Film Emulation - Trying to make DSLR video as close to film as possible

cyzercraft

And here is the last but not least! The main presentation of the thing I have been working on .. and yes it has a name now: EASY FILM

jpbturbo

That looks like instagram, not film.

maxr

@jpbturbo Man... lately you have this thing for telling the truth :P

@cyzercraft you did wrapped them well 4 sure.
Best of lucks

flablo

@cyzercraft maybe it's Vimeo compression or your encoding bitrate, but I see a lot of fuzziness

ahbleza

This seems relevant....

http://www.tjtfilm.com/Film_Emulation_LUTS.html

shian

^^^ that is exactly what I've been talking about. It hit every point and then some.

rNeil

It's that bit @shian mentioned about needing BOTH the film stock data AND a print-stock data that most "film-emulaters" miss whether in still images or video. Coming out of a still-shooter background, I've seen SO many presets for say Lightroom that are things like "Tri-X @400", "Tri-X @ 1600", "HP-4" or whatnot.

Huh? No one ever looks at just the flipping negative, you look at the print! So ... that's always a combination of film and paper. Tri-X @ 1600 printed on say a grade 1 high-silver-content paper will have a particular look it will NOT share with the same negative printed on a grade 5 RC-paper. Maybe you like the second, fine ... whatever. It isn't the same at ALL. This has driven me nuts for the last decade.

Let alone considering that MY "pushing" of Tri-X to 1600 will never be exactly what YOUR pushing of Tri-X to 1600 will be in that negative film only.

After some of shian's comments here and a few others, I've found and read some about this ... including one dude who noted the amazing variability in how he uses his "chain" of home-made LUT's in his professional colorist work. Starting with a 'correction' LUT for a particular camera (sometimes even camera/lens), THEN a film-stock emulation LUT, then some correction work but as he's a hold-over from the "old" days and did corrections on film workflow, only the things he could have done there ... THEN a print-stock LUT with a few touches after that's put in place to overcome inevitable non-sequiturs in the process.

And depending on the cameras used in a project and the look the DP wants, it can be an amazingly complex chain of choices to get all the footage 'there'. So far past my knowledge/depth of understanding to be essentially Magic.

Great stuff ... but like shian, years of experience, accurate and useful knowledge, and access to all the stuff old and new to create valid comparisons. And this is just as lacking in most stills shooters/processors today as for video folk. Sadly.

But dang it looks pretty when done righteously.

Neil

Oedipax

Those TJTFilm LUTs sound promising, but I don't see any info on whether they plan to sell them or not!

BurnetRhoades

@rNeil print stock is not a given part of the process, especially now. You aren't always looking at print or the influence of any print stock.

When you see a motion picture that has been shot on film and released as a DCP you're not seeing any release print stock. It's not part of the equation. It does not contribute to the look you see in the theater. Negative is scanned for DI, the digital interpositive. No print stock. No print stock lut.

Print stock luts are useful when you know you are going to be going back to film, to preview on a digital display before making that costly trip to the printer and lab. Apart from that I suppose it would be useful for emulating certain kinds of telecine material but, again, film to tape transfers aren't exclusive to print material.

Maybe there's something that I'm missing here but it's effectively anachronistic to use them in this way, for digital distribution and viewing. Unless you're simulating something like Technicolor's ENR, which is an enhancement to the look of a film exclusive to the printing process itself (which isn't what these print stock luts are doing), I'd need serious convincing before I believed the additional LUT for print stock emulation is anything but pointless and mildly destructive.

If you read the notes on the ultimate edition of Ridley Scott's Legend for BD release, they're basically apologizing for having to use an answer print to do this version of the film, because it was simply the highest quality material available. Answer Prints are generally the highest quality print version of a completed motion picture! They are what's used to make the internegative that release prints are struck from.

goanna

Answer Prints are generally the highest quality print version of a completed motion picture! They are what's used to make the internegative that release prints are struck from.

There are fewer and fewer cinemas still using optical projectors. AFAIK the 4K (or 2K!) digital copy we now see projected is produced straight from the DI (Digital Intermediate not Digital Interpositive).

Personally I find a film shot on film stock simply looks better; i.e. clearer, more apparent resolution, smoother highlights, better shadow detail and, of course, they tend to be better lit and have bigger budgets and better scripts and everything.

By contrast, I have always found that the "looks like film" claim refers to a result which is often quite poor and looks like a poor film (ie so bad you can't tell the difference).

With a movie like Upstream Color there's something good going on which I can't quite put my finger on and I'd like to. Shot on a hacked GH2, it's got the right look for me and somehow I don't care if it looks like film - and this quality, for me, is the future for independent narrative film making.

maxr

@cyzercraft if you feel this is not pertinent to the "discussion" I'll take it out with no sweat ,-)

Original footage

Recreation from Balibo film

Trailer from Balibo film

SPOILERS
In this movie, other than a very nice OST - though at the end of the film can't really enjoy Lisa Gerard's epicness stuff while indonesian sons of bitches are shooting in cold blood ordinary timorese citizens - and superb camera work (double challenge as they nail 70's medium and close up handheld shot style too), there are very interesting grading decisions. I.e. the blueish tonalities and harsh vignetting for scenes where Roger East imagines/recreates/discovers what could had happen to the five ausie journalists. Or a parallel narrative.

