shooting sunsets with prime lens

[Radim Schreiber]

I am using RED one to record 1-2hr long landscape and sky overhead shots (similar to time lapse, but it is recorded in real time). We use prime lens for best quality results.

There are two challenges:

1) When light conditions change rapidly, such as sunset or light peeking through forest on a creek, I have to be very alert and keep adjusting iris manually. Is there another way to do this more effectively, with less human error and participation?

2) Sunset changes exposure over time and it goes out of the lens iris range. Ultimately I would like to record 8hr scene (2k on a red drive), without having have to change ND filters. Is that possible?

Thanks
Radim

 

[Tony Brown]

Why do you want to change the exposure? The whole point is that the exposure changes, let your shot reflect this...... If I'm shooting a sunrise I'll set the exposure for around T5.6 100asa @50th and let the rest just happen. After T5.6 the show's over anyway......

Same for Sunset, maybe shoot at T2,8 to postpone the 'death' a little....

 

[Larry Gebhardt]

I would think that you would want to control the aperture so that the shot last longer than 20 minutes? Shooting sunsets, with a digital sensor instead of film, the latitude gives you very little time between blown out highlights and the noise floor. With film you have more latitude, but still not 2 hours worth when the sun us going down. I think the answer is a motorized control for the aperture...and that is getting into the motion control realm.

Try Tom's Timelapse forum www.timescapes.org I think that there are some people over there experimenting with home grown devices to move the aperture ring during a long shot where the light is changing...even from daylight to night time and stars, etc. Very cool effect if you can make it work.

Larry

 

[Radim Schreiber]

Why do you want to change the exposure? The whole point is that the exposure changes, let your shot reflect this...... If I'm shooting a sunrise I'll set the exposure for around T5.6 100asa @50th and let the rest just happen. After T5.6 the show's over anyway......

Same for Sunset, maybe shoot at T2,8 to postpone the 'death' a little....

This approach works maybe for 10 minutes. Scene gets very dark quickly and very grainy due to underexposure. Not sure why after T5.6 the show should be over(?). Because of DOF or diffraction?

 

[Tony Brown]

This approach works maybe for 10 minutes. Scene gets very dark quickly and very grainy due to underexposure. Not sure why after T5.6 the show should be over(?). Because of DOF or diffraction?

Well I confess I have no idea why you would want to cover a sunset over 8 hours (?) or even 2 as the poster mentioned, thats not really 'shooting a sunset' which is what I responded too, or indeed any prolonged period of time. Light levels change, why not reflect that?

DOF has nothing to do with shooting sunsets, if you are shooting a foreground scene AT sunset thats a completely different thing, if you're shooting a sunset you'd be at short infinity so DOF is not an issue. Shooting a scene at sunset means adjusting your foreground balance to match the background as it drops, one of the most difficult tasks to do well if its to work in an edit. Actually its not that difficult, you just need to be very organised with scrims etc.

"Scenes get dark very quickly" - well not in screen time they dont, 10 minutes would be an eternity. 9 times out of 10 you would undercrank to accelerate the magic moment and the fact that it gets dark is really the point, its supposed to, so why not allow the visual to just do what its supposed to do, get dark. If you chase the exposure, you're merely prolonging the period of normal exposure. The blacks should never get grainy as you shouldn't be trying to dig detail out... they're blacks

Sorry - I dont understand your reference to diffraction

I'm sure it would not be that difficult to rig an iris motor to some form of meter to allow the iris to be controlled by the light level if you so wished, I just dont really see the point. The fact that in 30 years I've never seen one probably reinforces that view.

 

[Dan Hudgins]

Ways to control exposure automaticly...

Ways to control exposure automaticly...

I am using RED one to record 1-2hr long landscape and sky overhead shots (similar to time lapse, but it is recorded in real time).

1) You can used crossed polarizers, one fixed and one that rotates, there can be a DC servo loop to the motor that rotates the one that changes the exposure with a photo cell behind, or you can use the average of the video monitor out to get the level for the servo loop. The main disadvantage of this is that there can be a small hue shift because the crossed filters tend to rase the K value, Schneider filters may work better, you can try them crossed to see if there is a viloet tint. Not sure but if you stack 4 filters, with two moving you might get a wider adjustment range, etc.

2) You can use a capping shutter if the camera is set to a long enough exposure. This can run its exposure time from photocell feedback also, you need a small computer program to run the capping shutter.

3) You can use electronic light valve like used in shutter glasses for 3D, same as #1 but no mechanical parts so it can work faster like for an explosion or something that flashes.

4) Shoot into a variable density silvered mirror disk, you need a microscope lens type snorkel lens or pinhole lens to get close enough, when the mirror disk rotates you can get exposure adjustment.

5) Same as #4, but you shoot through the variable density mirror disk using it as an adjustable ND filter with a pinhole lens. You can make a high quality "pinhole" lens using a microscope objective or eyepiece and relay optics. If you use two mirror or ND disks that are complimentary (reverse gradent) then you can use a larger than pinhole lens to shoot with.

==

Since the REDONE does not have a shutter over the sensor, I would be careful about having the sun focused on the sensor for hours at a time, the dyes in the Bayer filter may fade or burn depending on the iris setting, maybe?

