Help, I need to make a tough decision: 1080p finish or 4K finish?

[Tom Lowe]

I am currently shooting a Baraka-style film and I am several months into the shoot. The idea of the film is to mix timelapse shots on Canon cameras at 5K+ RAW with overcranked Red RAW footage. About 50/50.

My initial plan was to finish at 4K RGB 16:9 and have one of the first "Baraka" or "Planet Earth" style films ever shot and finished at 4K. But the severe delays in Epic have caused me to have no choice to but start shooting on Red One now. Of course, in order to overcrank, I am having to shoot at 3K RAW, which is not nearly enough resolution for a pristine 4K RGB finish. All of my timelapse will work at 4K RGB, but the achilles heel of my production is this 3K Red footage.

Up until this point, I really have not shot very many epic shots on the Red yet, but rather have been concentrating on the timelapse. In other words, even if I lost all my Red footage right now, it would not do significant damage to the film.

Question: Should I stop shooting 3K 50fps and start shooting 4K 16:9 right now, with the idea of simply using AE Timewarp, Twixtor, etc, in post to interpolate the 4K footage to 48fps, for example? Another benefit of this would be twice the amount of light available when I am shooting at dusk.

The question is, can Timewarp, Twixtor, etc, save my ass and allow me to finish "Southwest Light" at 4K RGB? Has anyone ever shot 50fps 3K 50fps and then compared it to 4K 24p interpolated to 48fps? How do they compare?

I need to make this decision soon. It's an actual setting on the camera that I need to make a hard decision about, because I am getting ready to really ramp up the amount of Red footage I am shooting.

FYI, I am already reserved for Epic X Stage 2 on #352, so obviously once Epic drops, this will no longer be an issue and I can shoot at 4.8K 16:9 96fps. :sifone:

 

[michael zaletel]

Question: Should I stop shooting 3K 50fps and start shooting 4K 16:9 right now, with the idea of simply using AE Timewarp, Twixtor, etc, in post to interpolate the 4K footage to 48fps, for example? Another benefit of this would be twice the amount of light available when I am shooting at dusk.

Hi Tom:

You know more about all this than most of us but I may have an idea. I believe the key to getting 48fps 4K out of the 24fps 4K is fast shutter speed. If you can eliminate as much motion blur as possible while shooting perhaps you'll get the best results out of Twixtor and can add the motion blur afterwards. Theoretically, with motion blur minimized, Twixtor could probably turn 4K 24fps into 4K 120fps amazingly well. The problem with this is it would require more light.

Again, not really advice or an answer just food for thought.

-michael zaletel

 

[Noah Kadner]

Seems like there's little point in a 4K finish when all your slow mo would be faked in post. If you really need super slow mo I'd suggest renting a Phantom HD- far faster than a RED can go. And I'm not sure you'd be laying claim to the first of anything in 4K at this stage in the life of RED especially in terms of time/lapse scene movies. That title was likely claimed years ago. I'd go for the best image quality you can get in camera and not so much worry about the 4K marketability. Projection in 99% of the world today is 2K or lower...

Noah

 

[David Battistella]

Tom,

Test.

For the kind of film you are describing, which is all about the quality of the image, I think that your decision is simple. Finsih 4K. You will never regret finishing 4K but you will regret finishing 1080.

OR

Finsih 1080 now and then do another pass later at 4K (if it warrants).

My gut as a film maker tells me to suggest to you that you do it all at the highest possible quality now and do what ever you have to do to aquire at that quality.

I suggest that if you can run some tests to see what 4K results are like, then you will at least be able to make an informed choice.

4K all the way. Go for it. You can always make a 1080 from the 4K but you will have a hard time making 4K out of the 3K, but you should do some blowup tests too.

This is a creative decision, not a technical one. Ask your self this: Is the speed more important than the quaility of the image? That pretty much gives you the answer.

Very best of luck. I really enjoy watching your work.

David

 

[Graeme Nattress]

Tricky one - will 3k 60p look better than 4k 30p + a 2x slowmo. I think on some material you'll be just fine with the software magic. However, if you're capturing complex overlapping movements, I'd go with the real 60p, because it's the overlapping movements that the software tends to get confused by.

