Difference in DOF between 1/3" 1/2" 2/3" 35mm and Mysterium sensors?

[Ursa]

Hi!

I hope the topic says the most: I'm wondering what the difference in DOF will be changing from one system to the other. I'm currently on the HVX200, and know that the 1/3" is not my cup of tea when it comes to working on fictional stuff.

I think I read somewhere that the Sony EX1 (1/2" sensor) at the same aperture will have 1 stop shallower depth of field than a 1/3" sensor camera. This sounds reasonable.

But how great is the difference from a 1/3" sensor to a 2/3" sensor? And then from 2/3" to the Mysterium (S35)? And then from Mysterium to "normal" 1:1,85 35mm?

1/3" is roughly DOF-equivalent to 8mm film?
2/3" is roughly DOF-equivalent to 16mm film?

And the most important of all: will there be a significant change in DOF going from 2/3" cameras to the Mysterium?

sorry for typos and/or bad language:weight_lift:

[Alexander Christ]

Hi Ursa,

have a look at Brook's RED ONE FAQ to DoF + FoV: http://www.reduser.net/forum/showpost.php?p=60414&postcount=12

[Prem Edpuganti]

Ursa:
Please go to this link and download the charts. You have the answers you are looking for.

http://www.cookeoptics.com/cooke.nsf/product/dfc

[Ursa]

OMG I just wrote the longest, most complete and most data-intensive DOF post ever!

Then I heard a "click" from my mouse.... PANIC! too late... I lost it all in the accidental "back button incident". Lost 1 hours of work. I am so very, very.... not happy.

So the rest of the post will be very short and blunt. I'm sorry for that.

look to this page for info on film sizes: http://gcmstudio.com/filmspecs/filmspecs.html

and here:
http://www.dvinfo.net/conf/archive/index.php/t-6348.html

at least I remember that the Mysterion had a crop factor of 1.54x compared to 35mm still (27,98mm diagonal vs 43,3mm) (which adapter lads like me use as reference). In that group we also roughly have S35 and Nikon DSLR (except D3).

super16 had 14,55mm diagonal size, compared to 11,0mm diagonal size on 2/3" cameras. I guess that means the DOF impression will be noticeable, but not by the common watcher (1.32x difference).

Mysterion sensor has a diagonal of 27.98mm

Yeah and I know that the DOF terms are full of confusion and different ways of measuring. I'm just after a DOF impression, the same impression the watchers get when seeing a 35mm film against flat HD video.

So I guess what I asked in the first place: "will there be a noticeable difference between 2/3" cameras and RED?" -is still unanswered.

But if (this is only wild calculations) doubling/squaring the size of the sensor is roughly equivalent to 1 stop difference in DOF (as it "doubles/squares" field of view from a given focal length), then the chart should look like this:

1/3" - 6,00mm diagonal - stops difference from RED: 3,66
2/3" - 11,00mm diagonal - stops difference from RED: 1.54
S16 - 14,55mm diagonal - stops difference from RED: 0.92
RED - 27,98mm diagonal -

Hmmm so being at F2.0 on the HVX (1/3") is roughly the same as F8.0 on the RED? hmmm doesn't sound INSANELY unlikely, however I actually would've thought the difference should be greater. Feel free to complement any own data on that field.

Again: sorry for bluntness and very likely wrong maths

Hope you find this as interesting as I do :) giving the impression of a "higher gauge shoota-instrument" seems to do wonders for the production value of the finished product.

[chuck colburn]

http://www.vanwalree.com/optics/dof.html

[Tom]

OMG I just wrote the longest, most complete and most data-intensive DOF post ever!

Then I heard a "click" from my mouse.... PANIC! too late... I lost it all in the accidental "back button incident". Lost 1 hours of work. I am so very, very.... not happy.

LOL, welcome to the Internets. :)

[Zack Birlew]

Happens to the best of'em, posts I mean.:weight_lift:

[Stephen Williams]

OMG

1/3" - 6,00mm diagonal - stops difference from RED: 3,66
2/3" - 11,00mm diagonal - stops difference from RED: 1.54
S16 - 14,55mm diagonal - stops difference from RED: 0.92
RED - 27,98mm diagonal -

Hmmm so being at F2.0 on the HVX (1/3") is roughly the same as F8.0 on the RED? hmmm doesn't sound INSANELY unlikely, however I actually would've thought the difference should be greater. Feel free to complement any own data on that field.



Hi,

Normally I assume 2.5 stops from 2/3inch to 35mm.

Stephen

[Eki Halkka]

Hmmm so being at F2.0 on the HVX (1/3") is roughly the same as F8.0 on the RED?

Yep.



Sensor Lens Aperture
1/3" 12.5 mm f2
1/2" 18.5 mm f2.8
2/3" 25 mm f4
S35 50 mm f8
SLR35 100 mm f16

All the above combinations produce visually similar images. Maybe not exactly the same, but very, very close.

Horizontal field of view is roughly 20 degrees, and you get the same framing and depth of field characteristics.



http://eki.pp.fi/temp/Eki/HV20/35mm_adapter/footage/F-stops/CompareDOF.jpg

The above is from my DIY SLR 35mm adapter tests - the DOF is rather similar. The adapter shot naturally suffers badly from loss of light at f16.

[David Mullen ASC]

Actually it should be 4.5 stops worth of difference between 1/3" and S35. It's 2.5 stops between 2/3" and S35, so another 2 stops difference between 2/3" and 1/3". The equivalent field of view of a 50mm lens in Super-35 is not a 25mm lens in 2/3" photography. It would be closer to a 20mm lens. It would only be 25mm if a 2/3" sensor was half the width of a Super-35 sensor.

