"True parabolics" can you really see a difference?

There seems to be an increasing number of so-called "true parabolic" umbrellas and reflectors being marketed

lately. Being a skeptical old so-and-so, I'm yet to be convinced of any practical advantage. Also, some basic

optical theory would seem to contradict the maker's claims.<p>Firstly, a parabolic reflector should follow a perfect parabola to be called "true" and thus have a single point of focus, which a foldable umbrella or strutted reflector obviously cannot do.<br>Secondly, the silvering of umbrella fabrics isn't anywhere near perfectly polished, and so can't reflect perfectly.<br>Thirdly, these umbrellas are used with flashheads that aren't a true point source.<br>Fourthly, it would be quite tedious to have to set the flash head at the exact focus of the parabola - assuming it actually had one.<br>Fifthly, the flash would have to radiate light evenly over nearly 180 degrees to fill the parabola.<br>Sixthly, these things are often used with a diffuser over them or as shoot-throughs.<p>This last point is the one I have a real issue with. A good diffuser is supposed to completely break up any directionality of light and scatter it over a wide angle; so what's the point of carefully shaping a light beam in a parabolic reflector and then diffusing it? That's like carefully arranging a train of dominoes, and then randomly kicking them all sideways.<p>Anyway, the question is, can anyone whose used or seen the results from a "parabolic" brolly really see any difference between it and an equivalent sized, but with no claim to being parabolic, ordinary brolly?<p>P.S. Has anyone ever seen a wraparound softbox anywhere? That is, one with a concave front.

<blockquote>

<p>Firstly, a parabolic reflector should follow a perfect parabola to be called "true" and thus have a single point of focus, which a foldable umbrella or strutted reflector obviously cannot do.</p>

</blockquote>

<p>Which is why the good ones have more struts. It's not about being perfect, it's about approaching an ideal behavior enough for it to work better than not bothering to approach it.</p>

<blockquote>

<p>Secondly, the silvering of umbrella fabrics isn't anywhere near perfectly polished, and so can't reflect perfectly.</p>

</blockquote>

<p>See above.</p>

<blockquote>

<p>Thirdly, these umbrellas are used with flashheads that aren't a true point source.</p>

</blockquote>

<p>Likewise.</p>

<blockquote>

<p>Fourthly, it would be quite tedious to have to set the flash head at the exact focus of the parabola - assuming it actually had one.</p>

</blockquote>

<p>Which is why, when you buy something system-oriented, and use the right mount, the light winds up right in the sweet spot, within the reasonable bounds of [see above].</p>

<blockquote>

<p>Fifthly, the flash would have to radiate light evenly over nearly 180 degrees to fill the parabola.</p>

</blockquote>

<p>Rigged right, they do, or close enough to [see above].</p>

<blockquote>

<p>Sixthly, these things are often used with a diffuser over them or as shoot-throughs.</p>

</blockquote>

<p>That's just frosting on the cake. Why carry two modifiers, when you can do multiple jobs with one?</p>

<blockquote>

<p>... really see any difference ... ?</p>

</blockquote>

<p>The ones that are more truly parabolic, and which are used correctly, do indeed seem to have much hotter output. That can mean the difference between the same light source and (for example) battery pack being able, vs. not able to keep up with the sun while doing fill work outside on a bright day. An extra stop might be make or break, that way. So ... you use the modifier that buys you that extra stop or more.</p>

<blockquote>

<p>P.S. Has anyone ever seen a wraparound softbox anywhere? That is, one with a concave front.</p>

</blockquote>

<p>Not per se. But then, that's what wrap-around reflectors, or multiple softboxes are for.</p>

But surely Matt, the whole point of using a large brolly is to get a softer and more diffuse light? Not to get a "hotter" spot or more light efficiency. For that you only need a small, deep and rigid polished reflector dish close to the light source. Then if you want it diffused, point it at a simple scrim closer to the subject - same result.<p>Besides, not all the brollies sold as "parabolic" are even silvered, some are translucent white, which I'm sure even you will agree is just ludicrous. A diffusing surface doesn't really care what shape it is, since it sends light away from itself in all directions. The same partly applies to a dimply or textured "silver" surface. My argument still holds that in order to be effective as a parabolic reflector its surface needs to have a near-mirror finish, anything else will clearly diminish its efficiency.

