[mcgregni]

It may just be then too many variables for the metering to cope with, for example extra light spill around the subject caused by the unusual bounce set-up. Generally I'd say you are best using an off -camera slave near the subject, with the light distribution and quality controlled with a modifier, such as umbrella or soft-box.

Try placing your flash right beside the camera, then bouncing, so that the light path is similar to bounce from an on-camera flash ... Make sure that your subject dominates in the frame, and that no direct light hits the subject (so for example pointing the flash 90deg from the front, or even slightly behind you). It will be interesting to see if this more usual bounce position and light distribution improves the metering.

[clackers]

You understood correctly. The flash was used off camera in slave mode and bouced off a wall. I totally agree that is more usual to bounce from on camera flash, but since the 360 v1 doesn't have a swivel head, it's not possible to bounce it off the wall, which was what I wanted to try... However, I've used it many time bounced off the ceiling with nice results. Although I'm often wondering if I would be better off buying a flash with a swivel head, maybe a cheap Bolt, or a Metz 52... Since I can't made up my mind, I stick with the 360, which does a great job most of the time. The Bolt flash looks good on paper, but feedback from users is almost nonexistent, and although it's cheap I don't feel like taking a bet on it.

This is getting to be an eccentric setup, Carl.

Shoot manual if there's anything unusual sbout your setup, environment or subject.

[CarlJF]

Okay guys, I've made a few more test. To make things clear, here's what the setup looks like:



The diagram is not to scale, but you should get the idea. It might be somewhat unusual, but certainly not funky. It's the basic 45° light setup and I could easily have replaced the wall by a reflective umbrella or other reflector.The flash is linked to the camera with a P-TTL cord so to avoid any problem with the wireless mode, It is set as master with no exposure compensation.

No, with this setup, with the flash in P-TTL, here's what I get:

DA55-300, 55mm (83mm eq.), F5.6, 1/160


I don't think I need to show you the histogram to prove you that it's heavily underexposed and not even near acceptable exposure. In fact, it's not much better than without flash at all. Now, I've tried Av mode, master, slave mode in wireless, different focal lenght or aperture, all with similar results.

Now, if I use the flash in manual mode, 1/2 power, I rather get this:


Which is exposed just fine and proving that good exposure isn't impossible to achieve with this setup.

And now, here's the interesting part. If I changed the K-500 for my Q7 linked to the flash with the P_TTL cord and the flash set on P-TTL, and here's what i got:

Q7, 06 lens, 17mm (79mm eq.), 1/160, F5.6


Exposure is spot on! I've put back the K-500 and got again the underexposed picture, and back again to good exposure when I plug in the Q7. So, it seems there's something wrong in the way the K-500, or at least mine, handles bounce flash. I've no idea what it is, or if it's the same with all Pentax DSLR (particularly the very similar K-50 and K-30), but there's obviously something wrong... I might try to reset the camera and see what happens...

[mcgregni]

So the Q7 exposes OK in P-TTL with exactly the same set-up and settings? Then it seems like the K500 has a problem .....

[clackers]

I've no idea what it is, or if it's the same with all Pentax DSLR (particularly the very similar K-50 and K-30), but there's obviously something wrong... I might try to reset the camera and see what happens...

Good luck with it, Carl, my K-30 doesn't do that. But your example at the moment has a real problem, your tests show that!

[CarlJF]

So the Q7 exposes OK in P-TTL with exactly the same set-up and settings? Then it seems like the K500 has a problem .....

That's exactly what I got. The Q7 exposes fine if I used the same settings (1/160, F5.6, ISO400, around 80mm focal length) that lead to underexposure with the K-500.

Good luck with it, Carl, my K-30 doesn't do that. But your example at the moment has a real problem, your tests show that!

Thanks for the feedback.So this really looks like a problem with my K-500. I will reset the camera and check how it goes...

EDIT: After resetting to factory default, the camera still underexposes when the flash is bounced from the wall (but strangely is fine if bounced from the ceiling). So, I guess the K-500 is like this. At least, I don't feel the need anymore to buy a flash with a swivel head! Well, until I upgrade the camera...

[CarlJF]

Okay guys, I've done a few more tests today and got some interesting results...

Try placing your flash right beside the camera, then bouncing, so that the light path is similar to bounce from an on-camera flash ... Make sure that your subject dominates in the frame, and that no direct light hits the subject (so for example pointing the flash 90deg from the front, or even slightly behind you). It will be interesting to see if this more usual bounce position and light distribution improves the metering.

For today, I've put the flash on a C-bracket and placed it as close as possible to the hot shoe. The flash was linked to the hotshoe with a P-TTL cord. Although it's not technically "on camera", it's very near to an on camera flash with swivel head. In this configuration, the 55-300 still underexposes a lot but the DA70 now works like a champ. Which is good news!

As an experiment, try zooming out to include a wider view of the scene .... your flash exposure should increase a bit as a result. Conversly, zooming in very close to a subject is likely to reduce the flash exposure value. The reason is that with wider scenes the meter sees more 'dead space', often not well lit dead space (like dull walls and shadowed areas), and so tries to light more with the flash. In any case, this test will confirm whether in general the P-TTL metering is working as expected.

What I've done is that I tested all my lenses. And the results are really interesting:

Good exposure: DA35 2.4, F50 1.7, DA70 2.4, Tamron 17-50 2.8
Underexposure: DA55-300 4-5.8, Sigma 10-20 4-5.6 (although not as badly as the 55-300)

So, all the fast lenses work perfectly. This also explains why I didn't noticed this problem before since I always use either the DA35, F50 or 17-50 for indoor casual portraits... All these lenses work fine with bounce flash. And the telephoto lens I used yesterday on the Q7 was also a constant 2.8 aperture zoom, which gave perfect exposure.

The two lenses leading to underexposure have in common that they're relatively slow and have variable aperture. I've no idea how this would affect P-TTL metering. But it allows me to draw this simple conslusion: if I'm planning to use bounce flash, better stick with fast constant aperture lenses...

Nikon or Canon are so much better there. I hope that Pentax will catch up with them at some point.

Maybe they're better, but a lot of Canon users complain that bounce flash in E-TTL leads to heavy underexposure... Just do a Google search with "bounce flash underexposure"... The 70D even seems to have built a reputation in the Canon community to simply be impossible to use with bounce flash for this specific reason...

[mcgregni]

Interesting results from your tests Carl. Just confirm please, your telephoto zoomed ones with the smaller apertures, did those shots fill the frame more with the subject? If so, as I said in my previous post, I'd expect underexposure when there's little space surrounding the subject, and so it needs plus flash compensation.

I would also try to increase the ISO on your telezoom ones, to compensate for the smaller apertures. When you used manual mode, you needed 1/2 power for correct exposure, so your flash is really working quite hard in that bounce situation. Using a higher ISO (say 800 - 1600) will take the pressure off the flash and give you longer battery life and quicker recycling.

[CarlJF]

Interesting results from your tests Carl. Just confirm please, your telephoto zoomed ones with the smaller apertures, did those shots fill the frame more with the subject?

It depends... Obviously, with the Sigma, the subject was relatively small because I could get the whole room on the shot... With the 55-300, the subject was filling the frame more, but not that different than with the F50 or DA70. I've also tested both zooms at their minimum and maximum focal length and the underexposure level doesn't seemed to vary much between these extremes. The Sigma was underexposed, but not as badly as the 55-300. With the Sigma, the shot could easily have been pull to an acceptable level in PP. And I probably could have been able to get it right by playing with EC.

I've thought about it and maybe it's because, with the slow lenses, there's enough light hitting the meter from the preflash for the camera to get an accurate measure. Although I would expect the camera to then send a signal to get full power from the flash, maybe there's some logics demanding that if the camera isn't able to reliably read the pre-flash, it rather fires it at minimum power... Maybe as a way to prevent an overblown exposure, since it's certainly better to have a bit less flash than way too much.

[mcgregni]

Its hard to say Carl. All I can really say is that fully automatic flash exposures can be influenced by so many variables that they need to be seen as an aid to the photographers input. The system won't be able to judge the intentions or remove all factors that could influence the calculations. Of course, full manual control will provide all the consistency needed, after the initial starting setting has been found and then perhaps chimped a bit. However, automatic flash can be a real help and speed things up in certain situations, but it can't do it all on its own.


I believe we photographers have to sort of gain an extra sense for how our gear will respond, and try and pre-empt things with our settings and positioning. You're certainly gaining this extra sense now with all the experimenting. I am sure it will prove invaluable in future and you will find the ways to get accurate P-TTL exposures first time. As I've said on another thread, I get around 7 out of 10 bounced and off camera exposures right first time .... the rest need that extra jiggling and re-thinking to figure out why they went wrong.

[beholder3]

It could be an issue with the lens, like a semi-stuck aperture lever or blades.

It could be the aperture motor in the K-500 as well.

You have to go thorugh normal bug fixing routine to narrow down things:
try another Pentax DSLR body with the lenses which underexpose.
try another similarly slow lens on the body wich underexposes.