Welcome to our community

Be a part of something great, join today!

  • 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 :)

Question: capture lightning with Scarlet / Epic

Patrick Grossien

Well-known member
Joined
Nov 20, 2008
Messages
2,987
Reaction score
0
Points
36
Location
Germany
Hi everyone, yesterday night I tried to capture the lightning of a thunderstorm (in total darkness) to see if it's possible. As for photography I'd just do a long exposure and be done with it. :) But I've never tried, to capture it with motion.

So as I was trying to keep it @5K with my Scarlet, I was stuck with 12fps. I tried different shutter speeds but - as I expected - the rolling shutter kept messing up most of my footage.
Some of the images - depending on the setting - came out as horizontal light streaks/ flares only.

Sadly it was all over much too quick because I wanted to try out different fps settings too, to see if that would have helped.
I can't wrap my head around it. So is there a way to approach that properly? Any storm chasers who can help me out here?

Thanks in advance!

Patrick
 
I'm interested in what others have to say as well. I was going to try this by setting pre-record to 10 seconds and running the camera at 5k 96fps at 1/192 to test. I think pre-record is awesome for this, just need to find out what's the optimal fps / shutter.
 
Good to hear I'm not alone in this :)
 
Good questions, Patrick! I would also like to get some tips.
There are thunderstorms coming tomorrow in the Alps.
 
Just tried another one @48fps. With a 1/96 shutter I get banding.

It was over too quick... Didn't capture anything with a faster shutter. No good lightning in front of my lens...
 
Try it with frame processing to "Frame Averaging" (on a tripod, not in movement).
Just a thought, I didn´t try it.
 
Also fully open shutter (360 degree) could help a lot capture all the light... I think adding frames in post or in camera can also be a good tip.

slow fps and pre-record will also be your friend (6 fps at 360 degree will start making a long exposure... Just be sure to blackshade for this setting).

Have fun and post any good results here :-)
 
Thanks guys for the great input. Now to find more lightning :)

Pre-record really is great for this. That's proven to be very helpful.

As for averaging I have to wait for the next storm - or go search one. Maybe Ryan can try it out! ;)

I shot 5k, 12fps and a 1/12 shutter which worked better at the time because of the darkness, but it still created banding. So 360° doesn't help to avoid the shutter problems but created the best images with low light.
Want to try the frame processing next.

And +1 for the motion mount. That's the first thing I was wishing I would have had in my tool kit! I'm 100% certain that would make it possible. Oh well.... someday :)
But for now I'm trying to find a way to make it work even without.
 
I set my scarlet to 4k 1fps at 360 shutter and got these: (Click here for full size)
All of these strikes averaged about 2.5 miles away from the camera
Wgr7xl.jpg

XnQIYl.jpg

gD21cl.jpg

qi9V4l.jpg

CGpApl.jpg
 
Todd, these are nice! Are the frames right before and after the ones you posted clean, too?

I'm thinking that maybe the extreme darkness ruined mine.
 
Correction, I shot those at 1FPS 360 shutter.

There were several rolling shutter artifacts in many of the strike I recorded during this storm, and a couple of strikes lasted up to three frames.

I would not try to capture this at night; Like trying to film the moon, the exposure needs to be set as if you are shooting in sunlight.

As the name implies, there is quite a bit of light in lightning ;-)

There is a lot of trial and error in involved.
 
As for averaging I have to wait for the next storm - or go search one. Maybe Ryan can try it out! ;)

I have access to one of the top floors (54th) in the tallest office building in our city so I'll try to head up there next time a storm rolls in at night and play around with some of the suggested settings here.
 
Yeah, if you're trying to capture lightning with FAST shutterspeeds, you're going to have issues with both ambient light, and banding/frame splicing, like you saw.

My uncle, about 15 years back, took some photos of lightning in Italy. Got 2 out of 3 to work. Really cool images. Won some contests with them. The thing that no one realized, is that of the 3, he took a 40 second, a 45 second, and a 50 second exposure. Lightning flashed during two of those exposures. Most people (non-photographers) assume he was just really lucky and snapped a shot at the exact moment that lightning struck. He was still "lucky" to be in the right place during a lightning storm, and even at that, "lucky" that lightning actually flashed in his frame during the exposure times. But aside from those bits of luck, knowing how he did it takes it from a 1 in a million shot to a 1 in 10 shot (or in his case, a 2 in 3 shot). Still looks pretty damn great, though.

You're running into 2 issues, and thankfully, they have the same solution. The first is that the ambient scene is way, way darker than the lightning. The second is that the lightning is so quick that it can get seen by only part of a frame thanks to the way CMOS sensors work (lack of global shutter). The answer is a longer exposure time. If 1FP isn't fast enough, you can go into the timelapse settings and use either Frame Summing or Frame Averaging. For this, Averaging is probably better, as it has been mentioned above. The added advantage of getting longer than 1fps exposures thanks to frame averaging, is that you'll have more leeway to stop down your iris, and use a lower ISO in post.
 
