New BMPCC owner, initial thoughts

[brianl]

I've been using the GH2 the last few years. Before that, EX3, HD100, a bit of fS700, 

Footage straight out of a GH2 is glorious, BMPCC footage is awful. And wow do I love it! As far as I'm concerned, the bmpcc gives you a film negatve.  Even ProRes is unbreakable in my NLE (sony vegaspro). And the BMPCC destroys the GH2 in terms of latitude. I still love my GH2s and keep  using them, especially because of the efficient file sizes, but for special footage, the BMPCC is badass. 

 

The battery life is a joke, but so what? for $50 I bought 5 bm batteries. I'm using the 32 gig cards which hold about 20 minutes. So fill a card, change your battery. This is a film workflow IMO and I'm stoked. 

 

Unlike, EX3's and other such cams, the BMPCC's menu is almost nonexistent. Love it, Going into an EX3 menu was like going past the event horizon of a black hole. Jeez, did I really need all those setting? I still don't know what "Knee" is. Again, the BMPCC emulates film. Ingest your negative into your NLE, calibrate your monitor, and you're now in control instead of stressing through endless camera menus on most pro cams. 

 

The LCD. Okay, it's not great. In fact it kind of sucks. There are a lot of options though from Cineroid to Zacuto loops, I'm sure I'll work it out, and I'll be bugging you guys here for idea and tips. 

 

My biggest concern right now is wide angle lensing. Hope to get up to speed on that. These speedboosters seem promising.

 

#1 gripe is mounting. The cam's body is 1/2" wide. I haven't been able to get a mounting that's satisfactory yet. I like using kwik disconnects. Probably going to need to upgrade.  I'm starting to think a cage is the way to go, too bad because I think cages are stupid and turn camera's into erector sets. But with the thin profile, I may have a cage in my future, the Wooden Cam half cage looks the least obtrusive so far. 

 

Davinci Resolve has been a big headache, lots are having trouble with video card compatibility. All I know is it doesn't work and lots of others have the same complaint. 

 

The camera is great through, it's probably not for everyone, but for people like me that love the 16mm film workflow, it's a dream. 

[Craig Marshall]

Forget Resolve unless you have a CUDA card installed. Use instead Cliptoolz COLOR. (major V2.0 upgrade out soon) and Cliptoolz RAW was especially desgned for BMPCC.

[brianl]

Forget Resolve unless you have a CUDA card installed. Use instead Cliptoolz COLOR. (major V2.0 upgrade out soon) and Cliptoolz RAW was especially desgned for BMPCC.

Thanks for the tip. I'm also going to try filmconvert. http://www.filmconvert.com/

[Axel]

Thanks for the tip. I'm also going to try filmconvert. http://www.filmconvert.com/

 

I know filmconvert is considered a great tool, but it doesn't convert raw. It also doesn't really grade, it just applies some retro look combinations (the science of whose is of no importance to any *film* produced within the last decade), and imho it takes you on a meander away from actually developing grading skills. Like it or not, Resolve is a hundred times as powerful, as are the tools in After Effects alone, only that they are not specifically dedicated to grading. 

[marcuswolschon]

I guess you ment to say awesome instead of aweful?

(That's the complete opposite.)

[Dr. John R. Brinkley]

I've treated my Pocket Camera like an old Nizo Super 8 camera. I got a pistol grip from PNC and this works well for most of my situations, otherwise I'll use a tripod.

I agree that it's like shooting with film and that mentality helps keep the shooting ratios manageable.

I'm sure there will be another next great camera always around the corner, but for me this is the camera I've been waiting a decade for.

Post production with color grading and it's never-ending possibilities is a slippery slope but it's good to learn new skills.

[brianl]

I guess you ment to say awesome instead of aweful?

(That's the complete opposite.)

 No I meant "Awesome" because of the film like workflow. I don't need it to be pretty out of cam. Pocketcam gives you a flat ugly negative but I like that. 

[brianl]

I've treated my Pocket Camera like an old Nizo Super 8 camera. I got a pistol grip from PNC and this works well for most of my situations, otherwise I'll use a tripod.

I agree that it's like shooting with film and that mentality helps keep the shooting ratios manageable.

I'm sure there will be another next great camera always around the corner, but for me this is the camera I've been waiting a decade for.

Post production with color grading and it's never-ending possibilities is a slippery slope but it's good to learn new skills.

