The video format for next 5-15 years

[Novim]

I'm a long time reader of this excellent site, now my first - perhaps a bit unusual - question.

Suppose during this winter you should shoot a series of interviews (some 20-30 hours with some 30-40 people) and that digital video material would be archived for use by others in the future (in videos, TV. films etc.). Interviews will be shot in improvised studio (mostly, but not always), the sound taken separately. What format - techically speaking - you'd use? What would be your preferences? 10-12 bit, 4-2-2 or better, what DR, etc? What gear you would buy / use for such a project (the gear budget will be 5,000-7,000€)? - Thanks in advance.

[Justin Bacle]

I think that as long as you export your footage in an open standard, there will always be some application that can open and edit it.
I would go with CinemaDNG video files and wave audio files.

Of course DR is important if you want to keep grading abilities. But it is not the most important.

Is your budget for all the gear (Camera+lenses+audio+support) or just for the camera body ?

[Xavier Plagaro Mussard]

Raw interviews?? Your budget goes on hard-drives!! A GH4 has been good enough for the aerial shots in Narcos. 

[MountneerMan]

I Know everyone likes to argue that resolution doesn't matter and 1080p is fine but IMO I think resolution will date your footage way more than color bit depth so lean towards resolution over bit depth.

It also might be worth shooting 60p. IMO I think we will see a big shift to 60p and away from 24/25/30p but this is just my opinion that has a lot to do with me personally liking the look of 60p

[jcs]

When one examines how the universe created DNA, what DNA represents, and how the brain stores and retrieves information, we can see that future video compression can be orders of magnitude more efficient. Resolution independent fractal & genetic algorithms (iterated/generative) are still be developed for video (nothing commercial viable yet), and perhaps with the assistance of neural networks (machine learning & AI), radically improved video compression will become commercially available in the next 5-10 years. Of course such systems will require more powerful hardware, however that doesn't seem like it will be an issue given how GPUs continue to more or less follow Moore's Law (possibly faster than Moore's Law with new processing designs: http://www.pcworld.com/article/3072256/google-io/googles-tensor-processing-unit-said-to-advance-moores-law-seven-years-into-the-future.html ).

[Tim Sewell]

Quote

however that doesn't seem like it will be an issue given how GPUs continue to more or less follow Moore's Law (possibly faster than Moore's Law with new processing designs

It's easy to forget sometimes that we're still only living through the very dawn of computing and digital technology. Just like a caveman with a flintlock, our great grandchildren will take things for granted that would seem like magic to us now.

[Jimmy]

You've really got two choices....

Shoot the best you can afford that, on paper, will be future proof on the devices we use now... TV, mobile, 2D screens etc

Or take a leap of faith on tech like 3d or 360 video.

I'd go for the first option and aim as high as possible ... maybe see if you can rent red weapon 6k or 8k 

[cantsin]

It doesn't make sense to store your project in a higher bit depth or color sampling rate than in the source material, unless the edit includes a lot of SFX or stills in a higher color depth. 

The pragmatic advice is to use h264 high profile, intraframe, 10bit if necessary, at the bitrates normally used for ProRes, DnXHD/HR , Cineform. The resulting quality will be better than ProRes/DnX/Cineform because of the more advanced codec. h264 also the advantage of being an open standard similar to JPEG, mp3 etc.; it will likely remain playable for a few decades if not longer. 

If you want absolutely highest archival quality, export DPX or TIFF sequences + uncompressed audio files. 

[Dan Wake]

tv will be 4k or 8k? imho 8k is ridicolous for tv.

[fuzzynormal]

8 hours ago, MountneerMan said:

I think we will see a big shift to 60p and away from 24/25/30p

Well, if you shoot in 60p, you can always down-covert it to 30p, so there's a little flexibility there.

I bet the 2D cinematic narrative storytelling is going to last a nice long time.  Multitudes of generations.  It will be supplanted with something incredible, yes, but there's always going to be people watching Casablanca, you know?

[Jimmy]

I'm not so sure.... things change quickly now. Ask a millennial kid if they would rather have access to books or emojis and alot will choose the latter.