 

[Radim Schreiber]

Well I confess I have no idea why you would want to cover a sunset over 8 hours (?) or even 2 as the poster mentioned, thats not really 'shooting a sunset' which is what I responded too, or indeed any prolonged period of time. Light levels change, why not reflect that?

DOF has nothing to do with shooting sunsets, if you are shooting a foreground scene AT sunset thats a completely different thing, if you're shooting a sunset you'd be at short infinity so DOF is not an issue. Shooting a scene at sunset means adjusting your foreground balance to match the background as it drops, one of the most difficult tasks to do well if its to work in an edit. Actually its not that difficult, you just need to be very organised with scrims etc.

"Scenes get dark very quickly" - well not in screen time they dont, 10 minutes would be an eternity. 9 times out of 10 you would undercrank to accelerate the magic moment and the fact that it gets dark is really the point, its supposed to, so why not allow the visual to just do what its supposed to do, get dark. If you chase the exposure, you're merely prolonging the period of normal exposure. The blacks should never get grainy as you shouldn't be trying to dig detail out... they're blacks

Sorry - I dont understand your reference to diffraction

I'm sure it would not be that difficult to rig an iris motor to some form of meter to allow the iris to be controlled by the light level if you so wished, I just dont really see the point. The fact that in 30 years I've never seen one probably reinforces that view.

Thanks for your response. I will try to shed little bit more light on why I need to accomplish this. We do not shoot commercials, documentaries or films, we shoot fine art - illusions of nature - wall and ceiling murals. LCD screen with a footage is installed in someones home, hotel, hospital etc. so the observer can experience nature inside. For this reason our scenes do not change, our angle of view doesn't change. What chances it the nature itself. That is a reason we are trying to do 16 hr scenes and that is why I would like to get 1 or 2 hr sunset scene at this moment because that is what seems to be possible technologically at 4k with RED.

DOF has nothing to do with sunsets in a distance, but we of course record foreground as well. Imagine scene that you could see from your window in a first floor of your house. So when we open iris, this decreases DOF and becomes potentially a problem.

"Scenes get dark very quickly" - yes they do. I completely understand your point. Form ILLUSION perspective it is completely normal to let the scene get darker as it does in normal life without adjusting exposure. This approach could work for about 20 minutes. There two reasons it doesn't work for longer periods of time:

1) Within 1-2 hours the scene brightnes changes so dramatically, that you would end up with completely black frame at the end if you started slightly overexposed.

2) I have noticed that as scene starts to get darker, RED footage becomes increasingly noisy. I have found that if I "overexpose" the scene, while maintaining enough detail, I can change ISO or exposure in post and have much less noise as a result.

Diffraction - my problem at the moment is continuous adjustment of an iris AND not enough latitude on each lens. When I start with fully closed iris at T22, difraction is a huge issue and frankly I do not want to use this small iris, so I start around T11, this way I loose about ten to twenty minutes of recording.


I'm sure it would not be that difficult to rig an iris motor to some form of meter to allow the iris to be controlled by the light level if you so wished,

Please give me more information and some directions about this system and how could I do it.

Thanks,
Radim

 

[Radim Schreiber]

1) You can used crossed polarizers, one fixed and one that rotates, there can be a DC servo loop to the motor that rotates the one that changes the exposure with a photo cell behind, or you can use the average of the video monitor out to get the level for the servo loop. The main disadvantage of this is that there can be a small hue shift because the crossed filters tend to rase the K value, Schneider filters may work better, you can try them crossed to see if there is a viloet tint. Not sure but if you stack 4 filters, with two moving you might get a wider adjustment range, etc.

2) You can use a capping shutter if the camera is set to a long enough exposure. This can run its exposure time from photocell feedback also, you need a small computer program to run the capping shutter.

3) You can use electronic light valve like used in shutter glasses for 3D, same as #1 but no mechanical parts so it can work faster like for an explosion or something that flashes.

4) Shoot into a variable density silvered mirror disk, you need a microscope lens type snorkel lens or pinhole lens to get close enough, when the mirror disk rotates you can get exposure adjustment.

5) Same as #4, but you shoot through the variable density mirror disk using it as an adjustable ND filter with a pinhole lens. You can make a high quality "pinhole" lens using a microscope objective or eyepiece and relay optics. If you use two mirror or ND disks that are complimentary (reverse gradent) then you can use a larger than pinhole lens to shoot with.

==

Since the REDONE does not have a shutter over the sensor, I would be careful about having the sun focused on the sensor for hours at a time, the dyes in the Bayer filter may fade or burn depending on the iris setting, maybe?

Hi Dan,
This seems to be very useful information to extend life of a sunset scene. To be honest I don't get most of it, but I will do some research and I may ask you some more questions.

 

[Tony Brown]

Please give me more information and some directions about this system and how could I do it.

Hi radim, what you do sounds interesting.

Crossing pola's would lose you 4 stops right off so doubt thats an option

I'm afraid I cant point you to a system but every SLR has aperture control so it must be possible to in some way link a Preston or Arri LCS to light sensor

Call Clairmonts - I hear Danny likes a problem to solve, they have some very clever bits of kit.