I'm with you Doc, I've had 16mm film envy all my life. This is the first digital cam that wets my whistle. 

[Sean Cunningham]

I know filmconvert is considered a great tool, ... it just applies some retro look combinations (the science of whose is of no importance to any *film* produced within the last decade)...

 

This isn't correct.  You're thinking of something like Magic Bullet Looks.  Film Convert transforms the measured response of a digital camera to the measured response of several actual film stocks.  It's essentially doing a very specific type of transform that you could do in an ACES color pipeline if you had an IDT for a given film stock.  The science being used is extremely contemporary and relevant.  It's the kind of thing that's going to eventually make all this sloppy, subjective "LUT jockey" business that's so endemic to raw photography an unfortunate little footnote of the past.

 

You could do a conversion and then consider this your "look", the same as filmmakers have for a hundred years or more shooting film and having it processed.  In the digital world the result of doing Film Convert is your inter-positive.  

 

If a filmmaker then has more money they could time their film to have a different look, instead of having just a "one light" processing.  That is the step that comes after Film Convert.  FC gives you your IP, now you can give it an artistic grade or not.  

 

They have some very basic controls for doing this but of course a dedicated grading tool will be better.  The ideal solution is to do your conversion and then grade.  Whether that is done is up to the sophistication of the end user, both their technical ability to grade and their ability to make the distinction between the two operations.

 

How it fits into a RAW workflow is another matter.  Their standalone tool, presently, I don't view as anything useful for more than a hobbyist that just wants to play around with it.  It doesn't even process AVCHD footage with sufficient accuracy or the quality you get by using the plug-in for After Effects and doing your conversion in a 32bit float environment.  For RAW you would definitely want to be using the plug-in and not the standalone, letting the host application handle the log-to-lin conversion (of course successful use of Film Convert is dependent on their "IDT" being based on the same linearized data).

[mtheory]

I know filmconvert is considered a great tool, but it doesn't convert raw. It also doesn't really grade, it just applies some retro look combinations (the science of whose is of no importance to any *film* produced within the last decade), and imho it takes you on a meander away from actually developing grading skills. Like it or not, Resolve is a hundred times as powerful, as are the tools in After Effects alone, only that they are not specifically dedicated to grading. 

 

This! 

[brianl]

I picked up the wooden camera half cage for the bmpcc. I believe it, or something similar is must-have. BH was selling the Wooden Camera for 50 bucks a while back so it represents a nominal investment. However I think it's mandatory, having a two point contact makes a big different, the bmpcc just felt too flimsy mounted cageless. The profile of the lens also could be problematic due to clearance issues. 

 

Kinotehnik loupe seems to be a necessary evil for budget shooters. It's not particularly well engineered, there's slight bowing in the plastic in the edge that mates to the bmpcc which can cover the cam's (too small) data info. You should considering biasing the mount AWAY from the data readout to avoid obstructing critical info like ISO etc. The company's tech support is non existent. Until Zacuto releases their loupe, I'm not sure if there are other, better choices. 

[Axel]

@BurnetRhoades

I will consider this. But what are the chances that the majority uses FC in that way? You are right about the highly subjective and not very succesful approaches to raw (and I don't single me out!), but isn't this what everyone asked for? Buy a $999 camera that brings you (Leonardo) DaVincis studio, complete with unmixed , "raw" pigments and empty canvases and next spend $200 - $300 for a can of fast sfumato? - No WAY! I could better stay with my filmmodes, cinestyles and all the other baked-in formulas I came to know with DSLRs.

For FC is a simulation of a bygone technique.

 

@brianl

I think there can be two ways to use the BMPCC: With it's 'stealth factor', without cage, but just with Dr. John R. Brinkleys pistol grip (I love pistol grips) and some pancake or so. Requires concentration, training and fine motor skills (play Mikado!), but it's your only chance to get shots, say, on the airport or in the subway without an explicit permission.

 

Or: Take the cage and screw everything to it that helps. Look in the mirror: Any chance you will go unnoticed?

[Sean Cunningham]

There is another tool that is being, or will be marketed by a professional colorist.  Shian Storm is also, in a more limited way, doing some empirically derived filmstock transformations in his ColorGHear suite.  It may not always be FC but there is a lot of "film conversion" going on, even in professional circles.  Film and the way it renders tonality and color will haunt digital acquisition for a long time to come.  They've yet to build a digital camera that can hold a candle to the best work being done in celluloid even today.

