Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Ask David Mullen ANYTHING

Hello David,
I'm almost to embarrassed to post this question. I have suddenly shockingly realized that I don't understand the relationship between sensor size and resolution. I was under the impression that the larger the sensor, the better the resolution. But then there are cameras like the F35 that have a full 35mm sensor but have a low 1920x1080 resolution, or some of the HDSRs that have even bigger sensors but again only shoot 1920. Are resolution and sensor independent of each other, and if so is DoF the only purpose of a large sensor. Or the HDSLR sensor capable of shooting 4k but the recording medium isn't, so that's how you get the 1920 resolution.
I had always understood that as RED released the EPIC and goes to higher resolutions that the sensor was getting bigger. Is this incorrect, and the sensor is staying the same size but there are more pixels on the sensor (and the pixels are getting smaller - like a tighter film grain?).
Thanks.
Nicola
 
In film, larger negatives using the same emulsion allow more resolution because more area is devoted to capturing detail.

But in digital, it's more complicated because you can put fewer but larger photosites into the same area on a sensor as more but smaller photosites, and to some degree, the number of photosites / pixels is one way of labeling resolution, though it's a bit misleading -- for example, you can have a 12MP sensor and a super sharp lens but put a heavy diffusion filter on the lens or simply shoot slightly out-of-focus and thus not record an image with a lot of resolution though it has a high pixel count.

Yes, the main reason for 35mm sensors is to be able to use 35mm lenses and get 35mm depth of field. But larger sensors also allow you to get more photosites on the sensor without having to make them too small to fit, plus lenses don't have to be quite as sharp because you don't need to resolve as many lines per millimeter when your image is captured on more millimeters total. So a lot of factors play into resolution and there are a number of reasons to make the sensor larger.

The F35 / Genesis sensor is something like 12MP actually, though the nature of its RGB striped color array filter means that it is specifically designed for a simple internal conversion process to 2MP per color, which is basically what 1080P RGB is (2MP devoted each to red, green, and blue should mean that the sensor is 6MP total, but there is an extra row for each color to record overexposure detail, which is why these cameras have a pretty good dynamic range.) The main advantage to this design was the extra dynamic range and the simple processing algorithm needed to create 1080P RGB, unlike the more complex algorithm needed to convert a Bayer filter pattern to RGB, which is why the Red camera saves debayering at full resolution for post, generating only a lower-resolution 720P live color image for monitoring purposes. But as processors get faster and better, this is less of a hurdle as the RedRocket card proves. (It's just that there is little reason to do a full debayer to 4K RGB inside the Red One since leaving it as a 4K RAW file saves on the amount of data to record -- converting it to RGB triples the amount of data.)

There is a healthy balance between having a lot of pixels to record an image versus using fewer but larger photosites; there are a lot of factors to consider. However as sensor design improves, they are able to get more efficiency out of the system with fewer tradeoffs in cramming more photosites on the sensor, which is why you are seeing 18MP DSLR's that seem more sensitive than ever, or at least, have less image noise, allowing more underexposed detail that is usable in order to shoot in low light levels.

DSLR's have the problem of a processor and recorder that weren't really designed for the burden of shooting 24 frames a second or faster at full resolution, so in order to get the 12MP or so sensor output down to HD levels quickly and simply, they use tricks like line-skipping and pixel binning to reduce the amount of data to handle, but this results in resolution and aliasing problems. Plus the compression and the rolling shutter artifacts are pretty heavy. All of this will eventually improve but it requires the DSLR makers to spend more on the motion image processing/recording technology inside the camera, but perhaps that doesn't make sense financially for them -- yet. Red cameras, on the other hand, are obviously designed from the ground-up for the purpose of capturing moving images.

The Mysterium-X sensor is basically a bit larger than the original Mysterium sensor, 30mm wide instead of 24mm wide, to allow more photosites without changing the size of the photosites. I also think that probably the odd size of 30mm x 15mm (non-standard in the cinema world at least) is possibly the result of working with, around, or from sensor technology for still cameras that was already in that odd size -- afterall, the Super-35 movie film format is similar in size to the APS-C DSLR and original Red Mysterium sensor, but this new M-X sensor is similar in size to Canon's APS-H sensor used in cameras like the 1D Mark IV, which is a less common size, falling somewhere between APS-C and Full-Frame (and Full-Frame, 36mm x24mm, is the same size as the movie format VistaVision.)

Dalsa, who made the first 4K RAW digital cinema camera, also used a non-standard sensor size slightly larger than Super-35 and close to the M-X sensor size -- probably because Dalsa already made CCD's for scanners/telecines and probably adapted one of those for the camera.

Either way, there is a design advantage to using a larger sensor area in order to add more photosites, rather than shrinking them in order to fit more into the same area (which is necessary in order to fit 3K onto the 2/3" Scarlet sensor, the size of the photosites will have to be smaller.)
 
Last edited:
nicola b marsh said:
Hello David,
Are resolution and sensor independent of each other, and if so is DoF the only purpose of a large sensor.
Nicola
Click to expand...
What's up, Nicola!

You are correct here. Sensor size and the number of pixels on that sensor are two independent design variables.

The sensor size is a design choice largely determined by which lenses you want to be able to use on the camera. For example, a camera with a Super 35-sized single sensor like the F35 is designed to accept 35mm film lenses.

On the other hand, the F23 is basically the same camera (both output 1920x1080 to HDCAM-SR), but the latter is a 3-chip camera that takes 2/3" lenses like the Digiprimes. So it comes down to which lens characteristics you want - Michael Mann prefers the 2/3" look, while you may prefer Cooke S4's and 35mm depth of field.

The choice of how many pixels that sensor should have is more complicated. Just because a sensor has more pixels does not mean that the pixel count you started with will match the final pixel count output by the camera, or in the case of the 5Dmk2, that all of them are being used to create the final image.

Without getting too complicated, single-sensor designs (F35, Genesis, D21, Red, Phantom, etc.) generally need to start out with more pixels than they eventually output to achieve their claimed resolution while still maintaining full color information in all three RGB channels. This is because the RGB photosites that make up each pixel have to be arranged either side-by-side (striped array) or interpolated via clever software after capture (Bayer). This process reduces the measured resolution of the image, so you have to start out with more pixels to make up the difference.

Also, the more pixels are crammed onto a sensor of a given size (say, Super 35), the smaller they need to be. Larger pixels theoretically should be more light sensitive and thus less noisy than smaller pixels at high ISO's because there is more surface area for photons to hit. (In practice, this is not always the case because the electronic image processing performed by the camera after it initially captures the image can effectively reduce noise to a greater or lesser degree, depending on how sophisticated it is). But sensitivity is a design consideration when determining how many pixels to put on a sensor.

*Dammit, David beat me to it. You, sir, are a typing machine!
 
thank you thank you thank you. I don't feel like an idiot, it is more complicated that I thought.
 
Aloha David -

I'm shooting aerials of the Big Island, volcanoes/lava from a helicopter (AStarr) next week. I am more than aware of safety.

Specifically wondering about shutter and fps. I have referred to multiple websites, people and ASC manual. My understanding is to use a 90 shutter, 36 fps. I would like to shoot both 72fps and 18fps. The show is for European television but I am able to use footage for myself as well. Ideal fps for both to utilize?

We will be airborne near sunset. I have been warned against MB usage and time lost screwing on filters doesn't sit well with me. Is there another option other than shooting @ T22?

Any suggestions outside of safety? I am not discounting safety at all. Very much aware of it and the fact that I'm not 18 anymore hanging out of a Blackhawk in Bosnia.
 
I have no experience shooting aerials so I can't really help you in regards to how to swap ND filters easily and safely. I hadn't heard the 36 fps / 90 degree shutter thing, but shooting slightly slow-motion would smooth out some of the roughness and bounce. I guess they are recommending 36 fps to smooth out the bounce but a 90 degree shutter to increase sharpness so that the bouncing doesn't soften detail due to blur during a bounce. Other fps, lower ones, sort of depend on how smooth & stabilized your rig is.
 
Nowadays it's done in post but in the past, it was done with a speed ramp, like in "Blade Runner" when Roy Batty ducks as Deckard fires at him when Batty enters the apartment.

You can create more blur in post, you can pick the moment to begin the speed-up and return to normal motion, etc. and you don't have to deal with the exposure change issues that comes from actually doing a speed-ramp in-camera.

It would be a good test though to shoot someone at 12 fps with no shutter to see how it looks. You may find that it's not sped-up enough, or maybe it's too much.
 
Getting IN the camera

Getting IN the camera

David,

I was wondering if you knew of any resources of places to find tips, tricks, or other suggestions on how to get various in camera effects? One of the things I enjoy about older films is how ingenuitive they were at getting effects in camera. I've picked up things here and there, but there does not seem to be a catalogue or place to find them, it seems to just be a matter of reading and listening to everything possible to glean what I can from others.

And I'm talking about everything from using forced perspective, to coordinating with the sets / props people to get everything in black and gray to get a B&W look with a Pleasantville style in camera ... (Hopefully I am make sense.)
 
