Hack for Feature Lenght Documentary Film for the BIG SCREEN - "Digital Cinema Package"

MisterJ

Hi!

It's been around a Year I signed in here for the very first time. So now, after many test shoots, I'm planning to choose GH2 for the next film project.

Film is about 85 minute long documentary. Films shooting style is mainly situation-based documentary (Cinéma Vérité aka Direct Cinema) with some "cinematic scenes". So some of the scenes are classic documentary and some more stylistic - fiction like - with lightning crew, dolly, directed action etc.

These situation-based events I'm going to shoot with GH2 attached to the Steadicam Pilot. And yes, no interviews. This means I don't have very long takes.

I really like Lee's Flow Motion v2 and Nick's Sedna (AQ1-A) or Cluster v6&7. These I've tested quite a lot. What I'm looking for is best combination in detail and in fluid motion (motion rendering - no strobo effect). And of course hacks reliability is a virtue too! =D

So does anyone has any extra pointers about taking a project to Big Screen? Or if You think there is superior Hack for this kind of project - please don't hesitate to speak out.

Thanx and all the Best, MisterJ

Aashay

Which one do you prefer amongst the four patches you mentioned? Cause you have tried most of the best ones.. Also try Golgop3

sherwood

You're juggling alot of unrelated film terms around, presumably to demonstrate you know what you're talking about. If you're actually referring to DCP (Digital Cinema Package) encoding and delivery, camera hacks and such have little to do with it. Spend some time actually researching the subject, then come back with some pertinent questions.

MisterJ

Thanks for the tip @Aashay. Never tested Golgop3. At this moment I prefer Sedna in detail and Flow Motion if I'm using camera in shoulder rig or in steadicam(it's so good in fast situations). But I think there must be many other Hacks too.

And miss/mr/mrs @sherwood, I don't now about you, but I think most of the terms I use here have a lot to do with filmmaking. I don't know what is your focus on filmmaking - but myself I'm more a maker than technical specialist. I've done commercials, music videos and films for TV - but this is the first one in Cinemas. I'm unfamiliar to DCP AND I want to get out the very best from GH2 for this project. Thats why I'm asking here some advice, before we start shooting this film in early 2013. If I sound arrogant - my apologies. My English isn't that good.

My point was to tell that there is a lot of scenes following a person with steadicam Pilot. In those situations, it's important to have a Patch like FM 2 - specially made for fast situations with moving camera around. "Cinematic scenes" in other hand are more done with steady and slow camera movements. In those scenes detail is the key word for me.

I'm going to take some testshot in January to post, DaVinci Resolve, and we will check those shots with colorist in bigger screen. So I really appreciate comments like Aashay gave.

Aashay

Cheers mate.. I really appreciate your view about Sedna and FM. cause I kinda feel the same way too. But since Sedna eats through the card, I also use Cluster 6. Try that as well if you haven't. FM is definately right up there as well. How many hrs of video can you get on a 64gb card with Sedna? I mainly make short films and since I wanna ultimately have it projected, I was wondering the same thing. I'm torn between FM and Cluster 6. Any idea which handles grading better? Keep us posted with your grading results.

hay

what is the desired delivery output of your files? what are the technical details of this "big screen" you refer to? is this digital projection, transfer to 35mm or something else? i think these factors should inform your choices re: equipment, patch, workflow, etc.

im working on a project to be digitally projected in a theater setting using a christie 2k setup. recently, as a result of evaluating workflow options, ive come to believe lensing (my friend's zeiss set looks much better than my nikon ais) is far more important than the patch, but this also seems to be quite a subjective consideration. for myself, in comparing graded results (400% crops) from stock firmware, fm2, v9b, intravenus and sedna - in aftereffects w/custom color space - the differences i see are slim - the files all hold up well to post work in a 32 bit environ (assuming i used 5dtorgb to convert mts to prores 4444, rather than editing just the mts files). at 400% in AE, 4444 looks ever so slightly better than 422hq to me, particularly in shadow areas, with less artifacts in out of focus parts of the frame...but improvements like this probably vary as output needs differ. i'm exporting slightly uprezzed 4444 files to be then further optimized according to the native resolution and preferred file format of a specific projector, so these decisions seem important. in terms of patches, my favorite of the bunch is fm2, mainly for its apparent smooth gradations. intravenus is growing on me, for its jaggier edge effects (reminds me of the results id get from rodinal). whatever the case, i had to pixel peep for hours under relatively controlled settings to determine what i wanted for this particular project. and as i wont have access to the projector until a week prior to the screening, im exporting various formats for further evaluation. im relatively new at this myself, so please post your experience and findings...