 

Anyway, yeah, people asked for "raw" but people don't ask for the highly imperfect and incomplete log-to-lin setup that's most commonly being used.  There is too much effort spent simply trying to manage up an honest and true rendition of what light the camera saw and the relative values present in a scene.  Getting to an appropriate linear representation so that you can then grade from a place of awareness and creativity, a place where you can make meaningful decisions that are more than just trying to get it to not look like shit anymore, shouldn't be an ordeal, it should just be.  It should be the first thing you see when you look at your footage.  

 

Seeing a meaningful representation of what you shot doesn't mean baking anything in or losing anything in the raw data. 

 

I got my first taste of that swapping over to ACES on this short I'm doing VFX for that was shot on the Epic.  It's shocking how a manufacturer IDT that's universally regarded as "horrible" is still so much better than their own ability to meaningfully render their own native footage to the monitor.

 

Anyway, the type of techniques that make FC possible are what will make mixing multiple cameras from multiple manufacturers on the same film as close to seamless as you will get until hitting their relative limits.

[Axel]

Reasonable. I see your point.

[Sean Cunningham]

And I'm in no way advocating using FC on everything or conversion to a filmstock look as something that needs to be done on everything or by everyone.  I love what it does but it's just another tool in the toolbox.

[chrisso]

There is too much effort spent simply trying to manage up an honest and true rendition of what light the camera saw and the relative values present in a scene.  Getting to an appropriate linear representation so that you can then grade from a place of awareness and creativity, a place where you can make meaningful decisions that are more than just trying to get it to not look like shit anymore, shouldn't be an ordeal, it should just be.  It should be the first thing you see when you look at your footage.  

 

Most experienced Resolve users can get there in a couple of mouse moves. I admit I'm finding it hard to learn and understand Resolve, but I tried Filmconvert on both my Pro-Res and converted Raw footage and didn't like what it did.

 

The best thing I've found is capture One, but it's a tiresome and slow workflow. I think it just needs a more dedicated attempt at mastering Resolve.

[Sean Cunningham]

They get to a look, or a representation, but they're still arriving at a subjective guess as to what it should be.   There should be no guessing and no series of mouse moves besides loading your clips.   When you first see your footage you should be actually seeing your footage properly transformed for viewing and then grading it from there. 

 

That's why Academy Color Encoding System is such an exciting development for RAW and why, hopefully sooner rather than later, there will be no reason to not be working this way.  Smarter, not harder.

[chrisso]

But everything about grading is opinion based - exposure, colour, sharpness.

If it's just going to come up 'done', then it's all going to be based on someone else's subjective taste. That's the opposite of what RAW is about. It sounds like you want jpeg with pre-made looks like the Sony Nex and GH series.

[Sean Cunningham]

Please re-read what I'm saying.  What the sensor sees, what the relative powers of light were in the scene as shot are not opinions.  Please, quote where I say it would come up "done".  In fact I specifically say it should come up with a correct representation of the scene as it was photographed.  From there you grade.  

 

RAW isn't arriving at chicken salad out of chicken shit. 

[Axel]

 In fact I specifically say it should come up with a correct representation of the scene as it was photographed.  From there you grade.

 

This correct representation is arguable in the first place. And if it was granted, you only drew near by primary CC, that is: Finding balance. You need to do this anyway.

 

From a glance at the FC demo, I'd guess under the hood RGB is split to channels and filtered to get the characteristic *cast* of  magenta/cyan/yellow resp. red/green/purple layers of different grain sizes and in different foci of the film stock in question. However scientifically exact this 'conversion' works, it can't be anything else than a distortion of values. If you put that first in the pipeline, you start with limited options. 

 

I forgot to bookmark one Resolve tutorial from youtube, which compared the MO to start grading with a LUT and 'from scratch'. The latter was proven to be risky for beginners but pushes the boundaries for the experts. 

 

Final argument for me against buying FC: It should be cheaper or free. The GUI (at least of the stand-alone-version I know) looks crappy, horrible, that "100% film grain" is the default position of the lever, for grain surely is only suitable for a very crude 'old film' effect. What is more, all filters could be build within Resolve or After Effects (in theory, of course. No one can measure this at home. But again, I fail to grasp the need to be precise, unless the task was to cut digital inserts in a DI, scanned from a specific stock). So this is one for the compulsive plugin collectors, imho.