580EX II Auto External Flash Metering surprise

[alan_chan4]

Just got this flash for the 40D a month ago and everything seems to work great.

Shots in E-TTL II is good. So today I was trying the radio slaves with this

flash and to my BIG surprise, it underexposes by 2 stops in Auto External Flash

Metering mode. Now I am no newbie to flash exposure and I can run any flash in

any mode from full manual to full auto in any situations. But when compared to

the Metz plain auto flash, the 580EX II consistantly underexposes by 2 stops. A

quick check with a Sekonic meter confirms it. A brief search reveals some other

580EX II owners had the same experience too. So what's the deal? Have Canon

knowledged this issue and had a fix for it? Man I was planning to use it off

camera too with radio slave and this is a total let down (yeah I know I can

alter the manual iso & aperture to compensate but this is not a cheap flash by

any mean). :(

[chrisjb]

Yup, same results on a 20d, have not bothered aith it on a 40d. all my metz,sunpaks and old toshiba far more accurate, when I rand canon thay would not admit there was a problem, maybe they should poke around some forums. have not heard a good word for this feature yet :(

[chrisjb]

BTW tried bouncing yet, seems another prob with distance reading lenses, (more FEC than ever, easy toget use to tho)I`ve now resorted to fong tupperware, if it don`t work better when sorted am going back to the 550`s.

[george_parris]

I thought that feature only worked with the mark III versions of the 1D level cameras? I will double check the manual.

[alan_chan4]

My bad. I mean Manual External Flash Metering.

[colinsouthern]

"I thought that feature only worked with the mark III versions of the 1D level cameras?"

 

The 1D3 was the first to support the automatic passing of ISO and Aperture information to the flash, but only whilst it was mounted on camera. Since then the 1Ds3, 40D, and STi have been added to the list.

[mark u]

You have to set the ISO and aperture combination if the camera isn't able to communicate them to the flash, either because they aren't connected or the camera doesn't provide the data in a form that the 580EX II can digest. I still find it peculiar that Metz can automatically pick up ISO and aperture with older cameras that the Canon flash can't handle at all. Maybe Canon will get it right eventually.

[bluphoto]

Excuse my stupidity/naivety, but are we talking about underexposure when in Manual mode here?

 

So if you're not using "auto" mode, then how can you say the flash is underexposing the image? Surely it is doing exactly what you're telling it to do by dialling the power directly into the unit.

 

From what I understand the only way you can determine that it underexposing is that you must be calculating the "correct" aperture using the Guide Number (58), subject distance and the ISO, right? If that's the case, then how does the "zoom" setting enter the equation? Obviously, you'll get less power on a spot meter like your Sekonic if your flash is on a wide zoom (say 35mm). At 105mm you'll get a much higher power at your metering point - kind of stands to reason really.

 

Sorry for my naivety in this matter, but I'm genuinely interested. If what you're saying is true then I guess you're also saying that its Guide Number is NOT ACTUALLY 58, as published by Canon. I would think if this were the case you could return it.

 

Is it the case that Canon maybe publish a table of Guide Number vs zoom setting - is 58 just the MAX guide number (at 105mm). Maybe check that your flash zoom setting is at is maximum 105mm, or alternatively, compensate (somehow) for the reduction in intensity produced by the wider beam spread.

 

Lastly, what units of distance are you measuring - feet or meters? I'm not sure what units you're SUPPOSED to be measuring in, but as there's around 3.3 feet to the meter, if you're using the wrong one you'll get about 1/3.3 (or 3.3x) the correct answer (which I think is just under two stops). Ie if you're using feet instead of meters then the subject is approx 3.3x closer than you're allowing for in the calculation, meaning you'll get 2 stops overexposure (I think), and vise versa, if you're using meters when you should be using feet, then you'll get approx 2 stops underexposure.

 

Hopefully someone can clarify this for me too, not because my flash isn't working, but just to clarify how all these parameters interact to come up with the correct aperture, and how zoom etc go into the equation.