<p>I have never used one of these new "parabolic light modifiers", so I can't make any statements about them based on personal experience. However, my gut feeling is that *both* Rodeo and Matt have brought up good points. Rodeo's statement that there probably isn't much difference between them and a conventional device is probably true, AND Matt's statement that even just one stop of improvement can be important (plus the convenience factor) is almost certainly also true. It's a matter of degree.</p>

<p>Here's an easy experiment (for those who have the items already on hand) that would speak directly to the change in light character by a PLM: Aim a properly set-up PLM directly (ie at right angles) at a white sheet . Photograph the sheet from the back side to look at the distribution of light intensity incident on the sheet. Next, hang a deep, large area honeycomb / grid between the sheet and the PLM, very close to the sheet. This will absorb most of the off-axis rays from the source, letting the on-axis rays continue directly to the sheet. </p>

<p>If there is a big reduction in light intensity when the grid is inserted, you will know that the PLM is not really directing the light in one direction. Repeat the experiment with the diffuser screen across the front of the PLM and there should be a huge difference in intensity with and without the grid.</p>

<p>There are certainly other ways to establish this, but this would be a very direct confirmation or refutation of Rodeo's concern.</p>

<p>Cheers,</p>

<p>Tom M</p>

As I see it, the only practical use for a parabolic reflector is to form a parallel or focused source of light. And the reflector has to be polished or nearly so for this to happen. Any diffusion, lumps, bumps, texture or deviation from a parabloid in the surface and it all goes pear-shaped (pun intended).<p>I'm posting a little diagram below to show exactly what I mean. You can't possibly get a parallel or focused beam from a parabloid with a diffusing surface, so you might as well start with a partly spherical or any other dish-like shape. I've only included a small number of ray paths for the diffused reflector, because otherwise the diagram gets very messy, not to mention laborious to draw!

<p>Don't know what happened there. Picture didn't upload.</p><div></div>

<blockquote>

<p>But surely Matt, the whole point of using a large brolly is to get a softer and more diffuse light? Not to get a "hotter" spot or more light efficiency. For that you only need a small, deep and rigid polished reflector dish close to the light source.</p>

</blockquote>

<p>Your quite off on your science. You can only get efficiently collimated light (all the rays parallel) from a reflector that is large in relation to the light source, so that the light source approaches a "point" source. Flash tubes are pretty big, so only large parabolas are efficient.</p>

<p>But that's not really the issue...</p>

<blockquote>

<p>Then if you want it diffused, point it at a simple scrim closer to the subject - same result.</p>

</blockquote>

<p>Only if you redefine "same" as "totally different".<br>

Say you have your "small" reflector 8 feet away from an 8 foot scrim. By the time the light reaches the corners of the scrim, it's 26 degrees from perpendicular to the scrim</p>

<ul>

<li>arctan(8ft diameter/2 / 8ft distance).</li>

</ul>

<p>So, if you have a scrim that scatters light over a 60 degree Gausian (less light farther from perpendicular) profile, by the time you reach the periphery of the scrim, you've essentially got zero light that can reach the subject. You've also got, in addition to the scrim's Gausian falloff, cos4 falloff. That's why an 8 foot scrim, lit the way you describe, doesn't light a subject as softly and uniformly as a decnt 4 foot soft box.</p>

<p>The second problem with the scrim is that, since you're lighting it with divergent (radiating away from one central point) light, the light is still mostly divergent after it goes through the scrim. All that preipheral light that doesn't reach the subject just scatters around your studio, lighting the walls, picking up colors, and making it either to your camera's lens (causing flare) or back to the subject, making your lighting uncontrolled and unpredictable.</p>

