avoid the risk of drop-outs tremendously.
Another important factor to maintain not running into drop-outs, is to
store the tapes properly.

>My capture software can stop on error -- never had it happen, though.

That's the advantage of just editing your own material :-)

cheers

-martin-
--
"If he can he'll smile 'cos he's a Royal Crocodile."

[Back to original message]