Flash metering help

[randy98mtu]

I took this shot with the 40D on-camera flash adjusted -1. Most of the shots taken this way came out great, but this one didn't get enough flash. The one before it was even worse. From reading the 101 tutorial, I'm assuming it's because of the white cat throwing off the flash metering pulled from the pre-flash. Is this correct? If so, I just need to pump my flash back up to 0 when the white cat walks into my shots? Thanks for any help. Yes I know I need the 430EX. My wife just doesn't know that yet. ;)

http://www.pbase.com/image/90734947/large.jpg (http://www.pbase.com/image/90734947)

[squigles]

Have you been shooting in 'M' mode :?:

[randy98mtu]

Yes. That's the only place I've been shooting lately hoping to learn more that way. Would I be better to use something else for flash shots?

[squigles]

Camera shutter speed on 1/60 :?:

I take it you don't have a separate flash unit then :?:

[jrsforums]

I took this shot with the 40D on-camera flash adjusted -1. Most of the shots taken this way came out great, but this one didn't get enough flash. The one before it was even worse. From reading the 101 tutorial, I'm assuming it's because of the white cat throwing off the flash metering pulled from the pre-flash. Is this correct? If so, I just need to pump my flash back up to 0 when the white cat walks into my shots? Thanks for any help. Yes I know I need the 430EX. My wife just doesn't know that yet. ;)


I believe your analysis is correct.

I assume you were in the default, evaluative, flash mode. Evaluative will produce consistent predicable results if you understand how it works. I use it all the time and get good results. It evaluates ambient vs pre-flash across all the metering zones to determine (to the extent possible) which is the nearest and most reflective. It will then set exposure to retain detail there. It does tend to err on the side of not blowing the highlights, but its quite easy to dial in the required FEC by monitoring the overexposure warning and histogram.

....so, yes, the white cat coming into the scene would have a major factor in lowering the flash exposure. You have to know your camera and how it will perform in different scenes/lighting. I do not use on camera flash...use a 580EX. I find, indoors, that I need a few clicks of +FEC, where outdoors needs -FEC. However, I shoot RAW, so want a bit of over exposure without blowing any highlights, then readjust exposure when converting.

[randy98mtu]

I was on 1/125 shutter speed. No, I don't have a separate flash.

Here is the exif from pbase. I know I have a lot to learn, that's why I'm here. I'm still a bit confused by metering for flash shots. Am I better to keep it at 1/60th to get the best overall exposure and then just have the flash bring it up to a 0 exposure? Thanks for your help.

Full EXIF Info
Date/Time 23-Dec-2007 21:47:01
Model Canon EOS 40D
Focal Length 17 mm
Exposure Time 1/125 sec
Aperture f/8
ISO Equivalent 400
Metering Mode (-1)
Exposure Program program (1)

[squigles]

I don't know shed loads, but I would assume 1/125 was rather too quick :?:

I use a 430 in 'M' and bounce the flash off the celling using 1/60 or 1/30 :!:

[jrsforums]

Yes. That's the only place I've been shooting lately hoping to learn more that way. Would I be better to use something else for flash shots?

EXIF from your image:

Exposure time 1/125s
F-Number F 8.0
Exposure program Manual
ISO speed ratings ISO 400

While someone else recommended 1/60, I would keep it at 1/125 as I find kids can move much faster than we realize. I have lots of blurred, slow shutter speed shots, to prove it.

Depending on the DOF you want you could slightly reduce (open) the f-stop....or, increase the ISO to get more ambient light in. However, I generally use 400 as my baseline with the current line Canons.

You might want to take multiple shots of a static, unchanging scene at different FECs to better learn how your setup reacts. Then change the distance and redo the same set.

[randy98mtu]

Thanks guys. I'll try staying at 1/60th next time.

I am also shooting RAW. I had never thought about over-exposing slightly to retain detail and then pulling it back in post. I have generally been underexposing slightly. Mostly because I am still struggling with flash and don't have a 430 to bounce. Thus I can't overexpose in natural light. My house doesn't have much natural light. :(

Edit:

Thanks John. That was why I was at 1/125th to start. I figured the kid and cats would be on the move and I wanted to freeze them. I plan to do more experimenting and figure on using the white cat some more to figure the whole puzzle out. :)

[squigles]

I am still struggling with flash

Join the club on that one, I find quite a few posts on here are very helpful in understanding the workings of using a flash :)

[rabidcow]

Shutter speed only allows more ambient light in when using flash, It has nothing to do with FLASH exposure.

[squigles]

I'm getting a headache now :(

So much to learn so little time :!:

[jrsforums]

Thanks guys. I'll try staying at 1/60th next time.

I am also shooting RAW. I had never thought about over-exposing slightly to retain detail and then pulling it back in post. I have generally been underexposing slightly. Mostly because I am still struggling with flash and don't have a 430 to bounce. Thus I can't overexpose in natural light. My house doesn't have much natural light. :(

Edit:

Thanks John. That was why I was at 1/125th to start. I figured the kid and cats would be on the move and I wanted to freeze them. I plan to do more experimenting and figure on using the white cat some more to figure the whole puzzle out. :)

Some reading for you:

http://super.nova.org/DPR/Canon/Exposure/ While specific to the 580EX, the concepts can be used starting with on-camera flash. BTW...all of his tutorials are worth reading http://super.nova.org/DPR/ .

He is not a fan of bounce flash as he wants control of the lighting/shadows. http://super.nova.org/DPR/Design/ I use both (not together of course)

Here is some basic info on "Expose to the Right" http://www.luminous-landscape.com/tutorials/expose-right.shtml Just be aware that you do not want to overexpose *at all*. Use the histogram and over exposure (black highlight indicators) on the LCD. If the over exposure indicators, for the "whites" you want (you may or may not worry about some specular highlights) are blinking, back off about 1/3 of a stop.

[John E]

For your situation, try shooting at: ISO 400, aperature at f/4-5.6, shutter 100-125, Flash Exposure Compensation (FEC) 0-+1, and see if your pictures improve a bit. By opening the aperature a bit, you will let in more light from the room (ambient light). If you shoot at shutter speeds of 60/sec, you may see some blurriness from the baby and the cats running around. 60/sec works good if people are standing still, not moving around like kids and animals have a habit of doing.

Also, avoid using flash when there is ANYTHING shiny or reflective in the background, such as a glass picture frame, chrome, mirrors, etc.

Also try reading the sticky notes on Flash tutorials in the "small flash and studio lighting" section.

[Wilt]

I took this shot with the 40D on-camera flash adjusted -1. Most of the shots taken this way came out great, but this one didn't get enough flash. The one before it was even worse. From reading the 101 tutorial, I'm assuming it's because of the white cat throwing off the flash metering pulled from the pre-flash. Is this correct? If so, I just need to pump my flash back up to 0 when the white cat walks into my shots? Thanks for any help. Yes I know I need the 430EX. My wife just doesn't know that yet. ;)



The issue is the metallic reflective surface just to the right of the mother! The cat did not affect the exposure, the ETTL simply saw the reflective item!

[BestVisuals]

You had the right ISO and the distance to the subject should've been just fine for a good flash exposure. The white cat shouldn't have made any difference if you were using any other camera than a Canon.

Since you haven't told us what the focusing points were, it's hard to make an exact judgement, however you probably just are suffering from Canon's absolutely horrible flash algorithm. Before I discovered the cause, I had 80% severe underexposure using Canon's ETTL flashes. By the way, switching to a bigger, more powerful flash won't make a bit of difference....it's the camera, not the flash unit.

I don't have time for a long explanation of what's wrong with the Canon E-TTL algorithm, but I can tell you how to fix it - turn it off! On my 20D, it's Custom Function 14, the "E-TTL" setting. Whenever you're shooting with 100% flash (i.e., flash is the main light, not a fill outdoors), set this function to 1:AVERAGE. This disables the E-TTL algorithm and averages the scene. I now get 95% dead-on exposure doing this. Outdoors remember to reset the function back to the default of 0:EVALUATIVE.

You should not have to constantly bias the flash exposure to get good photos! Nikon owners get dead-on flash exposure time after time after time. The poor algorithm in Canons would be the single greatest reason to switch to Nikons for me, I must tell you. In fact, the only real difference between E-TTL and E-TTL II is that E-TTL II lets you turn off E-TTL!

I've just saved you years of frustration with Canon cameras using flash.

[randy98mtu]

The issue is the metallic reflective surface just to the right of the mother! The cat did not affect the exposure, the ETTL simply saw the reflective item!

I didn't notice that. There was some gold wrapping paper and it looks like the wood table was reflecting quite a bit as well. Another thing to watch for with the flash.

Thanks for all the great tips! Most forums aren't so friendly to "beginner" questions. I try to do some searching and reading before I ask questions. It's nice to be able to get feedback on a specific shot/topic though.

[Wilt]

You should not have to constantly bias the flash exposure to get good photos! Nikon owners get dead-on flash exposure time after time after time. The poor algorithm in Canons would be the single greatest reason to switch to Nikons for me, I must tell you. In fact, the only real difference between E-TTL and E-TTL II is that E-TTL II lets you turn off E-TTL!

:lol: Har! har! har! Tell that one to my Nikon DLSR using friends, who had been using Nikon TTL film SLRs! They bitch as loudly about the Nikon dSLR flash! <roflmao>

[BestVisuals]

The amount of reflective wrapping paper shouldn't bias the flash exposure that much, but it might in a Canon if the selected focusing point was on it. The brightest & closest objects are what the focusing system keys to, and unfortunately, it meters the flash output on just what returns from those focusing points. Since we rarely shoot 18% grey subjects, this over-sensitive bias to the focusing points results in severe underexposure. See my previous post on how to fix this with Canon (and why Nikon's algorithm is better).

