TTL/BL tests , my results so far .

Discussion in 'General flash photography, lighting, and technique' started by Desmond, Apr 30, 2009.

  1. I though these results deserved a thread of their own since many ask about TTL vs TTL/BL .
    I'll start off by saying that I agree with anyone who says you should first learn to master TTL flash since that may be your only option at some stage .
    But so far TTL/BL has proved to be superior "On my D40, 18-105mm and SB800" , I'm not saying 'yet' that it will be so on every system since TTL flash does exactly the same as TTL metering on my D40 .

    I'll start off by posting the results that surprised me most tonight , while taking a picture in the [dirty] mirror to see how my flash diffuser behaved I decided to compare the two systems just for fun .

    [​IMG][​IMG]

    And just a few of the other 'exteme ' situation tests I've been doing - indoors , no ambient .

    [​IMG][​IMG]

    [​IMG][​IMG]

    [​IMG][​IMG]

    In general , with my D40 , 18-105mm and SB800 , TTL flash has been as erratic as TTL metering in the camera while TTL/BL has been very consistent in its results .
    TTL/BL has fired weaker in bright light , as it is meant to do because it is designed to take the ambient [background lighting ] into account when it decides how much flash power to put out , and stronger with a white background indoors with little ambient . So far I would consider both of these scenarios to be superior to TTL flash [on the D40] .
    When the subject is indoors with the background further away TTL/BL fires slightly weaker . Once we zoom in to the subject they look very similar with TTL/BL being very slightly weaker .

    [​IMG][​IMG]
     
    Last edited by a moderator: May 15, 2009
  2. WayneF

    WayneF

    Apr 3, 2006
    Texas
    Desmond, you seem to be using TTL BL mode indoors in dim light? BL mode is of course designed for use for fill outdoors in more usable bright light (or where there is something to be balanced). I would differ somewhat on the conclusions, when in outdoor situations with brighter lighting.

    http://www.scantips.com/lights/bl.html was about investigating if commander/remote uses TTL BL mode or not (it seems to), but it includes TTL and TTL BL, and includes a more diverse set of lighting situations (dimmer, brighter, backlighting, and none).

    See the last section (3/4 down page) about Thom Hogans comment on models older than the D200.
     
  3. I'll have a read of that when I get home tonight . The tests I did outdoors were also more pleasing though I am waiting for a sunny day for the 'extreme' test . If it takes the ambient into account when firing then when there is no ambient it should still produce good results and slowly back off as the ambient increases which is what you would want .
    The point I'm trying to make is that regardless of its specific purpose - it is giving better results indoors than TTL [ on the D40 and 18-105 combination]
     
  4. Tonight I tried something very different to try and work out how the TTL/BL does better with a white background and maybe I have cracked it this time . The most accurate way for flash to work would be to fire according to distance and not reflected light.
    I set it up with the reflector and black velvet again and tried two tests , one with normal focus and then another one with the focus set manually so that I could see how much distance affected both systems . It would appear that TTL didn't know the difference but TTL/BL reacted very strongly to the greater distance it thought the subject was at !
    I got this idea after reading that the Tamron 17-50 F2.8 can give incorrect results with TTL/BL because of incorrect distance info sent to the flash and wondered why it wouldn't also bother TTl if it also supposedly needed distance information .


    [​IMG][​IMG]

    And now the results that show TTL/BL relies very strongly on a lens that can send the correct distance info to the flash ...

    [​IMG][​IMG]
     
    Last edited by a moderator: May 1, 2009
  5. Hi Wayne , I just read that report and their conclusions are not quite corect compared to what I have found in my tests :smile:
    I see they used the D70 for the indoor shots - it is known to underexpose .
    They also made the same assumption that I originally made , that TTL/BL concentrates on the centre of the frame - this has proven to be incorrect .
    My tests have now shown that TTL/BL uses distance while TTL had no change when the distance was manually changed .
    The site mentions that the built in speedlight always uses TTL/BL which would make sense because it is proving to be more acurate than TTL .
    I don't quite agree with the statement that TTL/BL has no use when there is no ambient because if it is programmed properly as it seems to be with my D40 , zero ambient will mean full flash power according to distance for the subject and as the lighting increases it will back off accordingly .
    A 'proper' algorithm would have a starting point for ambient [zero] and progressively increase the compensation from that point .

