G3 to 20D with 420EX a downgrade, why?

[wibbly]

My 420EX worked great with my G3. I was constantly amazed how well and how natural the exposures looked. I mainly use the 420EX bounced from a white wall or ceiling.

On the 20D, I consistently get the infamous under/variable exposures. And yet ETTL-II is supposed to be an improvement on whatever the G3 has. The underexposure even occurs with 'full face' photos, where the face pretty much fills the frame. Surely this shouldn't confuse ETTL-II's strategy of weighting to the focus point? Changing between evaluative and average ETTL-II settings seems to make very little difference in these circumstances.

Any insights? Is it REALLY that much easier to get it right with a non-SLR?

W

[alan sh]

Two things here.

1. ETTL-II will not work in Bounce mode.
2. The 20D is known to underexpose by about 2/3 stop.

So, put the FEC up by 2/3. Set the custom function to run at 1/250th in Av mode. Put the camera in Av mode, choose an appopriate aperture and try it again.

Oh yes, put the ISO at 400. It gives you 2 more stops of distance for very little loss in quality.

I get excellent results from my 20D with a 420 or 580 - but it took me some time too.

Hope this helps

Alan

[tim]

I always shoot my 300D in M at F8 and 1/80th. ETTL (v1) does it's job pretty well, sometimes I use FEC. Bounce flash works fine, usually.

[Mike Panic]

its not that much easier to get it right w/ a p&s, its just that much easier to screw it up w/ a dslr :)

you have to remember that all the messing up you are doing is because the dslr w/ flash is much more versitle in the end then a p&s is

[PacAce]

Two things here.

1. ETTL-II will not work in Bounce mode.


This isn't an accurate statement. ETTL-II will work with the flash head in ANY position, as long as the flash is an EX or compatible series flash. What ETTL-II won't do in bounce mode is utilize the distance information from the lens if one is avalable.

[mr.photoguy]

Odd. I thought the 420ex and 550 ex were ettl not ettl II which is why you get the underexposure.
The 580 ex is ettl II compatable and won't give you the under exposure..

Anyhow I usually put my camera in M mode. Then I will choose my f stop.. say f-8 then choose my time exposure based on my focal length, so 100 mm would be 1/100 or a little more.
Then if I am using the 420 ex I have to up the FEC by 1 stop or a little lower.. depending.
With my 550 ex I don't touch the FEC.
Also make sure you shoot in RAW, so you can always correct the exposure.

[PacAce]

Odd. I thought the 420ex and 550 ex were ettl not ettl II which is why you get the underexposure.
The 580 ex is ettl II compatable and won't give you the under exposure..


The ETTL or ETTL-II logic is built into the camera, not the flash. Therefore, any EX flash wll work in ETTL-II mode on an ETTL-II camera such as the 20D.

[mr.photoguy]

The ETTL or ETTL-II logic is built into the camera, not the flash. Therefore, any EX flash wll work in ETTL-II mode on an ETTL-II camera such as the 20D.
Is there any reason that the exposure always comes out a stop underexposed. I never really understood that.

[CyberDyneSystems]

Be sure and check out the -=EOS FLASH=- sticky at the top of the EOS forum. I belive this subject is discussed at length. Perhaps Scottbergers excellent research on this subject will "shed some light on it" for you ;)

[wibbly]

Be sure and check out the -=EOS FLASH=- sticky at the top of the EOS forum.

Ah. Hadn't noticed that. Thanks. As it happens I'm getting "There was a problem looking up the post in our database. Please use your back button to return to the previous page." on those thread links at the moment. Will try later.

Still, it amazes me that the G3 + 420Ex subjectively seems to get it right more frequently...

W

[wibbly]

Is there any reason that the exposure always comes out a stop underexposed. I never really understood that.

Indeed. Seems a bit naff, if so many people feel the need to fix this with generic +2/3 FEC or whatever... Would it not be a better compromise to have the camera do that, and have people set -2/3 FEC when necessary!?

