Question on 550EX Exposure

[Skids]

Hi.
I have been using my 300D and 550ex flash for taking pictures of my nephews and nieces on Christmas Day and I have set the camera to Manual mode at 1/60 Sec and F8.0 (using bounce flash off of the ceiling with a stofen omnibounce attached) following helpful advice from another post.
I am pretty pleased with the results but I noticed that the camera exposes at -2 stops (?) I have worked around this by increasing the FEC on the flash to get the histogram across to the right more and also by shooting in RAW to allow me to set the exposure aftwerwards.
I was just wondering if it is possible to stop the camera underexposing in the first place?

Could you also tell me how the sample pictures here look to a more trained eye as I am not sure whether I could get them better or not by changing any settings (Do I need to increase/decrease exposures etc when importing the RAW data) Images have been cropped and reduced considerably in size.

Many Thanks

Skids

[PacAce]

I'm not sure how far or close your subject was from the camera but let's go through the math and see what we get. The 550EX has a guide number of 42 meters or 138 feet at ISO 100 (50mm). At f/8, the flash will cover a subject within 17 feet from the camera (assuming a hotshoe mounted flash).

The Stofen further decreases the distance covered by the flash by a factor of 2.5. Hence, the distance covered by the flash with the Stofen attached is now down to 6.8 ft, at ISO 100. If you shot at ISO 200, the distance will increase to 9.6 ft. At ISO 400, 13.6 ft. and at ISO 800, 19 ft.

If you were shooting at ISO 100 or ISO 200, and the subject was farther than 7 or 10 feet from you, that may explain why you were getting underexposed pictures. If that's the case, you can increase the ISO to get a farther reach from the flash. Or increase the aperture. Or do both. You can also remove the Stofen at the risk of getting harsher lighting.

[DaveG]

Hi.
I have been using my 300D and 550ex flash for taking pictures of my nephews and nieces on Christmas Day and I have set the camera to Manual mode at 1/60 Sec and F8.0 (using bounce flash off of the ceiling with a stofen omnibounce attached) following helpful advice from another post.
I am pretty pleased with the results but I noticed that the camera exposes at -2 stops (?) I have worked around this by increasing the FEC on the flash to get the histogram across to the right more and also by shooting in RAW to allow me to set the exposure aftwerwards.
I was just wondering if it is possible to stop the camera underexposing in the first place?

Could you also tell me how the sample pictures here look to a more trained eye as I am not sure whether I could get them better or not by changing any settings (Do I need to increase/decrease exposures etc when importing the RAW data) Images have been cropped and reduced considerably in size.

Many Thanks

Skids

I think that this is just another example of how bad the E-TTL is. With a Stofen in a normal room (so you get bounce off of the walls and ceiling) you should easily have enough juice to get a correct exposure.

As for a solution, I play the +/- on my 10D/550 like a piano. Don't be afraid to make a test shot and if you need to go to 2 & 1/2 + then just do so. And trust your histogram not the review screen. Fly by instruments and you won't crash.

[toddb]

Yep, increase distance, increase the ISO. Just get fimilar with how to change the ISO quickly. That's what I do anyway, the ISO 400 looks great, it's more important to get that exposre right.

[davidwegs]

With the 300d I found it will underexpose by default by around 1 stop. That added to a bounced flash and white (ish) frame, (that is most of the shot includes white or near white fore and background) will result in the flash trying to make all matter an average 18% grey. In the examples you have there is a great deal of bland color and therefore I suspect it is using the average grey to calibrate flash output.

You should tell the flash (in that scenario) to expose @ +1.5 stops and that should give you a .5 stop under.... perfect? well no but close without blowing your highlights.

Another thing is to use 800 iso to get less flash and more ambient. F8 is also a little less than ideal too (IMO), I would select f 5.6 or less and isolate the subject from the background a little.

Regards. :)

[robertwgross]

I think that this is just another example of how bad the E-TTL is. With a Stofen in a normal room (so you get bounce off of the walls and ceiling) you should easily have enough juice to get a correct exposure.


This is not true unless you are in a small room, or unless you crank up the ISO a lot. Go through Leo's math.

---Bob Gross---

[Skids]

Thanks for the information guys.

