[Class A]

The flash is not aimed at the ceiling above the subject.

Yes, and while the subject should receive some light from the reflector it won't be much and it would be mainly reflected back from the white background, creating a backlit situation.

Amoringello has reported that moving back with the camera solves the issue, which makes sense because then some light from the ceiling will hit the subject from the front. It also makes sense that removing the white background solves the issue as well because then the backlit scenario is avoided.

I also see how the spot metering and the -1.5 flash compensation are not helping. The former means that the camera will attempt to get the white of the background down to 18% gray and the latter means that the flash is firing at lower power than it would be normally.

In any event, with product shots like these, I'd suggest a fully manual approach. Set the camera to an aperture you need for your DOF requirements and a shutter speed that admits as much ambient light as you want (e.g., for fill). Then manually adjust the flash power to a level that gives you a proper exposure. Takes only a few attempts and from then on you'll be fine. If you run out of flash power, see whether you are bouncing the flash effectively (at the moment you are not) and/or increase the ISO.

With a fully manual approach the exposure won't change just because you move the subject creating a specular highlight or uncover some of the background, or whatever may cause the P-TTL approach to believe the exposure needs to be adjusted.

[amoringello]

The flash is not aimed at the ceiling above the subject. Instead, it's aimed at the ceiling well behind the subject. Additionally, the card on the flash head is preventing any of the light from illuminating the ceiling above the subject. I can see how this would be an impossible setup for proper lighting. Hope that helps.

er, what?
Think about the case of a warehouse with 20 foot ceilings. The light pointing straight up is effectively going nowhere.
This also has work flawlessly with the K5 for the past several years.
So, far from impossible.

---------- Post added 03-26-14 at 11:29 AM ----------

Yes, and while the subject should receive some light from the reflector it won't be much and it would be mainly reflected back from the white background, creating a backlit situation.

Amoringello has reported that moving back with the camera solves the issue, which makes sense because then some light from the ceiling will hit the subject from the front. It also makes sense that removing the white background solves the issue as well because then the backlit scenario is avoided.

I also see how the spot metering and the -1.5 flash compensation are not helping. The former means that the camera will attempt to get the white of the background down to 18% gray and the latter means that the flash is firing at lower power than it would be normally.

In any event, with product shots like these, I'd suggest a fully manual approach. Set the camera to an aperture you need for your DOF requirements and a shutter speed that admits as much ambient light as you want (e.g., for fill). Then manually adjust the flash power to a level that gives you a proper exposure. Takes only a few attempts and from then on you'll be fine. If you run out of flash power, see whether you are bouncing the flash effectively (at the moment you are not) and/or increase the ISO.

With a fully manual approach the exposure won't change just because you move the subject creating a specular highlight or uncover some of the background, or whatever may cause the P-TTL approach to believe the exposure needs to be adjusted.

Wow, sometimes I have to ask myself if some people even read the thread or have tried to reproduce the issue and see what results they get for themselves.

Again, go back to the warehouse scenario where there are no walls or ceilings that will contribute to the lighting. Then re-think how those comments would apply to that situation...

I believe I also mentioned at the start about using -1.5 in most situations but I get the same results on the K3 at full power. if you think about it, the reading from the pre-flash is so powerful that the main flash is calculated at such a low power that it cannot be decreased any more. So -1.5 or full give the same or close low output. Thus it never gets to 18% gray.

I would probably agree, a manual approach might be better in these situations, but because of what, where and how I shoot, I am just too darned lazy to go in and out and in and out of manual. I am not in a studio but wandering through rooms with different layouts, ambient lighting and amount of room with which to work. Many shots may require slightly different manual settings.

Regardless when something is working year after year and then suddenly fails with the change of *one* part, it is a good sign that something ain't quite workin' correctly with that *one* part.

[Class A]

Again, go back to the warehouse scenario where there are no walls or ceilings that will contribute to the lighting. Then re-think how those comments would apply to that situation...

Please accept my apologies, but I won't rethink anything because I have no interest in further investing time trying to help someone who uses "Wow" and a tone that borders on the condescending.

You wrote

While removing the bounce card from the flash head (so all light is reflecting off the ceiling-only), the exposure was "perfect" (def: perfect in terms of PTTL, meaning it is pretty darned close for an automated system).

So that apparently does not apply to your "warehouse situation" and readers of your posts are also supposed to figure out which aspects of your "setup" shot apply to situations where you experience the problem and which do not.

Good luck with making progress with this, but it may pay off to be a bit friendlier to people who are trying to help.

Regardless when something is working year after year and then suddenly fails with the change of *one* part, it is a good sign that something ain't quite workin' correctly with that *one* part.

You'd be forgiven to develop that hypothesis, but it is very well possible that what was working before was a curious combination of a faulty light setup with a faulty P-TTL implementation (the K-5 has a known overexposure bug). Not saying this is the case but you'll have to do a lot more to demonstrate that something is wrong with the K-3.

[amoringello]

Please accept my apologies, but I won't rethink anything because I have no interest in further investing time trying to help someone who uses "Wow" and a tone that borders on the condescending.

Fair enough.
With the lack of effort towards understanding the problem, If you feel that is condescending, so be it.

[lister6520]

I'm wondering if it could be a case of bounce light entering from the viewfinder. Admittedly though there are some specific cases described here where that can probably be ruled out but then again maybe there is more than one cause.

I have just managed to reproduce the problem but in a somewhat more extreme (and possibly unlikely) situation. I am using 250mm lens, shooting at the end of a 25ft room with the flash pointed straight up to the white ceiling and with a white wall about 3 feet behind the camera. The ceiling is about 11 feet above the ground. The camera is on a tripod at about 4 feet above the ground and I am not standing behind it so light reflected off the ceiling to the nearby wall can easily get into the viewfinder.

I get a shot underexposed by about 3 or 4 stops (estimate based on histogram). If then I cover the viewfinder but keep everything else exactly the same I get a good exposure, or possibly half a stop under depending on what one actually wants.

[amoringello]

I'm wondering if it could be a case of bounce light entering from the viewfinder. Admittedly though there are some specific cases described here where that can probably be ruled out but then again maybe there is more than one cause.

Certainly could be a problem. And it is one I am constantly forgetting about when shooting outdoors on a tripod and with two second timer or remote release. For this latest issue, I am usually looking through the viewfinder, and my big head pretty well blocks incoming light.

But that is very clever thought ... I do recall that some light is also reflected back into my eyes from the large bounce card. I have ignored the issue as it has had no effect on the K5.

Perhaps the K3 is extra sensitive from that light bouncing off my forehead and into the viewfinder or simply as glare across the viewfinder as well. I'll have to try again with extra care to adequately cover the viewfinder and see if that alters the behavior. hmmm......

[IchabodCrane]

Fair enough.
With the lack of effort towards understanding the problem, If you feel that is condescending, so be it.

By actually looking at your setup, reading what you wrote, and then taking time to respond with theories and suggestions, one would think you might want to try some of these things instead of just fire back about how everyone else is dumb. It pretty much shuts off the conversation.

[amoringello]

I'm wondering if it could be a case of bounce light entering from the viewfinder. Admittedly though there are some specific cases described here where that can probably be ruled out but then again maybe there is more than one cause.

Nope, I covered the viewfinder completely and no change.
But I did find that a lens with a smaller base aperture like the 18-55 f3.5-5.6 seems to also have a slightly more difficult time reproducing.
So I'm leaning more towards pre-flash light through the lens where the aperture is still wide open for metering, focusing, etc...

---------- Post added 03-26-14 at 05:29 PM ----------

By actually looking at your setup, reading what you wrote, and then taking time to respond with theories and suggestions, one would think you might want to try some of these things instead of just fire back about how everyone else is dumb. It pretty much shuts off the conversation.

Sorry, which things were not tried? Anything referring to avoiding walls, ceilings, reflections and similar, were covered before the suggestion.

The setup was just an example in response to a request to see the bounce card and general setup. It does NOT cover ALL shooting scenarios.

Anyway, I am working with Pentax on this as well. So if you feel I did miss something specific, let me know.
I am happy to gather as much information as I can, and am open to suggestions that have not already been repeated multiple times.