    I have done some "overcast" tests but still need a sunny day [ hopefully this weekend ] to do the ''harsh lighting '' tests .

    [​IMG]

    [​IMG]
     
  6. WayneF

    WayneF

    Apr 3, 2006
    Texas
    Since that is my report, then I will defend it. :smile: I thought it reported the facts to be much more obvious than that. But I think it did assume we realized that TTL BL was a fill flash system.

    "Known" in TTL BL mode in incorrect situations for TTL BL perhaps, but it is not just the D70 that screws up when attempting to use TTL BL in insignificant ambient. Users who know, who have actually seen and reported on many models say it is the "older models", specifically since before the D200, and then specifically, when using TTL BL in conditions of insignificant ambient (which like it or not Desmond, is a no no). I repeated their words, and seem able to duplicate that result. The newer models have since been made a little smarter to more quickly ignore any insignificant ambient when the unaware user attempts to try to balance it anyway. Otherwise, how exactly would that balancing work anyway? :smile:

    I normally do use Center Metering myself, but which is a different issue. The belief that TTL flash only meters a center spot was true in previous TTL technology, but is not true of i-TTL. i-TTL uses the same prism meter sensors as non-flash, which is 420 point metering on the D40. This is 1005 segments for the D300. It is full frame (or rather, can be full frame), not just center spot.

    D40 page 120 says: "TTL: TTL flash control by 420-segment RGB sensor. i-TTL balanced fill flash for digital SLR and standard i-TTL fill flash for digital SLR available when CPU lens is used with built-it flash, SB-400, SB-600, SB-800."

    Note its words are: "balanced fill flash" and word CPU. The BL concept is Fill flash Desmond. It will help to work with that system.

    TTL BL "Requires" use of a CPU lens, which is limited to the range of 18-300 mm (See D40 manual page 98).

    Obviously not true. :smile: We gave up guide numbers to instead actually meter things to know exactly how much light is actually in play. My own notion is that the lens distance information is used to help the system determine where (which pixels?) is the subject as opposed to what is the background. It is to determine which light should be metered, so to speak. How real is it to rack out the lens focus to 20 meters as a way to compensate our flash picture at 2 meters? :smile: You specifically told the camera that your important subject was literally at 20 meters, and the camera said OK, Sure, since it had no clue anyway.

    TTL BL is fill flash, with the goal to monitor and balance both subject and background, but of course, it has absolutely no clue what or where is our subject. It does not have a brain to know what it is looking at, and I do think it tends to assume the subject is either centered pixels or close pixels, but it can meter all the frame. The TTL metering is still very much key for flash, but TTL BL does use distance to try to figure out what light (pixels) is the subject, as opposed to what is background background.


    Perhaps if you pay attention to Nikons words more closely?

    Nikon makes absolutely no secret of that TTL BL is a fill concept (page 117 D40 manual):

    "i-TTL Balanced Fill Flash for Digital SLR: Flash output is adjusted for a natural balance between the main subject and the background.

    Standard i-TTL Fill Flash for Digitial SLR: Flash output is adjusted for the main subjec; the brightness of the background is not taken into account."

    This seems quite clear. Note their use of their word Background and their word Fill Flash. They appear to be very careful with their exact words. This result should be no surprise, and my clear opinion is that you are trying to use the system wrong. Which is your right, but also consider working with the system Desmond. :smile:

    As to "accurate exposure flash", there really is no such concept. Any reflective metering depends entirely on the amount of light reflected from the subject. This does depend on the amount of light incident on the subject, but it also very much depends on the degree of reflectivity of the subject itself. If you meter a white background, you WILL GET underexposure. If you meter a black background, you WILL GET overexposure (and you will get middle gray for both, no matter how much light was on it). Just how life is. Knowledge is in knowing to expect this. We need to know a few basics about how things work. See the metering link below.

