• Stormtrack's forum runs on Xenforo forum software, which will be undergoing a major update the evening of Wednesday, Feb 28th. The site may be down for a period while that update takes place.

Stereo photography of storms

I received this e-mail regarding stereo photography of tornadoes from Chuck with the NWS and I thought I'd post it (after his permission):
It is an interesting proposal. There maybe an easier way with a stereo camera (if those are still available). I'd love some 3D tornado images.

Bill Hark

As for viewing a tornado from different angles, I thought a nice project would be to try to get stereo pix of one in action. That would take a bit of coordination but would be the last frontier in tornado photography, since most of the ultracloseups have already been done. I've read that the best distance for normal stereo viewing of objects is about 10 feet (120"). Since eyes are about 3" apart, that is a 40:1 ratio. If a tornado is 3 miles away (~16K feet), the ideal separation perpendicular to the line of sight would be about 400 feet (just under a quarter mile). You can either be in radio communication with a partner in a separate vehicle and coordinate that way or agree on zoom setting and time (synchronize watches or camera clocks) and snap on the quarter minute or whatever. Just keep the horizon near the bottom of the frame and it should work unless there is really rapid change (close in debris, etc.). Stereo movies would be even better, but I think maybe that should be left to National Geographic. ;-)

This would be completely cool. If you could get it done. I think it is do-able, but you would have to be ultra-coordinated to get it accomplished. My thoughts - pick a day with the right dynamics so that the storms won't be moving quickly ... back-building supercells that sit and spin in the same spot would be ideal so that both cameramen won't have to continually coordinate movements to re-position throughout the duration of the event. The fact that we end up with multiple views of an event these days isn't due to coordinated efforts on the part of the chasers for the most part, but just comes down to who manages to get into position, which totally comes down to chance. I don't know - interesting to think about. Another idea would be to somehow figure out how to work with GPS coordinates to keep track of each other's position somehow.
A while ago there was a discussion about doing this with a lightning strike, namely a strike to a TV tower where the lightning would be predictable enough to distribute a team of photographers. After bringing the shots together, you could do things like stereoscopic imaging and 3-D 'Matrix' style animations of the bolt.

I remember several years ago two monsoon chasers in AZ coincidentally got the same bolt on film from different angles. Susan, wasn't that you and Brian Mayeux?
In stereo!

Actually, the photog coordination shouldn't be a major issue. 400 feet is certainly line of sight in most cases. The real issue would be getting the shutter to fire at the same time. Some of the newer digital camera (and film cameras) allow for RF remote triggering. Unfortunately, they don't have that kind of range for the RF.

You could certainly be very close with the "1, 2, 3, click" over the radio and it would most likely work out pretty well.

Getting the images put together in stereo would be a challenge. I'm not quite that savvy though if someone were to ask me and I was available, I'd certainly participate in a venture like that.
I think this would be extremely difficult in practice. Aside from all of the synchronization issues, both photographers would need to be on an even plane with the tornado - i.e. exactly the same distance away - but of course you'd need to be set up for this moment in advance - so you'd have to anticipate the tornado being at a point where it is oriented 90 degrees from the road alignment that the two photographers are on, and then take the picture at that moment. Further, the two images would need considerable overlap of the view, since you'd need to crop away the sections of the image which do not overlap (except for a small edge on the left/right), and the exposures would have to match extremely well in order for the composition to make sense to the human eyes. Also, you couldn't have an obstruction between the two views which is not viewed by both cameras - or it won't work right.

Difficult indeed I would think.

Margin of error?

I don't believe your distance would be that critical. Call it a margin of error of about +/- 50 feet. Not insurmountable by any means. Of course, 400 feet of string would do the trick. Manual exposure corridinated by radio to have the camera settings the same and a standard 28mm lens. Foreground would be an issue, but not too bad. Just pick and choose your spot.

We're only talking about two cameras here. You would need two chasers that dedicated themselves to producing this type of image. That may be more of a challenge than the actual technical details.
... or one chaser with two cameras and a radio-remote shutter release (and enough set-up time, which is probably the biggest challenge).

Another approach would be two prosumer DVcams. One photographer could then run out to a point of visibility for both cams with a little strobe flash to give a synch point. The rest is computer processing.

Some of you may know about the LDAR project already, but it was developed to detect lightning in 3D... you guessed it, every zig and zag of each strike in a small range (7 miles I believe) could be detected.

I had the opportunity to develop software to display that data in realtime, as well as display archive data. The project was through the original Global Atmospherics but since the Vaisala takeover, not much has been done to continue that...at least from what I have seen.

Here is a quick snap (minus city names) of the Dallas area... this is a quick snap without the points connected of one bolt of lightning.

It wouldn't be the same as a REAL picture, but the data is interesting, none the less. http://home.kxan.com/weather/ldar2view.jpg

Again, not as interesting without some other features turned on, but the data is... out there.
I am the one Bill Hark quoted to start this discussion, as I had not yet had my application to join the forum cleared. Now that I can speak for myself, I wanted to add a couple of thoughts. If a tornado is some distance away (3 miles in my example), split second synchronization may not be necessary. I suggested putting the horizon near the bottom of the frame to reduce the problem of clashing foregrounds.

How difficult is it? Well, you might practice with supercell towers and non-tornadic wall clouds where there is even greater margin of error allowed for synchronization. Since the best view might be a bit more than 3 miles away, the separation would have to be more than 400 feet -- just keep that approximate 40:1 ratio. A stereo time-lapse of a rotating wall cloud would really be cool.

I once did a "stereo" of clouds from an airplane by just taking a couple of shots a few seconds apart. It works! In this case I had two prints and used the "cross-eye" method to view them.