If and when VR is perfected.... 2D narrative might become something young people lose interest in. Within 15 years, we could be very far removed from how entertainment currently is served.

[Tim Sewell]

Quote

I'm not so sure.... things change quickly now.

They do - but there are still plenty of theatres (actual theatres, not cinemas) showing a healthy profit. Some art forms, once invented, do such a good job of capturing the human imagination that they will likely last forever.

[Jimmy]

Again... I'm not so sure.... though I work in VR a fair bit, so might be biased in my opinion.

A couple of years ago it was hard to see the end of traditional gaming ... but I would bet my house that in 15 years, very few people will be sat playing 2d games on a screen.

Narrative has more challenges than gaming... but 15 years will see those challenges diminish. At the very least... the way we consume 2d content will change... I can already sit in oculus cinema and watch a film on an imax sized screen (in VR) even when sat on an airplane. 

This has shifted my thinking away from 1080p towards as high a resolution as possible. VR and AR will need 4k, probably 8k

[bunk]

26 minutes ago, Jimmy said:

This has shifted my thinking away from 1080p towards as high a resolution as possible. VR and AR will need 4k, probably 8k

So you end up with 720 or 1080 per point of view. Doesn't sound like you really shifted away, more like you are tired of looking at 640 and less.

Btw isn't this thread about what codec best to use to archive stuff.

[Jimmy]

13 minutes ago, bunk said:

So you end up with 720 or 1080 per point of view. Doesn't sound like you really shifted away, more like you are tired of looking at 640 and less.

Btw isn't this thread about what codec best to use to archive stuff.

?

4K VR would be 2K x 2K per eye.

You would need 2K vertical resolution... eg, shooting 4K video.

Btw - I've already given a view on codec (i'd go REDraw).... Discussion are allowed to develop though.

[AaronChicago]

3 hours ago, Jimmy said:

I'm not so sure.... things change quickly now. Ask a millennial kid if they would rather have access to books or emojis and alot will choose the latter.

 

Off topic a bit but I watched a Ted Talk explaining that studies show Gen Z actually prefers books and physical objects over digital consumption. Gen Z is the generation after millenials.

[Jimmy]

They might prefer it, especially comparing like for like (a book v an ebook) .... just like someone might prefer film to digital.... But for practicality, actual usage rates will keep going the digital route and books will become even more of a novelty.

[bunk]

2 hours ago, Jimmy said:

?

You say 1920x1080 isn’t good enough anymore, but it’s still the standard. Let’s create a VR rig with this standard in mind.
Screen 16:9, 50mm lens turns out you need 12 cameras to complete a 360 circle. That adds up to 22K to get anything descend and that just the equator.
Back to your 8K with 12 cams 50mm lens. You end up with 640 per cam,i.e. 640x380.
With an 21mm lens you only need 6 cams, on 8K you end up with 1280x720 per cam or HD. a quarter of the standard you just trashed. But it’s worse as you also include the poles within the 8K.

[Jimmy]

If you read the post you quoted, I was actually talking about viewing 2D content within VR (in an IMAX style VR cinema).

[Bioskop.Inc]

Whatever you do, DO NOT shoot in 8-bit H264 or 8-bit whatever. In fact don't shoot digital!

Personally, if you really want to future proof anything you shoot for up to 15yrs (or longer), then use film & not digital. The one thing that no one has mentioned is storage & it'll be more economical/reliable to store film than digital & the scanning techniques will just get better throughout the years. Spend a little more now & then you'll have a reliable archive, for yourself or others, to delve into for years to come. Digital will change/morph quickly & in 15yrs time you might not even be able to play the digital footage you film today.

Why do you think some of the top film directors still use film to shoot their films on? I'll lay money on one major reason being that the majority of films, up until this day, have been shot on film & that future generations will want to preserve these archives, no matter what. I'll also lay money on the fact that more digital films will get lost/destroyed for future generations than those shot on celluloid - and you'll end up reading about people remembering seeing this or that film/tv series, which no longer exists.

It's something to seriously think about & that goes for the project that you are thinking of embarking on - how valuable is it going to be for future generations? And do you really need to worry about how long it will last?