What then would the TTL/BL do when you are overexposing the background, i.e. when you have a backlit subject outdoors, would it not underexpose horribly?
williamkazak wrote:
A very interesting discussion. It kind of suggests a starting point for making our own tests using TTL-BL. I am wondering what Desmond suggests for indoors/outdoors when things are moving along rather quickly for an event with people.
I would call TTL/BL the 'P' mode of flash - it tries to do the thinking for you so that when things are moving quickly you are less likely to blow the exposure .Of course no system is perfect but TTL/BL will bring you closer to a good starting point if you have time for adjustments as well .
By all means use whatever works for you - it's just good to understand all the modes if possible . TTL will need more compensation adjustments for varying ambient exposures .
bitmaker wrote:
Perhaps a better description would be "focus/out-of-focus" features within the scene.
Could you please do another test? Please shoot the round pouch against white background with your D90. Duplicate the exposure settings (including lens FL) you used with the shots done with the D40. Curious to see if the D90's 11 focus points will deliver different (as in: more averaged due to the 8 additional focus points) exposure evaluation data than the D40's 3 focus points.
I know this wasn't in exactly the same conditions but from what I have seen of the D40 metering it's almost as if the three focus points are elongated with regard to the way they meter - sort of stretched out rugby-ball shapes because they aren't affected by horizontal differences .
These were taken with the D40 and show how the metering along the focus points extends up and down the screen .
Beni wrote:
What then would the TTL/BL do when you are overexposing the background, i.e. when you have a backlit subject outdoors, would it not underexpose horribly?
How also do you dial in fill factor with TTL/BL?
I did a video regarding the metering mode you select and a back-lit subject
As you can see TTL/BL reacts to the metering mode you have selected and the decision it makes with regard to how well the ambient is exposed .
When you have the ambient correctly /over -exposed you don't want the flash pushing out too much light anyway - TTL/BL seems to work to preserve highlights by the fact that it meters for the brightest focus point receiving a pre-flash reading .
By the fact that matrix metering tends to decide on exposure quite strongly based on the active focus point I would say matrix will generally encourage a stronger output from TTL/BL if that's what you are after while centre weighted metering wil over-react to the backlighting and tell TTL/BL there's enough light .
Have a look at my tests of matrix metering with the D90 - no change in the scene - simply moving a focus point seems to suggest that "3D matrix metering" means that it exposes in favour of the focus points that light up on the same plane .
Avi B wrote:
No, this test should be redone with a D2 body or D3 body or something along those lines. Too bad you're not near, I'd loan you my D2H
BTW this testing seems pretty good to me.
Anything newer than a D100 should yield similar results - but remember that TTL/BL output is affected by a computer program in the camera body - and it is like xp vs vista where the next model doesn't always work exactly the same as the previous one .
DesmondD wrote:
Anything newer than a D100 should yield similar results - but remember that TTL/BL output is affected by a computer program in the camera body - and it is like xp vs vista where the next model doesn't always work exactly the same as the previous one .
It should be similar of course. I thought that the higher end bodies have more segments for metering, and may give finer control over the exposure.. Although frankly speaking, your findings are generically applicable and I guess I'm just splitting hairs
I doubt a vertical shot would change since it uses the focus points especially if you are doing close up head shots as you mention . Myself and someone who is a professional photographer have done a range of tests and both come to the conclusion that TTL/BL +1/3 flash exposure yields more consistent , pleasing ,results than TTL flash - the main variable being that TTL/BL is using all the focus points while TTL is metering for the centre of the frame .
Avi B wrote:
It should be similar of course. I thought that the higher end bodies have more segments for metering, and may give finer control over the exposure.. Although frankly speaking, your findings are generically applicable and I guess I'm just splitting hairs
Splitting hairs is what refines a theory . Mike , who is a professional photographer ,duplicated all my tests with his D3 and D300 and found that TTL flash varied between the two bodies while TTL/BL gave consistent results between them . He now shoots in TTL/BL+1/3 with both bodies because he knows he will get the same results from them that way .
So in a way it is good to ' split hairs ' in some aspects to put the finishing touches on a theory .
Outstanding report, Desmond. Your willingness to discuss and to share your findings will improve the entire Nikon community of photographers. I thank you. Constant practice and testing is very important in keeping up with changing technology.
The 3 pairs of shots in post "p.2 #5" are an excellent example of the i-TTL's difference between TTL (system uses flash sensor(s) only) and TTL/BL (system uses flash sensor(s) and the camera's light metering system), Desmond. In the first pair (Stinky's head) the TTL and TTL/BL delivered very close to the same flash exposure (the tiny bit more FE in the TTL shot could be from the FE sensor seeing and wanting to add exposure to the ambient background). In the second pair (Stinky's upper body) more of the room is in the shot, therefore the TTL system wanted to expose for the ambient background again. In this pair I actually prefer the TTL shot due to the white in his shirt being purer. The final pair really shows how TTL/BL handled that scene better. In the TTL exposure the ambient background made the system really crank up the flash output, to the extent that Stinky is starting to blow out. TTL/BL nailed the shot.
Since your pair of shots done with the D40 of the black-disc-on-white-background showed such a strong flash exposure shift in dealing with the change in DR when changing active focus brackets I decided to do the test using the D2X and D80 I have here. For my set-up the disc was cut from (almost) black construction paper and the white is a sheet of 1.5mm thick white HDPE I use for product photos. Of course, both cameras wanted to turn the white background gray. The jpgs are larger than I usually post but I wanted to make sure the shooting data was easier to read.
The D2X didn't show much difference in exposure between active focus on white and black. Don't know if this has to do with a difference in Matrix meter algorithms or the fact that the D2X has 2 flash sensors (I suspect it's more the former). The D80 showed only a bit more difference in flash exposures.
I found some of the results in post "p.2 #1" intriguing... specifically the pair of no-flash shots of the house in bright sun and shade. These shots are a good example of how even in ambient light the Matrix meter will change EV based upon where the active focus bracket is.
Was the D90's Active D-Lighting enabled for the house shots?
I knew the Matrix meter in my D2X brings "tweaks" to the EV based upon where the active focus sensor is (the primary reason I rarely use the Matrix meter in high DR shots), but the 2 f stop difference your D90 gave caught my attention. While I hadn't done this test before I have over time come to "feel" that in such scenes the D2X might have 1/3-2/3 stop difference when placing the active focus sensor on parts of the scene with alot of dynamic range. So I did an ambient high DR test as well, the results being only a 1 f atop difference in meter reading.
I had set-up to do this same test with the D80 but the difference in location of the focus brackets varied enough from the D2X that I would have had to make a drastic change in perspective. So, I simply adjusted the focus brackets and camera position of the D80 to check what its Matrix meter gave me for the roof and bush. 2 f stops, just like the house shots from the D90.
Personally, I prefer a Matrix meter that is weighted more for brightness and contrast than for the active focus bracket. I'm gonna go hug my D2X .
bitmaker wrote:
Personally, I prefer a Matrix meter that is weighted more for brightness and contrast than for the active focus bracket. I'm gonna go hug my D2X .
Greg
Thanks for doing this. Very helpful. I just went and hugged my D2Xs after your post
williamkazak wrote:
Outstanding report, Desmond. Your willingness to discuss and to share your findings will improve the entire Nikon community of photographers. I thank you. Constant practice and testing is very important in keeping up with changing technology.
Thanks , it helps me remember as well when I share my findings - along with opening myself to correction if I have made a mistake
bitmaker wrote:
The 3 pairs of shots in post "p.2 #5" are an excellent example of the i-TTL's difference between TTL (system uses flash sensor(s) only) and TTL/BL (system uses flash sensor(s) and the camera's light metering system), Desmond. In the first pair (Stinky's head) the TTL and TTL/BL delivered very close to the same flash exposure (the tiny bit more FE in the TTL shot could be from the FE sensor seeing and wanting to add exposure to the ambient background). In the second pair (Stinky's upper body) more of the room is in the shot, therefore the TTL system wanted to expose for the ambient background again. In this pair I actually prefer the TTL shot due to the white in his shirt being purer. The final pair really shows how TTL/BL handled that scene better. In the TTL exposure the ambient background made the system really crank up the flash output, to the extent that Stinky is starting to blow out. TTL/BL nailed the shot.
Since your pair of shots done with the D40 of the black-disc-on-white-background showed such a strong flash exposure shift in dealing with the change in DR when changing active focus brackets I decided to do the test using the D2X and D80 I have here. For my set-up the disc was cut from (almost) black construction paper and the white is a sheet of 1.5mm thick white HDPE I use for product photos. Of course, both cameras wanted to turn the white background gray. The jpgs are larger than I usually post but I wanted to make sure the shooting data was easier to read.
Thanks for the input - good stuff ! I have found that the focus mode I use also changes the results - I prefer 'dynamic auto-focus' where you select the focus point but the others are still working to help the active focus point , and that seems to induce the 'expose for focus point' concept in the camera - as I've mentioned in the past , each camera's matrix metering is a computer program and they differ like xp and vista sometimes .
It seems that matrix metering 'exposes for shadows' while TTL/BL 'works to preserve highlights' so perhaps they complement each other in a way . The trouble is that matrix metering often overdoes it and blows the highlights . When I get home from work I'll upload my metering test shots .
DesmondD wrote:
I prefer 'dynamic auto-focus' where you select the focus point but the others are still working to help the active focus point , and that seems to induce the 'expose for focus point' concept in the camera.
I rarely use dynamic AF... never had consistent focus results with the action stuff I do. Tomorrow I'll configure the D2X for Dynamic AF and go back to check the EV change on the composition of the house/yard.
bitmaker wrote:
I rarely use dynamic AF... never had consistent focus results with the action stuff I do. Tomorrow I'll configure the D2X for Dynamic AF and go back to check the EV change on the composition of the house/yard.
I never really do any action stuff but like to choose my focus point which is why I use it .
The trouble is with landscapes and matrix metering exposing for the shadows = blown highlights . After a series of tests I use matrix-2 as a starting point for high-contrast landscape shots .