There's no single resource for all those old tricks. Some are more special effects-oriented while others are just stylistic effects to create a mood. And some are both, such as the deep-focus tricks in "Citizen Kane", some of which involved effects work such as in-camera double-exposures and post optical printing.

The pre-digital age of visual effects are well-covered in the Cinefex issues of that period, starting in 1979, plus they occasionally ran something on an older movie such as the original "King Kong". "American Cinematographer" is a good resource, interviews with old DP's such as in Maltin's book, and there are some history of visual effects books out there.

"The Technique of Special Effects Cinematography" by Fielding is a good resource:
http://www.amazon.com/Techniques-Sp...r_1_18?ie=UTF8&s=books&qid=1267399329&sr=1-18

But this is more about visual effects work.

Interviews with old DP's from the studio era can reveal a number of in-camera trick shots since these guys liked to tell stories like that, ingenious solutions they found -- they weren't so comfortable talking about the art aspect of their jobs.

But I learn about these tricks from all over the place. For example, I watched some of the Kodak cinematography videos shot in the Australian film schools and John Seale showed this neat trick to hide a light from the camera lens in a small dorm room set from "Dead Poets Society" where he painted a bit of showcard to match the color of the wall above the window alcove and put it in front of the lens to hid a Tweenie he had mounted just above the window frame. It blended with the background and erased the lamp from the shot, in-camera.

In other interview books Jack Cardiff tells a story about creating a fake low sun and sky for an exterior landscape created on a soundstage (for "War and Peace") -- he spray-painted a sheet of glass in front of the lens to hide the soundstage ceiling with a fake hazy sky above the treeline, and then reflected a lamp in the glass to look like the sun was peaking out through the mist.

Those sorts of stories are scattered throughout dozens of books.

There are lots of tricks involving mirrors and glass, or foreground miniatures, that are fun to read about.
 
Ryan E. Walters said:
David,

I was wondering if you knew of any resources of places to find tips, tricks, or other suggestions on how to get various in camera effects? One of the things I enjoy about older films is how ingenuitive they were at getting effects in camera. I've picked up things here and there, but there does not seem to be a catalogue or place to find them, it seems to just be a matter of reading and listening to everything possible to glean what I can from others.

And I'm talking about everything from using forced perspective, to coordinating with the sets / props people to get everything in black and gray to get a B&W look with a Pleasantville style in camera ... (Hopefully I am make sense.)
Click to expand...

It would be worthwhile digging up some old FX guys, if there are some in your area. I was shooting a short with a lot of practical FX, and we found this great guy in TO that had been doing FX for decades, and had a lot of experience with practical and in-camera FX.

He's in Toronto, but you could look him up anyway:
http://www.keillorfilm.com/

Good guy.
 
Diopters allow you to focus closer which then causes DOF to get more shallow, but I don't really see why they'd allow you to shoot wide-angle shots where the focus is set in the distance, but the look is more shallow-focus. I don't think it's as simple as that in terms of reducing depth of field. I could be wrong.
 
O.K. Thanks. I was also sceptical about this myself, it's just that arri seems to advertise that exact property (shallow DOF on wide angle) on every page of their site that features the master diopters.
However, could maybe the fact that you loose infinity focus help reduce DOF with these - or do the diopters only work for focussing very close to the lens?

Alternatively is there any way to get very thin focus on wide shots other than using large format aquisition?
I'm seeking a look where you'd have people or objects basically swim into focus only when at a certain distance from the lens (1 meter or so) and almost immeadiately disappear into oblivion once they get just a tiny bit farther from the lens. Basically a sort of hyper myopic look.
Something similar to some photos here
http://www.flickr.com/groups/wide-angle_bokeh/pool/
but even more extreme.
 
Just to chime in, diopters only let you focus closer to the subject; they don't do anything else (except sometimes introduce chromatic aberration). It's just that when you focus closer, the depth of field decreases incredibly fast.

But a close up shot with a diopter should look identical to a close up shot without one; it's just that with the diopter you can focus on something a few inches away (but not on infinity) and without the diopter you can focus on infinity (but not on something so close). If you want that effect and have nikon lenses a really nice alternative is a macro tube, which you can get on ebay for like $20 and which is optically perfect if irritating to use. But diopters do not alter depth of field at a set focus distance, focal length, and stop.

For the effect you're looking for, there are many options but no one "solution:"

•The easiest option is to green screen it and do it in post. Just bring a huge green screen on location and then shoot the scene, shoot plates out of focus, etc. so everything matches. Then use some basic garbage mattes (or roto if the characters cross path) and a good lens blur effect in your compositor. This is harder if there's camera motion.