Could you tell me how the attached pictures look after I have imported them from RAW and increased the exposure.

I still do not have an 'eye' for this and don't really know if they look OK or not. I do think the colour seems a bit muted though.

Please be as critical as you wish as I need to learn.

[scottbergerphoto]

ETTL/II works very well if you pay attention to the idiosyncracies of EOS Flash and the new algorithyms of ETTLII that separate flash metering form the active AF points.

1. When using flash in ETTL(300D)/II(20D, 1DMarkII, 1DS MarkII), always put the camera in Manual Mode and the Flash in ETTL/II. That is the only way to avoid the Auto Reduction of Fill Flash. This is one of the major causes of underexposed shots.

2. Flash metering is subject to the same kinds of influences that fool ambient metering. Bright white subjects tend to fool the flash system into underexposing and black subjects tend to overexpose. ETTLII has made this a little better by separating the AF points from the flash metering and by ignoring the very brights and darks in evaluative mode.

3. In ETTLII, you can get more consistent exposures sometimes by using CFn. 14-1(averageing) for interior shots.

4. Use your flashes Guide Number as a general guide to whether you need to increase your ISO or move closer.
The relationship between the distance your flash can cover and the Guide Number is:
Distance(meters/feet) = Guide Number(meters/feet) / f stop at ISO 100
At ISO 200 multiply by 1.4
At ISO 400 multiply by 2

5. When you use a diffuser like the Omnibounce, be sure to add the stops you lose to the equation.

6. If you are using bounce flash, be sure to include the distance to the ceiling and from the ceiling to the subject. Also, watch your flash angle and distance from the subject or you could be bouncing behind the subject.

I have found ETTL and ETTLII even more so, to be very reliable flash systems.

Here are some flash pictures I did last weekend with a 1D Mark II and a 550EX tilted to bounce off the ceiling:
http://www.pbase.com/scottbergerphoto/christmas_2004
Happy Holidays,
Scott

[snibbetsj]

skids, I think those two photos are just a little under exposed but not by much, nothing that you couldn't compensate for in EVU or DPP (or ACR, whatever you use). The advice here is right on the money about the bounce/diffuser screwing up your ETTL-II. It's normal. BTW, even if you up the ISO, you'll still have to do the FEC because ETTL takes the ISO into account.

Practice makes perfect. You're doing pretty good right now.

Happy shooting :)

[DaveG]

This is not true unless you are in a small room, or unless you crank up the ISO a lot. Go through Leo's math.

---Bob Gross---


My definition of a normal room IS a small room. Eight foot white ceilings and pastel coloured walls. More to the point I've lots of images at EXACTLY that setup where the histogram is right in the middle.

[scottbergerphoto]

...... even if you up the ISO, you'll still have to do the FEC because ETTL takes the ISO into account.
:)What does that mean?
Increasing the ISO, increases the distance the flash can travel and provide accurate exposure by way of it increasing the sensitivity of the camera sensor to light. It does not in any way change anything in the ETTL system. If the reason the image is underexposed is because the subject is highly refective, then FEC may still be needed in addition to raising the ISO.
Perplexedly yours,
Scott

[DaveG]

ETTL/II works very well if you pay attention to the idiosyncracies of EOS Flash and the new algorithyms of ETTLII that separate flash metering form the active AF points.

1. When using flash in ETTL(300D)/II(20D, 1DMarkII, 1DS MarkII), always put the camera in Manual Mode and the Flash in ETTL/II. That is the only way to avoid the Auto Reduction of Fill Flash. This is one of the major causes of underexposed shots.

2. Flash metering is subject to the same kinds of influences that fool ambient metering. Bright white subjects tend to fool the flash system into underexposing and black subjects tend to overexpose. ETTLII has made this a little better by separating the AF points from the flash metering and by ignoring the very brights and darks in evaluative mode.

3. In ETTLII, you can get more consistent exposures sometimes by using CFn. 14-1(averageing) for interior shots.

4. Use your flashes Guide Number as a general guide to whether you need to increase your ISO or move closer.
The relationship between the distance your flash can cover and the Guide Number is:
Distance(meters/feet) = Guide Number(meters/feet) / f stop at ISO 100
At ISO 200 multiply by 1.4
At ISO 400 multiply by 2

5. When you use a diffuser like the Omnibounce, be sure to add the stops you lose to the equation.

6. If you are using bounce flash, be sure to include the distance to the ceiling and from the ceiling to the subject. Also, watch your flash angle and distance from the subject or you could be bouncing behind the subject.