Graeme

 

[M Most]

The question is, can Timewarp, Twixtor, etc, save my ass and allow me to finish "Southwest Light" at 4K RGB? Has anyone ever shot 50fps 3K 50fps and then compared it to 4K 24p interpolated to 48fps? How do they compare?

I think you need to do some tests yourself and make up your own mind rather than asking theoretical questions here.

I also think you already know that.

 

[J. Eric Camp]

If finishing 24fps consider shooting 30fps 4k. It's not much but it is 1/4 of the frame rate to get you to 48fps. It would give twixtor less to make up.

 

[David Mullen ASC]

3K RAW to 4K RGB for finishing is fine, it doesn't stand-out dramatically from the 4K stuff. If your time-lapse footage is going to look more impressive in 4K, then finish to that format rather than dumb everything down to 1080P just because it will make the lowest resolution stuff blend better with the best stuff. Remember you can always make a 1080P master from a 4K master.

I agree with Graeme that true high-speed photography is going to look better than slowed-down normal-speed photography.

Remember that the perception of resolution is just as much a factor as true resolution. If your 3K slow-motion stuff is shot in crisp light, or in silhouette, or is rather contrasty in general, it will look sharper than stuff shot in soft light. Also, if it is a close shot with the background in soft focus, it will look sharper in comparison. So all of these tricks can make the 3K stuff blend with 4K footage. Or even, a mild bit of post sharpening may also do the trick.

 

[Alexander Ibrahim]

Well if you have the budget and the gear, finish 4K.

You can then dump your 4K down to 2k or 1080p for delivery.

If you ever have to redeliver 4k... well you'll know its 3k in places... but if its good 3k your viewers will never notice.

Just like most viewers don't notice 720p footage in an otherwise 1080p show.

Finishing 4K keeps the best parts of your footage as good as they get, and gets the most of the footage that isn't quite up to par.

Regards

 

[Bruce Allen]

Tom - nice meeting you at Coachella!

First off, I'd shoot 3K 48fps.

If you want to test the other way - I just bought the full licenses of the latest versions of everything made by re:vision (Twixtor, etc) for another feature film project. So give me a shout if you want to do some tests of say 4K 30fps -> 60fps.

David's notes are spot on as usual.

Also, bear in mind most Red "4K" footage is slightly out of focus and motion-blurry. If your 3K is sharp as a pin, you should be good - might actually contain more detail than average.

I'd also run noise reduction on it with Neat Video, then sharpen it up in post. I also own those licenses, so let me know if you want to test that.

That said: if you did the whole finish at 3K RGB and then did a blow up to 4K right at the end, I don't think anyone would notice. If fact, it'd probably look better because your render times would be lower, so you'd have more chance to re-render stuff to get the color, contrast, etc JUST RIGHT. But don't tell anyone

Bruce Allen
www.boacinema.com

 

[Kris Anderson]

Projection in 99% of the world today is 2K or lower...

I think that comment is the anti-everything that RED stands for!

:nono: tut, tut Mr. Kadner! :nono:

 

[Roberto Lequeux]

4k!

Red One's strong suit is not as an overckrank camera, but as it is a 4k camera. Even more so with M-X 4.5k and the new magic.

Epic may come in time for you to add a few 120fps 5k shots to your project, and if it doesn't, you are still an artist that will still create something beautiful with whatever tools you have available.

Then you can easily make a 2k out of that, and if you feel really adventurous you could replace a few shots with 3k shots.

 

[Sanjin Jukic]

Tom,

I would go with 2K as a compromise way.

Why?

Your potential revenue is probably a limited theater release and maybe later a wider BRD/DVD distribution.

Do the following calculation:

Just because you are able to acquire on Canon DSLRs and RED MX at the moment...

Time-lapse you can acquire at the highest as you can with Canon DSLRs @ 5K.

Resize it then to 2K.

With RED @ 2K you could shoot even max 120 fps @ 2K.

Somebody would say it will look soft but all it depends what you shoot, how you shoot and also on the MTF of your lens.

