LED fresnels?

[JB Earl]

Has anyone seen or used the Mole Richardson led BabyLED / TweenieLED etc or anything similar like the Fotodiox DY200? They're very enticing for the obvious reasons (output and light quality) .

 

[Phil Rhodes]

I've seen stuff at trade shows, but never really used it for anything very critical.

 

Within those limitations, I thought the Zylight F8 was pretty good, and amazingly compact.

 

They all have the same problems in that they struggle to output a perfect white. Some are better than others - the Arri L7 series is apparently very good, but this is secondhand information - and it's worth looking for TLCI data as well as the normal CRI number.

 

P

[Adrian Sierkowski]

I haven't used the Moles, but on the whole the LED fresnels I have used have been ok, not great, not horrible, just ok. They don't "feel," fight for me for most scenes but sometimes you have to use them and were I working in an urban environment, I'd not have much of a qualm with it.

[Peter Mosiman]

I just recently used a whole bunch of the Mole's. I ran 19 babyLEDs off of 2 20a stingers. It was really great given the constraints of the location we were shooting in. I was using all daylight ones so I can't comment on the "white-ness" of their tungsten brothers, but the daylight color seem consistent and pretty nice looking.

 

My only qualm with them is that they don't pack quite the punch/per fixture you would hope or think they would have. It's fairly close to their tungsten "equivalent", but not quite.

 

As you can tell below, I was pretty stoked on how much power all that light pulled.

 

[Dino Giammattei]

We weren't using our Arri L7 for much other than background lighting so I can't comment on its flesh tone use. It was getting a good deal of use until it died right in the middle of a shoot. We just packed it up for its return to Arri. I'll let you all know what happens once they take a look at it. We loved it while we had it though. Checking the internet, we couldn't find another case of one failing so hopefully this is an anomaly.

[JD Hartman]

The low power aspect doesn't hold my interest. If they could make a 5k LED Fresnel, now that would be something.

[Phil Rhodes]

Only the very best LEDs - barely out of the lab - are actually any more efficient than the fluorescent tubes we've had for decades, give or take the efficiency of older iron ballasts. For this reason, the common flat panel LED softlights, while useful, are really something that could have been made some time ago. What's new is the ability to make efficient hard lights without relying on metal halide (that is, HMI) technology. The best HMI gear has very few caveats, although even the really nice stuff has a tendency to make noise when flicker free, and won't always hot-restrike with complete reliability. It's also considerably more physically fragile than LED. An HMI ballast is a bulkier and necessarily more expensive piece of technology than the relatively straightforward current regulating DC power supply required to run LEDs. Very high end LEDs may include optical feedback technology to improve long term consistency and colour management, although that isn't intrinsically expensive to replicate - it's mainly software and microelectronics. LED could turn out cheaper than the equivalent HMI, long term.

 

So, while it's not a completely clear-cut choice, there are certainly reasons LED technology may be desirable. High efficiency soft light, which ironically represents the majority of current LED applications, is not one of them.

 

It is not currently possible to make single LED junctions above a few watts. All devices above a few to maybe fifteen (in the lab) are arrays of smaller devices, which makes ellipsoidal or profile lights difficult as they rely on something close to a point source. LED is perhaps four to maybe five times more efficient (in terms of luminous efficacy, with good engineering) than tungsten, so you may be asking for a 1000-1500W LED device. It is probably possible to do this - people are mass-producing single components representing two hundred one-watt LEDs in a single package, so if you wanted to put five or ten of those inside a fresnel housing, you'd have something approaching a 5k tungsten fresnel. They're individually about 3.5x2.5 inches, but then 5K tungsten filaments aren't exactly tiny, so it might work. High power LEDs at this level, despite their efficiency, require aggressive cooling and this would be something of an engineering challenge to achieve silently, although the average 5k is already sufficiently big that you might find a workable solution that wasn't too impractical.

 

I think in the long term, though, all of the current technology will be seen as rather inadequate from a colour quality perspective. This is not solvable without fundamental advances by the manufacturers, and by "manufacturers" in this context we're talking about people like Cree and Nichia who make the actual LED semiconductors themselves. There are obviously huge financial rewards for cracking this particular egg and a lot of R&D is currently being done.

 

It'll happen soon enough.

 

P