    So, yes, we do always need to watch and compensate our pictures, especially flash pictures. Those who do get great results. Those who don't are always mystified. You seem to want to use some other BL trick (yes, which may try to boost an insignificant background, except limits often prevent it), instead of the straight forward watching and compensating. Your porch picture obviously should have been BL of course, that is exactly a balanced fill flash picture. I am wondering why the BL nearest porch post is not bright? It must have though this near thing was the subject? We have to balance this fill manually ourself in TTL flash mode, and you failed to do that. -1.7EV flash compenastion is often a good starting point.

    "i-TTL Balanced Fill Flash for Digital SLR: Flash output is adjusted for a natural balance between the main subject and the background.

    Standard i-TTL Fill Flash for Digitial SLR: Flash output is adjusted for the main subjec; the brightness of the background is not taken into account."

    Also page 37 SB-800 manual, page 33 SB-600 manual.

    Why is that context such a problem to accept? These are the tools provided by Nikon. It just has to be helpful to realize that.

    As for metering flash, an incident flash meter can be instantly gratifying, but it does require a manual flash system and some extra time. It is NOT for TTL, not for Commander, and it is not instant.
     
    Last edited by a moderator: May 1, 2009
  7. Fair enough on the older models , none of which I use , it is a no-no to use TTl/BL . I'm using a D40 in my tests and will be doing the same tests with my D90 .
    At the moment I am taking the manual as a guide only , regardless of what they say about when we should use TTL/BL I am getting better results from it in all situations [maybe slightly darker results in bounce use which I'll admit I will be using most of the time ]
    With regard to zooming out to 20 metres it wasn't to control the flash , it was to prove that TTL/BL uses distance more in its calculations while TTL never seemed to notice the distance variation .
    I've reached the stage where I no longer let the manual misguide me without doing my own tests . I don't really care how many times Nikon says I must only use TTL/BL outdoors for fill flash , you can't argue with the superior results indoors which is all that counts to me .

    [​IMG]
    [​IMG]

    When you get closer , and there is very low ambient , the results are almost identical so who cares what the manual says if you can stay in one flash mode ? [ I'm not saying I will until I have done more tests with the D90 ]
    Look at the results in the mirror shot , even in the shots of the black and white test "bride and groom outfits" with almost no ambient . TTL/BL with its disance calculations produces better whites than TTL .

    [​IMG][​IMG]

    [​IMG][​IMG]

    [​IMG][​IMG]

    The manual is a guide for beginners , tests are the real deal . The scenario above is an extreme situation which shows that under pressure TTL is simply a reflective reading while TTL/BL uses the distance and doesn't blow the whites or underexpose them . This was a situation where there was very little ambient - just enough for auto focus to work .
     
    Last edited by a moderator: May 2, 2009
  8. NateS

    NateS

    Oct 11, 2007
    Missouri
    I've noticed that when using matrix metering and flash in ttl I always get WAY too much flash. Switch from Matrix to center or spot metering and you will get nearly perfect flashes every single time with the TTL setting. TTL-BL only gives me an advantage (as you are seeing) when in matrix metering.

    Of course, I never do photos of people in matrix metering because I don't care what the background is exposed to (well I do, but I control it so it's a stop underexposed) so I only want the center or spot (main subject) to be metered correctly. With or without my flash I never shoot in matrix for people. It sounds like you do and that you must be in matrix metering all the time. I'd suggest you venture into the world of center weighted or spot metering with your flash, full manual and expose the scene a stop or to underexposed, then use flash to light the subject. You'll get much nicer photos and you will be able to use TTL consistently. Of course, if you prefer your way, that's fine, but doesn't hurt to mess around with ideas.
     