As Noah Kadner already said that 99% theater projection is actually @ 2K ready today and also Blu-Ray Disk is basically 1080p

so than you are just fine for your own postproduction and toady to sell your first project much easier, get some money back to fund maybe the next

project that could be acquired @ 5K and then finished in 4K...

Make your life easier today with thoughts that you are also ready to jump in the future 4K wagon right away....

Don't forget that you do not sell 4K or any other kind of technology just because can sell your idea of art content only...

 

[Jay Gannon]

Yeah I'd go for a 2k finish aswell.

 

[Uli Plank]

You can always add some sophisticated sharpening in post to 2K out of the R1, not the video-style some HDTC cameras do. But I'd always advise real slo-mo instaed of fake.

We did extensive tests with one of the best (Twixtor, IMHO). But as soon as there are complex backgrounds, they'll tend to look like jelly. You can get away with non-overlapping movements and out-of-focus backgrounds only, or it will show on the big screen.

 

[Jeff Kilgroe]

I think for the finish, I would go for 4K. Even though you may be up-rezing a lot of the RED stuff for now, I don't see a reason to throw away any information that may be there and definitely any detail you have gathered with the SLR timelapses. So, only fall back to 2K finish if you have to for other reasons -- budget, etc..

Shoot in the highest resolution and frame rate you can. I would try to not shoot 2K unless you just absolutely must do so. Fake slow-mo can work sometimes and sometimes it just can't do the job you want. Each shot will differ.

I'm hoping you're on the very short list of Tattoos!

 

[Jaime Vallés]

Finish in 4K. You can always make another 1080 or 2K master from it by downrezzing.

As for slo-mo, is there any chance you could shoot the scenes twice? Once in 4K 30p and once in 3K 50p?

 

[David Mullen ASC]

I say 4K, make 2K and 1080P from that as needed... but that's assuming you can afford it. If not 4K, then 2K instead of 1080P if the end result is for theatrical release. However, it's not much of a difference, 2K vs. 1080P.

 

[David Battistella]

Tom,

Let us know when you have decided to finish 4k.

David

 

[Anson Fogel]

I have performed similar tests

I have performed similar tests

Tom,

We have dealt with almost exactly the same dilemma, slow down 3K in post vs. shooting 2K in our case. For a whitewater short film we're working on between paying gigs, we are shooting almost everything overcranked. 2K is, as we all know, soft, but often 50FPS (3K) is not slow enough (fast enough?). So we have done tests, shooting similar action sequences at 3K 50FPS and 2K 100FPS and using twixtor and other methods to slow the 3K down.

The results are, unfortunately, not definitive, but do point towards using post slowdown only in certain cases. If we know the we desire not more than 60FPS final, we will shoot 30FPS 3K and slow in post, but only if its not a contrasty scene with lots of layers. We do find that Twixtor is generally a better tool than other methods, specifically Pro as an FCP plugin versus other options in AE or Motion, etc. - but that's not quite what you need to know.

With Twixtor slowdown versus in camera, some notes:
a) 50% slowdown (double framerate) is the limit using Twixtor in our experience before it begins to look false, strange, wrong, etc - but 50% slowdown can be quite good
b) The "look" of the twixtored footage can be a bit dreamy feeling, with some artifacts actually in our case adding an otherwordly, strange feel that is not undesirable in some cases. In camera only is often more objective feeling to me. Twixtored 50% footage will never look as clean and objective as that shot in cam.
c) Often in editing we find we want it even SLOWER - so shooting at the highest possible framerate in camera, in our case, became critical - as we often took 100FPS 2K and used twixtor to create 200FPS. I suspect with your style, you too will find that at times, you want it slower, but not often faster. This suggests you shoot 3K.
d) When targeting post slowdown, we shoot faster shutter than we would otherwise.

We also find that 3K 50FPS footage cuts well with 4K when you do some post sharpening on the rendered twixtor output. 2K is a different story of course.

On a related note - I hear you are shooting on some rivers - we took RED down the Grand Canyon last year on river, and shoot a LOT of water and rivers, happy to relate some hard learned lessons.