I borrowed this from rec.video.production:

If you use a camera with a "SteadyShot" function, such as the VX2000 or TRV900, you might want to disable the "SteadyShot" function, as it could result in a bunch of tiny reflections around each candle, sort of like a "firefly effect." I have a vx2000 and turn off this function whenever I am shooting around candles or similar lights.

Has anyone seen this come and go with steadyshot on/off?

I've had the phenomenon but attributed it to a poorly coated filter. Now I'll have to go back and test.


> I will be shooting Christmas Candlelight service with 2 DV this coming Christmas.
> I will be doing NLE with the footage. I have done a single camcorder shooting last year for the music service with good result.
> However I need help for all of you to help me to do a better job. For example, the light level will be very low due to the candlelight event, so how can I overcome it? Any other hint or suggestion?
>
> Ted

i have had the firefly effect happen to me while shooting at night at a local carnival all the tents had lights strung around the top that would flare i figured it was due to shaky camera work only to find out that the steadyshot could be causing the problem on another message board. i havent shot at night since but will give it a try also. shooting done with VX2k

Read this post http://www.dvinfo.net/conf/showthread.php?s=&threadid=3299 and see if there are similarities. My understanding is that Sony licenses the IS technology from Canon so they work in the same manner. I've shot candles with the IS on (XL1s, no filter) and no problems (or fireflies).

Jeff

This is a well-known quirk of Optical Image Stabilization. It's a very real effect, and I experienced it with the XL1 many moons ago at a night shoot on a well-lit football field.

OIS has its trade-offs like all other features do, and this is one of them. And it's true that Sony licenses the OIS technology from Canon, the patent holder (there's a lot of trading of technology between camera manufacturers -- for instance, all Canon camcorders have a LANC jack, which is a Sony protocol).

Any time you see Super Steady Shot on a Sony camcorder, it's actually Canon's OIS at work. The "dancing lights" effect is definitely something to look out for in certain shooting situations. The remedy is to turn off the OIS. It's not considered a defect or a problem... it's simply reflections from the vari-angle prism in the OIS mechanics. A side effect of an otherwise amazing and effective image stabilization technology.

One would think that the surfaces would be multi-coated to prevent just such an occurance. Must be something preventing the application of the coatings.

Most, if not all, coatings work at very specific angles of incidence. Could be that the angles involved with the OIS system are too shallow for effective AR coating to work.

The IS uses a variable angle prism. The angle will vary with the amount of shake present. Zoom in, more shake present, larger angle prism. Combine that with the angle to the specular light source and you have fireflies. The use of a filter may make the effect more pronounced. I tend to shoot specular light sources centered, out of experience. I know the center is the best corrected, sharpest part of the lens. I do it to help control flare, not realizing the added effects of the IS. Not good for composition, but combine that with the fact I don't use filters explains why I don't get fireflies.

A further thought. Thinking of 2 situations:

If the camera were locked down, then the stabilization system, though active, should not be moving its mechanism.

If the stabilization system is off, assuming it doesn't move the plates to a parallel postion and 90 degrees to the lens axis.

In both situations, wouldn't one see internal relections anyway, they just wouldn't move around?

Have to try this tonight. I have a 10:30 PM highway sightline shot for a lawsuit.

After 2 successive nights of shooting traffic signals for the local PD, I can attest to the 'Fireflys' presence. Steady Shot on or off!

The fireflys simply stop moving when one turns off steady shot. They do not disappear.

This is not acceptable to me and I will pursue this with Sony through the WEVA Camera Committee.

Bill, I know you stated that coatings are only designed for a specific angle of incidence. That may be. But they are apparently effective over a wider angle as witnessed by any multicoated filter user. Surely there must be some other reason that Sony )Canon) does not multicoat the glass surfaces.

AR coatings are relatively inexpensive when the incident light is normal to the surface. As the angle becomes more and more oblique, the effectiveness of AR coating diminishes, and the demands on coating layers increases, as well as cost to manufacture. Each layer is VPD deposited and baked for an extended period of time. Coating ovens cost $$$ / hr to operate. The prisms used in IS systems, by design, require rather oblique angles, as well as reflections that occur internally to the glass prism. Kinda hard to coat the inner glass surface, ya know?

I'd guess the real reason they don't coat the glass is that the transmissive losses increase as the number of coatings go up?

They could certainly coat the inner surface of the glass before they assembled the module, I'd think.

Strange that I never notice the issue until the PD-150. But then the older Hi8 systems cannot hold a candle :-)))) to its low light performance.

Never more pointed out than when I tried the stop light shots with a really good Hi8 camera 2 nights ago (they wanted analog because of legal challenge issues) and I could only record the color with the aperature at a minimum. And of course, there is nothing but black and a color disk on the screen.

The 150 shows the trees, the lamp posts, the lights, the street, details in cars, etc.

About the only thing I didn't like was the fairly strong halo around the lights. But it isn't a $30,000 FIT camera, is it?