For example, the info I have says that a 2/3" sensor is 9.58mm wide in a Sony camera, versus the RED sensor, which is 24.4mm wide, a difference of 2.55X. Of course, that's ignoring the actual image size extracted from the sensors.

[David Mullen ASC]

Actually it should be 4.5 stops worth of difference between 1/3" and S35. It's 2.5 stops between 2/3" and S35, so another 2 stops difference between 2/3" and 1/3". The equivalent field of view of a 50mm lens in Super-35 is not a 25mm lens in 2/3" photography. It would be closer to a 20mm lens. It would only be 25mm if a 2/3" sensor was half the width of a Super-35 sensor.

For example, the info I have says that a 2/3" sensor is 9.58mm wide in a Sony camera, versus the RED sensor, which is 24.4mm wide, a difference of 2.55X. Of course, that's ignoring the actual image size extracted from the sensors.

[Julio Quintana]

I know that this is a technical thread but I just wanted to address the practical application of this information.

From just looking at the math, it looks like perhaps it isn't that difficult to imitate S35 DOF with a 2/3" camera. "I can just open up the lens a couple of stops and it looks like 35mm". But on a practical level, it isn't quite that simple.

Typically, if you have a small, high-res chip (f23, Viper, etc.), you need a very good, sharp lens to get the most out of it. So a nice digiprime that can open up to a f2 will be very expensive, and thus, for many, unattainable. By contrast, I can throw my $300 Nikon (50mm f1.4) on the front of my Red, and although the lens has its imperfections, it is not as detrimental to the image as it would be on a small chip (of course it's not fair to compare an expensive digiprime to a still lens, but we're talking about how to achieve a shallow depth of field).

This is my understanding of the situation. Please correct me if I'm mistaken.

Julio

[Jaime Vallés]

http://www.vanwalree.com/optics/dof.html
This is a great link. Thanks for sharing!

[Stephen Williams]

I know that this is a technical thread but I just wanted to address the practical application of this information.

From just looking at the math, it looks like perhaps it isn't that difficult to imitate S35 DOF with a 2/3" camera. "I can just open up the lens a couple of stops and it looks like 35mm". But on a practical level, it isn't quite that simple.

Typically, if you have a small, high-res chip (f23, Viper, etc.), you need a very good, sharp lens to get the most out of it. So a nice digiprime that can open up to a f2 will be very expensive, and thus, for many, unattainable. By contrast, I can throw my $300 Nikon (50mm f1.4) on the front of my Red, and although the lens has its imperfections, it is not as detrimental to the image as it would be on a small chip (of course it's not fair to compare an expensive digiprime to a still lens, but we're talking about how to achieve a shallow depth of field).

This is my understanding of the situation. Please correct me if I'm mistaken.

Julio

Hi,

You understand well, wide open with a Viper or F23 with a digi prime needs a good focus puller. I have known a focus puller need over 100 takes to get focus at T2.8, I have done similar shots on 35mm at T5.6 without issue!

Stephen

Stephen

[hunterrichards]

Hi,
I have known a focus puller need over 100 takes to get focus at T2.8...
Stephen


What take did he get fired on?:)

[Eki Halkka]

The equivalent field of view of a 50mm lens in Super-35 is not a 25mm lens in 2/3" photography.

I used a highly unscientific method to check my numbers...

First i set up a virtual camera in LightWave 3D, set to "Super 35 1.78:1 3 perf" format, 50mm lens, f8.

I then changed the camera to the other formats and let the software do the math for me. I rounded up numbers a little for clarity (which i should have mentioned i guess). The software i.e. gave 24.6 mm for 2/3" lens with the same field of view, which i rounded up to 25 mm.

I also compared the results visually by rendering a test scene with each of those settings - they were quite similar, as expected.

The difference in our numbers might be i.e. due to the fact that i didn't change the aspect ratio when changing the camera - added to my rounded numbers that could explain the half stop difference (which is my error i'm sure).

Edit: Changing the image aspect ratio didn't affect the calculated lens lengths in LightWave. Interesting - if my numbers are wrong, it might be a bug in the software? Or maybe the proper way to compare imagers isn't to do it by width, but diagonally...

Anyway, for practical DOF characteristic "rule of thumb" 2 stops between each format is pretty close...

[Dan Hudgins]

I used a highly unscientific method to check my numbers...

First i set up a virtual camera in LightWave 3D, set to "Super 35 1.78:1 3 perf" format, 50mm lens, f8.

I then changed the camera to the other formats and let the software do the math for me. I rounded up numbers a little for clarity (which i should have mentioned i guess). The software i.e. gave 24.6 mm for 2/3" lens with the same field of view, which i rounded up to 25 mm.

I also compared the results visually by rendering a test scene with each of those settings - they were quite similar, as expected.

The difference in our numbers might be i.e. due to the fact that i didn't change the aspect ratio when changing the camera - added to my rounded numbers that could explain the half stop difference (which is my error i'm sure).

Edit: Changing the image aspect ratio didn't affect the calculated lens lengths in LightWave. Interesting - if my numbers are wrong, it might be a bug in the software? Or maybe the proper way to compare imagers isn't to do it by width, but diagonally...

Anyway, for practical DOF characteristic "rule of thumb" 2 stops between each format is pretty close...

The subject distance needs to be taken into account, in ratio to the focal length, i.e. if you use a 12.5mm lens vs. a 50mm lens and the subject is 2000mm away from the "focal plane", then there is a 37.5mm difference in the lens to subject distance, so there are small errors unless you take into account the ratio of focal length to subject distance, this ratio becomes more of an issue as the subject gets closer to the focal plane.

(The focal length also gets longer as the subject gets closer, making the DOF less deep. More so with large format lenses do to the focal length to subject distance ratio.)