|
I got this off the C-cube thread and would like some comments about this, if possible...I'm sorry it's so long but I don't really know what everything is he is talking about; I left it unedited. But, it seems to me that he is saying that tape storage for broadcasting will be dead in ten years (something I find hard to believe). Also, since he is only comparing what the big boys had to offer, could Ampex be sneeeking up on them with a super quality product at NAB? Just an interesting thought after reading the good posts regarding TVondaWEB's latest infocast. I'm glad to see that it sounds like many things will be announced now and in May...Finally gonna be able to show my faith in this company has been worth it.
Nursing a broken ankle, unable to do much webbing....
Miss you all and the great posts.........
Crutches are not good for lurking...
CMM
New NAB twists in formats war
by Ken Jones
(March 31, 1999) Before the Great Compression Debate came the Great
Format War; D1, D2, D3, and so on to D9, the newest standard known
previously as Digital S-VHS (JVC's Digital-S). Now the debate has
spawned yet another format, and fighting on the tape front is set to
flare up again at NAB. But this time it's a three-cornered contest, and
testing shows the new kid on the block matches or even outperforms
anything the 'broadcast' manufacturers can throw at it. Welcome to a new
chapter in the Great Compression Debate. Twelve months ago the debate
looked like a duel to the death-Sony and its MPEG-based tape format in
one corner, with the M-JPEG based formats led by Panasonic in the other.
Then last autumn at IBC98, the EBU/SMPTE Task Force on digital transfer
standards reported its viewing tests. Both Sony's SX and Panasonic's
DVCPRO25 got the thumbs down for general broadcast operations use and
the Task Force pointed the industry at 50Mbps I frame-only compression.
By accident or design, JVC (the industrial sibling of Panasonic
Broadcast) had just such a tape format (50Mbps, I frame-only
compression) well on its way to formal recognition by the SMPTE. And
that was none other than the latest manifestation of good old VHS,
established in most homes with electricity and running water. The guys
at the Victor Company of Japan couldn't believe their luck when the Task
Force suggested (in effect) that Digital S-VHS (or D9 as we now know it)
was the only digital tape format with a compression system robust enough
to use for material exchange and processing via SDTI.
At the time, Panasonic had DVCPRO50 in prototype and was working on
DVCPRO100 for American HDTV applications, but Sony had placed its faith
in SX and could only offer Digital Betacam as an alternative. The latter
is robust and high quality, but even Sony admits its compression is
unsuitable for transfer via SDTI, which of course was what the Task
Force was looking for.
So at about the time you read this, Sony will be unveiling the prototype
of its 'studio format' MPEG-2 video tape machines. Designed as the coup
de grace for both the Debate and the War, it will (eventually) record
and replay 50Mbps I frame-only MPEG, and replay Digital Betacam into the
bargain. But it's a year behind DVCPRO50, and a couple of years behind
D9 ...
Between the eyes
So let's recap and try to see what all the fuss has been about. For over
a year in these columns we've been debating the pros and cons of digital
compression for station operations. At one end of the spectrum are the
purists, believing signals should remain at ITU-R601 'for ever'-full
bandwidth switching and processing, with storage on D1 tape. (In
practice they're generally happy to compromise with 8-bit processing and
Digital Betacam which uses mild Motion JPEG compression of about 2:1).
At the other end are the MPEG fans who believe it should be used
everywhere because it has to be used for distribution. They know long
GOPs and 8Mbps compression aren't suitable for acquisition or station
operations, but argue for expanded profiles within the MPEG family, pin
ning their faith on smart transcoding to switch between them.
In the middle, as usual, are the operational people who have to make it
work without busting the company. Most have learnt the hard way to be
deeply suspicious of 'one size fits all' solutions and their stations
tend to be messy compromises with islands of different technologies more
or less interfaced to one another. Perhaps surprisingly, these are the
operations which tend to stick around longest.
Most of the 'messy compromise camp' are heavily into Motion JPEG in one
form or another. Avid, Panasonic, Tektronix, Quantel and even some Sony
products, all use the principle of intraframe compression where the
whole picture is processed every twenty fifth of a second. Sure, it's
not as storage efficient as temporal compression systems which process
one picture and then send change data, but the picture order can be
rearranged repeatedly without loss, and everything happens to the
familiar beat of the station clock.
On the acquisition side, the compression debate crystallised on
Panasonic's DVCPRO25 (M-JPEG) versus Sony's SX (MPEG), at least until
the EBU/SMPTE Task Force Report hit the industry squarely between the
eyes last autumn. Shorn of the usual diplomatic dithering, the Report
said neither system was good enough for general broadcast use, although
both were acceptable for origination.
Based on rigorous viewing tests, the results showed both the new digital
formats were inferior to analogue Betacam SP at four and seven
generations, typical of news and general broadcast production chains.
While the degradation at fourth generation was present but generally
below normal perception thresholds (making the formats acceptable for
news use), at seventh generation it was unmistakable even to the
untutored casual viewer. Incidentally, the tests also confirmed Beta
SP's poor performace at seventh generation, but the big surprise was
DVCPRO25 and SX turning out worse. These results were conclusive enough
for the Task Force to look at other solutions, so they tested 'studio
MPEG' and D9. At the time there wasn't a 'studio MPEG' tape format so
they used servers. Again the results were surprising. Although not as
conclusive as the DVCPRO25/SX tests, they established D9 was
indistinguishable from Digital Betacam at first generation, similar at
fourth generation and superior at seventh in some cases.
Even more of a bombshell was the comparison with 'studio MPEG'. The
tests showed that even if there was a tape capable of recording
MPEG422P-50Mbps, D9 would outperform it.
On the way out
So where are we now and what should you be looking out for at NAB? Video
tape is on the way out as a station processing medium, but will be with
us for a decade for acquisition. SX and DVCPRO could be considered
transient formats, likely to be replaced by D9 or possibly disk cameras
when station operations become fully server-based. And most crucially,
MPEG will remain a 'store, forward and distribute' mechanism; the 'messy
compromisers' will opt for 50Mbps I frame only for station operations.
Whether that's M-JPEG or MPEG won't matter-the differences are slight
enough for smart transcoding to treat them as one.
Author's note: This will be my last Opinion column for a while. As
regular readers will know I've been advocating digital terrestrial
broadcasting for a while and now I've joined the pioneers, ONdigital, as
Head of Production. I hope to bring you news from the cutting edge of
television's future later in the year.
|
