RAW vs. other recording formats

my last (and only serious) three features have been raw..
mini LF x2
red epic and kimodo x1

it seems that must have pro res is dissapearing

both DOPs spent a lot of time looking at the monitor and getting the a/cs to go from 4200 to 4800.. which was odd - Id have just got shooting.
 
There is/was an old interview with Rodney Charters, then working as a DP on the" Dallas" reboot, where he basically said, "Raw? We're shooting ProRes on Alexa and that'd it".

I take his point. Just look at outliers like one of the UK's top-rated series, "Death in Paradise", now heading into its twelve series. All shot using the ARRI Alexa Plus shooting in Avid DNxHD at 25p. If shot and posted correctly, not using RAW doesn't mean a show's end result can't look perfectly fine. Red Planet Pictures was nominated for, and won, the "Diversity in a Drama Production Award" for Death in Paradise. The reason they opted to go with DNxHD is they had a well established Avid workflow with an experienced Avid team. An added benefit was the Avid workflow offered an efficient, seamless workflow for the turnaround of a weekly show. It was also a simpler workflow with decent budget benefits. I heard they looked at shooting RAW, but in the end it really didn't look that much better than what they were delivering in 709 masters to the BBC when shooting DNxHD in camera. Production values, lighting, a decent script, a good DP and direction is what holds viewers over many seasons. "Billions" was shot in camera in 709 and looked fine. We, the technical and artistic people in production, can be our own worst enemy sometimes, using unnecessary workflows when the end product doesn't necessarily justify that workflow. It's all about the bottom line for a lot of productions.

Chris Young

https://www.youtube.com/watch?v=WR3p9eWnuOU
 
To get deeper into the raw 'argument' Death In Paradise is a good start point.

Yes you can shoot it dnxhd pro res (you need a lot of fill lights)

or you can look if raw will give you a couple of stops of bend.

couple of stops? not much?

In such a harsh sunny place this might mean going from 8k to 4k lights.. smaller gennies, smaller crew, faster company moves, less lunches, less hotel rooms, shorter midday shutdowns, less days shooting.

Saving of hundreds thousands of £ over a series.. but of course pissing off the crew and rental houses as you cut thier meat.

----

In my experience shooting raw stills one could shoot with lower fill levels and jank in post.. which basicallly led to more natural scenes. Thinner DR means blasting fill which creates a fake look, something that DIP may suffer from?
 
My last feature we shot Red but the (ex film) DP wanted a moody dark look so under exposed the whole film by 3 stops (I was watching the scopes) - none of the younger camera team stepped in as the DP is god.

The film will be riddled with noise and another nail in the Red/raw coffin

The young camera team should have prepared a 'moody lut' and then the DP could expose by eye and fill the data bucket at the same time.

They don't even need a LUT, just turn the preview down to ISO 100. Presto, you're lighting by eye, you can expose three stops under, but your noise is 3 stops lower too.

Drboffa - I remember Brawley and Papert saying they shoot everything prores 4444. I know Deakins went over to ARRIRAW. Not sure why. Maybe because why not? But Papert and Brawley are also well into "Why not?" budgets and support

I think it's hard to estimate percentages of these things because there are so many nooks and crannies of our industry. So many times on dvxuser, I've said something like, "You can't sell a camera without feature X because everybody needs it," or, "Nobody ever uses feature Y" only to have people fire back that they never use feature X or they can't get by in their line of work without feature Y.

I don't think the juice is worth the squeeze for me to get RAW out from my Sony cameras. But I completely agree with Sam that if you can handle the data and have a good workflow, there's no reason to use anything else.

JPEGs in photojournalism are a self-imposed limitation. I don't think they do it for any other reason. (Although if you need to submit images quickly from the field, low-res JPEGs are obviously the way to go. But you could do dual format capture.)
 
Drboffa - I remember Brawley and Papert saying they shoot everything prores 4444. I know Deakins went over to ARRIRAW. Not sure why. Maybe because why not? But Papert and Brawley are also well into "Why not?" budgets and support

Yeah. I think I've heard others on various cinematography or film podcasts talk about this as well. In some cases I have to imagine that the budgetary savings might be worth it even on a high end production.

I don't think the juice is worth the squeeze for me to get RAW out from my Sony cameras. But I completely agree with Sam that if you can handle the data and have a good workflow, there's no reason to use anything else.

Sony "RAW" (from the FX6/a7s3) is such a hassle I've never bothered to use it aside from testing, especially given how good the internal codecs look for most stuff. Even the proxies look good enough in most cases (before being pushed around in post).

If there were better compatibility across NLEs (i.e., if I could use ProRes RAW in Resolve) I might be tempted to shoot with it, but right now the juice is most definitely not worth the squeeze, as you note ahalpert.
 
When you say you "never touched ArriRAW" but you "prefer ArriRAW for sure", do you mean you never shot it on your own camera but you prefer grading it? (Just wondering.)
 
They don't even need a LUT, just turn the preview down to ISO 100. Presto, you're lighting by eye, you can expose three stops under, but your noise is 3 stops lower too.

Now here is a thing... poor information and knowledge.??