W

[PacAce]

Is there any reason that the exposure always comes out a stop underexposed. I never really understood that.

I guess it depends on how you define "under exposed". If you look at the histogram of these "under exposed" shots very closely, you'll notice that the shot was set up so that the hightlights did not get blown (or not much, if it was). In other words, ETTL-II makes it really difficult to get shots with blown highlights. The exception to this seems to be if the ETTL-II logic determines that the "highlight" is a significant specular highlight which can negatively affect the exposure. In this case, the algorithm ignores it in determining the exposure.

There are people who don't care for the pictures taken with ETTL-II flash thinking they're too underexposed. They use +FEC to compensate.

In case you haven't seen it yet, I posted some examples a week or so ago with histograms of how ETTL-II behaves in comparison to center-weigthed average metering, including how it handles specular highlights (the last two samples before the Sunpak samples):

http://www.photography-on-the.net/forum/showthread.php?t=51698

[Edit: BTW, I think I should also add that, from what I've read, ETTL-II tries to make the shot look as natural as possible so that it looks almost like how you saw it in it's ambient lighting condition.]

[slin100]

I guess it depends on how you define "under exposed".
I once did some testing with my 10D and 550EX. I compared shots taken of a uniformly colored, off-white wall with and without flash. With a uniform background, the metering pattern is taken out of the equation.

The shots taken with flash were always underexposed between 2/3 and 1 stop. IOW, the only way I could get the histogram of the shot taken with flash to match with that taken without flash was to bump up FEC by +2/3 to +1.

I would call that unexposure. Results with ETTL-II may vary but I, like many others, suspect Canon intentionally designed underexposure into their flash system.

[PacAce]

II would call that unexposure. Results with ETTL-II may vary but I, like many others, suspect Canon intentionally designed underexposure into their flash system.Isn't that what I've been trying to say? Well, maybe in not so many words, but the point was, that's how Canon designed ETTL-II to work.

[slin100]

Isn't that what I've been trying to say? Well, maybe in not so many words, but the point was, that's how Canon designed ETTL-II to work.
Sorry, I didn't get that impression that you were trying to say that. FWIW, ETTL-I works that way, too.

Anyway, I think many would agree with me that they would expect their exposures of the foreground to turn out the same with or without flash. That flash exposures turn out darker relative to non-flash exposures seems awfully counterintuitive to me.

[PacAce]

Sorry, I didn't get that impression that you were trying to say that. FWIW, ETTL-I works that way, too.

Anyway, I think many would agree with me that they would expect their exposures of the foreground to turn out the same with or without flash. That flash exposures turn out darker relative to non-flash exposures seems awfully counterintuitive to me.

Just out of curiousity, exactly how did you set up that test? Since the lens won't autofocus on a blank white wall, I'm curious to know how you managed the focusing? The reason I'm asking is becaues the 10D uses ETTL and not ETTL-II. The two behave differently. With ETTL the exposure metering is AF point dependent unlike ETTL-II which is not. And, if AF won't focus on a blank wall, then you probably focused on it manualy. And, if you did, then the camera was not in ETTL mode but in center-weighted average mode.

[scottbergerphoto]

Leo has done a great job of batting away the misconceptions expressed about ETTL/ETTLII. I will just summarize:
1. The camera body is programmed to provide either ETTL or ETTL II with a Canon Ex Series Flash. Currently only the 1D MarkII, 1DS MarkII and 2D DSLR's have ETTLII.
2. ETTLII never reverts to ETTL. If a lens doesn't provide distance information, or the flash doesn't have all the features of the 580EX, the camera does the best it can.
3. Tilting the flash doesn't affect the flash mode.
4. You can avoid much of the problem of underexposure in ETTL/ ETTLII by using the camera in Manual Mode and the flash set to ETTL/II. This eliminates the Auto Reduction of Fill Flash in all the other modes (Auto,P, Tv, Av).
5. ETTLII is more consistent then ETTL because it is not affected by the active AF points.
More detail in the EOS Flash Sticky.
Regards,
Scott

[slin100]

I used C.Fn 4-1, which is just about the same as manual focusing. True, flash exposures will be center-weighted averaged, but that doesn't mean ETTL is disabled. It's still ETTL as far as determining exposure. The point of my test was to factor out the metering pattern and ascertain how the histogram is affected by using flash.

ETTL-I is very biased to the active AF-points and only a uniform background could make this a non-factor. I'm reasonably certain that if I had taped a small grid on the wall to aid in AF that the result would have been the same.

[PacAce]

I used C.Fn 4-1, which is just about the same as manual focusing. True, flash exposures will be center-weighted averaged, but that doesn't mean ETTL is disabled. It's still ETTL as far as determining exposure. The point of my test was to factor out the metering pattern and ascertain how the histogram is affected by using flash.

ETTL-I is very biased to the active AF-points and only a uniform background could make this a non-factor. I'm reasonably certain that if I had taped a small grid on the wall to aid in AF that the result would have been the same.

I hate to say this but I disagree with your assessment of what ETTL is. The E in ETTL stands for Evaluative, as in "evaluative metering", the same one the camera uses when metering ambient light in Evaluative metering mode. This is in contrast to the center-weighted metering which is a completely different mode of metering. When the 10D evaluates the flash metering, it is using the same "brain" and electronic circuitry as what the camera uses for ambient light metering. So, you see, when the flash metering is based on center-weighted average metering, it does indeed mean that ETTL, i.e. using evaluative metering, is "disabled' because evaluative is not being used. :D

But, I must admit that when shooting a blank white wall, you would get the same result whether you are using evaluative metering or center-weighted metering.

[slin100]

4. You can avoid much of the problem of underexposure in ETTL/ ETTLII by using the camera in Manual Mode and the flash set to ETTL/II. This eliminates the Auto Reduction of Fill Flash in all the other modes (Auto,P, Tv, Av).

Scott,

Auto Reduction of Fill Flash works in all modes, including M. Auto Reduction is a function of the metered ambient light level and preflash. Unless disabled by custom function, it always kicks in at EV 10 and brighter.

[slin100]

I hate to say this but I disagree with your assessment of what ETTL is. The E in ETTL stands for Evaluative, as in "evaluative metering", the same one the camera uses when metering ambient light in Evaluative metering mode. This is in contrast to the center-weighted metering which is a completely different mode of metering.
I understand your reasoning. Maybe we'll have to agree to disagree as gentlemen. I think of the "E" in ETTL to mean more than just the weighting of various areas in the scene. It also encompasses the calculation of the flash output needed to achieve a desired exposure. On a 10D, changing the flash metering mode to center-weighted just changes weighting assigned to areas within the scene. The backend part of ETTL which involves flash output determination is unaffected. I admit that the 20D incorporates additional mechanisms to influence its calculation, like distance info.
When the 10D evaluates the flash metering, it is using the same "brain" and electronic circuitry as what the camera uses for ambient light metering. So, you see, when the flash metering is based on center-weighted average metering, it does indeed mean that ETTL, i.e. using evaluative metering, is "disabled' because evaluative is not being used. :D
I don't believe this is entirely correct for a 10D. ETTL metering on a 10D is heavily biased to the active AF points, which is definitely not the way evaluative, ambient light metering works.

But, I must admit that when shooting a blank white wall, you would get the same result whether you are using evaluative metering or center-weighted metering.Thanks!

[scottbergerphoto]

Scott,

Auto Reduction of Fill Flash works in all modes, including M. Auto Reduction is a function of the metered ambient light level and preflash. Unless disabled by custom function, it always kicks in at EV 10 and brighter.
Slin, I believe you are in error on this issue. Photnotes.org is also wrong on this issue. Chuck Westfall of Canon has advised in Rob Galbraith's Forum, that Manual mode eliminates the Auto Reduction of fill flash.
In addition, regarding CFn. 4-1, when you are depressing the * button, there is an active AF point and that point will be used to bias the exposure. It is only when you release the * button, before you take the picture that the camera will change to manual focus. In addition, when manually focusing, the flash metering is not Center weighted. All the focusing points are averaged.
Scott

[drisley]

I went from a G3 to a 300D and then a 20D and used a 420EX as bounce flash with all cameras.
Im my experience, exposures improved significantly with each camera upgrade.

