Is PTTL worthless?

OK, maybe someone can tell me if I'm right on this, or if I or my equipment is at fault...

To start, I have the AF540FGZ (same goes for the 360) mounted on my K10D (similar behavior on my istD).

Now, when I take a shot at something, oh lets say 4 meters or so away the subject appears to be exposed properly.
But once I get closer, lets say 2 meters the subject is often majorly underexposed. By about two to three stops!!

I did some tests for about an hour before I think I found the reason....

The camera is focusing on a rather dark subject with a relatively light background.
The camera is exposing for the subject and possibly nearby surrounding areas - whatever is in view, just to make things simple. It is expecting the flash to compensate to some degree.
The Flash is exposing for some tiny dot straight out from the flash sensor which apparently is seeing the bright background and is TOTALLY missing the darker subject I am trying to photograph. So it independently meters to a different exposure than the camera.

WHY!!???!?!?!?!?
Is that the "blessing" of P-TTL v.s. TTL?
ARGH!! and Double ARGH!!!!

As I understand it, the camera does all the 'exposing', so it shouldn't be that the camera thinks one thing and the flash another.

Do you get similar results with the built in flash?

Dan

Built in flash works perfectly, if with a bit less light. But it is consistent and predictable.

I wouldn't mind adjusting the 540FGZ + a stop as necessary, but it is totally unpredictable and I often loose a once time shot.

P-TTL is very good most of the time, but you have to beware of highly reflective items in the frame. When a bright light (the flash) bounces off them, it can throw out the metering in a big way.

Mirrors, polished surfaces, anything highly reflective, can make a huge difference.

G

Yeah, I understand that, but it is not an issue with specular highlights or overly light areas.

Try this. With the flash on the camera. Take a shot. Now cover the entire red area with your hand and take another shot. The exposure will be way different. Not exactly the same thing, but I think it shows that the camera is not involved in the metering of the flash.

FYI, the particular test I had was of my dark haired dog on a dark carpet.
Of course he lays in the corner next to the off-white wall.
The camera viewed only the dark carpet and dog. Very little strip of the white wall behind near the very top of the frame.
The exposure was two stops below what I expected.

I aimed the camera down slightly so that I missed most of the dog, and the shot was exposused much more as I would have expected.

So it seem clear that the camera is seeing and metering separately from the flash.

I guess now that I know this I can try to look out for it, but it is a major irritation. I wish there was some way to avoid it, but it is likelya hardware issue and not avoidable.

Argh!

I'll double check this with my AF360 and AF540... but bear this in mind:

You may be hitting the minimum distance for the flash. The external guns are 3x and 5x more powerful than the popup. Now you might think they can "throttle down" to nothing, but that's not the case.
I reckon what's happening is this: preflash fires, camera says to itself "whoa that's bright", and throttles the flash back to nothing because even 1/32 power would be too bright.

Now that may be off track, but I've been caught by it with TTL and the Z1 with the AF360

If you want to go really close, I suggest you point the flash straight up, and flip out the catch light panel. That'll cut the power down by a looooong way and it should give just enough light for close work (either that or use wireless P-TTL).

But, as I said, I'll double check when I have a minute
Matt

are you absolutely sure you are set to PTTL mode? this sounds more like the result one would expect from a flash in plain auto mode to me, which I suppose is your point, but the whole point of TTL and PTTL is the camera does all the metering.

Hmmm, I'll have to play with it a bit more. I only changed the angle of the camera between awful and great image. The change in distance would be minimal.
But you may be on to something where the flash is reflecting more off the brighter wall behind. But still, I thought that was what pre-flash was all about... so the camera could tell more about the environment during the flash and adjust accordingly.

I am still confused but I'll have another go at it and see if I can better narrow down what is going on.

I've just tried it on my K10D and I can't make it go wrong. Holding my hand in front of the AF360's red area makes absolutely no difference.

(I did also notice that it's tricky to put the flash in "A" mode. You can do it though, by powering off the K10D, setting the "A" mode, then powering the camera - the mode seems to stick. Now that may be useful...)

I would double check that the flash actually says "P-TTL" on the LCD.
Incidentally, I was using the K10D in Av mode, with ISO100 forced, and tried f/8 at about 0.5m taking a shot of a book.

(You may also want to check which version of firmware you have. I have the latest 1.20. I know that Pentax updated flash control in the early days with the *ist-D when the exposure was all over the show, and they fixed it without declaring it on the version history.)

HTH!
Matt

OK, here is a link to an example I just did. (not the same circumstances as desribed earlier, but must be close enough). I know the center spot is a little off, but I cannot see anything there that would cause such a large difference in exposure.


All settings are identical and taken a few seconds apart. I merely changed angle of the camera a few degrees and tried not to change position as much as possible.

Spot metering (although I've seen similar behavior in both center and multi-segment metering.
1/60s, f5.6 at 50mm (18-55mm lens).

I was about 1-2 meters away. Although I have seen similar behavior up to 4 meters away.

To me it looks as though the flash didn't fire on the first shot, but it's hard to tell from small pictures.

G

Exactly. Although it did fire, I saw the flash. When the flash does not fire the image is almost totally black. (When there is no flash on camera, the exposure is compensated sufficiently, but shutter would be too slow for shake reduction to do much good.)

Anyway, sorry about the small images. I thought they would sufficiently indicate what I was trying to say. Larger ones are here if interested:

http://moringello-east.no-ip.com:8080/test/_IGP3431.jpg
http://moringello-east.no-ip.com:8080/test/_IGP3432.jpg

looks like a problem Gerri Gets from time to time, when she doesn't have the flash properly seated in the hotshoe.....maybe a repair is in order if checking the contacts doesn't fix.
On the positive, at your dog looks cute. I don't think mine has ever stayed still for thirteen minutes in her life. My three year old isn't even sure what she looks like...she's just this blurred, black whiriling dirvish, that blows through the house once in a while....

Yes, the flash did fire, you can see the shadow behind and to the right of the dog's head when you boost brightness and contrast.

The only difference I can see, and I wouldn't have thought it would make that much difference, is that you're getting more of a reflection off the chrome ring on the dog's collar on the dark shot. You could try covering that up and see if you get the same results. The dog's fur is covering it more in the brighter shot.

Seems to me though that a reflection from that wouldn't be bright enough, but you never know...

Dan