Posted by PTravel on 07/14/06 03:22

"Toby" <kymarto123@ybb.ne.jpp> wrote in message
news:44b704f8$0$74470$bb4e3ad8@newscene.com...
>
> "PTravel" <ptravel@ruyitang.com> wrote in message
> news:4ho6s0FhtiuU1@individual.net...
>>
>> "Toby" <kymarto123@ybb.ne.jpp> wrote in message
>> news:44b6e10a$0$74478$bb4e3ad8@newscene.com...
>>>
>>> "PTravel" <ptravel@ruyitang.com> wrote in message
>>> news:4hnbilFdriuU1@individual.net...
>>>>
>>>> "Toby" <kymarto123@ybb.ne.jpp> wrote in message
>>>> news:44b5b15e$0$65941$bb4e3ad8@newscene.com...
>>>>>
>>>>> "PTravel" <ptravel@ruyitang.com> wrote in message
>>>>> news:4hfsbtF1rcvsvU1@individual.net...
>>>>>>
>>>>>> "Richard Crowley" <richard.7.crowley@intel.com> wrote in message
>>>>>> news:e8ucis$8o5$1@news01.intel.com...
>>>>>>> "PTravel" wrote ...
>>>>>>>> We're still talking about apples and oranges -- data reconstructed
>>>>>>>> using ECC is accurate and bit-for-bit what was recorded (or
>>>>>>>> supposed to have been recorded).
>>>>>>>
>>>>>>> No. That is the description of how ECC works for computer
>>>>>>> data. There is another MAJOR error recovery layer used by
>>>>>>> audio CDs, DVDs, DVD tape, etc. It is how they manage to
>>>>>>> be cost effective.
>>>>>>>
>>>>>>> Audio/video data that cannot be accurately *corrected* gets
>>>>>>> *mitigated* by extrapolation. This is the factor that is present
>>>>>>> in audio and video digital recording but NOT in computer data
>>>>>>> recording. It probably comes into play more often than we think,
>>>>>>> but if it is good enough we don't notice. But multiple repetitions
>>>>>>> of it amount to something approaching "generation loss".
>>>>>>
>>>>>> Can you site me to something that says how often this happens? I've
>>>>>> never heard of DV video transfers being described as anything other
>>>>>> than lossless.
>>>>>>
>>>>>>>
>>>>>>>> Replacing pixels (or scan lines or averaging blocks) results in
>>>>>>>> lost data.
>>>>>>>
>>>>>>> I think you have that backwards. Replacing pixels, etc. is one
>>>>>>> of the major migitating actions taken as a result of lost data.
>>>>>>
>>>>>> I meant that the resulting data is not the same as the original,
>>>>>> hence there is data lost. But how often does this happen?
>>>>>
>>>>> This reminds me of a story. How often do you think earthquakes happen
>>>>> in Japan? Once a month? Once a week? We recently did a doco about
>>>>> earthquakes here and visited one of their seismo centers. It turns out
>>>>> that the rate of measurable earthquakes is about 100 an hour.
>>>>>
>>>>> Do you see what I am getting at?
>>>>
>>>> Yes, which is exactly what I've been getting at. Borrowing Martin's
>>>> terminology, what is the ratio of mitigated errors to fully-corrected
>>>> errors? If it's comparable to the ratio of major earthquakes in Japan
>>>> versus small tremors that no one notices, then we're looking at
>>>> Richard's guestimate that works out to 1 mitigated error per 666 hours
>>>> of video. With that kind of rate, miniDV is, for all intents and
>>>> purposes, lossless.
>>>>
>>>> However, we remain in the realm of speculation -- without hard data,
>>>> this is all pointless, and I'll stand by my position, i.e. there is no
>>>> generation loss for D-25 transfers.
>>>
>>> OK, here are a couple of clues. First a quote from an ad for Sony tape:
>>>
>>>
>>> "DVCAM tape has a 50% lower dropout rate vs DV resulting in a four- to
>>> five-fold improvement in the error rate margin, which yields a better
>>> picture."
>>
>> Irrelevant (and marketing).
>
> Why is this more irrelevant than the Adobe claim, which is also marketing?

