[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: High speed Tesla spark photographs

Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>

Hi Dan,

At 09:17 PM 9/15/2006, you wrote:

Terry,

Since you are using a camera that produces considerable noise with even
relatively short exposures (a few seconds), it would be important to use as
short an exposure as possible to reduce noise build-up on the frame.
Remember, you may only be capturing say a 10us arc (i'm guessing on the
actual duration here), but if you are leaving the shutter open for even 1
second,  you are still capturing noise for that entire second which just
integrates over time.
Unlike Peter, I have four mirrors mounted very close to the camera so I have a much higher probability of getting a "good" picture with far lower exposure times. But I am happy to hear I won't have to use the very high speeds like 1/800 sec!! So normally I should be fine as far as sub 1sec exposures go. I might have to block and black out the non productive image areas since they might roll past unneeded ambient light areas...
What I am worried about is that a good picture will just show the really bright power arc without much else. I really want to see the streamer activity that "leads up" to the power arc. I think the frame will need very high light gathering ability. The power arc might be very bright but the lead up streamer very dim. Peter's pictures show "after" arcs only, which are sort of a bazaar subject right there(!) But apparently the lead up arcs are "dim". A BIG lens telescope might be a "better" lens since light gathering is where it's at. But you have to look into a mirror... If the speed can be low, that can be a pretty big mirror ;-)) I suppose one could spin the focal point mirror in a Newtonian reflector telescope for the ultimate contraption ;-)) Oddly, speed of light propagation might be a concern then if it was all too big!!
Software can "subtract" a "black" picture from a real one to take out sensor noise. It can also "despeckle" based an mathematics to reduce the noise far further. I suppose I could also freeze the whole camera in the refrigerator too ;-))
So the worry is that the interesting stuff will be very dim and hard to register on the sensor array even with all the software tricks... Astronomy folks have much more sensitive sensors... 

http://www.starlight-xpress.co.uk/SXV-M25.htm
Of course, one has to pan down to the very bottom of the page to find it cost $6000 =:O But we are looking for ideas not hardware ;-)) 

There are cheaper ones ;-)))

http://homepage.ntlworld.com/molyned/web-cameras.htm
But all this is just in case we run into a road block. Hopefully, just the typical rather nice digital camera will work fine. But I am thinking I will have to "pull" the stuff out of a normally black noisy background. The sparks are visible to the eye, so it should be "in range" ;-))
Just have to try and see and fix the issues as they come up. Right now the problem is just getting those *%&$^& mirrors just to get started.... 



Also, you mention color balance and all that jazz, but if you really want to
be serious with these images, you REALLY do need to shoot RAW images and
convert them to TIFF files using 16-bit linear mode.  With RAW images, it
doesn't matter what your color balance is as  you can select this when you
do the RAW conversion.  Also, when shooting JPG you only get 8-bit of
information vs. 12-bit.  Of course, you could argue that that added
resolution might not be that important when shooting a bright object like an
arc.
The camera happily produces 9MB tiff images. But I was playing with it today and it did not seem necessary. The JPG 500K files seemed to be just fine. I did discover an obnoxious digital smoothing function in the menus which I turned off to great advantage. Color balance is not a big deal at all as long as the "Do these colors look right to you?" test works ("old folks" remember that one ;-)) 

http://hot-streamer.com/temp/070805111846.jpg
So once the mirrors get here, I'll just have to play around with it to see what big problems come up and go from there. But my feeling is light gathering should work out with software tricks. Once "figured out", better cameras can be called upon... The software is very happy playing with "raw" images already ;-) 

Cheers,

        Terry



Dan