[slin100]

Leo, and others who are still paying attention,

If you still don't agree that ETTL/ETTL-II is still active even when the flash metering is center-weighted, please read Canon marketing director Chuck Westfall's post on ETTL-II where he includes an excerpt from the 1D2 white paper. http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB18&Number=248058&Forum=All_Forums&Words=%2BETTL-II%20%2Bdistance%20%2Baverage&Searchpage=0&Limit=25&Main=247684&Search=true&where=bodysub&Name=&daterange=0&newerval=&newertype=&olderval=&oldertype=&bodyprev=#Post248058
Nowhere does Chuck say that ETTL-II becomes disabled when switching to center-weighted. It's still Evaluative. It's just that the evaluations are slightly changed! ;)

I just realized that perhaps were having a disagreement over the semantic meaning of the word "evaluative."

[PacAce]

Leo, and other who are still paying attention,

If you still don't agree that ETTL/ETTL-II is still active even when the flash metering is center-weighted, please read Canon marketing director Chuck Westfall's post on ETTL-II where he includes an excerpt from the 1D2 white paper. http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB18&Number=248058&Forum=All_Forums&Words=%2BETTL-II%20%2Bdistance%20%2Baverage&Searchpage=0&Limit=25&Main=247684&Search=true&where=bodysub&Name=&daterange=0&newerval=&newertype=&olderval=&oldertype=&bodyprev=#Post248058
Nowhere does Chuck say that ETTL-II becomes disabled when switching to center-weighted. It's still Evaluative. It's just that the evaluations are slightly changed! ;)

I just realized that perhaps were having a disagreement over the semantic meaning of the word "evaluative."

I guess we are talking about two different things. If by ETTL, you are referring to the mechanism by which the camera fires a preflash, evaluates the data returned by the sensors and then sets the appropriate exposure value for the shot, then you are correct, ETTL is not disabled unless the flash in put into manual mode. I thought you were talking about the method used by the camera to evaluate the preflash reading, which would either be evaluative or center-weighted.

[PacAce]

I don't believe this is entirely correct for a 10D. ETTL metering on a 10D is heavily biased to the active AF points, which is definitely not the way evaluative, ambient light metering works.
Thanks!

I can most definitely assure you that evaluative metering for ambieant light is also AF point biased. The easiest way to test this is to tape a black lens cap on the blank white wall and point the leftmost (or the rightmost) AF point at the lens cap. Set the camera to wall distance so that the lens cap covers an area the size of the partial metering circle found in the middle of the viewfinder. Now, set the camera to P, Av or Tv and manually set AF point to where the cap is and take a shot. Now, move the AF point (without moving the camera) to the middle AF point, or even to the AF on the other end opposite where it was first set to, and take another shot. If the metering is not AF point biased, the two shots should look exactly alike but you'll find that they're not. With the AF over the black cap, the image comes exposed more (imsge is lighter) than when the AF point is not over the black cap (image is darker).

[slin100]

Slin, I believe you are in error on this issue. Photnotes.org is also wrong on this issue. Chuck Westfall of Canon has advised in Rob Galbraith's Forum, that Manual mode eliminates the Auto Reduction of fill flash.
Notwithstanding Chuck's advice, my experience suggests otherwise. I ran a quick test and I clearly see the effects of toggling Auto Reduction of Fill Flash (C.Fn 14 on a 10D) while my camera is set to M mode. I'll try to post pictures of my results tonight.

BTW, I discovered that the 1D Mk II does not have Auto Reduction of Fill Flash at all. I'm not sure about the 20D, but it may be the same as the 1D Mk II.

[PacAce]

BTW, I discovered that the 1D Mk II does not have Auto Reduction of Fill Flash at all. I'm not sure about the 20D, but it may be the same as the 1D Mk II.

I think you will find that there is more than a superficial difference between the way ETTL and ETTL-II work and that one statement that may be true for ETTL does not necessarily apply for ETTL-II, and vice versa. Sometimes it gets very confusing, especially when the discussion of both is going on simultaneously and in the same thread. :D

[slin100]

I think you will find that there is more than a superficial difference between the way ETTL and ETTL-II work and that one statement that may be true for ETTL does not necessarily apply for ETTL-II, and vice versa. Sometimes it gets very confusing, especially when the discussion of both is going on simultaneously and in the same thread. :D
I never suggested anything to the contrary. I'm well aware of many of the differences between ETTL and ETTL-II.

[PacAce]

I never suggested anything to the contrary. I'm well aware of many of the differences between ETTL and ETTL-II.

I wasn't suggesting that you were. All I was implying is that you have practical experience with ETTL. However, I have not seen anything to suggest that you also have any practical ETTL-II experience, other than quoting other sources, like from other forums. And, because you have seen it quoted elsewhere by someone else, if it supports what you are saying, then you assume that it must b e true. However, if it doesn't, then it must not be true, at least based on your own experience.

Case in point:

If you still don't agree that ETTL/ETTL-II is still active even when the flash metering is center-weighted, please read Canon marketing director Chuck Westfall's post on ETTL-II where he includes an excerpt from the 1D2 white paper. http://www.robgalbraith.com/ubbthre...rev=#Post248058
Nowhere does Chuck say that ETTL-II becomes disabled when switching to center-weighted. It's still Evaluative. It's just that the evaluations are slightly changed!

and:

Notwithstanding Chuck's advice, my experience suggests otherwise.

So, what I was trying to say, in a nutshell, is that just because a statement made by something is not true based on yhour own experience, don't not make the original statement not true. That's all I was trying to say.

[slin100]

All I was implying is that you have practical experience with ETTL. However, I have not seen anything to suggest that you also have any practical ETTL-II experience, other than quoting other sources, like from other forums. And, because you have seen it quoted elsewhere by someone else, if it supports what you are saying, then you assume that it must b e true. However, if it doesn't, then it must not be true, at least based on your own experience.

I do not have any practical ETTL-II experience and, at this point, I can only rely on the word and evidence of others, including yours. I'd like to think, however, that I'm willing admit when I'm wrong. I try not to always see things in black and white. Instead of saying something is true or false, I'll relate my experiences.

As far as the case in point, I haven't been able to find Chuck's post.

So, what I was trying to say, in a nutshell, is that just because a statement made by something is not true based on yhour own experience, don't not make the original statement not true. That's all I was trying to say.
Point accepted. Unfortunately, when it comes to Canon's flash system, Canon has opted not to be very forthcoming with details, so it's really impossible to authoritatively say whether or not anything is true. Chuck Westfall is about as close to authoritative as we can get at the moment. Other than his words, we can only rely on the preponderence of evidence and experience as reported by all of us to draw conclusions.

Thanks for the discussion.

[slin100]

I can most definitely assure you that evaluative metering for ambieant light is also AF point biased. The easiest way to test this is to tape a black lens cap on the blank white wall and point the leftmost (or the rightmost) AF point at the lens cap. Set the camera to wall distance so that the lens cap covers an area the size of the partial metering circle found in the middle of the viewfinder. Now, set the camera to P, Av or Tv and manually set AF point to where the cap is and take a shot. Now, move the AF point (without moving the camera) to the middle AF point, or even to the AF on the other end opposite where it was first set to, and take another shot. If the metering is not AF point biased, the two shots should look exactly alike but you'll find that they're not. With the AF over the black cap, the image comes exposed more (imsge is lighter) than when the AF point is not over the black cap (image is darker).
Yes, I see. However, there's still a difference between evaluative flash and ambient metering, at least on a 10D, since I have direct experience with it. :)

With flash metering, exposure is determined solely by the active AF points (a kind of spot metering, if you will). With ambient metering, exposure may be influenced by the active AF point, but the readings from the 35 zones also influence exposure.

Now, on a 1D Mk II with ETTL-II, flash metering uses only 17 out of the 21 evaluative zones available. So, ambient and flash metering are different, and in different ways between cameras. Do you know how a 20D works? I'm interested to know.

[PacAce]

Yes, I see. However, there's still a difference between evaluative flash and ambient metering, at least on a 10D, since I have direct experience with it. :)

With flash metering, exposure is determined solely by the active AF points (a kind of spot metering, if you will). With ambient metering, exposure may be influenced by the active AF point, but the readings from the 35 zones also influence exposure.


Again, I have to disagree with your assessment of how flash metering works. Here's a direct quote from Chuck himself on that topic from the following link:

http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Number=137317&page=&view=&sb=5&o=&fpart=1&vc=1

The reason you are seeing an improvement in the consistency of your E-TTL flash exposures with the 10D has to do with the fact that its E-TTL metering algorithm has been signficantly modified from previous EOS digital SLRs, especially D60 and D30. The active focusing point still determines which portion of the scene will receive the most emphasis in terms of flash metering, but the primary flashmetering area is much larger with the 10D than it is with the D60 or D30.


and


When the lens is set to manual focus, all 35 zones receive equal weighting for flash metering in E-TTL. When the lens is set to AF, a large area centered around the active focusing point receives more emphasis than the remainder of the metering area. Incidentally, check your 10D manual on page 75 to see a diagram of the 35-zone metering area. Be sure to note that there is a significant portion of the picture area (all of the edges) that is never metered at all.


I do have to correct myself, though, based on what Chuck stated above. I falsely assumed that average flash metering was center-weighted but that is not the case according to Chuck. The average metering is weighted equally for all the 35 zones.

[scottbergerphoto]

BTW, I discovered that the 1D Mk II does not have Auto Reduction of Fill Flash at all. I'm not sure about the 20D, but it may be the same as the 1D Mk II.
Once again, I believe you are in error, or once again Chuck Westfall and all the pros that use 1D Mark II's are wrong. If you read the threads at www.robgalbraith.com on "Underexposure and the 550EX", it is discussed extensively.

If you have evidence that the 1 Series cameras do not have Auto Reduction of Fill Flash, kindly post a link to a reliable source to back up your "discovery".
Scott

[Tom W]

Awwww, man. An E-TTL vs. E-TTL II discussion and I missed most of it.

Actually, a good read. The discussion has brought out another piece of information for me, as most discussions do.

For the record, my experience with E-TTL vs. E-TTL II strictly from a user's standpoint. Both require +2/3 stop of flash exposure compensation to expose what my eyes (and histogram) judge to be good exposure. BUT, the old E-TTL required +/- 2/3 stop from that point depending on subject color. White (as in wedding dress) required a full = 1 1/3 stops, while black (as in tux) required -1/3 stops. E-TTL 2 is much more consistent. Unless the entire scene is very dark or very bright, no changes are needed.

My estimation (without hard data) is that E-TTL II still uses the primary focus point as part of the algrorithm of evaluating exposure. But it does so at a much smaller degree than E-TTL did. Plus, if available, distance information is taken into consideration.

The II version is definately better, but I'm not going to be happy until I understand it as intuitively as I do f-stop, shutter speed, and ISO. I like to be able to visualize all the variables in my head, and I can't do that with E-TTL I or II.

[slin100]

Once again, I believe you are in error, or once again Chuck Westfall and all the pros that use 1D Mark II's are wrong. If you read the threads at www.robgalbraith.com on "Underexposure and the 550EX", it is discussed extensively.

