Archived from groups: rec.video.desktop (
More info?)
In article <113kr4j95en5078@corp.supernews.com>,
"Alpha" <logos1@trip.net> writes:
>
> "John S. Dyson" <toor@iquest.net> wrote in message
> news:d1dqcu$2gka$1@news.iquest.net...
>> In article <113knkahghg3qec@corp.supernews.com>,
>> "Alpha" <logos1@trip.net> writes:
>>>
>>> "PTRAVEL" <ptravel88-usenet@yahoo.com> wrote in message
>>> news:39um6aF5f9jj9U1@individual.net...
>>>>
>>>> "Alpha" <logos1@trip.net> wrote in message
>>>> news:113k515dk1qoo35@corp.supernews.com...
>>>>>
>>>>> "MrTobor" <tobor8man@hotmail.com> wrote in message
>>>>> news:1111101564.390811.167540@o13g2000cwo.googlegroups.com...
>>>>>>I have tons of old 8mm tapes that I'd like to bring into the digital
>>>>>> age (specifically to DVD). What I'd like to do is transfer all my raw
>>>>>> footage to DVD (or CD) at the maximum quality level with no
>>>>>> editing--just straight transfer.
>>>>>
>>>>> Buy a stand alone recorder...Pioneer or Panasonic.
>>>>
>>>> Do NOT buy a stand alone recorder.
>>>>
>>>> The OP wants to archive analog video that will later be edited.
>>>
>>> He wants DVD. DVD. Brain dead?
>>>
>> He did say that he wanted to capture to DVD, subsequently edit
>> and then produce a final DVD. Actually, it isn't a really good
>> idea to do two concatenated (term of art, means multiple compression
>> decompression) of low rate MPEG2 -- DVD is a little too limited
>> at about 8-9mbps for the highest quality.
>>
>>>
>>>
>>>>He wants the video at highest quality.
>>>
>>> DV 25 mbit second is highest?
>>>
>> DV 25mbit with good quality (current) encoders is damned good -- but
>> not quite as good as the nearly transparent DV50 or DigiBeta. Claiming
>> that DVD is 'highest quality' would be utilizing alot more poetic
>> license than claiming that DV25 is 'highest' quality.
>>
>> One big quality issue is that neither DV25 nor MPEG2 like to see much
>> noise
>> in the video, but DV25 with current generation encoders can deal with
>> more random content. It would make sense to use a super high quality
>> DV25 encoder (e.g. Canopus ADVC300), and then subsequently use a
>> non-realtime
>> (non-causal filtering) MPEG2 encoder with lots of noise reduction. You
>> can get the benefits of noise reduction with almost NO artifacting and
>> no quality loss relative to consumer tape formats. Almost as good as
>> using
>> 'best quality' DV25 encoding, using almost unlimited MPEG2 bandwidth (e.g.
>> 15mbps or faster) can be useful to allow the MPEG2 encoder to deal better
>> with totally random input (noise in consumer video.)
>>
>> * When recording noisy video onto MPEG2, if there isn't a heap of
>> noise reduction, there is much less advantage to the inter-frame
>> compression. It is still more effective than DV25, but also more
>> artifact prone than just to ignore inter-frame information and just
>> use DV25 compression. (AFAIR, DV25 still tries to take advantage of
>> intra-frame similarity by optionally taking advantage of the similarity
>> of two fields, but the payload capability of DV25 allows for more
>> noise data.) When ignoring the noise in DCT domain compression, there
>> is some artifacting that makes it sometimes look worse than other
>> kinds of information limiting.
>>
>>
>> Perhaps the biggest theoretical loss of quality when mastering DV25 onto
>> MPEG2 is the mismatch of the 4:1:1 to 4:2:0 color sampling. This is
>> almost
>> a non-issue given composite or color-under (consumer analog) video
>> sources.
>>
>> Basically, using DV25 before MPEG2 does limit the chroma response from
>> 3+MHz to 1.4-1.5MHz... This is significantly greater chroma than what
>> can be supplied by Hi8, 8mm, BetaMax, SuperBeta, EDBeta, SVHS, laser disk,
>> OTA video, CED Video or any other common consumer video source. Consumer
>> video sources tend to be limited to approx 500kHz (or less) of color
>> data per component, except LD which can theoretically approach 1.2-1.5MHz
>> in both components, if some rules are bent. Otherwise, the chroma
>> bandwidth would tend to be limited to approx 1.2MHz and (afair 0.5MHz?)