The more you look for them, the more deficiencies you can find w/current prosumer DV's.

But if a few convincing and succinct letters are read by the right engineering people at the manufacturers, at least maybe the problem gets addressed, sooner or later.

Actually mult-coating of optical surfaces increases the transmission of light. It was the advent of multi-coatings that help make zoom lenses less expensive and perform better (lower maximum aperatures and less flare).

I reviewed my candle footage last night and I really don't see fireflies. The scene was staged after a friends wedding. One large candle, flame dead canter, and two hands coming into frame to light the wick. I let the tape roll on the flame for maybe a minute. The camera was on a tripod, no IS, no filters. The ambient light is fairly bright, so I don't have an extremely contrasty scene. I wonder if the larger size of the XL lense allows for better baffeling in the lense. Another possibllity is the coating of non-optical surfaces. Inside a better lens the different barrels, sleeves and mechanical parts are coated an opaque black to absorb reflections. for example the aperature blades are always a dull, flat black. this minimizes reflections. I've seen some lenses with fairly reflective internal surfaces.

Now, I'm not saying the XL lenses don't produce fireflies. Chris has obviosly encountered them and they do exist. I'm going to shoot some test this weekend (lights in a parking lot at night) and see why I haven't been bothered by them. I'll also do the same shots with the 14x manual lens so I have a baseline. From all this maybe I can figure a way to minimize the firefly effect.

Jeff

Yes I encountered the fireflies with the standard XL lens, but switching off the OIS eliminated the problem.

I watched the fireflys off of car headlights settle just below the headlights. Really a fine point of light so they would be hard to find if they were not moving.

Since the Sony lens are probably made by Canon, I'd guess that they are reasonably well baffled and coated. Certainly not as well as say, a $6,000 lens, but OK.

The halo around a light sounds to me like the digital edge artifact. But it could be a type of lens flare.

It's been an interesting read so far.

With my 150, I once shot a an orchestra that used ultra reflective drums where the reflected light (pretty hot) shone the effect of a star filter. Very mild (small enhancement), but the light point looked like a star. I thought it looked very nice. Then again if it were too blaring, it would begin to look tired but for now, everyone's gonna dig it!

I'm working on some tests I've done concerning fireflies and the hallow effect i've obtained. Mike, what was the weather like when you shot your tests? Was it humid? We just had a little cold snap here, went into the lower 70's overnight and less humid the last two days. I reshot some tests with the halo (street lights). I think the halo is humidity related. Light refracting off moisture in the air. It is less noticeable in the footage I shot last night.

Jeff

The dancing firefly effect is not always apparent and I think it might be more to do with the 'type' of optical stabilisation.
The problem can be found in older Hi8 cams - my old TR2000 had this problem - I distinctly remember that by turning off the OIS the problem went away...
My knowledge of OIS systems is very limited but AFAIK, there are two types - one uses a moving or vibrating prism and the other uses a silicon lens that is moved or deformed to compensate for shake. IIRC, the silicon oil filled lens is perhaps the older type of OIS - my TR2000 had warnings in the manual about the possibilty of air bubbles showing on the lens after long, high altitude flights etc..

Jeff, it was a dry night but Vallejo has SF bay on 2 sides but plenty of wind. I'd say it wasn't very humid for California. Under 50% (just a guess).

The “mysteries” described in these posts are known and have been described and documented in the past and except halo, belong to the “lensflare” domain.
If we suppose “good “optics, halo or optical blooming (not to be confused with the CCD blooming effects) resulting from small and strong lightsources, is caused by diffraction.: The (intensity clipped) lightsource shows its “airy disks” just around the lightspot if the F-number is relatively low (large aperture). At higher F-numbers, the same diffraction phenoma generates the “star” (6 point stars when 6 blade diafragm is used).
Fireflies or pointsource ghostings are generated by surfaces (mostly) in the frontside of the optics path. Parallel flat sufaces are the best candidates for generation of fireflies. Hence fluid prism OIS is a problem candidate. Moving lens OIS concepts are more robust in this respect. Even perfectly coated surfaces (still reflecting .5% of the light) cannot resolve the problem. We should know that lightsouces often have a surface light intensity (nit) which can easely reach 10^5 times the (dark) scene intensity. If someone wants to verify the (dancing) fireflies, just keep the camera close (under about 45deg) to a white surface (in a dimmed room) and point a laserpointer under 45deg, into the cameralens and the wonderfull ghost are there.
Deeper reflections are possible too and depend on the conbined surface nonflatnesses of the internal optics . Shifting (internal) zoon parts have their influence too. Reflection raytracing (Zemax) can predict all this…
Besides the overalll flare effects, which tend to add extra (haze) on the whole scene, and thus reducing the contrast, the aperture form itself can occur in the picture as multiple color patches.(e.g. oblique shooting into the sun) This effect belongs to the internal reflection generated behind the aperture and close to the CCD (prism)