[jbinpg]

Re. bounced light into viewfinder: simple test would be to use the removable viewfinder blind that comes with every K-3. You would need to remove the viewfinder "eyecup" to use it.

[amoringello]

Re. bounced light into viewfinder: simple test would be to use the removable viewfinder blind that comes with every K-3. You would need to remove the viewfinder "eyecup" to use it.

Did yours come with one? I didn't see that in the box for my K3. I know it has been in prior models so I was sort of surprised to not see it... I just figured it saved $0.02 on costs. :-)

Anyway, I dug out an old one and used that.

[NotaxPen]

I noticed something new on this issue. If I use a manual focus lens, the problem never happens. I get consistently exposed pics every time regardless of distance from the subject (provided I'm not blocking the bounced light). If I use an auto focus lens, I still get very underexposed shots nearly every time if I'm real close to the subject whether it's switched to auto focus or manual focus. This seems to suggest that the K-5 uses more of the focal distance data to calculate the exposure than the pre-flash since a manual focus lens cannot communicate distance to the camera.

[amoringello]

Interesting.
I don't have any manual focus lenses other than LenseBaby... but it doesn't even register to the camera as having a lens on it so I can't duplicate the behavior. :-(
I'll send the info to Ricoh and see if they have anything to say about that.

Thanks for the info.

[Class A]

If I use a manual focus lens, the problem never happens..

What exactly do you mean by "manual focus lens"?

Note that unless your lens is at least an "A-lens", (supporting the P/KA mount) P-TTL won't be used at all.

[stevebrot]

I noticed something new on this issue. If I use a manual focus lens, the problem never happens. I get consistently exposed pics every time regardless of distance from the subject (provided I'm not blocking the bounced light). If I use an auto focus lens, I still get very underexposed shots nearly every time if I'm real close to the subject whether it's switched to auto focus or manual focus. This seems to suggest that the K-5 uses more of the focal distance data to calculate the exposure than the pre-flash since a manual focus lens cannot communicate distance to the camera.

What exactly do you mean by "manual focus lens"?

Note that unless your lens is at least an "A-lens", (supporting the P/KA mount) P-TTL won't be used at all.

My question as well. The focus mode should not impact P-TTL function since it is not part of the equation. The camera does not need to know focus distance in order to measure the intensity of the pre-flash burst and it is never part of the equation. That is, after all, why we use P-TTL. What is more, the AF system does not calculate actual distance, it merely establishes best focus regardless of distance.

Your underexposure at very close distance is not unexpected depending on what you mean by "close". Most on-camera flash units do not have wide enough coverage to illuminate a subject at closer than about a 18 inches. That is why specialized ring flashes and off-camera flash are the norm for close-up and macro work.


Steve

[NotaxPen]

OK, I guess I should repeat. I'm using an AF540FGZII on a K-3. The problem only happens with bounced flash when the camera is less than about 3 feet from the subject. It is very repeatable and never happens when I am further from the subject or if I use a manual focus lens. With my A 28/2.0 I can shoot as close as it's min focus distance of 1 foot and get accurate bounced flash exposure every time. With 21mm Limited, Sigma 30/1.4, 40mm Limited, DA* 55/1.4, DA 18-135, and DA* 16-50/2.8 the underexposure problem is consistent at less than three feet. All my other lenses don't focus closer than 3 feet except 70mm Limited which is the odd one of the bunch. The problem seems to be hit and miss with the 70mm. The manual focus lenses are "A" lenses. I read somewhere in this forum that the camera uses focus distance in some instances to determine proper flash exposure, that's what it seems to be doing in this case rather than calculating exposure based on the pre-flash. That is what seems strange. The underexposure is not a case of not being able to illuminate part of the frame due to not having enough coverage. It is almost like the flash thinks you are shooting the flash straight at the subject rather than bouncing. So lets say the bounce distance is about 7-8 feet from the flash to the wall to the subject, and it's putting out enough power if it was 2 feet. About 4-5 stops underexposure. With my K-5 I get overexposure with bounced flash in the same conditions.