Why does a larger chip size equal shallower depth of field [Archive]

Evin Grant

04-25-2007, 01:11 PM

Walter is correct, assuming you match your original composition just moving back and zooming in will not change your DOF. Only the amount of background in your shot. Most shooters however do not match thier original frame, they punch in a little or the also move thier subject farther from the background. This does impact the DOF but it's amazing how many people don't make the connection. I just wen't through this on a shoot with a very experienced Director/DP. We were shooting on the HVX and he said he wanted the background fuzzier, let's move back and punch in. I said that won't help you need to move your subject farther form the background and change the optical relationship. We argued for a minute then I just showed him on the monitor. "Oh" was the response. Some things just get so ingrained it's hard to communicate any other way.

An easy way to think about it is if you move back and punch in your subject and background become father away and this increases your DOF because DOF works in an exponential way as you focus farther away more and more of your image comes into focus.

Stephen Williams

04-25-2007, 01:19 PM

Hi,

Walter is correct.

As a larger sensor/lens combination has more resoloution, and that IMHO is the reason for less DOF. If lenses used with a 2/3 inch sensor had 3 times the resolving power of 35mm lenses then the DOF would be the same.

Stephen

Chris Kenny

04-25-2007, 01:31 PM

Hi,
As a larger sensor/lens combination has more resoloution, and that IMHO is the reason for less DOF. If lenses used with a 2/3 inch sensor had 3 times the resolving power of 35mm lenses then the DOF would be the same.

Not this again. We just went over this a few weeks ago. What you're saying here (which is not what Walter is saying, really) is simply not true. If it were, 35mm footage shown at SD resolutions would appear to have the same DoF as footage from SD 1/3" camcorders. It doesn't. Yes, the fact that the format has much less resolution does make focus appear a bit deeper, but not nearly enough to account for this effect.

Consider also that DoF adaptors for small chip cameras (which produce shallow DoF despite the fact that they reduce resolving power) actually do, in fact, work.

David Mullen ASC

04-25-2007, 01:36 PM

I don't think that's true, Stephen -- the high-end HD lenses like the Digi-Primes already have to be very high-rez and they don't give 2/3" photography the depth of field of 35mm, even if you used an old lens on the 35mm camera.

It's the focal length, f-stop, distance focused, and Circle of Confusion that affect depth of field. 35mm photography uses over twice as long a focal length on average to achieve the same field of view, hence why it has less depth of field typically than 2/3" photography. Making a 2/3" lens any sharper isn't going to close that gap, and besides, we're talking about the real world here -- it would be hard to design a 2/3" lens much sharper than the Zeiss DigiPrimes anyway.

Even my own experience shooting 35mm anamorphic shows that anamorphic has less depth of field than 35mm spherical because you use focal lengths nearly twice as long for the same horizontal field of view. The sharpness of the lens is only a small mitigating factor. You can shoot on a new 25mm Zeiss in Super-35 cropped to 2.35 or an old 50mm anamorphic lens, get the same field of view more or less, and the anamorphic image will exhibit less depth of field.

Trouble with depth of field calculators is that you can adjust the Circle of Confusion figure and get whatever results you want. Practical experience tells you that smaller target areas use shorter focal lengths on average and exhibit increased depth of field characteristics, and the only solution is to shoot at wider apertures to compensate.

I've noticed that it's often 2/3" video shooters who make the strongest claims that the depth of field difference is minimal with 35mm, and 35mm shooters who make the strongest claims that they are radically different -- the truth is somewhere in between as usual. I still think it is about an effective 2-stop difference in practical terms.

LighthouseMEdia

04-25-2007, 01:37 PM

The other factor people don't take into fact is when you move the camera back and zoom in you cut the light entering the lens and then most people will compensate by opening the aperture even further which does in fact inherently affect DOF that combined with the other factors that Evin and Walter both mention lend towards the DOF illusion. Stephen sorry but I'm agreeing with Chris on this one, resolution somply just doesn't factor in on DOF.

Brook Willard

04-25-2007, 01:41 PM

I could've sworn I wrote the FAQ for threads just like this one...

Stephen Williams

04-25-2007, 01:42 PM

I don't think that's true -- the high-end HD lenses like the Digi-Primes already have to be very high-rez and they don't give 2/3" photography the depth of field of 35mm, even if you used an old lens on the 35mm camera.

Hi David,

My point is that Digi Primes on a Viper have rather less DOF than a standard zoom lens on a 16x9 DigiBeta. If you agree with what I am saying how do you account for the difference?

Curious as always,

Stephen

Stephen Williams

04-25-2007, 01:48 PM

Consider also that DoF adaptors for small chip cameras (which produce shallow DoF despite the fact that they reduce resolving power) actually do, in fact, work.

Hi Chris,

When using a SuperSpeeds on a P+S adapter @ T1.4 they show far more DOF than they would on 35mm @ T1.4 due to the lower resolving power of the 'system'.

FWIW on 35mm people often avoid using them wide open as they are not really sharp and have very low contrast until stopped down to T2.5.

Stephen

Stephen Williams

04-25-2007, 01:52 PM

Stephen sorry but I'm agreeing with Chris on this one, resolution somply just doesn't factor in on DOF.

Hi,

Ever wondered why different COC are used? It's due to resoloution, so without a doubt resoloution is a major factor of DOF.

The latest Zeiss master primes have less DOF than Cooke S2's from 1940 due to their higher resoloution.

Stephen

David Mullen ASC

04-25-2007, 01:53 PM

Sharper, contrastier lenses do make the actual point-of-focus more obvious and the fall-off look faster, thus reduce apparent depth of field -- but nowhere at the level needed to compensate for the difference in target size between 2/3" and 35mm. You're exaggerating the importance of the lens sharpness.

You'd have to imagine some theoretical super-sharp lens (many times sharper than current market products) on a 2/3" camera and a super-fuzzy lens on a 35mm camera, which doesn't have anything to do with practical reality.

Besides, as the other post says, those groundglass adaptors for 1/3" cameras make the image softer yet the depth of field is reduced, so clearly the focal length being used matters a lot more than the sharpness of the image being produced.

Zooms also have some affect on depth of field because the physical length of the lens creates some odd issues, the focal plane being so much farther than the front end of the lens in relation to the subject. I don't know exactly what the issue is though but it isn't just due to the sharpness of the prime versus the zoom.

Brook Willard

04-25-2007, 02:00 PM

The number of variables eventually become so large that people will eventually have to stop, take a deep breath, and add "ish" to the end of a statement. There are many, many factors that can have an effect on DOF. The three basics, as everybody here clearly knows, are focal length, subject distance and stop.

There are, of course, the finer points. The points in discussion right now. What sort of imager is being used? Bayer? 3-chip? Film? Are the lenses calibrated for that particular medium? Do the three wavelengths of light converge at the same plane or at slightly different planes, depending on the media? How sharp are the lenses? How well do they perform at a given stop? How physically long are the lenses? What is the condition of the anti-reflective coatings? What is your acceptable COC? What is the resolution of your imager? The list goes on.

While it is crucial to have an understanding of how these things work, there is a point at which we start chasing our own tails. :)

Stephen Williams

04-25-2007, 02:00 PM

Zooms also have some affect on depth of field because the physical length of the lens creates some odd issues, the focal plane being so much farther than the front end of the lens in relation to the subject. I don't know exactly what the issue is though but it isn't just due to the sharpness of the prime versus the zoom.

Hi David,

That one is easy to explain!

DOF is calculated from the front nodal point of a lens. On a zoom that may well be 12 inches in front of the film plane. The image is also larger on the film plane until one backs off the camera 12 inches, then the image size & dof will match at the same F stop.

Any physically thick lens will have front nodal point in front of the assumed point with standard DOF tables. Individual DOF tables for a specific lens will take this into account.

Stephen

Chris Kenny

04-25-2007, 02:08 PM

When using a SuperSpeeds on a P+S adapter @ T1.4 they show far more DOF than they would on 35mm @ T1.4 due to the lower resolving power of the 'system'.

Yes. The effect you describe (lower resolving power leading to deeper DoF) does exist. It just isn't the primary effect leading to shallower DoF on larger formats.

Stephen Williams

04-25-2007, 02:13 PM

Yes. The effect you describe (lower resolving power leading to deeper DoF) does exist. It just isn't the primary effect leading to shallower DoF on larger formats.

Hi Chris,

My theory is that small sensors with low quality optics have huge DOF due to their low resoloution, and not just the size of the sensor.

Stephen

Alex Boothby

04-25-2007, 02:20 PM

Hi Chris,

When using a SuperSpeeds on a P+S adapter @ T1.4 they show far more DOF than they would on 35mm @ T1.4 due to the lower resolving power of the 'system'.

Stephen

I guess I see what your saying - that is an interesting way to look at it. By this rational Citizen Kane posted on the web at 320x240 will have a similar apparent DOF to a Tony Scott movie posted @ 320x240. Likewise a half blind person will not appreciate the shallower DOF of 35mm vs 2/3'. Can't argue with that really, but it has more to do with apparent (or perceived) resolution than classic DOF characteristics.

Chris Kenny

04-25-2007, 02:30 PM

My theory is that small sensors with low quality optics have huge DOF due to their low resoloution, and not just the size of the sensor.

Not just the size of the sensor, true. But mostly the size of the sensor. Or, rather, the fact that you use longer lenses to achieve the same FoV with larger sensors.

Yes, if you use a 50mm lens from 20 ft., you'll have the same nominal DoF as with a 25mm lens from 10 ft. (This is the issue Walter raises. There's more to be said about this, but I won't say it here.)

However, this isn't really relevant to what happens when you change formats. If you switch from 16mm to 35mm, you'll be switching from a 25mm to at 50mm lens at the same distance to maintain your angle of view. A 50mm lens focused at 10 ft. obviously has shallower DoF than a 25mm lens focused at 10 ft.

Stephen Williams

04-25-2007, 02:31 PM

By this rational Citizen Kane posted on the web at 320x240 will have a similar apparent DOF to a Tony Scott movie posted @ 320x240.

Hi,

I would be interested to see a 320 x 240 of both scanned back to film and viewed on a big screen, then we could talk further about it.

Stephen

Stephen Williams

04-25-2007, 02:37 PM

Not just the size of the sensor, true. But mostly the size of the sensor.

Hi Chris,

That's the bit I am thinking too hard about! Just a theory, dosen't mean in practice that it's of any use! LOL

Stephen

Alex Boothby

04-25-2007, 02:41 PM

Hi,

I would be interested to see a 320 x 240 of both scanned back to film and viewed on a big screen, then we could talk further about it.

Stephen

Orson rolls in grave... :blink:

Martin Drew

04-25-2007, 04:56 PM

Hey. This old chestnut again.

I believe Walter is not correct. he points out that 2 object are equally blurred when you blow one up and compare it to the other, but as soon as you magnify the image you are changing the COC. Image magnification is a factor in DOF too. When you use a longer focal length lens you will need to move further away from an object to make it appear the same size in the frame. This changes the perspective so a second object behind the first appears larger than it would with a shorter focal length lens. So in a situation where the second object was just acceptably sharp with the wider lens it will not be acceptably sharp with the longer lens because it has been magnified in the frame. Therefore the DOF is reduced with the longer lens.

The important point to remember is that (assuming theoretical optics) there is only point perfectly in focus, Depth of Field is just the amount of distance in front of and behind that point where the amount of blur is small enough that it cannot be perceived.

M

Roxco

04-25-2007, 04:59 PM

While it is crucial to have an understanding of how these things work, there is a point at which we start chasing our own tails. :)

If you want to learn how to do brain surgery you got to hang out with a few surgeons and learn their lingo. Filmmaking is not too much different and even tail chasing is a good way to listen in on the surgeons prior to making my first cut.

The only thing I can ask is what affect does wave interference have on focus?

Is something out of focus because the waves of light are colliding on the sensor? Or are they just spilling onto multiple pixels?

Rosco

P.S. Hopefully the only dumb questions are those not asked?

Anders Holck

04-25-2007, 05:06 PM

When using a SuperSpeeds on a P+S adapter @ T1.4 they show far more DOF than they would on 35mm @ T1.4 due to the lower resolving power of the 'system'.

One reason for this is that the matte screen on the pro35, mixes in an arial image in addition to the image exposed on the screen. This makes a compromize between a film looking boke and less light loss.

10s

04-25-2007, 06:39 PM

Why does a larger chip size equal shallower depth of field?

I don't know if this helps but, the answer is lengthy so I suggest one goes to pg 166-168 of the ASC Manual, 8th edition, it's referred to as the "Lagrange Invariant," a law of optics.

Yes larger formats, i.e., 65mm do produce shallower focus (DOF) compared to smaller formats, i.e., 35, 16, 2/3, 1/3.... when employing comparable field of view, constant aperture, focal length.

J. Bernard Vallon

04-25-2007, 07:07 PM

My understanding is this:

There is no physical difference between a 50mm f2 lens, and a 100mm f2 lens, other than size. Imagine the following:

on s16 film: a 50mm f2 lens photographing a subject 2 meters away and

on 35mm film: a 100 f2 lens photographing a subject 4 meters away

Double everything, INCLUDING the size and distance of the subject and scene, and the image, INCLUDING the depth of field, will be identical.

In effect, when you use larger formats with larger lenses to achieve the same field of view, you SHRINK the world around the camera, thus operating closer to the 'macro' realm of the camera, where depth of field is exponentially shallower.

This is why longer lenses have 'shallower' depth

Brook Willard

04-25-2007, 08:08 PM

For the original question of the thread: RTFFAQ (http://www.reduser.net/forum/showthread.php?t=1487) :)

Phillip Hudson

04-25-2007, 08:59 PM

More than you ever wanted to know about DoF:
http://en.wikipedia.org/wiki/Depth_of_field

This link is more interesting for DoF though:
http://www.cambridgeincolour.com/tutorials/depth-of-field.htm

Poi Boy

04-25-2007, 09:09 PM

In the end the different theories really don't matter, you light and compose your shot with whatever combination you are working with and give it the depth of field you want, be it shallow or deep. With Red we will have that 35mm dof control we all love ( more light or less light ).
Aloha
-A

Jorge Díaz-Amador

04-26-2007, 06:52 AM

Hey folks, just joined here, first posting.

I have studied Depth of Field extensively. It's no wonder this one issues causes so much confusion. It can be difficult to understand, because it's deceptively simple, but complex in practice.

Essentially, depth of field (DoF) is only affected by only two things: Image magnification and aperture size (iris or f-stop). But you have to keep the following in mind:

Changing focal length = Changing magnification
Changing film aperture or sensor size = Changing aperture (f#)

The reason bigger sensors have less DoF is that they act like a larger aperture (iris). They allow a larger difference in angle between the light rays that are in focus and the light rays that are out of focus. All these rays have to pass through the lens nodal point, but from there they diverge on their way to the image plane. Bigger image plane (sensor) size = more divergence angle.

Also, image plane magnification is higher because for any given field of view, the image magnification on a larger sensor will be greater. Hence longer focal lengths for the same field of view.

Circle of Confusion (CoC) comes into play, but this is a subjective issue. The CoC is just a way to arbitrarily define the point at which an in-focus detail would be blurred enough to seem out of focus. If the detail size is below the CoC limit it would appear to be in focus, even though it is not as well resolved as another image detail that is sitting right at the point of ideal focus.

The more resolution your film or sensor has, the more image details that are in focus will appear to be in focus and "sharp". Therefore, details that are not as well focused as those will appear more blurred by comparison.

So:
Higher resolution = Less SUBJECTIVE depth of field

To follow up on something David Mullen said, I agree that 35mm Anamorphic has less DoF than 35mm spherical (either Academy or Super 35). Along with the longer focal length of the anamorphic lens (which would actually only cause less DoF in the _vertical_ plane, since the _horizontal_ focal length is shorter by 1/2). But the image plane size is larger because you are using a 18.6 x 22 mm aperture size, rather than the 10.2 x 24 mm effective aperture of Super 35mm at 2.4:1 AR. The final aperture (film plane) is twice as large with Anamorphic.

Also, to quickly compare DoF between different formats, I like to use the horizontal image plane size and just compare the proportion (without taking into effect the subjective resolution issue). I use the horizontal rather than the diagonal because the diagonal changes with aspect ratio. But you can use the diagonal if you compare identical AR's.

So:
S35mm ANSI (24.9) to Super 16mm (12.4) = 2.0:1
S35mm ANSI (24.9 mm) to 2/3 HD (9.8 mm) = 2.54:1 (about what Panavision claimed).
S35mm ANSI (24.9 to 1/3 CCD (4.8) = 5.2:1

S35mm @ f2.8 = S16mm @ f1.4 (easily doable)
S35mm @ f2.8 = 2/3 HD @ f1.1 (not possible with 3CCD prism)
S35mm @ f2.8 = 1/3 HD @ f0.54 (clearly impossible)

I hope this helps,

Jorge Diaz-Amador
CinemaTechnic
Miami, FL USA

Martin Drew

04-26-2007, 07:09 AM

An excellent post Jorge, very clearly put, great to see you here.

M

Álex Montoya

04-26-2007, 07:16 AM

Stephen Williams

04-26-2007, 07:25 AM

That may be a dumb question but was wondering why 35 mm SLR lenses haven't been used before with cinema cameras... Is the quality difference so abismal? They shouldn't be since they can perfectly capture 10 Mpixel pics, right?

So, what am I missing?

Hi,

They have been used for motion control, VFX background plates & Stunt cameras for many years.

Not really suitable for narative work IMHO.

Stephen

Johan Lindgren

04-26-2007, 07:31 AM

Chris Kenny

04-26-2007, 09:05 AM

Rehoused 35 mm SLR have been used but not loved. For a number of reasons SLR lenses are not proper for achieving quality cine work. For example the sensor area is different on both systems. Making a 35mm prime for cinematography a "normal" lens where 35mm would be wideangle on SLR. The optical construction differs between the two types. Then there are all the other reasons. Colormatching, handling, breathing etc.

Well, many photo lenses now are designed to work well on digital SLR cameras that have sensors pretty much the same size as Red's. And photo lenses will have the resolving to support 4K... they're used on photo cameras with that sort of resolution all the time.

The issues with photo lenses are:

1) Breathing; that is, small changes in focal length when racking focus. Lens makers try to design this out when making cine lenses, since it can make in-shot focus racking look odd. Photo lens makes mostly don't bother. However, some photo lenses don't breath much.

2) Accuracy of barrel markings for focus. Cine are individually tested and tweaked, and their barrel markings are placed based on how they actually perform. Photo lenses don't go through this process, and barrel markings might be a little off on some units. In these cases, it might be necessary to establish focus by eye and then use marks on a follow focus, or to just figure out how far off the barrel markings are and take that into account when setting focus.

3) Distance between barrel markings. Most photo lenses these days are designed with auto focus cameras in mind. In such a case, you actually want to have very little distance between focus barrel markings, as it leads to faster auto focusing. When you're pulling focus by hand while looking at the barrel markings, in contrast, you clearly want those barrel markings to be as far apart as possible, so you can make more accurate adjustments.

4) With zooms, a lot of photo lenses don't maintain the same f-stop throughout their zoom range, which would be a major hassle on-set. Higher end photo zooms do tend to maintain f-stop, though.

5) A similar issue, with primes. If you want five primes between, say between 18 and 100mm, that all open up to the same stop, I don't think anyone in the photo lens world offers that. Also, as far as I know, nobody in the photo lens world sells prime sets that have been tested and tweaked to match for color, sharpness, etc.

All of that said... anyone who has used a digital SLR camera knows it should be quite possible to produce a good looking image with photo lenses on Red, and anyone looking to save money this way probably isn't doing the kind of shoots where every second counts because you're paying three dozen people union wages to stand around while you mess with your lenses. So, photo lenses are going to be a very good option for some folks.

Álex Montoya

04-26-2007, 09:10 AM

Thanks a lot, Chris. That was very instructive.

David Mullen ASC

04-26-2007, 09:14 AM

A lot of the problem with SLR lenses is mechnical, not optical. If you're pulling focus on a tight, long-len close-up of an actor and you measure that they move from 6' 7" and then lean in to 6' 3"... well, try and find those distances on a typical SLR lens!

It's a similar problem too with ENG lenses versus cine lenses. On an ENG and still camera lens, an operator pulling focus usually wants to go from infinity to minimum focus with one rotation of the wrist. But if the lens is physically small, this means that the distance marks are incredibly close together. With cine lenses, it can take two full rotations of the lens to go from infinity to minimum focus.

Plus some still camera lenses rotate in the opposite direction of cine lenses for near to far, which can annoy focus-pullers.

Also, cine lenses tend to be more physically robust in construction.

Then when you talk about zooms, it gets even more complicated. With a still lens, you don't have to worry about a zoom breathing when you rack-focus, you don't have to worry about the image tracking (staying centered) as you zoom in, and you don't have to worry about the exposure dropping at the extreme end of the zoom -- all because you are only taking a single frame. When you do all of these things in real time on camera with a moving image, then you have to design all of these problems out of the zoom.

Last night I was talking to someone about this topic and he said "the difference between a good still camera lens and a good cine lens happens in the last 10%." In other words, 90% of the time, you're probably not going to see a big difference in quality. But improving the cine lens that last 10% to be better and solve all those problems that a still camera lens has, accounts for the big difference in price. That, and the fact that economics of scale favor any mass-produced lens.

Stephen Williams

04-26-2007, 09:16 AM

Well, many photo lenses now are designed to work well on digital SLR cameras that have sensors pretty much the same size as Red's. And photo lenses will have the resolving to support 4K... they're used on photo cameras with that sort of resolution all the time.

The issues with photo lenses are:

1) Breathing; that is, small changes in focal length when racking focus. Lens makers try to design this out when making cine lenses, since it can make in-shot focus racking look odd. Photo lens makes mostly don't bother. However, some photo lenses don't breath much.

2) Accuracy of barrel markings for focus. Cine are individually tested and tweaked, and their barrel markings are placed based on how they actually perform. Photo lenses don't go through this process, and barrel markings might be a little off on some units. In these cases, it might be necessary to establish focus by eye and then use marks on a follow focus, or to just figure out how far off the barrel markings are and take that into account when setting focus.

3) Distance between barrel markings. Most photo lenses these days are designed with auto focus cameras in mind. In such a case, you actually want to have very little distance between focus barrel markings, as it leads to faster auto focusing. When you're pulling focus by hand while looking at the barrel markings, in contrast, you clearly want those barrel markings to be as far apart as possible, so you can make more accurate adjustments.

4) With zooms, a lot of photo lenses don't maintain the same f-stop throughout their zoom range, which would be a major hassle on-set. Higher end photo zooms do tend to maintain f-stop, though.

5) A similar issue, with primes. If you want five primes between, say between 18 and 100mm, that all open up to the same stop, I don't think anyone in the photo lens world offers that. Also, as far as I know, nobody in the photo lens world sells prime sets that have been tested and tweaked to match for color, sharpness, etc.

All of that said... anyone who has used a digital SLR camera knows it should be quite possible to produce a good looking image with photo lenses on Red, and anyone looking to save money this way probably isn't doing the kind of shoots where every second counts because you're paying three dozen people union wages to stand around while you mess with your lenses. So, photo lenses are going to be a very good option for some folks.

Hi Chris,

Good post, I remember you were thinking of buying some cine primes, did you do so, or are you still going the still lens route?

Stephen

Alex Boothby

04-26-2007, 09:20 AM

Hey folks, just joined here, first posting.
Jorge Diaz-Amador
CinemaTechnic
Miami, FL USA

Nice to see you on this list Jorge - you do great work. I hope to buy some lenses from you shortly.

Welcome
Alex Boothby

Roxco

04-26-2007, 09:25 AM

I just want to keep this thread going with a good question...

Rodney Charters recently said that the reason they did not choose to use a Digital Film Camera for 24 was the delay in feedback to the monitor on systems he looked at back then.

Are they any concerns for focus pulling with the RED based on his comments or will we still be dependent on distance markings?

How will this be different for Steadicam Operations?

Thanks in advance for the brain surgery, ;)

Rosco

Jorge Díaz-Amador

04-26-2007, 10:01 AM

Martin and Alex thanks for the welcome, it's greatly appreciated!

There has been quite an intense debate here about Cine vs. Stills lenses here, and I'm not going to get into that just yet, since I have not read the previous posts. But I do have to say something, although it may burst some bubbles.

After having seen "Crossing the Line" at NAB, I can say that RED is going to be one of the most demanding cameras in terms of lens quality. The resolution of the camera is so high that it will reveal lens aberrations that would not have been noticed on, for example, a 2K camera.

Based on the data I have, the RED Mysterium sensor has a pixel pitch of 184 pixels per millimeter. If there is no optical low-pass filter, the sensor's MTF response may extend all the way to it's theoretical max of 93 line pairs per mm (lp/mm).

ARRI did some 4K film scans of resolution targets that were shown at the User Group meetings at NAB. They determined that the current color negative filmstocks have a resolution of just above 80 lp/mm, but not quite 100 lp/mm (there was no 90 lp/mm target on their test chart.

So, the RED is perhaps the first digital cine camera that has resolution equal to the best that color negative film can deliver (again assuming no low-pass filter). [The Panavision Genesis has a 12.4 M Pixel sensor but it does not output 4K AFAIK].

This little camera is going to be very tough on lenses. I plan to work testing to determine which lenses will give the best performance on the RED.

Jorge Diaz-Amador
CinemaTechnic, Inc.

Martin Drew

04-26-2007, 10:05 AM

There definitely is an OLPF so a few bubbles probably remain unburst.

M

Chris Kenny

04-26-2007, 10:36 AM

Hi Chris,

Good post, I remember you were thinking of buying some cine primes, did you do so, or are you still going the still lens route?

I mentioned during the photo vs. cine lenses discussion over on Cinematography.com a couple of months back that I thought a lot of the vastly higher cost of cine glass was due to low volumes, and making design decisions that pushed quality well past the point of diminishing returns for the budget conscious, and that it would probably be possible to build pretty respectable cine glass that only cost a few times as much as photo glass (rather than a few dozen times as much), if there was a large enough market.

It's now pretty clear that Red is going to create that larger market (if things pan out according to their tentative schedule they're going to be making 1000 cameras per month by next year), and their lens pricing seems to be along the lines that I imagined would be possible. So, we'll probably be buying Red's glass.

David Mullen ASC

04-26-2007, 11:06 AM

Are they any concerns for focus pulling with the RED based on his comments or will we still be dependent on distance markings?

Focusing by taking measurements will always be a part of 35mm-style focus-pulling. Trouble with focusing by watching a big HD monitor is that you are reacting to something that happens after-the-fact (an actor moves) whereas half the time with focus-pulling, you are anticipating where they might move to, partially based on measurements you took on the set after a rehearsal. Or just based on object in the room -- for example, an actor crosses to a table that you know is "x" number of feet and inches from the lens. You don't want to be "chasing" a moving actor around too much by watching the image on a monitor and focusing by eye.

mike70

04-26-2007, 12:16 PM

For those of us who are not on the same technical plane as most of the gifted folks who posted in this thread and who are still a bit dazed and (circle of ) confused, would it be possible to get a consensus in terms of the practicalities of DOF with red :

1. If shooting 2k with the glass that comes with the camera, is there greater depth of field (presuming, as per Brook's helpful faq, that we are talking framing the same shot, i.e., the same field of view)?

2. If shooting 4k with a 2/3 or s16 lens, same fov, is there greater DOF?

3. If there is more DOF, is there any way to say roughly how much more for either of the above, like David's "2 stops". (Not sure what Jorge's 2.0 & 2.5 proportions were referring to).

Brook Willard

04-26-2007, 12:25 PM

No glass comes with the camera. If you shoot 2K through a S35mm lens, no, you would get S16mm "DOF" characteristics. As described in the FAQ.

You cannot shoot 4K with either a 2/3 or S16mm lens. As described in my FAQ, you would be limited to their respective "DOF" characteristics.

Look for DOF charts online or in the ASC manual to find equivalent DOF between different formats at the same FOV.

Stephen Williams

04-26-2007, 12:41 PM

Look for DOF charts online or in the ASC manual to find equivalent DOF between different formats at the same FOV.

Hi Brook,

With respect that's quite complicated & I am not sure I really understand what you are saying.

Stephen

Brook Willard

04-26-2007, 12:55 PM

My apologies, allow me to elaborate.

Let's assume a 50mm on S35mm has approximately the same FOV as a 25mm lens on S16mm.

A 50mm lens @ f/4 with a subject distance of 10 feet has acceptable focus from 8'11" to 11'5".
A 25mm lens @ f/1.4 with a subject distance of 10 feet has acceptable focus from 8'6" to 12'1".

While it is not exact, the DOF and FOV characteristics of these two shots would be very similar.

Another example:

S35mm - 100mm @ f/4 @ 10' = 9'9"-10'4"
S16mm - 50mm @ f/1.4 @ 10' = 9'7"-10'5"

Again, approximately the same DOF and FOV between these setups.

So S35mm DOF could be "emulated" to an extent with S16mm assuming the proper lensing decisions. It won't be exact and it'll require fast, sharp lenses for S16mm... but it'll look more "right" to the average audience. Approximation is key here.

mike70

04-26-2007, 01:05 PM

I guess that's what confused me--Here you cite examples of the DOF's being almost the same, whereas in the FAQ the disparity is quite large:

85mm lens @ F: 2.8 @ 10' subject distance has a DOF of 7" (S35mm)
40mm lens @ F: 2.8 @ 10' subject distance has a DOF of 2' 9" (S16mm)

Brook Willard

04-26-2007, 01:06 PM

Correct. That's because here I'm exposing the S35mm camera at f/4 and the S16mm camera at f/1.4

mike70

04-26-2007, 01:21 PM

So, keeping fov and f-stop the same, does that mean that you do get greater DOF shooting 1. with the 2/3 or s16 lenses or 2. shooting at 2k with the 18-50 zoom that is offerred with the camera?

Brook Willard

04-26-2007, 01:25 PM

Well, as described in the FAQ, you would need different lenses to maintain the same FOV between formats. Accordingly, as the focal length would change... the DOF would change.

But yes, if you shot 2K with the same lens, you're basically using the S16mm area instead of the S35mm area. This is described in detail with diagrams in my FAQ.

The 18-50 is not offered with the camera, it's a separate product.

Jorge Díaz-Amador

04-26-2007, 01:58 PM

S35mm - 100mm @ f/4 @ 10' = 9'9"-10'4"
S16mm - 50mm @ f/1.4 @ 10' = 9'7"-10'5"

So S35mm DOF could be "emulated" to an extent with S16mm assuming the proper lensing decisions. It won't be exact and it'll require fast, sharp lenses for S16mm... but it'll look more "right" to the average audience. Approximation is key here.

Brook,

You're using a 2.8:1 ratio here. Did you calculate this using the same CoC for both formats? Usually a smaller CoC is used for smaller formats since the higher magnification (on display) reduces the apparent DoF.

I do think that the 2:1 ratio S35=S16 is accurate. I have shot S16 films using that rule and the DoF characteristics of shots done that way do have much more of a 35mm feel. They just have more of a texture from the grain and do not have the ultimate sharpness of 35mm, but the three-dimensional look is there.

Ideally we would want to use a CoC for S16mm that is half the size of that used for S35mm but in practice it does not work out. That's because the film negative is limited to the same resolution per mm in both formats, and also because the lenses need to work "harder" in S16mm. A lens that resolves 100 lp/mm will look as good in S35mm as a lens that resolves 200 lp/mm in S16mm.

-JDA/CT

Brook Willard

04-26-2007, 02:03 PM

I based those conversions off of the charts in my ASC manual. COC is .0010 for both.

Tim Young

04-26-2007, 02:10 PM

i just wanted to post a link to a quicktime tutorial that my brother-in-law made that discusses the factors that affect DOF. he talks about aperture, focal length, subject-camera distance, and i'm pretty sure he talks about target size. i think it's really helpful. he used after effects to make some motion graphics that i think really make the concepts make sense visually. anyway, i thought someone would find it useful:

http://snodart.com/tutorials.php

casey warren

04-26-2007, 03:11 PM

Tim Young

04-26-2007, 03:37 PM

thanks. i helped with it, but it was mostly my brother-in-law's work. he was planning on making a dvd about how to make a DIY 35mm adapter but decided not to finish it, so he posted what he had done so far. i'll pass on to him the wikipedia suggestion.

snodart

04-27-2007, 12:22 AM

I posted this on DVXuser some time back. It answers the original question:

http://www.dvxuser.com/V6/showthread.php?t=71373

and as Tim mentioned, the video tutorial might be helpful as well. (The FAQ thread here is great, but maybe hearing it presented differently will help.)

There are many factors that will determine how the finite gradations of focus will be perceived in the already captured image: the eyes that are viewing the image, the medium on which the image is being viewed (projection screen, computer monitor, TV, photo paper, etc) the medium used to capture the image (ie: film or ccd), the distance that the viewer is from the image, the size of the image being viewed, etc.

There are so many variables... and the resulting differences in apparent DOF are slight enough that I'm not sure it is worth the brain power. The bottom line is that aperture size, focal length, and subject-to-camera distance are the primary means by which DOF is altered.