Thank you Nick for your insight. And I will definitely try it out next time I can shoot a thunderstorm.
For a single image I think we can make it work with slow shutter and averaging. I mainly tried to see if there's a way around the rolling shutter issues when recording motion. Right now I believe there's only the upcoming motion mount which could solve this. But is there really no other way for recording lightning with motion?
 
...But is there really no other way for recording lightning with motion?

Of course there is...

[video]http://en.wikipedia.org/wiki/File:Lightning_Trumbull_County_Ohio.ogv[/video]

You just need to use higher frame rates...
 
Ok, thank you Todd. Do you have any hints as with what frame rate this starts to make sense? Probably more than whats possible with scarlet !?

Thank you for helping me out here :) I know it's trial and error. But it's not always easy to replicate the situation, so I'm trying to be prepared as good as possible. ;)
 
Hi All,

That slow mo is really cool. I too would like to know if a Scarlet is capable of capturing lightning in motion. It does not have to be at 6,000 plus fps. I'd be happy in the 23.976-29.97 range. As I posted in the scarlet section I was shooting heat lightning about a week ago and despite testing different shutter angles I could not eliminate the banding issues in camera.

My end solution was to take the footage and post process it by hand removing bad frames and replacing them with good individual flash frames adjusting the levels to match the defective footage. I then keyed out the lightning itself with a luma key and added it back applying an edge glow to the bolts with a series of filter blurred layers. The background levels were augmented by using a clean flash frame and adding it on top of the footage to fix the general ambient level that was almost totally black during non flash segments. All this done in AE. Not exactly what I had hoped to be doing after I finished shooting. Would definitely love to know if an alternative solution is available for in camera capture.

Bill

PS... Thanks again Ryan for the redirect.
 
Hi All,

That slow mo is really cool. I too would like to know if a Scarlet is capable of capturing lightning in motion. It does not have to be at 6,000 plus fps. I'd be happy in the 23.976-29.97 range. As I posted in the scarlet section I was shooting heat lightning about a week ago and despite testing different shutter angles I could not eliminate the banding issues in camera.

My end solution was to take the footage and post process it by hand removing bad frames and replacing them with good individual flash frames adjusting the levels to match the defective footage. I then keyed out the lightning itself with a luma key and added it back applying an edge glow to the bolts with a series of filter blurred layers. The background levels were augmented by using a clean flash frame and adding it on top of the footage to fix the general ambient level that was almost totally black during non flash segments. All this done in AE. Not exactly what I had hoped to be doing after I finished shooting. Would definitely love to know if an alternative solution is available for in camera capture.

Bill

PS... Thanks again Ryan for the redirect.

The issue is just the nature of how CMOS sensors work. The sensor is organized into lines horizontal of pixels. Over the duration of one frame, light is continually hitting each pixel of the sensor. At some point, the camera goes "Ok, that's the length of time I wanted to sample. Time to take a reading." and then it reads out the voltages on the individual photovoltaic cells (i.e. pixels) one line at a time, starting at the top and working it's way down.

Let's just throw some imaginary non-science based numbers at this, just so you can get the concept. Let's say it takes 5 milliseconds for the camera to sample and record from the top of the frame to the bottom. Now let's say that a Lightning flash is exactly 5ms. If the lightning flashes at the exact precise nanosecond that the camera starts recording, you'll get the whole flash in frame. If the camera is 1/4 through sampling when the lightning flashes, you'll get lightning in the bottom 3/4 of the frame. If the lightning flashes 2ms before sampling starts, you'll get lightning in the top 3/5 of the frame.

Or, if sampling takes 5ms and lightning flashes for 1ms, then the lightning would have to flash right as sampling started, or just before, in order to appear in the whole frame. Too much before, and it'll appear in just the top part of the frame. A little after, and it'll appear in just the lower part. If it's too fast, it could appear in just the middle of a frame, which means that the event you're trying to capture is just too quick.

So, if a flash can be too quick, why would a longer exposure be the definitive answer? Well, if you're taking very quick readings, the sensor is sort of "resetting" it's readings at a fast interval. If, instead, you have it look at a larger chunk of time, it's letting those photovoltaic cells gather more photons, which translates to more voltage before a reading and reset. Which translates to "Hey, as long as this lightning flash is contained within this 1 second (or 15 seconds) then I'm good. I've got it. I know that the rest of my scene is exposed properly for this exposure time per frame, so if the lightning hits, I'm good." You *could* also grab lightning with a shorter shutterspeed and faster framerate, like you tried. The problem is, nature doesn't like to sync up with your camera in terms of the precise moment of lightning in relation to your camera taking a sample. Plus, this is likely to make the rest of your scene so stupidly dark as to make it unusable anyway.

By the way, you don't have to wait for a lightning storm to play around with this. I had a shoot once with camera flashes going off (it was supposed to look like a faux fashion shoot) and the flashes never blew out a single frame in the way intended. They just blew out 2/3 or 1/2 or 1/4 of a frame, because the flash was never synced with the sampling of a single frame.
 
Back
Top