I have found ETTL and ETTLII even more so, to be very reliable flash systems.

Here are some flash pictures I did last weekend with a 1D Mark II and a 550EX tilted to bounce off the ceiling:
http://www.pbase.com/scottbergerphoto/christmas_2004
Happy Holidays,
Scott

You must be the only person in the world that found E-TTL to be reliable.

[scottbergerphoto]

If you must know Dave, the fact that Canon thought of me when designing ETTL/II makes me feel very special;) .

All I can tell you is what I experience. The pictures I linked to in my post below, were shot with a 550EX in ETTL II, and a 1D Mark II in Manual mode with Cfn. 14-1.

With regards to ETTL, here is a set of shots using (2) 550EX's and (1) 420EX in wireless ETTL, ratios on, using a 10D in Manual Mode :
http://www.pbase.com/scottbergerphoto/ayana_bailey

If I had problems with it, I would be the first to bitch. Bitching is one of my hobbies.
Regards,
Scott

[planesh00ter]

I was looking at the pics.
I am new to flash and please indulge a basic question..
1) I was looking at Scotts bounce pics, I have been trying verticle bounce and a notch clicked fwd and like the backlite and "softness" on the faces, ( i have a 420ex,drebel) and would like to know how to fill a little in the front, does this Lumiquest 80-20 Pocket Bouncer work for this?
thanks

[DaveG]

If you must know Dave, the fact that Canon thought of me when designing ETTL/II makes me feel very special;) .

All I can tell you is what I experience. The pictures I linked to in my post below, were shot with a 550EX in ETTL II, and a 1D Mark II in Manual mode with Cfn. 14-1.

With regards to ETTL, here is a set of shots using (2) 550EX's and (1) 420EX in wireless ETTL, ratios on, using a 10D in Manual Mode :
http://www.pbase.com/scottbergerphoto/ayana_bailey

If I had problems with it, I would be the first to bitch. Bitching is one of my hobbies.
Regards,
Scott

E-TTL in wireless seems to work very well. In fact wireless TTL is the only reason that I didn't switch back to Vivitar 283's.

But I've got HUNDREDS of shots where E-TTL failed. And these were just were things like mundane corporate cocktail party shots. I'd shoot and it'd be a stop hot. I'd turn around and it would be two stops under! Of course I would get great exposures too, but it was like the flash was just guessing and guessing badly.

E-TTL II seems to be working better. It still requires some care but at least the exposures seem to be within 2/3 of a stop (over or under) now. In any case I thought that Canon invented E-TTL II for ME!

[scottbergerphoto]

In any case I thought that Canon invented E-TTL II for ME!Feel free to share my delusion!:D
Scott

[PacAce]

You must be the only person in the world that found E-TTL to be reliable.

Errr. Make the two. E-TTL works quite well when you understand how it works and know it's limitation or idiosyncrasies and apply the proper compensation (in both directions) when called for.

[AMG]

I agree with skids. I have a 420 ex flash with my 35mm film camera and my indoor shots looked like this too. I find they lack the pop you would get even with a cheap camera with a built in flash, but obviously you get that harsh flash look. I guess my question which I`m sure skids will appreciate, is when we are shooting young kids we really need to shoot fast, we can`t play around too much with the settings. We can however take the time to take a couple of test shots beforehand to see the result and keep the camera at that setting. I guess my question is what exactly is happening in this pic, is flash output basically too low ? What are the best steps to take beforehand to ensure that all the pics taken later will have a nice exposure with good lively color. p.s. I am about to buy the 350d so I will be able to take these test shots, something you can`t do with film minutes before a party.

[Scottes]

It probably should have been done 10 months ago, but I'm moving this to Small Flash...

[LittleG.]

When I initially started with E-TTL, I also had problems with exposure control. Sometimes underexposed and sometimes over. After a time I realised it was not the distance, the E-TTL system, or me that was causing it. It was the function of evaluative metering. If I use spot meter (or centre-weighted if you don't have spot), measure a neutral mid-tone (at the same distance as my subject), and use exposure lock on it, then all is well.