They're both irrelevant. The only relevance of the Adobe claim is that I
will accept it before an unsupported assertion by a Usenet poster because
Adobe is on the hook for FTC violations if it's wrong.

>
>>
>>>
>>> So somebody is lying, is it Adobe or Sony?
>>
>> "Better picture" is one of those subjective terms that are legally termed
>> "puffery." However, Sony's material could easily be interpretted as
>> saying, essentially, DVCAM is better than DV because the 50% lower drop
>> out rate means one mitigated error per 1332 hours of video, versus one
>> mitigated error per 666.
>
> So Sony is lying? Or is it Adobe?

I don't know. Neither do you.

>
>>
>> Hard evidence. Not speculation.
>
> Where is the hard evidence on the Adobe claim?

There isn't any. Context is everything -- I've explained the context in
which the Adobe claim was relevant.

>
>>
>>>
>>> Here's a quote that perhaps expands on that theme:
>>>
>>> "Finally, banding or striping of the image occurs when one head of the
>>> two on the scanner is clogged or otherwise unable to recover data. The
>>> image will show 10 horizontal bands (12 in PAL countries), with every
>>> other band showing a "live" picture and the alternate bands showing a
>>> freeze frame of a previous image or of no image at all (or, at least in
>>> the case of the JVC GR-DV1u, a black-and-white checkerboard, which the
>>> frame buffers appear to be initialized with). Most often this is due to
>>> a head clog, and cleaning the heads using a standard manufacturer's head
>>> cleaning tape is all that's required. It can also be caused by tape
>>> damage, or by a defective tape. If head cleaning and changing the tape
>>> used don't solve it, you may have a dead head or head preamp; service
>>> will be required.
>>>
>>> This sort of banding dropout occurs fairly often; about once per DV tape
>>> in my experience.
>>
>> You can't be serious. I've never experienced that kind of banding
>> dropout. Do y ou keep your equipment clean? Do you mix dry and wet lube
>> tapes?
>
> This is a quote from a professional. Note that he says: "banding or
> striping of the image occurs when one head of the two on the scanner is
> clogged or *otherwise unable to recover data*."(emphasis mine). And
> clogging isn't always an issue due to inadequate cleaning. Ever hear of
> oxide flaking? It happens routinely even on the best tapes.

I didn't say it was the result of inadequate cleaning. It can result from
over-reusing tapes, from a bad batch of tape, from humidity and other
environmental conditions, etc. Notwithstanding what the professional, I've
never seen it with my only-reasonably-well-cared-for prosumer camcorder. Of
course, I don't reuse tapes, stick to one band, and don't use it in harsh
conditions.

>
> Take note that a figure for this kind of unrecoverable error is given in
> the last line, and it is more than an order of magnitude larger than your
> contrived one.

I didn't contrive it -- I'm merely repeating the guestimates that have been
ventured in this newsgroup. My number is pure fiction. The number in the
article doesn't comport with my own experience. Both numbers are anecdotal
and not based on hard data.

>
>>
>>
>>> Usually it isn't even noticeable -- a single frame of banding due to a
>>> momentarily clogged head won't be visible unless there's motion in the
>>> scene to show off the frozen stripes.
>>
>> Since I rarely shoot with a tripod, my video is always moving.
>
> Irrelevant.

Hardly irrelevant. You: it won't be visible unless there's motion. Me: in
my videos, there's always motion, yet I've never seen it.

> It means that there is lost data due to dropouts, hence the ECC is not as
> perfect as you seem to so desperately want to believe.

Then why haven't I experienced it?

>
>>
>>> Have a look through your old tapes frame by frame (on a slow day, of
>>> course!) and you might be surprised how often you'll be able to find a
>>> single, subtly banded frame.
>>
>> I don't go through my video frame by frame, but I do look at an awful lot
>> of single frames in the process of editing a project. Never seen it.
>
> Which means that it doesn't exist?

Nope. It means that my experience is contrary to the anonymous author of
your article. I'll trust my own eyes.

> And BTW, how many is "an awful lot"? Just for reference, there are roughly
> 30 frames/sec, which translates into approximately 108,000 frames on a 60
> min. tape. Would you say that generally you carefully view 1% of all
> frames shot? 10%? Give me a figure here...