  9. WayneF

    WayneF

    Apr 3, 2006
    Texas

    Yes, there are so many variables. One size is not going to fit all, we need to pay attention, and understand some basics. :smile:

    I fully agree with you about Center metering. No one knows what Matrix is going to do. There is much to be said for simple direct overt compensation, which I can control, without an unknown hand stirring the pot.

    Possible Variables:

    TTL vs TTL BL

    Center metering or Matrix metering

    Bright backgrounds or insignificantly lighted ones.

    Back lighted or front lighted subject (large affect).

    Black vs white backgrounds of course, always a biggy.

    Actual subject filling most of the frame, instead of just a tiny area in it.

    Camera model

    And user notions too. :smile:
     
  10. NateS

    NateS

    Oct 11, 2007
    Missouri
    I agree completely and the following are MUCH less of an issue usually when using center metering:

    Bright backgrounds or insignificantly lighted ones.

    Back lighted or front lighted subject (large affect).

    Black vs white backgrounds of course, always a biggy.

    ...because the camera isn't as interested in what is behind the subject. Matrix metering for this type of shooting takes way too much control out of the photographer's hands and I'm definitely not a big fan of that.
     
  11. Tonight I decided to do the same tests with my D90 and the Tamron 17-50mm F2.8 because apparently there have been issues with this particular lens giving incorrect distance readings for the TTL/BL system .
    The results were very similar to the D40 results . Matrix metering made the same mistake with the white and black as before . [centre weighted would make this same mistake in such an extreme situation ] . I've just posted the two shots without flash because as with the D40 , the results from TTL flash looked exactly the same with regard to the histogram [ inaccurate ]
    [​IMG][​IMG]

    Then I switched to TTL/BL [SB800] and got the same results as with the D40 and 18-105mm lens . TTL/BL has produced superior results indoors in all my tests so far . It has used distance in its calculations and ignored the effects of too much white and too much black in each situation .

    [​IMG][​IMG]

    [​IMG][​IMG]

    [​IMG][​IMG]

    These results look less extreme than those from the D40 which meters very strongly for its centre focus point . It has handled the extra white in the scene quite well . But as with matrix metering when there is a lot of black in the scene matrix metering , and TTL flash which relies heavily on matrix metering , blows the whites while TTL/BL has maintained a fairly constant exposure .
    Just to demonstrate the effects of focus distance on TTL/BL [ something which doesn't affect TTL ] I first manually changed the lens to close up focus .....

    [​IMG]

    And then focused on 20m ......

    [​IMG]

    This is not to suggest using focus adjustements to control flash exposure , just to demonstrate the effect distance information has on TTL/BL and the importance of being sure that a lens you use in this mode gives the right information to the flash system .
    You may master TTL/BL and then change lenses and get the wrong results so check that it works properly with each lens you use .
     
  12. Now that I know distance affects TTL/BL flash but not TTL flash I decided to set up a test for the two systems to see which of them handled a tricky situation better .
    Granted , we may never come up against this scenario but if we do which system handles it best ? TTL/BL came out on top once again .

    Not too much difference when the subject is fairly close to a white 'wall' ....
    [​IMG] [​IMG]

    But as the distance of the subject increases TTL is confused by the white 'wall' while TTL/BL concentrates on the distance of the subject and ignores the white 'wall' , continuing to expose the subject correctly .

    [​IMG][​IMG]

    [​IMG][​IMG]

    The background is getting brighter with TTL/BL because the subject is getting closer to the wall at the back , TTL has been totally confused by the reflector .
     
  13. daniyarm

    daniyarm

    500
    Jan 2, 2009
    Tucson, AZ
    Based on my limited experience with SB-900/D300 combo, I much prefer TTL/BL for indoor shots. TTL for me darkens the background way to much, although I haven't used it for macro yet.
     