 

Of course, if you're NOT calculating the correct exposure via the guide number, then can you tell me another way to determine "correct" exposure?

 

thanks

 

Guy

[bluphoto]

Sorry, but when you say you're no newbie to flash exposure, I assume that you realise and understand that there will be no communication of the aperture and ISO between camera and flash when used via (traditional) radio triggers, and that you will have to set both your camera AND your flash power manually.

 

You're not under the impression that your 580ex2 will do the above calculation for you, are you?

 

I've only been using off camera flash for a few months, so still consider myself a newbie, so please don't think I'm being rude for suggesting the obvious!

 

cheers,

Guy

[anov]

Guy;

 

Alan is talking metered manual not full manual. This is the auto mode or thyristor mode set by custom function.

 

There's a sensor on the flash that detect light. You enter aperture and ISO on the flash. Once enough light is reflected back to that sensor the flash will shut down.

 

Alan:

 

I have that flash too and it underexpose by about 2 stop as well (eyeballing histogram, not measured by light meter).

[bluphoto]

Ooooh, I must have missed that chapter in the manual - in fact, I don't think I read the manual at all!

 

Metered manual, eh? Sounds like it warrants some further investigation. Thanks for the heads up.

[bluphoto]

I guess if the flash is metering the light (metered manual) then there must be a metering pattern (whether it's spot, average, center weighted etc). Is this the same pattern as used by your Sekonic meter (spot.... right)? I guess you're bound to have discrepancies if the metering pattern is different?.?.?

 

I may be wrong, but as people generally point their flash at the subject they want to light, then common sense would suggest that a center weighted approach to the flash metering would prevail. Whether this is the case or not I'm not sure. I wonder if the metered area covers the same area as the illuminated area, no matter what zoom setting you're using.

 

Lastly, I assume you're not trying to illuminate a pale subject, and your flash is dialling down accordingly, or that your flash metering pattern is not also including some lighter areas of scenery (bright sky, for example) which would also cause it to dial down.

 

Really, really lastly, I assume you're using the reflective metering mode on your Sekonic, rather than the incident mode?

 

Just throwing some more reasons out there- I'd still like to know where zoom comes into the guide number calculation, if anyone wants to interject.

 

many thanks,

 

Guy

[alan_chan4]

Thanks everyone. After collecting more opinions, it is clear that the 580EX II is "normal" to underexpose by 2 stops in this mode. Something is fishy (I mean Canon).

[jimstrutz]

Yeah, something's wrong. Every time someone mentions the 580EXII's auto flash feature they complain about the 2 stops underexposure. It seems to be a universal issue. Anyone have one that actually meters correctly in auto mode?

 

Since this appears to be the case, can you override the flash's ISO setting when mounted on a 40D or 1DIII, or do you have to remove the flash from the camera body?

[chrisjb]

To do that Jim it would problably have to have CF for manual auto exp. will try it tomorrow on 40d

[colinsouthern]

"I still find it peculiar that Metz can automatically pick up ISO and aperture with older cameras that the Canon flash can't handle at all. Maybe Canon will get it right eventually."

 

Maybe Canon want you to upgrade your camera? For camera-mounted 580EX II's, it's becoming less and less of an issue as new models all support it (1D3, 1Ds3, 40D, 450D to date, and presumable the 5D2 when it's announced sometime in the next ten minutes/years).

 

Mostly it's a moot point though - the whole idea of that mode is to allow the flash to operate autonomously from the camera (in which case the Meta suffers exactly the same limitation); using Auto External Flash Metering on camera is generally counter-productive in my experience, considering that more sophisticated mechanisms are available (ETTL II).

[mark u]

Actually I think that the (firmware) design team for Speedlites has made several mistakes - for example, the incorrect zooming of the original 580EX with crop factor cameras. I hope it isn't regarded as a trainee area before being unleashed on camera design. Nikon makes their autoflash mode operate correctly: an SB800 is fully backward compatible with earlier bodies. Canon could and should produce an equally capable, equally fully featured, and equally compatible flash for their own system to strengthen their competitive hand.