<p>Large scrims, flying flats, etc., are perhaps the most misunderstood of all light modifiers, with the possible exception of overly small "Fongian" diffusers, which are typically ascribed all sorts of magical powers and ability to defy physics.</p>

<p>So, collimating the light, which takes a parabola or lens as large as the diffuser, is necessary to make the diffuser act like the light source you desire. That is why the "octabanks" are such popular light sources these days. The conventional "soft box" has four panels, and there's only so much you can do to approximate a parabola in 4 panels, so the lighting is not as well collimated as you'd like it. Octas use 8 panels, and cut the mean error in half. 16 panel parabolics cut the mean error in half, again.</p>

<p>As for the diffused parabola being an "oxymoron", if you take the time to draw a decent picture, tracing the rays, instead of a cartoon that supports your overly sarcastic POV, you'd see that it makes perfect sense. Take an 8 foot parabola, one that really collimates the light, so all the rays are parallel leaving the parabola, then add that same 60 degree Gausian profile as the large scrim. At 8 feet from the subject, you've now got every point on the reflector scattering some light onto the subject, but simultaneously, you've got less stray light escaping the subject area. Same reason one puts a $300 grid on a $300 6 foot Chimera softbox.</p>

<p>RJ, you really need to kill the sarcasm, because you do it a lot, but your positions are almost always technically very weak, and the combination of poor attitude and poor information makes you look quite the prat.</p>

<blockquote>

<p>A good diffuser is supposed to completely break up any directionality of light and scatter it over a wide angle;</p>

</blockquote>

<p>No. A "good" diffuser is supposed to break up light only enough to scatter it over the angle desired by the photographer.</p>

<blockquote>

<p>so what's the point of carefully shaping a light beam in a parabolic reflector and then diffusing it?</p>

</blockquote>

<p>That's part of the principle that computer types call GIGO, "garbage in, garbage out". In order to keep in control of the light, instead of letting the randomness of the "good diffuser" that's "supposed to completely break up any directionality of light and scatter it over a wide angle" control you, you need to control both the diffuser itself and the light that goes into it. "Good stuff in, good stuff out."</p>

<blockquote>

<p>That's like carefully arranging a train of dominoes, and then randomly kicking them all sideways.</p>

</blockquote>

<p>Wrong game. Think dice, more random. Control the light, and control the picture. Let something "scatter it over a wide angle", an undefined angle, and you're at the mercy of the roll of the dice.</p>

<p>OK. I can see I'm only going to get an argument from here on out. When what I actually wanted were some objective opinions of whether the <em>finished look</em> of a picture was at all affected by the shape of reflector used.</p>

<p>I'm posting some comparison shots below with the light modifiers I have to hand. The issue I have with silver umbrellas, parabolic or otherwise, is the ugly light they put out. Even P.B's advertising shots can't disguise the fact that the shadows are all over the place - and look much the same as any bog-standard silver umbrella to me.<br>

The silver brolly used in the picture below was an 8 strut 48" diameter (that's true opening diameter, not measured across the arc of the struts!), the white brolly was a translucent 42" used reflectively, and the softbox was about the same area as the white brolly. Personally I much, much prefer the softer shadows of the white brolly and softbox, together with the less specular highlight reflections. These lower specular reflections would mean far less time spent in makeup for a human model. As far as I'm concerned the shape of catchlights in a full-length shot is not particularly relevant.</p>

<p>I'm pretty sure that being more closely parabolic would have made almost no difference to the appearance of the silver brolly shot, and would have improved the light output of the white brolly and softbox only marginally, if at all.</p>

<p>I'm not disputing that the major advantage of any silver brolly is its light efficiency, but that's bought very much at the cost of light quality. I'll use one for fill if I have to squeeze all the available power out of a light, but I'd rather not use one as a key light.</p>