  14. WayneF

    WayneF

    Apr 3, 2006
    Texas
    It is neat that TTL BL uses the D lens focus distance to help determine subject. It surely can be helpful in that situation, but there are many different situations of flash pictures.

    One common opinion is that a hot shoe flash indoors ought to be bounced at every opportunity. Most people willfully shun direct flash after they discover bounce, except when the no-ceiling situation forces it (and not all people ever discover bounce). Are you going to keep using hot shoe direct flash all your life?

    So try a few other situations Desmond, try bounce flash (warning, this is a trick).

    There are only four things that affect exposure of our flash pictures... ISO, aperture, shutter speed, and flash power level. Shutter speed is not a factor for flash, but it is a factor for the ambient portion of our balanced pictures. For outdoor daylight fill flash, we probably use camera P mode, or maybe A mode, which gets the ambient daylight exposure close. The TL BL flash mode matches that ambient exposure with appropriate flash power, for a nice balance of the two light sources. This is the definition and the entire idea of TTL BL. Good stuff too, in its place. This is pretty easy to do manually with TTL flash mode of course, but TTL BL adds a degree of automation.

    But indoors, the ambient room light is normally insignificant, which is normally why we are using flash. There typically exists nothing to be balanced. Setting TTL BL mode is doomed to fail, it cannot possibly work. There is no ambient light, and the manual camera mode will not permit aperture/shutter adjustments to match it anyway. The best we can hope for is that the newer models of cameras are smarter about giving it up more quickly in hopeless inappropriate situations.

    So indoors, we often set camera Manual mode, which determines three of these factors (or alternately, automation limitations set them), leaving only flash power level to contend with. Indoors, the flash is the only light source of any significance, and flash power level is the control for it (and aperture and ISO of course, but those are set now, so for TTL flash, overtly the control is flash power level).

    So when automation does not get this exposure just right (and it is a very difficult problem which only the human brain can really recognize and compensate), we learn to help it with a bit of flash compensation. Learning that automation cannot always get it right is normally our first clue, and learning flash compensation is very often necessary (and always should be verified), is the very first step of learning to use our flash with success. No big deal, we merely observe what is happening, and help a little as necessary. It is the same deal that we do for either TTL or TTL BL compensation... no difference, just a little tweak. All either flash mode can do is to provide one specific flash power level which will determine the entire picture (the hardest part is to realize how easy it is). Flash power level is not affecting the subject and background differently, not at all. That is to say that it is the one same flash power level for everything in the scene. It is true this level can be correct at only one distance, and wrong at all other distances, but it is the same one flash level which we must get set right.

    Anyway, as we learn more, we learn that bounce is good. And umbrellas are good. Mainly, just getting the flash off of the camera is good. We learn many things, including that hot shoe direct flash is not our best tool, and that compensation is one of our best tools.

    Unfortunately, this distance factor you have discovered does not work for bounce flash, or off camera flash like Commander/Remote mode. The flash must be on the hot shoe with flash head straight forward ahead level for this lens distance info to work.

    Which is sort of the pits Desmond. :smile: We need more tools in our bag. The foremost of those tools for flash is to be proficient at flash compensation, to know how to get the results we want in any situation. This is quite easy, we simply watch what is happening, and tweak it a bit (and we come to know what to expect in advance too). Realizing that this is true seems difficult for many.
     
    Last edited by a moderator: May 2, 2009
  15. Maybe my D40 is smart enough to make TTL/BL work ?
    TTL direct compared to TTL bounce flash ....
    D40 , 18-105 lens B800 .
    [​IMG]

    [​IMG]

    TTL/BL direct flash compared to TTL/BL bounce flash ....

