Super geeky digital cinema talk: Michael Cioni prepping for a 4 K World. Man, the tech at the top end (Hollywood) is amazing. If I'm understanding right, Girl with the Dragon Tattoo was shot in 5K and the final product (not dailies, etc..., just the final film itself) was 51TB! And holy cow, I'd never heard of the Clipster before - what a monster.

- jim 9-01-2012 6:08 pm

Back of the envelope calculations

5120x2700 = 13.8 Mpixels per frame

assuming 12-bit RGB, 36 bits per pixel, round up to 5 bytes per pixel

yields approx 70 MB per frame

24 frames per second yields approx 1.7 GB/second

3600 seconds per hour yields approx 6 TB per hour

two hour film: 12 TB

Hmmm, I'm off by a factor of 4 somewhere. But yes, uncompressed video is huge. Distribution to theaters is in compressed format.
- mark 9-01-2012 8:14 pm [add a comment]

Did you watch? Pretty interesting. They shot in Sweden and sent H.264 over the net to Hollywood every day. They would start editing that immediately, and then when full res arrived in the mail a few days later they could just swap it in and preserve all their cuts.

He also threw out a one liner that 4K was going to be the death of broadcast television. He didn't elaborate though.
- jim 9-01-2012 8:22 pm [add a comment]

12 TB = 480 blue ray disks, or about 10 million 3.5" floppies.
- mark 9-01-2012 8:36 pm [add a comment]

Most broadcast HD is really crappy HD, and people seem satisfied. As with audio, it appears that choice >> quality.
- mark 9-01-2012 8:39 pm [add a comment]

i watch soccer on a couple of different channels. one broadcasts at 720 and the other is 1080. at first i was thrilled with the 1080 except now i realize my 120mhz tv cant handle the speed of the action and i get digital artifacts. *apropos of very little*
- dave 9-01-2012 8:45 pm [add a comment]

For sports 60 frames per second is better than 60 fields per second, thus 720p60 is better than 1080i. (My opinion. Also the opinion of ESPN.) Given constrained bandwidth, 720p can be lower in artifacts than 1080i. Interlaced video is just harder to compress.

1080i is a good choice for HBO. Most of their content is movies at 24 frames per second. Mapping this to 60 fields or 60 frames is roughly the same. The 24 fps screws up motion a lot. The choice of i or p isn't going to make much difference.

De-interlacing well is hard. Many TVs have good de-interlacers. Some don't, which impacts the quality of interlaced video.

Another reason I'm skeptical about 4k in the home is that 1080p60 hasn't caught on. It's not very hard in 2012. Just lots of little problems that could be addressed easily if there was enough motivation.

- mark 9-02-2012 6:44 am [add a comment]

Just watched the whole thing. Awesome talk.

5k for 4k makes a lot of sense. It's one of those things video guys wouldn't necessarily understand about film workflows. But apparently Red has some smarts.

Using non-linear editors on compressed video, and using the edit decision list on full res is an idea that's been around for a while.

My TB estimates were off because of visual effects. Perhaps they carry around additional data with effects, rather than just rendered results, making it easier to make changes.

His comment about 4k killing broadcast TV is interesting. I assume he's talking about bypassing traditional distribution chains, which are technologically rigid. Given enough power in the client computer, that computer doesn't give a crap about resolution, aspect ratio, compression format, etc. It can just deal. Digital cable TV, however, is still primarily MPEG-2. Yes, they use that compression standard from twenty years ago.

So on the one hand, the switch to 4k can be an element which accelerates the ongoing revolution in television delivery.

Riffing on his "transition vs. transformation" discussion, most digital TV in the US is really "digitized TV". Everything stayed the same in terms of the value chain and the topology of the technology. What happened was an almost 1 for 1 transition of devices from analog to digital. Analog tapes went to digital tapes (then discs). Analog modulation went to digital modulation (plus compression). Analog set tops went to digital set tops. Analog video connections to analog TV sets went to HDMI and digital flat screens.

But there are new delivery models which bypass the legacy value chains and infrastructure. E.g., netflix could get to 4k way faster than a traditional cable service could. (However, the cable companies are putting together their own "alternative delivery" services. So they won't be caught entirely flat footed.)

On the other hand, will 4k really make a huge impact? The TV manufacturers are drooling over this. But many consumers would probably be wowed by HD that didn't suck. Will they really demand true 4k in the house? How many people are happy with upsampled DVDs and won't spend the extra bucks for Blu-ray? Broadband internet in the US is another issue. Most people buy broadband from a company that wants to sell them video services (which are more lucrative). Will the data divisions of the cable co's and telcos facilitate the death of their video divisions?

As you can tell, I have more questions than answers on 4k in the house.
- mark 9-02-2012 7:52 am [add a comment]

One of the tools he mentioned: episode
- mark 9-02-2012 7:55 am [add a comment]

add a comment to this page:

Your post will be captioned "posted by anonymous,"
or you may enter a guest username below:

Line breaks work. HTML tags will be stripped.