>> (Using widerband LD chroma might introduce some artifacting on decoders
>> that aren't prepared to handle it.)
>>
>> It is true that the realtime MPEG2 encoders (used in set-top DVD
>> recorders)_
>> can be pretty good, but to get the maximum, noise free quality, it can
>> be better to use offline (often very slow) noise reduction. This slowness
>> can gain significant quality because of the potential for non-causal
>> filtering. (Offline noise reduction can look forwards and backwards in
>> time, while real time processing TENDS to be more limited.)
>>
>> It isn't WRONG to use a set-top DVD recorder, and if one is used, then
>> just max-out the quality settings, and be careful about comet-tailing
>> (or other motion artifacts) when enabling the (very much needed)
>> online, real-time video noise reduction for MPEG2 encoding of noisy
>> consumer video.
>>
>> John
>
> I do not disagree. But this individual is not going to research all of the
> data, spend months of testing Codecs, etc. The set top boxes have various
> forms of noise reduction geared for NTSC. Many of these are equal to very
> expensive proprietary approaches. I presumed these variables in my
> statement.
>
I have seen some examples of motion artifacting when using real time noise
reduction before MPEG2 compression... This is one reason why I tend to resist
that scheme. It seems like alot of people are tolerant of noise reduction
artifacts that I am not tolerant of. When I made my suggestion, I was
trying to make sure that the 'master' copies have the least damage possible.
If a practical master is made to 'DVD', then the noise reduction artifacts
(LARGE amounts of noise reduction are often very necessary for MPEG2 encoding)
will be imprinted on the signal forever.
If this wasn't essentially a 'master' copy, then noise reduction artifacts
onto a distribution copy is of less consequence.
Indeed, if the original poster is interested in a 'quick' solution, then
a set-top DVD recorder will necessarily be the best (but non-optimal
quality) solution. On the other hand, using a compression scheme that
can withstand some video noise (definitely not MPEG2), and with reasonably
limited real-time video noise reduction (e.g. the ADVC300) and timebase
correction, then the slightly imperfect, but minimally damaged master
images can be DV25 format. (Of course, a practical set top DVD recorder has
to have lots of noise reduction and some kind of tolerance of time
base errors, not claiming that a set top DVD recorder is necessarily
lacking features.) One could take the suggestion to an extreme, and
master onto DV50 or DigiBeta, where the master would be minimally
damaged. In fact, that is EXACTLY what I do before mastering onto DV25...
With a minimally damaged DV25 recording, then subsequent non-realtime
noise reduction can produce a nearly artifact free MPEG2 video, probably
significantly better than the MPEG2 master that is produced by the
set-top box. The 'sad' thing is that the damaged MPEG2 master with
a set top box will likely be used (after some editing/processing) to
produce an even more damaged MPEG2 copy. If minimal processing/changes
are made to the master copy while recording onto the distribution
master, then it is possible that the additional artifacts due to recompression
will be limited (hopefully.)
How much damage is the user willing to accept? This depends upon his/her
tolerance to spending time on finding/developing and implementing
a solution to the problem along with simple tolerance to video defects.
Truly, it isn't necessary that video always be perfect.
It is possible that the 'probable' lower quality option of multiple MPEG2
encode/decode cycles will be MORE DESIREABLE than the 'probable' higher
quality option of using reasonably well encoded DV25, and then production
of an MPEG2 product. It is definitely easier to use a set top DVD recorder,
and the results just might be adequate for the user!
In my own case, I cannot stand the motion artifacts that I have seen with
noise reduction before MPEG2 encoding, but on the other hand, my current
setup doesn't produce visible artifacts... (Interestingly, IMO, I can
take a moderate quality LD -- e.g. Cars Heartbeat City, with significant
chroma noise, and produce a DVD that ALMOST looks like NTSC never touched
it, and solid chroma... It looks like a slightly limited bandwidth DVD,
with no significant NTSC footprint or large area chroma noise!!! Most
importantly, no visible comet tailing or odd motion stuttering/smear
artifacting.) Likewise, such feats of magic can be done with original
tape formats also -- except with significantly greater bandwith limits
for 8mm, VHS or Beta/SuperBeta.
John