Aashay

Quality of the lens being a factor makes sense..however I'm confused about the transcoding process. Isn't it true that the whole reason you transcode is to make it easier for your CPU? If transcoding works better than mts files in terms of quality, why would Adobe spent so much money to enable native editing? Don't get me wrong sir, I'm merely wondering and not questioning what you said.

hay

i imagine adobe is thinking more about mass customer appeal than ideal workflow practices for the obscure demographic im part of. i made the choice to convert to prores 4444 solely because i saw a slight visible improvement, post grade, both in AE 32bit environ at 400% and the test 4444 output file. also, we may end up going with jpeg2000 compression - its too early to say so whether these very subtle improvements will be visible in the actual screening in may. i'm new to this so i reserve the right to contradict myself at a later date:-)

MisterJ

Hi!

And thanx @hay and @Aashay! All this is valuable info for me. Sorry for such a late reply, but I've been really busy last few weeks. Should been editing one commercial as I write this... :)

By "Big screen" I mean DCP deliver for Cinemas. Workflow will be digital from start to end. But what set up, I'm not sure yet. Common compression seems to be jpg2000, but I have to ask this from colorist, he is going to do final mastering. So in January we are about to do some testshooting, grading and output too - I will give a feedback at that moment how things went in grading.

I'm going to convert all material to proress 4444 before editing starts. And about Lens - Panasonic Lumix G X Vario 12-35mm F2.8 ASPH. I haven't use it yet, but I'm hoping it will work out great. I'm choosing zoom lens to shoot fast and easy, when situation is on. And specially when I'm using steadicam, there is no time to balance camera each time I change a lens.

Great to hear that You have liked FM2. It takes little less space in card and seems to be a best choice for a fast action and camera movements in Real Life Situations. Any thoughts about FM2 in low light? I have liked how noise looks in Sedna. And Aashay - Sedna in 24H mode gives You aprox 55 min. recording time.

Cheers, MisterJ

matt_gh2

What are you using to make the actual DCP files? I had heard about opendcp.org as a free option, but hadn't yet looked into it. What is everyone using for DCP files creation? Curious to hear different experiences preparing and screening of DCPs.

fredfred27

I have been dealing with open source DCP solutions for about two years now. The first open source tools that I used were Open Cinema Tools. They were great, but the creation of the Jpeg2000 files was very slow. With the introduction of another project, OpenDCP, multithreading was added along with a better GUI. I consult for a theater chain that does their own in house content creation, they bought Cineasset from Doremi. It is very fast, but has it's drawbacks. There is a great open source project that I discovered recently, Fnord's J2K plugin for Premiere, After Effects and Photoshop. It has been around for awhile, but just recently added support for the Digital Cinema specifications for Jpeg2000.

So, here is my new open source workflow:

Edit in Premiere

Color correction, effects, etc. in After Effects (Workspace color "sRGB IEC619...")

Export as Jpeg2000 files via Fnords J2K plugin (Output Color "DCDM X'Y'Z' (Gamma 2.6) 5900K by (Adobe)", J2K Digital Cinema 2K setting at 250 Mb/second)

Create audio/video MXFs in OpenDCP

Create DCP in OpenDCP

One of the key parts of this chain in getting content from editing and onto the screen is the conversion from RGB color space to the XYZ color space for DCPs. Cineasset has this option built in, but after viewing many DCPs created by that program, we noticed some problems with the final color levels. It seemed that the conversion was not working correctly, and tests bore this out. This lead me to testing Adobe's "DCDM X'Y'Z' (Gamma 2.6) 5900K by (Adobe)" color conversion. This looked much better. Now I had to create a sequence of 16bit TIFFs in XYZ space to use in Cineasset for DCP creation. This is is when I stumbled on the updated Fnord J2K with support for Digital Cinema. Have not looked back since.