[jimstrutz]

"...using Auto External Flash Metering on camera is generally counter-productive in my experience, considering that more sophisticated mechanisms are available (ETTL II)."

 

Yes, but...

 

There are situations where on camera autoflash can be useful. For instance where one wants to use multiple flash, and the other flash are older tech autoflash units, and you don't want the pre-flash to pre-trigger them. Setting all to autoflash metering can work in spite of what some have said about only doing this in manual flash mode. Another situation is when using an older E-TTL (I) DSLR (300D, 10D, etc.) with it's erratic metering pattern. I'm sure that if I thought about it I could come up with other situations where autoflash metering on camera is better or just preferred for some reason.

 

In any case, Canon offers it as a feature (one of the primary reasons they cost more than an original 580EX) and it doesn't work right. Obviously, not enough people are complaining about it or Canon would have addressed it by now. Personally, I'm waiting until it gets fixed before I buy one, because I'm one who would use it.

[colinsouthern]

"There are situations where on camera autoflash can be useful. For instance where one wants to use multiple flash, and the other flash are older tech autoflash units, and you don't want the pre-flash to pre-trigger them. "

 

Yes, but ...

 

... No-one's saying that those options aren't available. If you have a 1D3, 1Ds3, 40D, 450D (and presumably all new models) then it's a non issue. If you use an older model of camera then it takes all of 20 seconds to set it manually.

[jimstrutz]

Well, it wouldn't be an issue if it worked. Problem is it's 2 stops off. If Canon offers the feature and some people upgrade their flash to get it, it should work right. It's probably fixable with some sort of firmware upgrade or internal adjustment, but why are they all bad?

 

Of course the question stands: Are they all bad? Are there any good ones out there?

[colinsouthern]

"Well, it wouldn't be an issue if it worked."

 

I've only ever used mine in this mode a few times - and on those occasions it appeared identical to E-TTL II. Possibly an issue might occur because (unlike E-TTL II), the flash sensor (potentially) has a vastly different field of view to both the camera and the flash - so it has no way of knowing if parts of what it's receiving reflacted light for are supposed to be included in the scene or not - as such, I suspect that it probably only works well for specific types of scenes.

[jimstrutz]

Ahhh, so we now have evidence of one that works, or at least worked when tried. That's good to know. Now if Canon can only fix the ones that don't.

 

Autoflash metering on the whole is/was generally regarded as working well in most situations, but can be fooled with very long or very short lenses that see a different field of view. Years ago, autoflash was good enough that when TTL metering first appeared it was regarded by many as a solution in search of a problem. I always had good experiences with Sunpak & Vivitar flashes from the old days.

[colinsouthern]

"Ahhh, so we now have evidence of one that works, or at least worked when tried. That's good to know. Now if Canon can only fix the ones that don't."

 

I suspect it was more the composition at the time than the unit itself. Personally, I've only ever thought of using it off camera in a kind of quasi-manual type situation.

[model mayhem gallery]

<p>I only recently started using E-TTL flash because I have many Alien Bee and White Lightning Studio strobes. After investing in radio wireless triggers I wanted to start using my two 580 EX flashes as well. I simply put them in FULL MANUAL MODE, use the Cybersync wireless transmitter to do a test fire and meter with my Sekonic. Works 100% perfect everytime. If the light is too bright, I use the nob to lower or increase power to my taste. This is the exact same thing I have been doing with my studio strobes for years why would anyone want auto\manual flash? I have played around with that on my Canon 5D Mark II and 580 EX II, but as far as i can tell it doesn't add anything that i can't get to much faster manually. If all else fails and I am only using my E-TTL flashes I put one on the camera as master the second in wireless group C and then I can use the flash Compensation control to turn the exteranl flashes power up and down wirelessly, to get exactly what I want again manually. For me the 580 EX flashes have been fine but I very seldom use auto anything on my camera. OOpps, I occasionall use autofocus on my zoom lenses. </p>