DOes changing the ISO on Red change the core file.. afaik back in the day when I wanted a scarlet (and got and fs7) my tests said ISO was fixed and the numbers just changed the monitor and how the file opened in the NLE

In which case using 100 would have worked

BUT the camera team were suggesting in the dark we move the camera to 3200, suggesting that chaging ISO (or certainly certain steps of it) do change the core file.

the DP suggested 3200 'too noisy' and proceeded to shoot underexposed 800, the camera remained on 800 for the whole shoot and the histo topping at 50-60%

But there we go.. no one seems to do any tests on what things like ISO change actually do.

Of course Ive done some tests with my c200 and learned some stuff. Winding down to 100 iso would not work.. a moody lut would

to be clear on a 2015 scarlet at 100 would work.. today? im not sure.
 
ahalpert "I completely agree with Sam that if you can handle the data and have a good workflow, there's no reason to use anything else."

I think I was being a little subtle.

More like.. "if a raw workflow means you can save money on set then it may be worth the extra hassle in post"

Of course there is more blur being..
444 file are extremely flexible as much or more than raw.
444 files may be larger than raw at certain compression ratios
linear capture may not provide the best quality per byte as compression of highlights in a non linear manner will save data that can be used elsewhere for a given data rate.


so the trade off of post flex and data rate is not clear until one examines particular cameras/workflows, shooting top end 444 is neither bad quality or efficient with data as raw is also not bad quality or efficient with data
 
Now here is a thing... poor information and knowledge.??

DOes changing the ISO on Red change the core file.. afaik back in the day when I wanted a scarlet (and got and fs7) my tests said ISO was fixed and the numbers just changed the monitor and how the file opened in the NLE

In which case using 100 would have worked

BUT the camera team were suggesting in the dark we move the camera to 3200, suggesting that chaging ISO (or certainly certain steps of it) do change the core file.

the DP suggested 3200 'too noisy' and proceeded to shoot underexposed 800, the camera remained on 800 for the whole shoot and the histo topping at 50-60%

But there we go.. no one seems to do any tests on what things like ISO change actually do.

Of course Ive done some tests with my c200 and learned some stuff. Winding down to 100 iso would not work.. a moody lut would

to be clear on a 2015 scarlet at 100 would work.. today? im not sure.

The core file is unchanged. You're just looking at a preview of pulling it down in post. You will lose highlight latitude as a consequence of increasing shadow latitude. But that should be apparent from your display

Rating the camera at 3200 is a crazy suggestion here. He would never work in your town again if he rated at 3200 and underexposed 3 stops
 
But fwiw I worked on a show where the (youngish) DP shot Red rated at 800 and underexposed two or three stops I would guess (if I remember right) and then in post sort of boosted the image by up to two stops to keep it a bit underexposed but with better highlight retention and added grain. It looked good. I've worked on shows with some of the top young indie DPs out there and when they shoot Arri they tend to underexpose by around a half stop to add some "texture." The saying is "overexpose film, underexpose digital."

Right, but here we're moving away from a generic approach to exposure (eg retain highlight detail while minimizing noise) towards crafting a specific look that may call for some noise. If Sam's DP wants the noise he'll get with his exposure approach, that's all well and good. But I got the impression that he was making decisions that weren't carefully thought through.

I remember reading about how much testing they used to do in Hollywood. (Perhaps some still do?) Before starting principal photography, they would test the film stock, even if they'd use it a million times before. Have it processed by the lab they would later use. Test the lighting on the actors in make-up. Test the grade. View it all on a movie theater projection and make sure they were happy with the approach.

I don't think the method of Sam's DP was intended to produce noise, but just achieve a low-key look. You could have a noisy low-key look or a noise-free low-key look. But hey, maybe he knew what he was doing. If the assistants were suggesting to preview at 3200 so he could expose properly at that ISO, that would make sense, but it would be the opposite of crafting his look by eye. Seems like he wanted a dark preview
 
stillswise I shoot raw and jpg, often dont actually bother dinking with the raw unless im being paid.

----------

Im not current with red. Early reds had one 'iso' and any changing of settings just affected the monitor and the start point in post.
Im guessing this is no longer the case but I dnt know. Many cameras now have dual gain or whatever it is called
if this is the case then shooting at 3200 in the dark would be good advice, becuase you are turning the 'amplifier' on

"overexpose film, underexpose digital."

Pesonally I dont agree .

Words like over expose and ISO are really random. Use a WFM to keep you highs just below clip point is what is needed with digital raw
With log less bits are allocated to highlight stops so, having 'important data' in the 50-70 range is probably good advice.

What does over expose mean?
Do something different to your sekonic meter?
Allow lots of clipping on the non lutted WFM?
allow lots of clipping on the lutted WFM?
Make the monitor not moody? with a lut? without one?

Clearly its not agreed. what over exposure is.
 
I should have siad for stills I absolutely shoot raw professionally. In fact beyond that for locked off stills (hotel foyers) I would do multiple exposures to colllect maybe 20 stops of data ready for plonging together in post.