If After Effects is not in your workflow, OpenDCP can do the conversion from RGB to XYZ Jpeg2000 files, but the color conversion is not on a par with what you get by doing it in After Effects.

If you are interested in all of the parts that go into making a DCP, here is a link to the first open source DCP project:

http://code.google.com/p/opencinematools/

Here is a link to Cineasset:

http://www.doremilabs.com/products/cinema-products/cineasset/

More importantly, here are the parts that I use now:

http://www.fnordware.com/j2k/

http://code.google.com/p/opendcp/

matt_gh2

@fredfred27 This is great info. I use Premiere and After Effects, so this workflow looks great. Will definitely try this. Thank you.

willieturnip

@MisterJ I'm not sure what your experience with steadicams/glidecams/etc is but I have found that even zooming a lens requires rebalancing most of the time. To this end it makes no difference using primes or zooms..

vicharris

If you fart on the camera you need to re-balance the damn steadycam!!! Learning to use a Glidecam HD2000 right now.

Raysito22

@vicharris , have you read steadicam's operators manual? If you fart, you gotta learn how to burp all at the same time to rebalance it. There's other way but I find it even less appropiate and it's only for man and way too grose.

@MisterJ As willieturnip said, I really don't get the reason to use that zoom lens (which is a good lens, but not a cinematic one, my personal-view) as you WILL need to rebalance. So, you are not gaining time, but you might be compromising the look a tad bit compared to using few primes. I've read it's even sharper than Oly's 12mm, which I have and don't really like (will be soon selling it to buy a SLR Magic 12mm).

Even so, if you're sticking to the 12-35, you probably already know (as this post has been rescued back from december) it's a really sharp lens and also has some jitter if you zoom in during recording (no zoom-in or zoom-out shots is a real bummer, sometimes zooming in is cool, ask Tarantino about it hehe).

Also, being a sharp lens as it is, you should try some of those soft looking patches by Driftwood.

MisterJ

Hi!

And thanks for the info @willieturnip , @vicharris and @Raysito22 !

You guys are right. First of all, I spent one week learning how to use steadicam - and I was hopeless! In these learning hours I just can't master it - not even close. So I just cave up and returned back to square one. =D

I also changed my mind about using zoom lens. I tested nokton 17,5mm one weekend and fell in love with it. That's the one I'm going to use in shoots. And later 25mm too.

I tested 12-35 and it has quite nice and smooth zooming, but usually, I dislike shots with zooming.

I've also tested @Driftwood latest Cluster X Moon. I've already shot one day with trial 4 and love it. Next thing is to take couple of shots to Colorist and see how much better it handels grading than FM2.02 does.

Chers! -misterj-

Raysito22

I believe those are wise options you took in order to get more of a film-look. And yeah...operating a steadicam seems easy when you see proffesionals doing it but it's a nightmare to learn how to do it propperly, it takes a lot of time. I'm starting too with my Laing steadicam + vest. Patience and practice are the only tools to learn how to use it, that's what I've learned for the moment.

Keep updating this as you advance with the project, would be nice to know how it works out. Best of luck!

MisterJ

Thanks @Raysito22, i will keep u posted.

After we have released our Demo Trailer, I can post some material. We did one test with FM2.02. He haven't worked with hacked GH2 material before and liked it. Let's see next phase, when we do some under/over exposure tests. With moon.

Shoulder rig is my choice at the moment. About stedicams. It's not that difficult to get decent shot, when you are walking at the same pace, but doing a shot, where you start from fixed composition and then you move from point a to b and then ending into fixed composition. In a small apartment. My God! :D After this testweek my sincere respect goes to stedicam operators.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters