How do I get the film/cinematic look with a digital camera?

[Carl Looper]

is this a windup.. or just bolex :)

 

:D

[Gregg MacPherson]

Is there another windup in play....was that drawn without a pencil??

 

Is this camera of the female genus? Some unwinding on Freud's couch could ensue....

Edited April 21, 2016 by Gregg MacPherson

[Carl Looper]

Is there another windup in play....was that drawn without a pencil??

 

Is this camera of the female genus? Some unwinding on Freud's couch could ensue....

 

 

I'll leave it up to you Gregg to psycho-analyse the work. B)

 

Wasn't actually done with a pencil. Was done with a 0.4 black artline, on photocopy paper, scanned on a flatbed, scaled and contrast enhanced, and saved as a 1 bit png, using perceptual dithering. Very small file size - only 50KB

 

Here's a jpeg of the same scan, but in order to show it at scan definition, had to crop it (to fit 1024 x 768). Compression quality = 60% to get it under the attachment limit of 300 KB.

 

Edited April 21, 2016 by Carl Looper

[Michael LaVoie]

The film look is a combination of image acquisition (shooting) and grading (post-production). This advice applies to the image acquisition portion. When this whole digital as film thing started, 24P and the DVX-100 came out, American Cinematographer did a shoot out test with that camera against 35mm and 16mm and then listed a whole array of menu settings that they recommended to get the best "film look". Gamma settings, color settings etc. One important tip at the time, not sure if it was theirs or another article was to use a black pro-mist filter on the lens to cut the harsh highlights. This was Rec709 of course. Way before log or raw. It was one way during shooting to help achieve the film look.

 

Before DSLRs and modern cameras like the Red and Alexa, in order to throw a lens right onto a digital camera, you needed a Movietube, PS-Technick, or Letus35mm lens adaptor. Most of these 3rd party mounts had some sort of either spinning glass or vibrating ground glass that would help greatly in providing a cinematic look as it created a soft subtle texture to the image. So you could take off that pro-mist filter as the adaptors were doing that work for you.

 

Now that we have a straight up lens on the sensor with no need for any such device, we are stuck with very crispy images. I went right back to using the 1/4 black pro-mist immediately after the adaptor period was over. But I don't think most people remember that or know of those days. A black pro mist is a very subtle effect but it rounds out the image nicely. Some DP's swear by nets on the lens for a similar effect but I have found that a Tiffen 1/4 black seems to do the job without calling attention to itself.

[Nicholas Kovats]

Once again an elegant rendering to backup your statements. Touche!

 

 

 

I'll leave it up to you Gregg to psycho-analyse the work. B)

 

Wasn't actually done with a pencil. Was done with a 0.4 black artline, on photocopy paper, scanned on a flatbed, scaled and contrast enhanced, and saved as a 1 bit png, using perceptual dithering. Very small file size - only 50KB

 

Here's a jpeg of the same scan, but in order to show it at scan definition, had to crop it (to fit 1024 x 768). Compression quality = 60% to get it under the attachment limit of 300 KB.

 

Scan_20150914.jpg

[Nicholas Kovats]

Fascinating, Michael. This topic deserves a unique thread. On the various available techniques to "degrade" clinical digital sensors. Are there any serious efforts to introduce modular Optical Low Pass Filters (OLPF) that can fundamentally introduce "randomness" in sensor captures? And so on.,,

 

 

Now that we have a straight up lens on the sensor with no need for any such device, we are stuck with very crispy images. I went right back to using the 1/4 black pro-mist immediately after the adaptor period was over. But I don't think most people remember that or know of those days. A black pro mist is a very subtle effect but it rounds out the image nicely. Some DP's swear by nets on the lens for a similar effect but I have found that a Tiffen 1/4 black seems to do the job without calling attention to itself.

[Robin R Probyn]

But film can also look incredibly sharp too.. as clinical as anything shot on Alex/F55/RED or any other the other high end digital camera,s.. Im not sure this is a difference between the two.. ?

[Mark Dunn]

Robin: You can make a point without being a dick about it.

Americans are statistically much less likely to be able to understand irony, of course. It's a potential character flaw but not as bad as obscene language on a public forum.

BTW if film grains are "on/off" as you put it how do you account for continuous tone rendering and why are all images not soot-and-whitewash? (If you need to look up the former term, which I think you do, please consider doing so before replying. Don't bother with wikipedia, that definition won't help you. Try a real photography textbook).

[Aapo Lettinen]

cinematic look with a digital camera?

 

1. lighting

 

2. color correction

 

most of the other stuff is just trying to make the image to not look bad if you're shooting with mediocre equipment. good lenses and dynamic range may help if you want to use lots of flares and stuff or shooting wide open all the time for some reason.

 

I think it is usually not necessary to try to make digital to look like film. just make it look great on its own, to have its own look, and try to benefit from that.

 

but if someone is very short on money or wants to cheap on things, usually on shoestring/student/low budget level, one can try to emulate film if it is absolutely not possible to shoot on the real stuff.

I don't see a reason why one should do that but it may not necessarily look that bad though it's nothing like real film in the end. a little bit of grain is handy when trying to hide bad VFX however so it has its uses :lol:

[Carl Looper]

But film can also look incredibly sharp too.. as clinical as anything shot on Alex/F55/RED or any other the other high end digital camera,s.. Im not sure this is a difference between the two.. ?

 

One could even get a 48fps video-like look with film.

 

I've been experimenting with multiple 16mm film projection. I've managed to hack two 16mm film projectors, so that they run in sync, under computer control. They stay in lock through phase lock control of the motor, from feedback sensors on the blade rate. To obtain 48 fps on the screen, each projector has a two blade shutter, of which one of the shutter openings on each projector can be masked, and then both run out of phase, so that when one projectors shutter blade is closed, the other is open and vice versa.

 

Film would be shot at 48 fps and then two prints made where every second frame from one is printed to one print, and every other frame is printed to the other print. I have a programmable optical printer I've built that can do that. The two prints are then projected together, through a beam splitter, so that the result on the screen is basically the two films being interlaced, the result being a 48 fps screening.

 

I think it'll look quite interesting - video like but at the same time not. I'm curious to see how it differs from video.

 

C

Edited April 21, 2016 by Carl Looper

[Aapo Lettinen]

actually, by my experience, most of the real film originated stuff does not "LOOK LIKE FILM" , it is a very special narrow range of looks which people are considering "film look" and it is very easy to miss that with for example lighting or colour correction. shooting for technical quality (very sharp, minimum grain & aberrations, etc) + lighting decisions + heavy colour correction is the usual way to do that

[Carl Looper]

BTW if film grains are "on/off" as you put it how do you account for continuous tone rendering and why are all images not soot-and-whitewash?

 

Film grains actually are on/off but the light (and image) it passes isn't.

 

An image mediated by light, as much as the light itself, isn't to be found in any particular photon detection or the number of detections at any single point. Indeed in the grand scheme of things there may only ever be one photon detection or not at any given point in space, regardless of how such is created. For how small is a photon detection? What are the chances of two photon detections occupying exactly the same point in space?

 

Rather an image is in the distribution of the photon detections, which requires the concept of an area, rather than a single point, although we might ask of this area that it be vanishingly small (down to some limit determined by the physics). While photon detections are either on/off it's actually the wave function of the corresponding photons which physically characterises the photons that produce the detection, and even more so images. A single photon detection carries next to no information at all. Its only in relation to every other photon detection that one gets a sense of the light proper. Indeed this seems to be part of the very physics of light itself, as described by wave functions - which are very much continuous functions. They physically determine the distribution of photon detections. Or the distribution of photon detections require we represent them this way.

 

We can get a sense of this in the following image where each pixel is either on or off (black or white) but there is nevertheless a variation in tone here. The tone is in the distribution of the pixels rather than the specific value at any particular location. At any given location (pixel location) the information is just on or off. The tone we otherwise see is quite literally the distribution, and what we can otherwise call the image proper, and this is completely analogous to the continuous wave function that is said to determine the distribution of photon detections. Or we can just say the image determines them.

 

 

Edited April 21, 2016 by Carl Looper

[Carl Looper]

Its worth comparing the above image with this one. Here the pixels are exactly the same - on or off. But an entirely different result.

 

In film, if all the grains were exactly the same size this is exactly the sort of result you would get. This is why the size of the grains is just as, or more important, than whether they are on or off.

 

Edited April 21, 2016 by Carl Looper

[John E Clark]

Its worth comparing the above image with this one. Here the pixels are exactly the same - on or off. But an entirely different result.

 

In film, if all the grains were exactly the same size this is exactly the sort of result you would get. This is why the size of the grains is just as, or more important, than whether they are on or off.