IMHO this grading decisions come from a context and are used for a context, here 1975 and 8mm (I think) shoulder style camerawork. Not a masterpiece but worth the watch =)

PS
Ok found some words from the director, Robert Connolly:

I used ingenue lenses from the ‘70s, standard 16 lenses not super-16 lenses, I graded it – using Brett Manson, an amazing grader who also did Tsotsi – to make it look like reversal, we used a faster stock that had more grain in it and a whole range of things. We emulated the style of that time with the camera movements – handheld!

Viva Timor Lorosae!!!

BurnetRhoades

Very cool.

rNeil

@BurnetRhoades,

If you are talking only about a "current" practices/tools workflow for those choosing to shoot film and deliver digital, then yes ... there isn't a print-stock part of the process. But then, that's not what I was referring to, nor what Shian up above referenced either, is it?

What that part of the discussion was about is the notion of matching the total look of a say, late-60's/early-70's film as released ... which most definitely DID include both a film-stock used in the camera, AND the print-stock used for release.

Which is why the comments about someone trying to "match" the look of the film-stock only and thinking that they were then going to be able to match what the released-to-theatre PRINT looked like.

Neil

BurnetRhoades

@rNeil if you're matching the look of a low quality telecine from one of those late-60s/early-70s films perhaps. Otherwise it sounds like something for maybe the next installment of Grindhouse.

A good quality DVD or BD transfer of these films does not look the same as seeing a projected print of the same film in a theater and it may use negative material, resulting in a different aesthetic than what you might remember from seeing the film in theaters. Similarly, you're not getting the same experience from a DVD or BD release of David Fincher's Seven as you might have in one of the select theaters to get an ECC release print.

So that begs the question, when doing film emulation on digital footage with the target being one of these old films, are you trying to emulate what that film would look like after a high quality telecine, as viewed on your digital video display, or are you trying to emulate the look of that film as projected in the theater, while still viewing on your digital display?

The former makes more sense than the later, which would actually need a third LUT, the one that emulates how much of the light passed through the print hitting the screen would be reflected into your eye at some pre-defined distance from screen. Because, yes, you don't just look at negative, you look at print, but you don't just hold the print up to your eye and you don't watch the film through a loupe over a light table. So this can get rather silly.

Realistically speaking, 99% of potential users for film emulation are using as their benchmark the look of film origination as it appears on their video displays, be that their computer or television set, watching stream, digital disc or broadcast. Some material viewed this way might necessitate a two-LUT, dual film emulation. The ideal as expressed by deluxe, archival, special edition or restored releases of films distributed in the digital realm will not always have a print component and therefore require a one-LUT single film emulation strategy.

Oh, and:

But then, that's not what I was referring to, nor what Shian up above referenced either, is it?

He was responding, emphatically, to a posted article referencing a professional colorist's approach to film emulation that's similar to both his and that of Film Convert. In the body text of the article as well as the caption to every single image depicting film emulation he mentions two LUTs being used, one for negative and one for print. So, yes, it would be reasonable to assume this is what both you and he were referring to as "the way" and the missing piece of film emulation as a general practice.

maxr

There's a saying "each crazy with his theme" :P

3 more examples of recent films which happen in the 70's context and 3 different grading courses

Hunger, 2008. 1st of McQueen's films. Context: inprisoned IRA activists fighting for their rights.
This movie also features one unforgettable long dialog scene with phenomenal acting

Balada Triste de Trompeta, 2010. Very strange/macabre film by Alex de la Iglesia with great intro titles.
Context: Spain somewhere at beginning of the 70's • clowns are a BIG metaphor :o)

Le Havre, 2011. Though the film is supposed to happen nowadays I think Aki Kaurismäki is still living in the 70's
There is a care and (false) simplicity in the scenography only rivalled by the colour palette, just beautiful

This one though beautifully graded (with no hope), it's a bit out of - era - context

matt_gh2

@goanna Agreed that Upstream Color has nice look to it. I think these days, audiences are getting used to a variety of looks/textures/image types etc. It's all about whether "it works" for the film...or better put..."how it works" for the film. Not sure what hack setting, camera settings, and lenses Upstream Color guys used, but it's nice.

@maxr Le Havre definitely has that 70's look to it, both in texture/image type but also in choice of shot composition.

BurnetRhoades

Le Havre definitely has that 70's look to it, both in texture/image type but also in choice of shot composition.

While being shot on the same Vision3 stock as all of this year's non-Fuji film-origination. Astounding.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters