On Camera Bounce Flash

CarlS

Neil,

I have read much of the information on your web site
covering bounce flash techniques and watched your B&H presentation on the
use of on-camera flash.  Great
presentation!  Thank you for making so
much information available publicly. 

In your presentation, you discuss the two methods of flash
operation – TTL and Manual modes.  I’m a
little confused on some of the technical aspects of TTL and was hoping that you
would be willing to provide some clarification. 

In your presentation, you gave a great overview of using TTL
flash to allow the speedlight to automatically set and adjust flash output, and
then advised using flash exposure compensation to make micro adjustments in
output.  It just seems too easy… 

It makes sense to me that the speedlight is able to very
accurately set flash output levels based on ISO, aperture, shutter speed, focal
length, and distance to the subject. 
Allowing the TTL logic to set output would also make sense to me if we
were using the flash facing towards the subject.  However, you said one of the purposes of bounce flash is
to soften the light by making the light source larger via bounce surface.  When bouncing, why wouldn’t I want to set the
flash zoom as wide as possible to produce a larger light source on the bounce
surface (while keeping in mind your recommendation to shield the flash and
prevent light from falling forward onto the subject)?  If the bounce surface is of significant
distance away, it would make sense to me that I may want to zoom the flash to
keep the light concentrated, but this would be dependent on distance to the
bounce surface, not the focal length of the lens and/or distance to the
subject.

The same could really be asked for allowing TTL to set
output using aperture, ISO, and shutter speed, but the monitor pre-flash
probably helps augment these settings appropriately for individual bounce
situations.  I guess I’m wondering in far
too many words why we can’t get more consistent results through familiarity with
manual control of the flash output. 

Thanks for any insight..

dbruno

As an avid "forum reader", I'm waiting for a response to this myself.

Neil vN

Regarding the differences between manual flash vs. TTL flash.

With manual flash, YOU control the 4 variables: aperture, ISO, distance & power, to get to correct flash exposure. 

With TTL flash, te camera controls the power of the flash, as the ISO, aperture and distance changes. 

" It just seems too easy…"

That's the crazy thing - TTL flash makes it that easy. 

"It makes sense to me that the speedlight is able to vary accurately set flash output levels based on ISO, aperture, shutter speed, focal length, and distance to the subject."

The focal length doesn't come into play regarding accurate flash exposure. It does affect whether your flash covers the entire scene, but it isn't directly involved with flash exposure. However, as you zoom your lens, it will affect your flash head's zoom position, and that will affect power. But it is an indirect consequence. 

"Allowing the TTL logic to set output would also make sense to me if we were using the flash facing towards the subject."

TTL flash is calculated via a pre-flash which the flash emits, split-seconds before the actual flash output. The camera meters this returned light, and adjusts the main burst of light accordingly. 

This means you don't need your flash pointed directly forward - the camera will adjust output according to the pre-flash it is metering. 

The only auto flash system that I am aware of that adjusted flash output according to distance, was on some of the Mamiya 35mm cameras. Ingenious, but not practical for bounce flash. 

"When bouncing, why wouldn’t I want to set the flash zoom as wide as possible to produce a larger light source on the bounce surface (while keeping in mind your recommendation to shield the flash and prevent light from falling forward onto the subject)?  If the bounce surface is of significant distance away, it would make sense to me that I may want to zoom the flash to keep the light concentrated, but this would be dependent on distance to the bounce surface, not the focal length of the lens and/or distance to the subject."

You're confusing a few things here, but hopefully the explanation about the TTL pre-flash will clear this up for you - the actual distance is compensated for by the camera metering for the amount of light returned from the pre-flash. 

I keep my zoom mostly fairly tightly zoomed, otherwise the black foamie thing eats up too much of the light. 

"The same could really be asked for allowing TTL to set output using aperture, ISO, and shutter speed, but the monitor pre-flash probably helps augment these settings appropriately for individual bounce situations."

Not "augments". Calculates. 

That's a huge difference. Make sure you completely understand the role of the pre-flash. 

"I guess I’m wondering in far too many words why we can’t get more consistent results through familiarity with manual control of the flash output."

Manual flash most definitely has its place. I even use it sometimes for bounce flash. 

eg:  high-ISO bounce flash photography 

But mostly I use TTL flash when bouncing flash because the camera takes care of the distance problem. 

I hope all this make more sense now. 

dbruno

Thanks, Neil. Although I'm not the original poster, I am still a bit confused about the pre-flash calculating flash power for a flash unit set to bounce: I'm assuming the pre-flash fires directly from the flash head to the subject, and calculates power based on this distance. But if the flash is angled up at say 45 degrees - just for example - the distance the main flash has to travel is different from that used by the pre-flash for its calculation. I'm just using very simplistic ray-tracing here, but how does the pre-flash know it's a bounce situation, and take this into account?

Thanks!

Neil vN

No, you're still thinking in terms of the distance from the flash to the subject.  The TTL pre-flash by-passes that idea. 

Here is what happens. 

1. You trip the camera's shutter. 

2. During that brief interval before the main burst of flash goes off, and the sensor is exposed to the light (i.e., before the 1st curtain moves) ...

3. ... the flash (controlled by the camera), emits a short blip of light of KNOWN levels. i.e., the camera knows how much light it emitted. 

4. The camera then meters that same blip of light that is reflected (or diffused in a softbox), and then, compared to what the camera knows it sent out, will calculate how much light to dump during the actual exposure!

5. In this way, the aperture and ISO are all taken into consideration, as the camera figures out how much light it needs to put out for correct TTL exposure.  The DISTANCE doesn't directly come into play. It is implied by the level of light returned from the pre-flash. 

6. So if you bounce your flash off the wall behind you, the camera have NO idea of the distances involved. It would be an impossible calculation for a camera. BUT, via the pre-flash it knows much light was returned, and from that the TTL flash level is calculated. Genius technology!!

So with all of this ... let go of the idea of distance. It doesn't directly come in to effect in calculating the TTL flash. 

dbruno

Excellent, thank you. Cleared it right up - Dave

CarlS

Neil,

Thank you for the detailed response. The pre flash makes very good sense to me, as well as how TTL adjusts flash output based on the metering is sees from the pre flash. It is probably inconsequential, but I do wonder if this metering takes place in the camera or in the flash unit itself.  The SB-800 flash body always points with the camera and the flash head itself is the only part that tilts, so I'm wondering if the metering takes place in the lower portion of the flash body rather than via the camera's meter.

The reason I am going down this rabbit hole is because when I'm in manual mode (on the camera) and exposing for a darker background, the camera meter almost always tells me that I am under exposing the photograph - so I'm thinking that the flash metering must take place independently.

Anyhow, back on topic.  You also said that you keep your flash zoom fairly tight, because otherwise too much of the flash output gets absorbed by the black foam shield.  I am usually shooting with either a 35mm or 50mm prime lens.  Should I be allowing the SB-800 to correctly identify each focal length and just leave it that way, or should I be manually zooming the flash in a little tighter?  I don't have my camera in front of me at the moment, but I think I can remain in TTL while manually controlling the flash head zoom.  This question here is ultimately one of the main reasons for the original post.

Thanks again Neil,

Carl

Neil vN

"It is probably inconsequential, but I do wonder if this metering takes place in the camera or in the flash unit itself."

If the metering takes place via the flash, it would be an Auto flash, or Auto Mode ... where the light is metered via a cell on the front of the speedlight. A thyristor circuit cuts the flash when there is enough exposure. 

This is done without a pre-flash.  Maybe the modern speedlights do it differently. 

Now, TTL flash is measured Through-The-Lens. The camera meters it, and controls it. 

"I am usually shooting with either a 35mm or 50mm prime lens.  Should I be allowing the SB-800 to correctly identify each focal length and just leave it that way, or should I be manually zooming the flash in a little tighter?"

If you are bouncing your flash, you are effectively creating a much larger light source that will flood your scene with light. So that restriction of the flash-head needing to be zoomed to the appropriate focal length, falls by the wayside. Not a factor at all anymore. So zoom the flash tighter. 

"I don't have my camera in front of me at the moment ... "

And that is part of the confusion here. You need to step through this, logically, with your camera and flash in your hands. 

CarlS

Neil,

Thanks for another prompt and in depth response.  It isn't that I haven't stepped through this logically with the camera in my hands, I was just stuck as a car passenger and didn't couldn't verify my assumption.  

In experimentation this morning, I see that I can manually position zoom head while keeping TTL-BL active.  I can go as tight as 105, but will experiment at 85 since it is one setting off of the extreme.

Regarding our previous discussion on how distance does, or does not, affect the TTL flash output, I found some supporting information that other readers might find interesting.  This site has loads of technical information on TTL (http://www.scantips.com/lights/ttlbl-d.html).  It will take me quite some time to make it through the whole site, but regarding the use of distance information from the lens:

The flash only uses distance information when the flash head is pointed directly at the subject.  It uses the distance as supporting information to make a guide number comparison while in BL mode.  The site has plenty of technical details, but essentially the distance information protects against over exposure resulting from too much flash output.  When using the flash out of horizontal or at an angle, the distance information becomes useless to the flash (As Neil mentioned a few posts up).  I can verify this by looking at the LCD on my SB-800 that displays distance only when the flash is pointed straight ahead.  As soon as I move the flash head to any other position, the distance display disappears.  Pretty neat!

Thank you again Neil.  Great site, great forum.  I (and I'm sure many others) appreciate you taking the time to help us understand this material.  

Carl

dbruno

This has been a really interesting and informative discussion, and I'm certainly glad I latched onto it.

biffy7

Want to see if distance has an effect on the pre-flash TTL? 

Scenario A

1. Set Up a bounce situation. 

2. Take the picture.

then...

Scenario B 

1. Stay in the same bounce setup. Don't change the camera distance to the subject.

2. Cover the lense with a "dark" filter (e.g., Neutral Density, polarizer, both).

3. Take the picture.

Notice how in Scenario B the flash output is much more than Scenario A. That's because the TTL system determined that more light is required for a proper exposure, AND the distance did not change.

Hope this helps.