[jarnos]

I took some test shots using same settings except some shots were made in AF-S focus mode and other in MF: ISO 200, multi-segment metering, +1.0 flash exposure compensation, 1/180 s exposure time, F5.6. I got different kind of results. Some of them (that were shot using AF-S), were significantly lighter than most of the others. The object is a white door. Sample images are in the web temporarily:



I hope the inconsistency is not due I used Starblitz 3600BTZ TWIN flash with the camera few times. I later measured its trigger voltage 53V (and temporarily significantly higher). I wish I had been warned about the possible too high voltage beforehand. On the other hand, I have not heard anybody had broken his/her DSLR by too high flash trigger voltage. Can anybody else make similar test shots by K10D?

I hope external flashes give more consistent results in P-TTL mode.

[stewart_photo]

It very much depends on where the camera meters and how much a strong reflection from the flash is included in that. Even if you were to mount the camera on a tripod, there would be variations since the camera considers each image individually - not always metering on the exact same spot on a mostly similar object (such as your white door). Add something specific in the image (a clear subject) for the camera to meter on and you should see more consistent results.

stewart

[RiceHigh]

Consistency and accuracy are not the strong point when it comes to P-TTL. The old Pentax TTL auto flash was, but the P-TTL is not.

As for the test you asked for, you can have a look at mine:-

P-TTL Vs TTL

As for the test you did, your door probably a white one and many flash systems might revert it to a grey one by underexpose it anyway.

I took some test shots using same settings except some shots were made in AF-S focus mode and other in MF: ISO 200, multi-segment metering, +1.0 flash exposure compensation, 1/180 s exposure time, F5.6. I got different kind of results. Some of them (that were shot using AF-S), were significantly lighter than most of the others. The object is a white door. Sample images are in the web temporarily:



I hope the inconsistency is not due I used Starblitz 3600BTZ TWIN flash with the camera few times. I later measured its trigger voltage 53V (and temporarily significantly higher). I wish I had been warned about the possible too high voltage beforehand. On the other hand, I have not heard anybody had broken his/her DSLR by too high flash trigger voltage. Can anybody else make similar test shots by K10D?

I hope external flashes give more consistent results in P-TTL mode.

[roscot]

For once I agree with Rice. My experiences with P-TTL have ranged from scratching my head and wondering what happened, to pulling my hair out when it completely blows a shot. I consider any good results I get in difficult lighting situations, a matter of luck.

What Stewart says is quite true. Any bright light or reflections will fool the system. Using a glossy white painted garage door for testing will guaranty a problem. Find a more normal scene to test in.

I find P-TTL works well within a limited set of circumstances. I no longer use it For more difficult situations. I use a combination of my Sigma 500DG Super on manual, or my Vivitar 285hv in auto mode or manual, and I'm getting much better and consistent results.

All this being said, I have seen work by people who get good results with P-TTL, but I don't understand how they do it.

If your Starblitz flashes had caused any issues, your camera would have fried by now. Just don't use them again without a Wein Safe Sync adapter.

[jarnos]

RiceHigh, as I mentioned the door is white, but exposure compensation was set to maximum i.e. +1 EV. Actually +2 EV would be better when you measure from white object. I wonder why Pentax has limited it to +1 EV. Thanks for the link to your test.

Roscot, thanks for the hint about Wein Safe Sync adapter. It is true, that light reflects brightly at certain small area of the white painted door in the test, and small change of angle of view could have big effect on the brightest area; I was not using a tripod. Another example is to shoot yourself in mirror using flash; your flash reflector will be about the only thing visible in the photo. I think there should be an option to let camera determine the exposure not by the lightest area of the preflashed view, but so that most of the image area has proper tones, as there may be highly reflective small areas in a random scene. That would not work well, if you shoot relatively small (angle of view) object in dark when there is nothing near behind the object.