<p>Joseph, a diffuser should <em>diffuse</em>. If it doesn't do that fully, then any directional light it lets through will degrade the softness of the shadows it throws. It'll also create a hotspot in it's light output and thereby increase specular reflections. Sure, you might sometimes want a semi-diffuse light, but that's something different again to what a softbox is supposed to provide, and is the reason that good softboxes provide an inner diffuser as well as the main surface. The direction of light hitting either a reflective or transmissive <em>full-diffuser</em> should make not a jot of difference to the light being reflected or transmitted. I'm not talking about quantity of light here at all, but about quality, and truly soft light is fully diffuse and directionless. Any spill control should be applied after diffusion with flagging or honeycombs.</p><div></div>

<p>And Joseph, at least I can disagree with someone without resorting to name-calling.<br>

BTW, Gaussian is spelled this way. If you're going to bandy technical terms at least spell them correctly!</p>

<p>I am not at all about the science. I only know what I see from personal experience. I own a PLM, and no, I do not see a difference in the 'finished' look of the light on a subject, but output is indeed increased. I use a long, skinny flashtube (on my Sunpak 120J, pointed into the umbrella, so coverage of all umbrella interior surfaces is excellent. My opinion only.</p>

<p>I'm with Nadine-I don't play much with the science, despite having been a science geek in high school. I'm much more for practical applications. I haven't worked with a parabolic umbrella myself, but I have worked with reflectors from 7-20 inches, and the power output and diffusion from a 20" stadium reflector on a Speedotron head is very nice. I don't use umbrellas outside much, it gets pretty windy where I am, so must of my outdoor lighting is done with bare reflectors and softboxes.</p>

 

<p>I own PLM's, conventional umbrellas, and softboxes. From a power perspective, the PLM's are much more efficient and do not waste much light ... my meter shows about a 2 stop difference between a bare PLM and conventional umbrella. From a quality of light perspective, the silver PLM looks most like a gridded softbox. With the front diffusion fabric on the light looks identical to the Photek Softlighter while being more efficient. Practically (not debating math/physics here), the PLM behaves very much like smaller parabolic reflectors.</p>

<p>I don't own any PLM's so can't comment about that. I know a little about diffusion though.</p>

<p>If we think about a bedsheet as diffusion and we look very closely at it we can see the threads in the fabric. Between the threads there is empty space. When the light passes through some of the light beams pass through undistrubed while some will hit a thread and scatter around. If the fabric is woven tightly less light will pass through undisturbed.</p>

<p>My point with this is that all diffusion is not the same. It comes in different grades from light to heavy and there may also be a slight difference depending how the material is diffusing the light. Heavy diffusion will cut out a lot of light while light diffusion will only cut a little. As some light is passing through the material undisturbed it also means that the directionality of that light is kept intact.</p>

<p>Just to show you how many types of different diffusion fabrics are available commercially:<br>

http://www.rosco.com/us/technotes/filters/technote_3fv.cfm</p>

<p> </p>

<p>What counts is the size of the source with respect to the subject. The bigger the source, the more gentle the shadow.<br>

Further; I have always found it advantageous to have a hotter centre - that is, a non-uniform source. The advantage being a penumbra that graduates even more gently.<br>

The model is veiled sunlight through a window<br>

The disadvantage of a uniform source, however large: Penumbra fall-off depends on the relative geometry of the source edge and the subject edge. When the geometries are similar the penumbra is abrupt. Thats why one has to twist softboxes to the diagonal for most subjects.<br>

Remember that natural penumbra is colder than the source. So fill (invisibly) with a coldish reflector. Lastolite , like Kilroy, never made it this far.</p>

<p>One more point to add for the parabolic umbrella is its ability to focus and de-focus at will. When focused the quality of light it directed is hard and harsh resembling sunlight and when defocused it direct soft light. The gradual change between hard and soft at will is an ability that other lightshaper cannot.</p>