If you have evidence that the 1 Series cameras do not have Auto Reduction of Fill Flash, kindly post a link to a reliable source to back up your "discovery".
ScottI don't own a 1D Mk II, so I'm not in a position to provide my own evidence. From Chuck Westfall's post (http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB18&Number=231700&Forum=,f18,&Words=%2Bauto%20%2Bflash%20%2Breduction&Searchpage=1&Limit=25&Main=231666&Search=true&where=bodysub&Name=&daterange=0&newerval=&newertype=&olderval=&oldertype=&bodyprev=#Post231700), however:
Back in March, I discussed the issue of automatic flash exposure level reduction:

...based on our testing, E-TTL II on the EOS-1D Mark II no longer applies automatic flash exposure level reduction. This puts more control into the hands of the photographer. You can still control the level of flash fill by adjusting flash exposure compensation, but the E-TTL II algorithm is so accurate that manual adjustment is typically unnecessary.
I find it somewhat amusing that Chuck has to garner this information through testing. Here's another post (http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB18&Number=220588&Forum=,f18,&Words=%2Bauto%20%2Bflash%20%2Breduction&Searchpage=1&Limit=25&Main=220452&Search=true&where=bodysub&Name=&daterange=0&newerval=&newertype=&olderval=&oldertype=&bodyprev=#Post220588) I found from Chuck:

I'm saying that based on my testing, E-TTL II with the EOS-1D Mark II doesn't apply automatic flash exposure level reduction at any EV level. My mention of "bright light" was intended to cover the light levels where the original version of E-TTL applies automatic flash exposure level reduction, which as you state starts at EV 10.

One of the frustrations that I had with the automatic flash exposure level reduction idea (which dates back to the original version of A-TTL with the T90/300TL combination back in 1986) is that it was intended for top-lit or front-lit scenes only. It was actually counter-productive for backlit scenes, especially at dusk and dawn when light levels become high enough to activate the reduction algorithm. That is why I strongly lobbied Canon Inc. in the early 1990s for the CF that let users shut off auto reduction mode. I'm glad they took my suggestion then, and I'm even happier that they developed E-TTL II.
If I'm not mistaken, Scott, you own a 1D Mk II, so you are in a position to refute Chuck.

[slin100]

Again, I have to disagree with your assessment of how flash metering works. Here's a direct quote from Chuck himself on that topic from the following link:

Leo,

I'll concede the point that flash metering uses the same circuitry as ambient metering, and that ambient metering emphasizes the active AF point. It would seem, however, that things are not quite that simple. I found this chart that describes 300D ambient and flash metering behavior. http://doug.kerr.home.att.net/pumpkin/300D_metering_DAK.pdf

It credits Chuck Westfall as the source of information. If it is correct (and I'm still skeptical enough not to necessarily think it is), then it suggests that ambient and flash metering behave very differently. In particular, when the camera is M mode, ambient metering is center-weighted averaged while flash metering is evaluative. I'm quite surprised to find that ambient metering is averaged in M mode. I'll definitely have to check it out.

BTW, everyone please accept my apologies for hijacking this thread. The discussion is about the 20D and many of the points made here don't apply to it. It's really mind-numbing how differently flash exposure is treated among all of Canon's DSLRs.

[scottbergerphoto]

In case anyone still has the energy or desire to continue reading this thead:
1. Page 28 of the 550EX manual: " To disable automatic balanced flash exposures, set the camera's picture taking mode to M".
There are many threads at www.robgalbraith.com addressing this issue. Unfortunately, each new one adds to the confusion over this issue. While in the post you list from 4/04, Chuck indicates that there is no Auto Reduction of fill flash in the 1D MarkII, that becomes unclear in June, when numerous posters disagree with that and indicate that only by going to Manual Mode do they stop seeing it. Personally, I shoot in Manual Mode with the flash in ETTLII and the flash exposure is correct the majority of the time.

2. ETTLII does not use the active AF points. If you don't believe this, take it up with Canon. Don't be confused by the term "subject based". That means that the camera makes a determination where the subject is based on the results of the preflash. That may be a completely different point then where the active AF point is . It is not the same as AF point based.

