70D 430ex II bounce under exposure - 2

I also recently upgraded from a t4i to a 70D, I also use the 430EX, and I also mostly shoot my kids indoors! But I haven't seen a major difference in flash exposure after the upgrade.

This might be because I almost exclusively shoot in M mode and I believe in that mode the flash fills in as much as it needs to in order to correctly expose.  Could you tell us the mode you're using, and if not in M, then give it a shot and see what happens?

I typically use M, 1/60s, wide open or 1/2 stop less, with 400 ISO.  I also typically use the 40mm f/2.8 when shooting indoors.

Good luck!

I face the same under exposure problem for about 2EV when using 70D, 18-135 lens, 430EXII or Metz 50AF-1 flash under bounce from ceiling. I would like to purchase a 600EX soon but want to know does it solve this under exprosure problem? Please kindly advise, thank you!

davel33 wrote:

I shot a Xmas dinner/party friday night with 70D/Sigma 24-70 2.8 hsm/ Sigma 18-35 hsm with a 430EXII flash. About 230 shots and about 3 under exposed and all 3 were my fault. I use AV, auto ISO with min shutter set at 1/60. Used bounce flash with a DEMB flash attachment set for about 70% bounce and 30% direct.

Good luck tracking this down

Dave

Haha, didn't realize there was a new thread....was wondering why all the activity died on the old one!

I think once we get a bit more info we might need to start a more general thread (without the 430exII in the title).  And in the first post we'll need to make sure we outline all known information.

For instance, davel33, the underexposure problem does not really manifest itself with your two lenses mentioned because they are both really fast lenses.  The problem doesn't show up until you are using much slower lenses f3.5 to f5.6...

So it doesn't surprise me that you didn't see a problem.

OK...so I've been doing a bunch more testing and have narrowed down some things. Some of them can be easily tested, so it would be good to have a few other people verify they see the same things.

1. The amount of under-exposure is definitely related to the max aperture when the exposure reading is taken. This means for a non-variable lens (i.e. 17-55mm f2.8) you should be able to expect decent performance and almost useable exposures because of the faster lens. And since the max aperture is constant, you should get that same type of exposure performance at any zoom. With variable max aperture lenses, the amount of underexposure should vary with the zoom (not because of the zoom but because the max aperture gets smaller as you zoom more). I did some testing with my 18-135mm and stepped my zoom enough to take a picture at every max aperture setting (3.5, 4.0, 4.5, 5.0, 5.6) and the under exposure got consistently worse with every step going towards f5.6 (and 135mm).
2. The amount of under-exposure also got worse with distance. So the farther away from the subject, the more under-exposure there is.
3. Both of the above point to under-performance that gets worse as you get less light coming into the camera.
4. My Xsi did great with exposure levels for any zoom/distance/max aperture setting.

Next...I was really curious what was actually happening. So I broke out my scope and a light sensor. I rigged the light sensor to about 4 inches in front of the flash output, setup a tripod and did some back to back tests with my Xsi and 70D using the same tripod/lens/flash/sensor/settings in Manual mode. The only differences being the body swap.

I actually found some very interesting results. I need to think about them a bit more before drawing any conclusions (and hopefully test another couple bodies as well to compare). I'll show my screen captures below...but here are just a couple of points to consider on these screen caps:

• First, for anyone that doesn't know, flash 'power' is actually increased/decreased by the amount of time the flash is left on. So the flash always outputs about the same light level...but to change the power you just change the amount of on-time.
• Also should note that in my screen captures below the output was inverted.  So when the signal goes 'down' the flash is on.
• I found the flash emitted from both 70D and Xsi to be pretty equivalent when FEL was used.
• At 135mm and using the 'shutter-only' (measuring the pre-flash...not the real flash) I found the Xsi put out a LOT more power...meaning the flash stayed on MUCH longer than the 70D.
• At 135mm, the Xsi actually put out a fairly similar flash power as it did when using FEL (probably fairly coincidental), but the 70D's flash power level was much less.
• At 18mm, the Xsi and 70D actually put out fairly identical curves.
• So for some reason when zoomed to 135mm (f5.6 max aperture), the Xsi output quite a bit more flash power during the pre-flash. The 70D actually output even slightly less than it did at 18mm.
• Need to test with a couple more bodies to see if Canon DSLR's typically follow the Xsi pattern of emitting a more powerful pre-flash when the max aperture goes down. It would almost make sense that it would...to make sure you get enough light back for proper exposure level readings...but need to test other bodies first to see if there is a pattern.
• To help compare the curves better, I'm putting the Xsi Curves on the left and similar 70D curve on the right side.

Xsi (left), 70D (right), FEL used, 135mm, M, iso-800, 1/100, f5.6

Xsi (left), 70D (right), 'Shutter-only', 135mm, M, iso-800, 1/100, f5.6

Xsi (left), 70D (right), 'Shutter-only', 135mm, M, iso-800, 1/100, f5.6

WilbaW wrote:

I have no comment except to say what a pleasure it is to see someone rigorously testing something, rather than relying on unenlightened guesswork, stubborn ignorance, and regurgitated myths. Keep up the good work Sir!

Or, instead of spending hours and hours of rigorous testing, one could do what normal photographers do and learn how to use your equipment, experiment, apply corrections as necessary, and move on with life.

TTMartin wrote:

HymanRoth wrote:

Or, instead of spending hours and hours of rigorous testing, one could do what normal photographers do and learn how to use your equipment, experiment, apply corrections as necessary, and move on with life.

This is an actual problem resulting from the use of a new technology (dual pixel sensor).

I think Canon needs to do a firmware update to correct this.

The only way Canon will realize they need to do a firmware update is if they get enough complaints about the bounce flash behavior.

Ignoring it and living with it isn't going to get it fixed. Everyone who has the issue reporting it to Canon Technical Support will get it fixed.

You're assuming this is an "issue".  My 70D exposes properly, and I suspect that what most people see as "underexposure" here is the camera avoiding blown highlights.

HymanRoth wrote:

Or, instead of spending hours and hours of rigorous testing, one could do what normal photographers do and learn how to use your equipment, experiment, apply corrections as necessary, and move on with life.

HymanRoth:

If your camera exposes properly...great for you.  That doesn't mean that many of our cameras do not.  Trust me...I have enough experience with using cameras and have tested enough equipment to know that (at least some) 70D's are NOT exposing properly.  There are many people also finding the same thing I am on this forum and others....so that also eliminates me as the single point of failure.

Not sure if you did what I suggested...but the test results you uploaded were not even valid for this problem.  You need to take some pictures with preferably at least a 55mm zoom on  your 18-135mm lens.  The problem is exacerbated with smaller max apertures.  ALL of the shots you uploaded were taken at 18mm which has a f3.5 max aperture and is in the almost useable range.  So you need to try some shots at least 15 feet away at f5 or smaller max aperture (55mm or longer zoom).

And as you can just see from my scope plots above there is something HUGELY different going on with the pre-flash from 70D.  At least when compared to an Xsi.  I need to compare with other bodies now and see if its normal for different pre-flash power levels for different bodies or if they all follow the same pattern minus the 70D.

The test results above are only interesting at this point.  I don't have enough data to prove anything...but I'm getting closer.

TTMartin wrote:

HymanRoth wrote:

TTMartin wrote:

HymanRoth wrote:

Or, instead of spending hours and hours of rigorous testing, one could do what normal photographers do and learn how to use your equipment, experiment, apply corrections as necessary, and move on with life.

This is an actual problem resulting from the use of a new technology (dual pixel sensor).

I think Canon needs to do a firmware update to correct this.

The only way Canon will realize they need to do a firmware update is if they get enough complaints about the bounce flash behavior.

Ignoring it and living with it isn't going to get it fixed. Everyone who has the issue reporting it to Canon Technical Support will get it fixed.

You're assuming this is an "issue". My 70D exposes properly, and I suspect that what most people see as "underexposure" here is the camera avoiding blown highlights.

What lens and aperture and other settings did you use when your bounced flash exposed properly?

Not sure if he's tried other aperture/zoom levels, but for all the pictures he uploaded he was using 18mm zoom (max aperture of 3.5).  So none of his test shots that he uploaded even really shows the problem very much.  At max apertures of f3.5 you get exposures that still tend to be slightly underexposed IMO...but are good enough that you could throw around the 'saved highlights' argument.

When using f5.0 and f5.6 max apertures and all of my shots show a full 1.5+ of underexposure on the histogram...that's NOT saving highlights.  That is wasting a serious portion of the dynamic range of my nice expensive sensor that I paid for.

I posted some images in the closed thread.  I used the 18-55 STM and the 15-85mm, but at different focal lengths and apertures.  Yes, there were minor variations in brightness, and the ETLII system is not foolproof, but its #1 goal, at least on its more "prosumer" models, is to avoid blown highlights.

bobn4burton wrote:

HymanRoth wrote:

Or, instead of spending hours and hours of rigorous testing, one could do what normal photographers do and learn how to use your equipment, experiment, apply corrections as necessary, and move on with life.

HymanRoth:

If your camera exposes properly...great for you. That doesn't mean that many of our cameras do not. Trust me...I have enough experience with using cameras and have tested enough equipment to know that (at least some) 70D's are NOT exposing properly. There are many people also finding the same thing I am on this forum and others....so that also eliminates me as the single point of failure.

Not sure if you did what I suggested...but the test results you uploaded were not even valid for this problem. You need to take some pictures with preferably at least a 55mm zoom on your 18-135mm lens. The problem is exacerbated with smaller max apertures. ALL of the shots you uploaded were taken at 18mm which has a f3.5 max aperture and is in the almost useable range. So you need to try some shots at least 15 feet away at f5 or smaller max aperture (55mm or longer zoom).

And as you can just see from my scope plots above there is something HUGELY different going on with the pre-flash from 70D. At least when compared to an Xsi. I need to compare with other bodies now and see if its normal for different pre-flash power levels for different bodies or if they all follow the same pattern minus the 70D.

The test results above are only interesting at this point. I don't have enough data to prove anything...but I'm getting closer.

I actually took a bunch of images using my 18-55 and 15-85 at all focal lengths and apertures, and all looked fine.  I didn't post them because I was getting bored with this "issue".  I see you've posted this issue on about 3 or 4 different photo sites, so you are single-handedly creating Internet Hype Syndrome for something that, frankly, I can't see.  If you do Google searches for "Canon underexposure", you'll see that people have had this same complaint about most Canon models going back to the original D Rebel.  This is just the way Canon meters.

davel33 wrote:

A little extra data Last night I did a bit of testing with the Sigma 24-70 hsm. sigma 18-35, Canon 18-135 stm and the Canon 40 2.8.

The Sigma 18-35 and 24-70 did not have any problem but the Canon 18-135 stm did have the problem but what was interesting was the Canon 40 2.8 was about half way between the the 24-70 and 18-135 so I dont think that its only the lens speed because the 24-70 2.8 does not have the problem at all and the Canon 40 2.8 does have it just not as bad as the 18-135 stm. The 430EXII and the 320EX gave the same results with AV set to 4.5 and TV @ 60 and iso @ 400 in AV mode. I will test the Sigma 50-150 2.8 tonight.

Dave

That is interesting.  Not sure if it has any affect, but the two sigma lenses you tested both seemed to work great...but the two canon lenses exhibited the problem (but to differing degrees).

Now it makes me want to test a non-canon, slower lens.  Do third party lenses report any distance/focal length type info?  Wonder if that factors in to the equation at all.