[Beowulf] HDTV video file sizes

Robert G. Brown rgb at phy.duke.edu
Tue May 29 13:57:35 PDT 2007


On Tue, 29 May 2007, laytonjb at charter.net wrote:

> Uncle, Uncle!!!
>
> Actually that was  a good answer. I see that I need to learn more :)
>
> So about 8-9 GB/hour....
>
> What I have in mind is a large number of hours of HDTV being recorded to
> storage.  I'm guessing that total number of hours, but I think the general
> number is over 4,000 hours (about 36,000 GB or 3.6 TB). Actually it's not

You're as good at arithmetic as I am, Jeff...

That would be 36 TB.  Not QUITE so easy or cheap to get.

BTW, I pulled my 3 GB answer from sites that indicated how video will be
delivered over the web, where bw is a scarce resource.  So one can
assume that it is highly compressed and similar in scheme to standard
movie-length DVD recording (HD or not).  You got more/better answers, of
course, but you could PROBABLY drop that to 12 TB or so if you use a
good recording/compression scheme.  Depending, of course, on what the
data looks like and what your image tolerances are.

    rgb

BTW, if I sent you my google string, you'd have to send me the recipe
for your ribs, right?  Although all the spoilsports on the list have
fixed it up so that there's no point anymore...


> that much data is it? Just a few hard drives and you've got it.
>
> Thanks!
>
> Jeff
>
>> At 08:49 AM 5/29/2007, laytonjb at charter.net wrote:
>>> Good morning,
>>>
>>> I was doing some thinking over the weekend (while cooking ribs on
>>> the grill :)  ).
>>> Does anyone know who much data 1 hr. of HDTV produces? Let's try 720 for
>>> now and perhaps 1080. I'm looking for the file size if you store the
>>> whole thing
>>> in a single file.
>>
>>
>> Are you asking about "as generated in the studio" or "as recorded" or
>> "as broadcast"
>>
>> the raw data rate is >1 Gbps (142.18 Mb/s for NSTC sampled at 14.318
>> Ms/s up to 1.486 Gbps for SMTPE 292M sampled at 74.25 Ms/s)
>>
>>
>>
>> There's several compression/redundancy removal steps in the chain,
>> and different HD broadcast media (over the air in US (ATSC), over the
>> air in Europe (DVB-T), direct broadcast satellite (DVB-S, and others)
>> , cable) use different bit rates, and different compression
>> schemes.  And, of course, the DVD (including the new BluRay and
>> HD-DVD) have their own encodings as well.
>>
>> In the US, HD is broadcast over the air in a 6MHz wide channel at
>> between 19-20 Mbps (3 bits/symbol).  However, that 20 Mbps stream can
>> be divvied up in lots of ways: 1 really HD channel, 5 SD channels, 2
>> SD channels plus a medium rate HD channel.
>>
>> Wikipedia has a lot of info on this..
>>
>> The appearance of the decoded output depends a LOT on how good the
>> encoding was.  You can cheap out and just do simple frame encoding,
>> with no frame-to-frame encoding, in which case you get high
>> resolution with lots of artifacts. Or, you can spend a lot more
>> effort on the encoding, and make use of the frame to frame
>> redundancy, and get a lot less artifacts.  The telling difference is
>> if you have something like a panning shot over a complex, but fixed,
>> background (e.g. a forest in the distance).  A good encoder will be
>> able to make use of the fact that big swaths of the image are
>> actually the same from frame to frame, just displaced.  A cheap
>> encoder will not.
>>
>> Cable TV and direct broadcast satellite use somewhat different data
>> rates (since they have different heritage), and different encodings, sometimes.
>>
>> Compressed digital video that is intended for further editing is also
>> compressed differently, because the "broadcast" compressions tend to
>> have unsuitable artifacts in the editing process. Squeezing a raw
>> data rate of >1 Gbps down into 20 Mbps or so always entails some
>> compromises, and the broadcast compressions are designed to allow
>> inexpensive decoders (and expensive encoders..you'll be making
>> millions of decoders and dozens of encoders) and for artifacts that
>> are visually unobjectionable to an end user.
>>
>> As you can imagine, there is much opportuntity for transcoding artifacts.
>>
>> These days, H.264/AVC is probably the leading candidate for compression
>>
>>
>> So.. for over the air HD broadcasts, 20 Mbps should do you, which is
>> well within the range of a variety of hard disks.   Converting to
>> GB/hr, I get 8-9 GB/hr
>>
>>
>> James Lux, P.E.
>> Spacecraft Radio Frequency Subsystems Group
>> Flight Communications Systems Section
>> Jet Propulsion Laboratory, Mail Stop 161-213
>> 4800 Oak Grove Drive
>> Pasadena CA 91109
>> tel: (818)354-2075
>> fax: (818)393-6875
>>
>>
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
>

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu





More information about the Beowulf mailing list