April 12, 2011

Manual and TTL

Filed under: 20-minute,photography — Mark Dalrymple @ 5:01 pm
Tags: ,


I went to the FlashBus yesterday, an around-the-US tour by two of the big names in small-flash photography: David “Strobist” Hobby and Joe “Numnuts” McNally. David is a manual shooter and Joe is primarily a TTL shooter. The dude sitting next to me, and several others at the seminar didn’t seem to get the distinction between Manual and TTL flash usage. And unfortunately this about the only part of the entire day where I felt the explanations were weak.

Manual is what it says: everything is manual. You have to tell each flash how much light to dump when it is triggered. You can adjust the flash by changing its power directly on the device, usually by fractions of power like 1/128, 1/64, etc or by using a CLS remote unit but telling each flash to be manual. The Elinchrome Skyport system also lets you adjust Big Flashes in manual mode remotely. You’re just telling each flash exactly how much light to dump when triggered. How that light looks, is entirely up to you. Want more fill? Increase your fill flash a stop from 1/32 to 1/16 and see.

How do flashes get triggered? There are a couple of ways – you can PocketWizard each of them. You can also optically slave them so that a trigger flash, say a 1/128 speedlite on your camera, will cause the other flashes to fire. Nikon calls this “SU-4” mode. A CLS commander can also trigger the lights.

TTL, on the other hand, has a back-and-forth conversation between the camera and the speedlight that figures out how much power each flash will dump when triggered. You configure the lighting ratios on the commander: the key light is 1.0 (baseline), the fill light is -1.0 stop, the kicker light is +1.3 stop, and so on.

When you press the shutter release button the camera, a complex dance happens. A preflash is emitted. The speedlite looks at a message encoded in the flash, and then fires itself. Each of the groups fires in turn, the camera sees how much light is being produced by them, by using a sensor inside the camera getting the light Through The Lens (hence the name TTL). In conjunction with the ambient light, the camera’s exposure settings, and an N-thousand database of exposure scenarios, the camera figures out on the fly what the absolute power each flash needs to dump. “Group A, you’ll fire at 1/4. Group B, you’ll fire at 1/16. Group C you’ll fire at 1/2.”.

Once each flash is told what power they’ll fire, they’ll wait until they get an “OK! Go!” signal from the commander, and then dump their light on to the scene.

In manual you have complete control over every light. You could have 37 lights each with a different settings. It is more labor and mentally intensive configuring each light, but because you have complete control, there are no surprises from shot to shot.

In TTL you give up a measure of control to the computer brains in the speedlites and the camera in exchange for convenience. You just specify the ratios of light between different groups. You can move from dark lighting conditions to brighter, and system will compensate by telling the flashes to dump more or less light. The exposures will generally look pretty good. You spend more time taking pictures than configuring lights. But then you also get to debug things when things go weird if TTL gets confused.



April 4, 2011

Measurebators and the Historgram

Filed under: 20-minute,photography,skills — Mark Dalrymple @ 9:07 pm
Tags: , ,

Bork portrait

The Grid Live is a new live videocast from the Kelby empire of photography and photoshop training. It’s a neat show with a nice format, easily my favorite of the Kelby TV shows. It’s a simple format, two or three guys talk for an hour on one or two topics. And that’s it. Kind of like Photoshop User TV crossed with DTown but without the obligatory software tutorials. It’s even more fun because they always have sound syncing issues, so it’s like watching a Kung Fu or a Godzilla movie.

Today’s program talked about “measurebators”, the folks who get off on the technology of photography, generally ignoring the artistic side. I have those genes in my lineage, having witnessed my dad use a transmission densitometer on samples of 4×5 film to determine the proper base fog density marklar involved in N+1 zone system weirdness I totally do not understand. My dad also took kick-ass photographs.

One part of today’s show was spent harping about histograms. A histogram is a chart that shows how much of the image is at a certain brightness. A histogram clustered to the left side is pretty dark, a histogram clustered to the right side is pretty light, and a histogram clustered in the middle is gray.

It’s possible to get too enamored with your histogram. Matt K was saying how someone asked him at a training session “what is the perfect histogram”, which of course is a nonsensical question. Low-key photos will have radically different histograms than high-key photos.

Many photographers never look at the histogram, saying “I just look at the blinkies”, which the highlight warnings that flash at you if you exceed the amount of light your sensor can handler. If you looked at the histogram with a blinky scene you’ll see it butted up against the right side.