No need for maths in my case ... ;)

[tim]

Once you learn how ETTL works it works fine, but if you want to point and shoot you should probably buy a point and shoot camera, not a pro/semi pro system, which assumes you know what you're doing.

[PacAce]

When I initially started with E-TTL, I also had problems with exposure control. Sometimes underexposed and sometimes over. After a time I realised it was not the distance, the E-TTL system, or me that was causing it. It was the function of evaluative metering. If I use spot meter (or centre-weighted if you don't have spot), measure a neutral mid-tone (at the same distance as my subject), and use exposure lock on it, then all is well.

No need for maths in my case ... ;)
This is very confusing, to say the least. First you mention ETTL, which is associated with flash, then you meniton evaluative, spot and center-weighted metering, which is associated with measuring ambient lighting. The "exposure lock" in your last sentence, is that AEL or FEL?

[LittleG.]

This is very confusing, to say the least. First you mention ETTL, which is associated with flash, then you meniton evaluative, spot and center-weighted metering, which is associated with measuring ambient lighting. The "exposure lock" in your last sentence, is that AEL or FEL? The evaluative meter in the camera is taking a reading from the ambient light from different areas in the scene via the multiple focus points, and averaging out so that for eg if there is a lot of light tones in the background but the subject is actually dark, it will result in underexposure of the subject (sometimes this is what you want eg silhouettes).

With spot metering only one AF point (the centre point) is used to measure the ambient light, so you can choose what to put that point over (usually neutral mid-tone of subject) and then use the FEL button to lock it. This fires a mini flash to record the exposure requirement for the distance/light of the area you have your AF point on.

The ETTL is basically calculating the amount of output it needs to achieve the exposure according to your aperture/speed/distance.

I hope this explains more clearly. I'm not sure if I am misunderstanding myself now!

[scottbergerphoto]

The evaluative meter in the camera is taking a reading from the ambient light from different areas in the scene via the multiple focus points, and averaging out so that for eg if there is a lot of light tones in the background but the subject is actually dark, it will result in underexposure of the subject (sometimes this is what you want eg silhouettes).

With spot metering only one AF point (the centre point) is used to measure the ambient light, so you can choose what to put that point over (usually neutral mid-tone of subject) and then use the FEL button to lock it. This fires a mini flash to record the exposure requirement for the distance/light of the area you have your AF point on.

The ETTL is basically calculating the amount of output it needs to achieve the exposure according to your aperture/speed/distance.

I hope this explains more clearly. I'm not sure if I am misunderstanding myself now!
Whoa! You're mixing apples and oranges here. Evaluative and Spot metering refer to ambient light and have nothing to do with flash metering in ETTL (see ETTL2 below). ETTL is a form of flash metering that biases exposure to the active AF point. You can force ETTL to average all the focus points by changing the lens to Manual Focus. FEL allows you to set your flash exposure and then recompose the picture but keep the same flash output. ETTL2 does not use the active AF point but uses the preflash not only to adjust the flash output, but also to identify where the subject is and bias exposure to the readings coming fom those sensors. It also throws out the high and low readings to reduce the chance that the exposure will be influenced by a very light or dark background. ETTL2 was supposed to make FEL unnecessary. You can also get ETTL2 to average all the flash sensors by switching the CFn.14 from Evaluative to Averaging. This has no effect on the metering of the ambient light.
In addition, Evaluative Metering, of ambient light, heavily biases the active AF point rather then averaging the sensor inputs.

[LittleG.]

Oh well, looks like I'm just as confused as Skids, who I was trying to help. :o

I might have a completely wrong understanding of the workings in the technology, but my methodology works. How come I get the exposure I intend if I've got it all wrong :confused:

[PacAce]

Oh well, looks like I'm just as confused as Skids, who I was trying to help. :o

I might have a completely wrong understanding of the workings in the technology, but my methodology works. How come I get the exposure I intend if I've got it all wrong :confused:
LOL ;) Despite your confusion regarding ambient and flash metering, your methodology for getting a good flash picture is still perfectly valid. It works because what you're doing is one method of working around exposure problems caused by subjects and/or backgrounds that are far from mid-tone in reflectivity. By pointing the partial metering circle at a neutral-toned object and hitting the FEL button, the camera can set the proper flash exposure settings for the picture. :)

[LittleG.]