Stick a fork in me, I'm done. On to the next thread.
Scott

[PacAce]

Leo,

I'll concede the point that flash metering uses the same circuitry as ambient metering, and that ambient metering emphasizes the active AF point. It would seem, however, that things are not quite that simple. I found this chart that describes 300D ambient and flash metering behavior. http://doug.kerr.home.att.net/pumpkin/300D_metering_DAK.pdf

It credits Chuck Westfall as the source of information. If it is correct (and I'm still skeptical enough not to necessarily think it is), then it suggests that ambient and flash metering behave very differently. In particular, when the camera is M mode, ambient metering is center-weighted averaged while flash metering is evaluative. I'm quite surprised to find that ambient metering is averaged in M mode. I'll definitely have to check it out.

BTW, everyone please accept my apologies for hijacking this thread. The discussion is about the 20D and many of the points made here don't apply to it. It's really mind-numbing how differently flash exposure is treated among all of Canon's DSLRs.

I think we're at the end of our discussion here. The original poster started out discussing the 20D and ETTL-II. You then changed the subject to the 10D and ETTL and now you bring up the 300D which is a totally different animal altogehter as far as how metering is selected on it. Have a good day. :)

[Steve Whittaker]

Anyhow I usually put my camera in M mode. Then I will choose my f stop.. say f-8 then choose my time exposure based on my focal length, so 100 mm would be 1/100 or a little more.
With my 550 ex I don't touch the FEC.


Can anyone explain how this works! I shoot mainly indoors, and testing this theory out at home it appears to work. Is it the ETTL 11 flash compensating or what. If this works 99% of the time indoors I'm staying with it.

Please test it and let me know.

Steve

[tim]

Can anyone explain how this works! I shoot mainly indoors, and testing this theory out at home it appears to work. Is it the ETTL 11 flash compensating or what. If this works 99% of the time indoors I'm staying with it.

Please test it and let me know.

Steve

We don't need to test it, we know how things work. Read this (http://photonotes.org/articles/eos-flash/) link, from the FAQ in this forum, and you'll know too.

[Steve Whittaker]

Tim

Sorry for posting in the wrong section, I'm new to the forum!!!!
Excused?

Thanks for the link to the notes, I'll spend some time reading through them.

In the meantime can you assist me with this confusion: If my hand held meter gives me a reading of say 1/15 @ 5.6 indoors
which I know should be correct. However when putting these settings into the camera and the the flash (550ex) on
auto it may overexpose. The theory quoted above would allow me to set as follows and give me a perfect shot set at
focal length 50mm, shutter at 1/50th (or closest) aperture set to say 5.6.

I've never come across this theory before, can you elaborate

Thanks for a good forum.

Steve

Steve

[tim]

No problem Steve, you're in the right place.

If your meter says 1/15th at F5.6 and you add flash it will overexpose, since you'd added more light than it said to use - it meters for ambient light. If you turn your flash off it'll be well exposed, but blurry since you've used 1/15th of a second hand held (I presume). I don't know where you got 1/50th from but it's not from anywhere I know about. Have a read of the EOS bible, it's easier than me explaining exposure, feel free to ask questions once you're done. Play with your camera and flash, since we're digital it's free and easy :)

[Steve Whittaker]

Tim

Thanks for the reply.

What started me on this trail of exploring 20D + 550ex flash settings was the comment made by another member, of which he stated:

Quote: ''Put my camera in M mode. Then I will choose my f stop.. say f-8 then choose my time exposure based on my focal length, so 100 mm would be 1/100 or a little more. With my 550 ex I don't touch the FEC''

As I mentioned previuosly I tried these settings along with many other variations and the results were pretty good, in fact I'd say I was very happy.

Expanding on this further if I was to set camera in M then 1/60 @ 5.6, focal length 60mm to allow me to hand hold, regardless of ambient lighting (darkor light) the results are very good. With this technique if I set focal length at 125mm and shutter at 1/125 then I think I'm right in saying that the results would be good in the same lighting conditions (dark or light)

If all of what I've said is right then it will help me tremendously as I shoot weddings and as you probably are aware, time to sit and think about settings isn't always available. Changing from film has been a steep learning curve.

Thanks again
Steve