Friday, July 20, 2007

Collective Mediocrity

The Live 2007 double-CD of the SFJazz Collective turned up yesterday. Consistent with the previous 3 offerings, this is another really polished performance from one of the tightest live acts I've seen.

But sadly, this year that's about the only compliment I can pay it.

One of the reasons this group was put together was to force a severely modern rethinking of jazz that we have become too comfortable with. However, with the exception of Miguel Zenón's two arrangements of Monk's Epistrophy and San Francisco Holiday/Worry Later, I was a trifle disappointed with the rest of the band's all-too-delicate handling of Monk's work. They sorely could've used Miles Davis' advice:
When you play music, don't play the idea that's there, play the next idea. Wait. Wait another beat, or maybe two, and then maybe you'll have something that's more fresh. Don't just play from the top of your head, but listen and try to play a little deeper.
Newcomer Dave Douglas saved my flagging attention on the second CD of original compositions with his three-movement San Francisco Suite which, along with Eric Harland's Union seemed to be the only pieces that really provided a platform for interesting musical discussion. Admittedly I've only listened to the album a couple of times since yesterday, and while it does take a couple of listens for me to grok a tune, with the rest of the pieces I had the feeling that I was simply embedded in a stream of somewhat incoherent background conversation. Nothing on the album had the sweeping vision of Collective Overture (on the Live 2006 album), or the joy of Development (on Live 2005).

Hopefully the injection of new blood this year with Joe Lovano replacing Joshua Redman, and Stefon Harris replacing Bobby Hutcherson will revitalize this group in the coming season.

Thursday, July 12, 2007

Second Light

I've always been a fan of those cool wine/martini glass shots, like this one over at Liquid Air Photography. Now that's all fine and dandy, but I just have one flash, and the setup clearly requires two broad sources of light (one on each side of the glass).

Now although I plan to get a second flash sometime in the near future, I'm putting it off for a couple of reasons: Firstly, I need to actually do other stuff with all that money like um, pay rent. And secondly, having just one light source forces me to think a little more creatively than I otherwise would.

Now before I tell you how it all came together, goggle at the result for a second or two:

See what I mean about the two light panels? So the problem for me, is to figure out how to illuminate two panels on either side of the glass with my lonesome FL-50. It's something of a tricky problem since I can't allow any extra light to contaminate the scene, but I still need to get those two panels lit up bright enough.

So here's a top-view schematic of how we could possibly achieve this:

The key realization here is that unlike the Liquid Air setup, we don't need both panels illuminated via reflection, but instead one of the panels can be constructed out of material that transmits light via diffusion. This allows us to use a single flash on one side of the setup, lighting one panel via diffusion, and the other panel via bare reflection. With a little bit of placement trickery, we can make sure that bare light from the flash doesn't contaminate the subject itself.

The nice thing about this setup is that we can control the relative illumination of the panels by adjusting the flash distance, as well as the horizontal placement of the glass. The final ghetto setup cobbled together with all the garbage that was within reach on the table, is shown below:

One point that's not obvious from the top-views of the setup, is that the flash is kept just a smidge below the table level. This prevents the light from contaminating the surface that the glass sits on.

A couple of things that make me really happy about the result: As a proof-of-concept, I'm thrilled that it works, and I've learned a lot just by restricting myself to using just one flash. I could certainly improve on it too --- the DOF could've been increased (the base of the glass is a little out of focus), and now that I know this works, I need to build a little DIY studio to replace that ghetto setup.

Monday, July 02, 2007

Depth of Illumination

It's been a busy month, but more about all of that in another post.

It's refreshing to be exposed to a piece of information that radically changes my way of thinking about something. Especially when that something is as mundane as the distance of a source of light from a subject.

We're familiar with the notion of depth-of-field when it comes to focus, but now think about exactly this concept, but applied to illumination instead. We know that light falls off in intensity as we get further from the source. In fact, elementary physics tells us that the amount of light falling on the subject will be inversely proportional to the square of its distance from the light source. Fine. That's easy-peasy. So if we want to double the intensity (1-stop up), we need to reduce the distance to the light source by a factor of 0.7071 (1/sqrt(2)). Conversely, if we want to halve the intensity (1-stop down), we need to increase the distance by a factor of 1.4142 (sqrt(2)). That's trivial. Now here's where it gets interesting:

We're normally illuminating more than just one plane. Typically the scene consists of a bunch of objects scattered about at varying distances from the light source, so if we expose correctly for the middle-distance objects, the ones closest to (furthest from) the light source will be over(under)exposed. But by how much? If we define +-1-stop of light as an acceptable exposure deviation, then from our calculations above we know that objects within the 0.707x to 1.414x (where x is the distance of the correctly exposed object from the light source) box are "acceptably" exposed.

The cool thing is that this gives us a useful creative knob to tweak. The diagram above shows that by moving the light source further away (and cranking up the light power to compensate), we can increase this box of "acceptable" exposure. Similarly, if we want the zone of acceptable exposure to fall off quickly, we just have to move the flash in really close (and decrease its power). The object in the middle-zone gets exactly the same illumination each time, but the gradient of illumination can be as smooth or as sharp as you like.

A practical application of this is when doing portrait shots against a background. By controlling the depth of illumination, we can easily control the relative illumination of the subject versus the background simply by changing the distance of the light source from the scene.

In a nutshell, the following 2 rules should help:
  1. Given a correctly metered subject, you'll know that 30% inside and 40% outside its distance from the light source, all objects are within a 1-stop illumination deviation.
  2. Conversely, if you've got a scene of depth x, then to illuminate it all within a 1-stop deviation of the central part, the light source has to be at distance at least x from the near edge of the scene.