LOL ;) Despite your confusion regarding ambient and flash metering, your methodology for getting a good flash picture is still perfectly valid. It works because what you're doing is one method of working around exposure problems caused by subjects and/or backgrounds that are far from mid-tone in reflectivity. By pointing the partial metering circle at a neutral-toned object and hitting the FEL button, the camera can set the proper flash exposure settings for the picture. :)
Phew. That's a relief. :D

I think I'll leave all the complicated stuff to the big boys. I'll never make a techie advisor ... :rolleyes:

[PacAce]

Phew. That's a relief. :D

I think I'll leave all the complicated stuff to the big boys. I'll never make a techie advisor ... :rolleyes:
What is more important is how you use the tool and and the results you get from it and, based on what I've seen on your site, you seem to be doing a very good job of it. I wouldn't worry too much about the techie stuff. :D

[bachscuttler]

based on what I've seen on your site, you seem to be doing a very good job of it.

Geez...and what a site! :p

That is a beauty! (I mean the actual site design, as well as the images)

Back on track, I just picked my 580EX up today.
The very first thing I noticed was that it underexposed, even after reading the manual.

I've picked up a few tips and settings in this thread and it now performs beautifully.

[LittleG.]

Geez...and what a site! :p

That is a beauty! (I mean the actual site design, as well as the images)

Hey, surely you guys aren't talking about moi petite website? Or should it be petite website de moi? Maybe I'll leave out the French grammar as well as the techie stuff ... :o

Shucks ... thanks guys :-D

[René Damkot]

I might have a completely wrong understanding of the workings in the technology, but my methodology works. How come I get the exposure I intend if I've got it all wrong :confused:
You've got it right (assuming you use a 1 series)... There's some confusion on the subject in the manual, but FEL on a 1 series does allow spot metering the preflash.

[PacAce]

You've got it right (assuming you use a 1 series)... There's some confusion on the subject in the manual, but FEL on a 1 series does allow spot metering the preflash.
From the tests that I have done, the FEL for the 1DmkII seems to be using the whole partial metering area instead of the AF point only. And it doesn't matter what metering mode I'm using, which is no suprise since flash metering is independent of ambient light metering. So, it would seem like all digital EOS camera use the partial metering area for FEL. I don't have a 1Ds to verify this but I'm assuming it's going to behave just like the 1D2.

[René Damkot]

I'm not going to go over this again, been there, done that... Probabely has to do with your setting of CFn 13. Take a look in This thread (http://photography-on-the.net/forum/showthread.php?t=107977)

[PacAce]

I'm not going to go over this again, been there, done that... Probabely has to do with your setting of CFn 13. Take a look in This thread (http://photography-on-the.net/forum/showthread.php?t=107977)
Yes, I've seen that thread. I'm not disputing that the FEL follows the AF point around the viewfinder. What I'm disputing is the notion that the FEL is a sort of spot metering for flash which my tests show that it is not the case. In other words, when you do an FEL, the area metering with the preflash is not the 3.8% of the viewfinder associated with spot metering but rather the 13.5% of the viewfinder associated with the partial metering area.

The following shots were taken with the same camera settings with the AF point over the center of the lens cover and the SD card (the SD card is slightly larger then the AF point). If the FEL metering were spot metered, then both images should have come out with the same flash exposure reading, but it didn't.

The FEL exposures set in the two images below are more indicative of a partially metered FEL flash exposure. And the only reason I'm pointing this out is because the FEL exposure will not come out as expected if one were assuming that the FEL were behaving as a spot meter (the AF point) instead of a partial meter (the AF point and a larger area around it).
.

[René Damkot]

Lol! That's more specific testing than I've ever done!
I don't know the % of coverage. Might be closer to the 13.5% than to 3.8%. I Do think the SD card is smaller then the spot metering circle in the VF. Then again, when using FEL I alway make sure to cover the subject by the metering point with some room for error, so most of the time it would be close to situation 2.
Would be nice to do a test using centre AF point, FEL and a zoomlens, to find out where the borders of the metering circle are. Might give it a try if I have some spare time...