•The other option is just to rent the fastest lens you can get and then cheat people a little closer to it, maybe. Most of the shots you posted look like 35mm at f2 or something, which isn't too unusual. I believe Panavision rents a 50mm t1.0, which would give you a great shallow focus look that would definitely stand out for daylight photography (just make sure to rent a lot of good NDs). Telephoto lenses can really give you this effect if you're willing to sacrifice the sensation of depth that wide angle lenses give.

•Get a tilt/shift lens. This is the best solution but you'll need to worry about how things line up both in terms of horizontal and vertical distance from the camera and it may take some experimentation to get a location that doesn't look too weird when shot with a tilt/shift lens (parts of the background may be in focus and parts may be out of focus depending on how you line things up so a busy background may look very weird). People use these (and fast-motion photography) for those cool shots that make everything look miniature, or for buildings shot at an angle, or for dual focus shots.

•Get a lensbaby or a bellows, which are extreme versions of the tilt/shift lens. This gives a really intense and subjective look; I think Kaminski uses it in The Diving Bell and the Butterfly if you want to check it out.

•There are mattebox filters that blur everything outside a small circle. Kind of cheesy and old school, but they work. There are split diopters, too, which were popular a few decades ago for putting two planes of action in focus. But those probably won't give the look you want because they have a pretty obvious split down the middle.

This is the type of thing where there's no easy solution so you need to figure out exactly why you want that look and what aspects of it are most important to the story--and then pick something (like one of the options I listed above) after discussing your choices with your DP. If this is something as simple as your 2/3'' future red camera not promising enough shallow focus, I don't think renting a master prime diopter (which is worth way more than the camera) is a good choice--just rent a fast lens (or a tilt/shift lens) and 35mm body for the day or green screen it.

Edit: Arri may very well be advertising that the quality of glass in their diopters allows you to use them wide open (t1.3), which would definitely provide very shallow focus at close distances, whereas with other diopters maybe you'd have to stop down before getting an acceptable image.
 
Last edited:
I once read an American Cinematographer article about a commercial or short film that made sort of a mega-version of a 35mm groundglass adaptor like a RedRock or something, but this used an 8x10 view camera, and the recording camera rephotographed the image off of the focusing screen of the view camera, so the depth of field was very shallow.

But other than large formats or really, really fast lenses, there is not much help in getting a shallow focus wide-angle wide shot unless you opt for efx tricks such as selective blurring in post.
 
Hey David,
Sorry to ask a gear question again, but I think gear questions can be a good window into the creative process.

You are very far along on the cinematographer journey, but are there any tools (lights, light tools, etc) that you have recently discovered or developed a new affinity for?
 
As a shooter of narrative cinema, I tend to use the meat & potatoes of equipment, the stuff that serves the widest range of shooting scenarios. Specialized tools tend to serve specialized shots, which is more the purview of commercials and music videos.

In other words, I expect to be using 650w Tweenies for a very long time in my career....

New tools... well, the obvious one is the new M-X Red One, which I think has pretty much solidified the camera's position as the best digital cine camera out there in the market, until Red (or someone else...) comes up with something better. Being able to work at 800 ASA as a base without any compromise is amazing, and to be able to not break a sweat and shoot at 2000 ASA is fantastic.

Being able today to use a DSLR for tiny spaces is also an interesting new trend, though I can't wait for the Scarlet and Epic to come along in order to avoid the picture quality issues that these DSLR's have.

Otherwise, my favorite new tools are pretty small-scale... I love the Woodylight, which is sort of like a Lowell Rifa but more flexible. I love the new HO (high output) versions of the Rosco LitePads. I love Source-4's, both the tungsten version and the ones adapted to use a Joker HMI.
 
Hi David I watched AVATAR in IMAX again tonight as Alice is out tomorrow so we lose it, and it is still striking to me how much of annoyance that stuttery motion blur is. Is this amplified by the stereoscopic filming do you think. To bring up an older question would it have been better to shoot at 48fps.

How would 48fps change that 'film look' with a 180 shutter? I might go and shoot some tests and see the difference in look. Do you think it will be something that the industry will consider - the move to 48fps? Do you think it would not help the initial image of Digital -whilst it is still going through the motions of proving itself as a excellent tool alongside film.

Shooting at 48fps is a favourable thing to do when a lot of effects are in play to decrease motion blur obviously helping with the post.

Also do you think it would make it easier on peoples eyes watchin Stereo at 48fps?

What is your take on it all?
 
You must log in or register to reply here.
Back
Top