 

face2.png

 

There's size and orientation. I didn't join the T-Max bandwagon, but I recall one of the benefits of T-Max was finer but more 'regular' grain structure. B&W motion picture stocks where never 'improved' with modern developments, so there isn't much in the way of examples.

 

In color, given usual processing, the 'grain' is not 'sharp' crystals of silver but fuzzballs of dye, again, with random orientation and depth in the 3-d emulsion.

 

What most people see when they see 'grain' is glumps of crystals or fuzzballs, or in any case, 2-d projection of a 3-d collection.

 

With still printing one could use a 'grain focuser' to determine the print focus for the 'sharpest' grain effect... but for me... I sometimes felt that the resulting image was 'less sharp'... so I tended to focus by eye on the 'larger' print scale rather than the 'microscopic' level.

Edited April 21, 2016 by John E Clark

[Nicholas Kovats]

Screw interlaced video! I want my interlaced film! Pushing the hybrid analog/digital envelope once again, Carl.

 

 

One could even get a 48fps video-like look with film.

 

I've been experimenting with multiple 16mm film projection. I've managed to hack two 16mm film projectors, so that they run in sync, under computer control. They stay in lock through phase lock control of the motor, from feedback sensors on the blade rate. To obtain 48 fps on the screen, each projector has a two blade shutter, of which one of the shutter openings on each projector can be masked, and then both run out of phase, so that when one projectors shutter blade is closed, the other is open and vice versa.

 

Film would be shot at 48 fps and then two prints made where every second frame from one is printed to one print, and every other frame is printed to the other print. I have a programmable optical printer I've built that can do that. The two prints are then projected together, through a beam splitter, so that the result on the screen is basically the two films being interlaced, the result being a 48 fps screening.

 

I think it'll look quite interesting - video like but at the same time not. I'm curious to see how it differs from video.

 

C

[Nicholas Kovats]

Bang on again, Carl. You just illustrated beautifully the rigidity of uniform pixel distribution (an x-y matrix) relative to the randomly dispersed and varying size of film particles and their distribution (x-y-z). Just astounding that these fundamental principals were discovered over 100 years ago.

 

 

 

Its worth comparing the above image with this one. Here the pixels are exactly the same - on or off. But an entirely different result.

 

In film, if all the grains were exactly the same size this is exactly the sort of result you would get. This is why the size of the grains is just as, or more important, than whether they are on or off.

 

face2.png

[John E Clark]

Screw interlaced video! I want my interlaced film! Pushing the hybrid analog/digital envelope once again, Carl.

 

 

There is the lenticular effect, often used as Cracker Jacks prizes for over a century...

 

Here's an image series of the effect, which could be done using these days with display panels that are long and narrow...

(I take no responsibility for the image sequence and what it may import to any one... any where...)

 

 

Edited April 21, 2016 by John E Clark

[Max Field]

 

For simplification and longer explanation, I have made a video all about the blackmagic pocket camera. It dives into all the questions you asked and then some.

This is a great educational piece, production on point. With that said, I need to buy you some new sneakers.

[Tyler Purcell]

This is a great educational piece, production on point. With that said, I need to buy you some new sneakers.

LOL! That's what happens when all your money and time goes into equipment and having fun, shoes aren't very high on the priority list! HA!

[Carl Looper]

 

There's size and orientation. I didn't join the T-Max bandwagon, but I recall one of the benefits of T-Max was finer but more 'regular' grain structure. B&W motion picture stocks where never 'improved' with modern developments, so there isn't much in the way of examples.

 

In color, given usual processing, the 'grain' is not 'sharp' crystals of silver but fuzzballs of dye, again, with random orientation and depth in the 3-d emulsion.

 

What most people see when they see 'grain' is glumps of crystals or fuzzballs, or in any case, 2-d projection of a 3-d collection.

 

In T-Max the grains (or crystals) are grown flatter than their predecessors - more like discs as distinct from their predecessors which were more ball shaped. And their flat side is engineered to face the direction in which light hits them. This provides for a more efficient light sensor in various ways. For normal contrast stock, the crystals still need to vary in size. The variation in size is that which provides for tone encoding.

 

And yes, in colour film the silver is eventually removed and replaced with dyes in the location otherwise occupied by the silver.