BLF - bracket like f--k
 
To me, overexpose means rate the camera slower than the stated ISO – so for 500T film I'll often set my meter to 320T or something. This might be an approach designed for beginners but it's what works for me. For digital, it's the opposite, I'll rate a camera at 1000 or 1250 ISO when the stated ISO is 800.

I believe that the first approach you described is correct. And that to overexpose on digital you would do the same thing.

This is all assuming that you are always delivering the same exposure levels to the camera (regardless of if you expose by medium gray, white, light meter, etc.) Thus, when you lower the rating on your meter, you must compensate by brightening your lights or opening your aperture. The result is higher intensity light hitting the sensor.

If you have a normal exposure at ISO 800 and then increase the in-camera gain to ISO 1250, you will end up overexposing. But if you then compensate with lighting or aperture changes, then you'll end up delivering less light to the sensor and underexposing for base ISO (relative to your own exposure target or manufacturer recommendations or whatever metric you choose. It's a relative determination)

Also, I don't think that the metering approach you described is for beginners. I'm pretty sure that top shelf DPs do the exact same thing.
 
not having been to college for some decades I cant even remember what correct exposure is!

a grey card put the meter in the middle if I remember?

wheras im always working down from clip in digital which is all I do nowadays.

i point the camera at the scene and close the iris until that there is no flat at the top of the wfm

(or I make a choice about clipping.. can the halogen spots go? can the windows go? the hairlight?)

Both situations dont really tak account of the DR in the scene

because the grey card might be at 5.6 but the view out the window might be f32 and the black cat in the coal hole 1.2. so what is the 'correct exposure' then?

Or if Deakins exposises like this is that correct exposure? Or is he incorrect? Say that to the face of his fanclub!

I consider incorrect exposure not to be a specific greay card on the wfm, but the fact the file wont go where you want it to in post because of clipping, noise or colour.
ergo corect exposture is anything that can produce the result you wanted to get.
 
because Ive been so busy crewing - ive not played with the C200 much.

ive chosen to keep the exposrue 'the same' if the windows are or are not in shot - this is a choice

this https://vimeo.com/690426517

maybe the bright side is a bit 'thin' and the nikkors dont match the canon

I think it is sharp

I went out and bought a stigma 18-35 after this thinking that best was to get colour consistency (on a budget) by sticking mainly to one lens.

ive not had time to try it!
 
Its a fair point that basically ETTR maybe not getting the best from these cameras - unlike a ture 14-16bit still camera where pure ettr rocks.

I certainly found that 'keeping the hump in the middle' worked well with my FS7 and here (Florence) im saying her cheek might be a bit bright and the texture a bit thin.

So yes a bit deeper than one imagines might indeed have value :)

There is also a vogue for shooting what is IMO 'too dark' Suprised by Oxford which I worked on seemed to be crawling around in the shade too much for my taste. Im not sure we ever got to properly enjoy the beauty of the starlet!
https://filmfreeway.com/SurprisedbyOxford
 
I still have yet to have worked on anything--TV series, commercial etc--that would allow me to shoot on RAW even if I suggested it. The extra step in workflow is not happily accommodated because it adds cost. It is probably testimony to the world that I find myself in professionally, but the kind of nuance that RAW would allow over a robust ProRes codec is sadly not something that anyone other than me would care about. If it was a particularly punishing exterior situation I might be able to make the case as far as holding detail in sky...but probably not.

As far as underexposing for digital, I don't personally have an overall rule on it. To me "proper exposure "is not pegging it where the meter says, it's how I want the individual scene to look. A couple lying in bed at night is going to look awfully lit if you expose "properly", I want their faces probably 1.5 - 2 stops under. If by underexposing we mean get the scene where you like it then stop down 1/2, yeah I don't do that. I'd rather get the "fat negative" or fill the data bucket as Sam says, we can always bring it down in post. I've suffered a few times by having producers decide they want it brighter and I've given them the room to do that than deliver a crushed camera original that they are stuck with...but hey, the checks still cash.
 
Depending on the camera, there might not be an extra step or cost - but most of the time it's not worth it for people if the camera offers other "robust" options.

But the days of quirky issues with RAW in NLEs, or ludicrous storage costs, or unreasonable data rates are really a non-factor for productions making these kind of considerations in 2022+ (unless maybe it's like 8K/60p at ridiculous numbers).
 
Depending on the camera, there might not be an extra step or cost - but most of the time it's not worth it for people if the camera offers other "robust" options.

But the days of quirky issues with RAW in NLEs, or ludicrous storage costs, or unreasonable data rates are really a non-factor for productions making these kind of considerations in 2022+ (unless maybe it's like 8K/60p at ridiculous numbers).

The thing is, a lot of the decisions being made aren't based off of what is possible today, but what was a decade or so ago when "posts" workflow was established. And a lot of these shows, the post team is set in their ways and those cats call the shots and run the show. It's based on what makes their lives easier. Side note: it's always irritated me how the people that sit in the climate controlled building in their comfy chairs with the luxury of getting an extra few moments to think about and check things, get to make all the decisions about how stuff is done in the field, regardless of how much harder it makes things, that are often in uncontrolled(less controlled) environments/situations.
 
Back
Top