1 or 2% is probably fair, though I don't "carefully view" each frame.
However, I do view them carefully enough to note any errors as obvious as
the banding described in the article. As an example, I had a title saved as
a style in Premiere that I used throughout a project. When I created it, I
inadvertently stuck a single pixel off to the side -- that's one pixel out
of a 720 x 480 frame. I had no trouble noticing it.

>
>
>
>>
>>> For what it's worth, I've only rarely found such a banded frame on any
>>> DVCAM footage I've shot, which indicates to me that DV is right on the
>>> edge of reliability. DVCAM, with its 15 micron track width, or DVCPRO
>>> with its 18 micron track, are sufficiently on the safe side of the
>>> bleeding edge so that this sort of droput is much less likely to occur.
>>>
>>> Bear in mind that analog BetaSP typically has several dropouts per
>>> minute; the last time I measured visible dropout rates on Hi8 and S-VHS
>>> I got numbers in the range of a dropout every 3-5 seconds (Hi8) and
>>> every 7-20 seconds (S-VHS). One visible dropout per hour-long tape, on
>>> average, is not something to get flustered about. But if it does bother
>>> you, shoot DVCAM or DVCPRO instead."
>>>
>>> "Digital Dropout: DV is highly resistant to dropout because the same
>>> data is recorded multiple times within the track. The playback
>>> electronics compare the multiple data streams, and reject those that are
>>> corrupted by an oxide particle dropping off the tape. The dropouts are
>>> still there, but you don't see them on the screen. But if the dropout is
>>> big enough, the correction circuits don't have enough data to work with,
>>> causing digital dropout, which results in a sudden blockiness in the
>>> picture. The worse the dropout, the larger the blocks, until the picture
>>> is lost completely. They are much rarer than analog dropouts, but
>>> they're much more noticeable when they do happen."
>
> Read the last part of this carefully, starting with: "But if the dropout
> is big enough..."
>
> No comment on this, counselor?

Do you see any quantification of this kind of error? I don't.

>
>>>
>>> Finally, view this pdf file if you want to see dropout rates of DV vs.
>>> DVcam graphically displayed with counts/min.
>>>
>>> http://www.pmdmagnetics.com/displayfile.asp?id=48140.
>>>
>>> It's pretty good, but it ain't perfect.
>>
>> It's completely irrelevant. The metric is: "mitigated" errors versus
>> perfectly corrected errors.
>
> Banding shows an unrecoverable error. Block replacements signify
> unrecoverable errors. I've presented you with evidence that such errors do
> indeed exist, and in fact one source (a professional, no less) has given
> the approximate rate of such errors on DV tapes, which he compares
> unfavorably with that of DVCam.

I never said they don't exist -- that's a Martin argument. I said that I
don't have any reason to believe that they occur so often as to result in
generational error. As I've said repeatedly, quantify these kind of errors.
Until you can, you're just speculating, and I have no interest in
speculation.

>
> The simple fact is that if DV were really error-free there would be no
> reason at all to have DVCam and DVCPro formats, and even if those formats
> existed only for the purposes of marketing, professionals would soon learn
> that they could save lots of money by using DV instead.

DVCam and DVCPro formats exist because they are more resistent to drop out.
What has that to do with this discussion?

>
> As a professional in this business for well over twenty years, I can
> assure you that not a single network or production company with which I
> have contact here in Japan (CNN, NBC, CBS, ABC, ABC Australia, ORF, ZDF,
> ARD, BBC, Reuters, APTN, Swiss TV, NHK, TBS, Fuji TV, TV Asahi, NTV, SBS,
> YTN, MBC, KBS, to name the main ones) uses miniDV as a primary aquisition
> source or within the production chain post-acquisition. There are reasons
> for that, and one of them is not that they like to spend more money than
> necessary.

I'm sure that's true. What has that to do with this discussion?

>
> Now you can huff and you can puff and make a lot of fancy noise and pull
> your lawyerly rabbits out of your lawyerly hat-- you might even win your
> case with clever sophisms--but that doesn't change the reality one whit.

Except that you don't know what reality is, despite my asking you to tell
me. Quantify the "mitigated" error rate. If you can't, you're just blowing
smoke. That's not a "sophism," but a simple fact. Evidence is compelling.
Opinion isn't.

>
> Toby
>
>

[Back to original message]