The blinkies are great. I use them all the time. I also use the histogram. It is very convenient, being only one click away on my camera’s selector. I mainly use it to make sure I’m notxposing a scene. As I mentioned earlier in Your Camera Lies To You, the LCD screen in a dark environment can make a photo look great, and then you go home to process it and it’s a stop or two underexposed. The blinkies will not help you in this case. You can happily shoot and chimp and come home to a mess to process. Sometimes when I’m shooting in manual mode, I forgot to check in with my exposure meter and use inappropriate settings. Again, chimping might not clue you in to an underexposure situation.

Case in point, when I shoot the local Hometown Christmas production I frequently check the histogram because the theater is dark and the lighting conditions change constantly.

Related to all of this is a school of exposure called “Expose to the Right”, or ETTR for those who like unpronounceable acronyms. Basically you try to overexposure your scene just enough where you don’t exceed the capacity of the sensor, but not going over. The histogram is vital in getting this perfect exposure. In post processing you pull the exposure back to where it should be. This concentrates the image’s pixels in the upper (brighter) end of the histogram, which has more granularity than the darker end due to there being more bits available. I use ETTR if I’m doing slow-shooting photography, arty or landscape stuff. If I’m shooting 800 frames documenting behind the scenes of a fast-paced video production department (like I did last weekend), then for my own post-processing sanity I wouldn’t use ETTR.

The take-away? Learn to use the histogram, but don’t be a slave to it. Check in on it every now and then like an old friend. On many shoots I never look at it. I do check-in on my histogram if I’m shooting indoors or somewhere that it is dark, or if I’m running purely in manual mode and I might have forgotten to check my exposure.

March 30, 2011

Your camera lies to you

Filed under: 20-minute,photography — Mark Dalrymple @ 1:57 pm
Tags: ,

A friend dropped me an email yesterday, asking “is there anything you can do with these photos?”.  Here is one of them:


Doesn’t look too bad, does it?  When you zoom in on it, it becomes a disaster:


Out of focus, and terribly noisy.

The image was shot with 12 MP point and shoot, f/5.something at 1/30th, ISO 1600.  Knowing how tiny the sensors are in point and shoot cameras, and therefore how tiny the image wells are for collecting light, I’m amazed that this shot is as good as it is.

Other shots  from the same event had similar problems.  One had the speaker totally out of focus while the audience was in focus.  But it was obvious that wasn’t the photographer’s intent.  So what happened?  Upon further quizzing, the photographer (who otherwise is an excellent DSLR shooter) said that things looked great on the LCD screen.

Camera screens lie to you.  Lying liars lying about lying.

There’s two things to remember when chimping your photos on the camera screen.  The first is that the image being shown is a reduced version.  The camera is downsampling a 12+ MP image into a small, maybe 1 MP screen.  Drastic image reduction makes images look sharper.

The second is you’re looking at a back-lit screen. This photo was obviously taken in a dark environment, so the bright backlight makes the shot look even better.  So it is really easy to understand what happened: a poor shot was made to look ten times better due to the reduction and backlight.

If you chimp (and I am a proud chimper), you should learn how to zoom in on photos to actual pixels at 100%, and a step out, at 50%.  This lets you accurately judge focus and overall sharpness.  On my Nikon DSLRs, I twiddle the menu settings so I can press the button on the multi-selector thumbpad dealie to zoom in to 100%.  I can shoot, chimp, button, check focus, and then move on.

The histogram can be your friend, especially with one of these “shooting in a dark environment” scenarios.  The backlight will make everything look great, even if your image is a stop or two underexposed.  Take a quick peek at the histogram to make sure you don’t have a lot of leftover headroom.

What could have been done to make this particular shoot a success?  A couple of things:

  • Use mechanical stabilization.  Even though the P&S had in-camera image stabilization, the shot was still blurry.  1/30th is hard to hand-hold well, and slower speeds even harder.  I would have tried to brace the camera on a table, chair, building support column, or something.
  • Lower the ISO a stop or two to gain image quality. This will put your shutter speeds into really slow range, but take enough and you may get one sharp one, especially if the speaker isn’t dancing around like a madman.
  • Use a better camera.  A DSLR, especially something that just drinks in the light like a D3[s] or a 5D would let you shoot at higher ISOs without too much worry about image quality.
  • Use a faster lens.  With a point and shoot you don’t have any choice.  On a DSLR, an f/5.6 lens means you will have to shoot at even higher ISO or a slower shutter speed than if you have an f/2.8 lens.

But if you have a bad image, and need to use it, you need to use it.  I made some adjustments in LightRoom to get this “after” picture:

Medium fixed

I used LightRoom 3’s noise reduction (which is most excellent), turned up the contrast and clarity.  You can see there’s a lot more definition in the speaker’s face, and a lot of the noise is gone.  Plus I healed-out the little white blobby thing.

Blog at