    [​IMG]

    [​IMG]


    I totally agree with you that we should be using bounce flash whenever possible but after doing my own tests , having read many other opinions on the subject I have proved , that for me at least -" TTL/BL gives me superior results to TTL flash when there is nothing to bounce off regardless of whether there is ambient or not "

    So far I have seen no evidence that low ambient light affects how well it works because most of my indoor shots have been in the dimmest light I could ever have to work in - and it has still proved superior to TTL .
    I feel I have done enough tests to be able to rely on TTL/BL in any direct flash situation and am only starting on scenarios where I will be bouncing flash so as the title of this thread states these are my results "so far" - not a final opinion .
    I'm going outside now - it's a nice sunny day for some flash tests :smile:
     
  16. WayneF

    WayneF

    Apr 3, 2006
    Texas
    OK, there always many ways to work, but FWIW, I fail to see the advantage you must perceive in it.

    For example, my first thought was that your two direct flash images 0030 and 0031 might have been a mistake - looks like the same image was inserted and shown for both... No perceivable difference, but the file name is different.

    The two bounce images are a bit different. Both are underexposed about one stop, but the TTL image slightly less so than the BL image. If we are rating them, I would rate it higher for that reason. But I realize it is a matter of compensation, which I understand you were not doing here (to show results).

    My point is that one flash compensation fixes either one, sort of a moot point which is chosen. Except one mode makes more sense logically, since you cannot explain balancing ambient here. :smile:

    I believe the D40 may have been slightly later than the D200 point of reference regarding BL algorithm shift. Which is good for BL, it should know more about when to give it up. And certainly the D90 is later.
     
    Last edited by a moderator: May 3, 2009
  17. glumetu

    glumetu

    499
    Apr 3, 2008
    Romania
    Sorry to bump in, but I read on several occasions that TTL/BL should be avoided.

    Why not let normal flash + slow sync (and the usual compensation on both camera and flash) work their magic in getting ambient light into the flashed frame?

    Just trying to learn, I don't really have much experience with flash (other than the pop-up :) )
     
  18. WayneF

    WayneF

    Apr 3, 2006
    Texas
    Yes, slow sync would allow slow shutters in dim settings to pick up more ambient. And of course, it seems simpler that camera Manual mode does that too, with more control, since it will allow any shutter speed we wish, and Slow sync only allows metered shutter speeds (which could be a plus too of course).

    Either way, very often, for pictures of people who move, we may not like shutter speeds down around say 1/2 second. We are probably more likely to just call this dim light insignificant and forget it (and pick a faster shutter less likely to offend).

    Regarding stopping motion, the flash is certainly very fast, fast enough to be used for high speed photography, like milk drops or hummingbird wings. But we have to keep the continuous light out of that for it to be true.. Otherwise, a slow shutter can still blur it, as seen in the continuous light.
     
  19. I'm no expert on the subject but one thing I've learned is not to believe everything you read , rather do tests and see the results for yourself .
    I'm not stating that TTL/BL is superior in every situation , just in almost every situation I have tried it [ mainly diect flash which is supposed to be avoided but sometimes is unavoidable] .
    My thoughts on using slow synch and compensation is 'whatever works for you and gets the results you want ' .
    If someone is proficient in using normal TTL flash and using the relative compensation then there isn't much need to use another mode .
    I'll post my results for sunlight later , obviously TTL/BL is superior in the situation it is designed for in these results but plain TTL is as clever or dumb as your cameras metering mode which is probably why many people us the more predictable centre weighted .
    With my latest tests outside [ back lit sun ] TTL fired full power from 2 meters away from the subject and still complained that it was -0.3 under exposed , in this case I would say why worry with TTL and compensation when TTL/BL does it automatically .
    But generally it is personal preference , one guy sets everything to manual and ends up with F5.6 13/00th iso 400 indoors , another guy dials in compensation in aperture priority mode and uses and slow synch flash and gets the same results - the customer doesn't know or care how they got the results as long as they are right .
     
  20. WayneF

    WayneF

    Apr 3, 2006
    Texas
    Yes, amen, testing and thinking it out are certainly big pluses. Only way we are going to figure out most of it.

    But when there is a computer involved deciding on how things will be done, it is always good to listen the explanation of those who wrote that software. :smile: Nikon actually tells us very little, but they do assume to guide us.