 

And yes, the graininess we otherwise see with our eyes is not really at the level of individual dyes, and even less so at the level of the silver atoms occupying the crystal site, due to how they overlap each other, and grain aliasing.

Edited April 22, 2016 by Carl Looper

[Carl Looper]

Screw interlaced video! I want my interlaced film! Pushing the hybrid analog/digital envelope once again, Carl.

 

 

Thanks Nicholas. The dual projector setup also allows for 3D projection. I'm currently making a short 3D film, to screen on the projectors, for the opening of our new lab. To mix things up a bit, the source imagery will be a mixture of material shot on film, video and otherwise computer generated. I'm a bit of an "anti-purist" in this way - or perhaps I'm more of a purist - in the sense that one could argue heterogenity (difference) is more fundamental or pure than homogenity (sameness). Either way "hybrid" is a good word for it. Or "mongrel". For the same reason that genetic diversity might be considered better than in-breeding.

 

C

Edited April 22, 2016 by Carl Looper

[Carl Looper]

 

There is the lenticular effect, often used as Cracker Jacks prizes for over a century...

 

Here's an image series of the effect, which could be done using these days with display panels that are long and narrow...

(I take no responsibility for the image sequence and what it may import to any one... any where...)

 

 

 

This brings up another experiment I've been working on, where I've affixed a lenticular sheet onto a computer screen (and could do something similar for rear projected film). The sheet cost me no more than $25 and is simply affixed to the screen. The sheet is diagonally affixed (the lenticules running at an angle) as this allows for much easier calibration and a better distribution of colours. There's no need to be particularly fussy with attaching the lenticular as whatever angle/position it ends up being is auto-calibrated using a test pattern on the display and a feedback camera trained on the display.

 

Once calibrated a multi-view video (8 views per frame) is spatially interlaced in real-time (using a GPU shader) according to where the lenticular sheet ends up being positioned on the screen. The result is glasses free three dee (bit of a tongue twister).

 

I want to try the same thing with film, which it is a bit more tricky due to projector lens distortions and other factors to take into account during prep, (and no way to change such once it's burnt into the film), but I can imagine a properly created setup would allow for some creative scope in terms of what could be done with such.

 

C

Edited April 22, 2016 by Carl Looper

[Carl Looper]

Well, getting cinematic looking digital images is challenging. It all comes down to a few simple things:

 

- Dynamic range (both of the imager and storage format)

- Shallow depth of field

- Type of shooting (smooth shots/dolly/crane, etc)

- Color correction/grading

 

For simplification and longer explanation, I have made a video all about the blackmagic pocket camera. It dives into all the questions you asked and then some.

 

 

I watched this entire video - which isn't my usual cup of tea (but very much enjoyed it) - however I was waiting for the part where it answers the question as to how one gets a film/cinematic look with a digital camera. Was looking forward to debating that :)

 

But not to worry, it was great to see Tyler in the flesh (so to speak) and doing a great job selling Black Magic cameras.

 

And they are great cameras. I was fortunate enough to play with one last year on a job we were doing in Indonesia. Was working with the Black Magic software development kit (SDK), writing software to tap the camera's live 10 bit uncompressed HDMI stream, for processing such in real time, and it was a delight to work with. Really well documented. Was able to write the software over a couple of days in my hotel room and have it working live on location without any issue at all. And not only that, was able to write a version for both Mac and PC which isn't typically that common in an SDK.

 

Was recently playing an acting role in a feature film (well, a support cast role, or lets be blunt: an 'extra'), that is being shot on the Black Magic pocket camera. Its so strange to see a serious work (which it is) being done on such a small camera. But that's the great thing about the cameras - that they are small, yet give you an exceptional image. They really are some form of black magic.

 

C

Edited April 22, 2016 by Carl Looper

[Landon D. Parks]

If you're looking for a very basic, nitty-gritty 'how to get a film look' tutorial, I have just the one for you. Courtesy of me.

 

https://youtu.be/WS9-y3hRkLI

 

You're welcome!

[Gregg MacPherson]

@ Carl, re the bolex drawing.

 

I wasn't being curly. I was trying for the collegial and humorous. I saw the memo from bill and was trying to do my bit.

 

I used to really enjoy photocopying pencil sketches, and drawings with biro. Even a biro has a range of values between simply existing or not on the page, and this did translate into interesting high contrast images on the Xerox page. I never really went for the Rotring thing. Was your Artline a fibre point or steel?