I didn't notice that. There was some gold wrapping paper and it looks like the wood table was reflecting quite a bit as well. Another thing to watch for with the flash.

Thanks for all the great tips! Most forums aren't so friendly to "beginner" questions. I try to do some searching and reading before I ask questions. It's nice to be able to get feedback on a specific shot/topic though.

[BestVisuals]

There are many, many posts with Nikon photos that have correct exposure. I saw one on this website where the Canon photographer showed a right-out-of-the-camera flash photo next to their Canon photo. The Canon photo was 3 stops underexposed, the Nikon dead on. I'll take the Nikon flash algorithm anyday.

:lol: Har! har! har! Tell that one to my Nikon DLSR using friends, who had been using Nikon TTL film SLRs! They bitch as loudly about the Nikon dSLR flash! <roflmao>

[jrsforums]

The issue is the metallic reflective surface just to the right of the mother! The cat did not affect the exposure, the ETTL simply saw the reflective item!

All the reason why we need to understand how our equipment reacts...i.e. practice, practice, practice :lol:

Actually, I find that ETTL (unbounced) reacts pretty well to specular reflections, whereas, if I am bouncing, it reacts by quickly & drastically quenching the flash exposure. In those cases, I can quickly get the shot by orienting the head to 'normal' position. I have no idea why it reacts that way, but it does. This is with 5D, 580EX, and evaluative flash mode, manual exposure.

In the OP's shot, at -1 FEC, I think the cat is properly exposed. AT least, it looks that way on my monitor. Part of the problem is the -FEC.

[BestVisuals]

By the way, folks, I shoot two 20Ds at weddings and usually take about 800-1200 photos a wedding. Probably 95% of those photos are with flash (either fill or main light). I use flash for EVERYTHING. I'm well acquainted with Canon's flash algorithm. My photos, for the most part, could go to the printers right out of the camera. The way I can do this is by lots of experimentation with the Canon and reviewing the results in-camera.

Turn off ETTL if you want any kind of decent flash exposures using a Canon OEM flash. Or switch to a third-party brand of automatic flash, such as a Metz.

[jrsforums]

You had the right ISO and the distance to the subject should've been just fine for a good flash exposure. The white cat shouldn't have made any difference if you were using any other camera than a Canon.

Since you haven't told us what the focusing points were, it's hard to make an exact judgement, however you probably just are suffering from Canon's absolutely horrible flash algorithm. Before I discovered the cause, I had 80% severe underexposure using Canon's ETTL flashes. By the way, switching to a bigger, more powerful flash won't make a bit of difference....it's the camera, not the flash unit.

Focusing points make no difference with ETTL-II

I don't have time for a long explanation of what's wrong with the Canon E-TTL algorithm, but I can tell you how to fix it - turn it off! On my 20D, it's Custom Function 14, the "E-TTL" setting. Whenever you're shooting with 100% flash (i.e., flash is the main light, not a fill outdoors), set this function to 1:AVERAGE. This disables the E-TTL algorithm and averages the scene. I now get 95% dead-on exposure doing this. Outdoors remember to reset the function back to the default of 0:EVALUATIVE.

Let's say we can agree to disagree. The problem with setting flash metering to "averaging" is that it averages the entire scene without taking into account (or trying to) what the subject is and getting that properly exposed. It does not try to differentiate if the thing it is metering is 3 feet away or 30 feet.

There is a very good tutorial with examples of testing all modes here: http://super.nova.org/DPR/Canon/TTL/

If find evaluative to be the smartest and most consistent. You just need to understand how it attempts to make it's "decisions" and know when it needs to be tweaked.

You should not have to constantly bias the flash exposure to get good photos! Nikon owners get dead-on flash exposure time after time after time. The poor algorithm in Canons would be the single greatest reason to switch to Nikons for me, I must tell you. In fact, the only real difference between E-TTL and E-TTL II is that E-TTL II lets you turn off E-TTL!

I've just saved you years of frustration with Canon cameras using flash.

You seem to think that a metering system which is set to meter to 18% grey can tell if the thing it is metering is actually white or black and automatically adjust for it. Also, by your statement it does not appear you understand what ETTL or ETTL-II do and their differences.

Frankly, if I were that much "in love" with Nikon, I would be shooting them.

[jamesdmo]

RVs, thanks for mentioning an approach I wouldn't have thought of. I've been struggling to get well-exposed indoor flash exposures with my 430EX and XTi. They've consistently tended toward being quite underexposed. After reading your "turn off E-TTL" advice, I made a few comparison shots. No contest - much better without E-TTL!

Now for the real-life comparison today and tomorrow with family visiting for Christmas . . .

[Curtis N]

John's (jrsforums) advice is the most useful in this thread. Those who have used different flash metering systems (traditional auto, TTL and E-TTL in its various forms) know that no flash metering system is fool-proof. All of them require the photographer to learn how the system works, predict its response to a scene and adjust accordingly.

With the 580EX II, I have the choice of three different flash metering systems, and I don't use any of them exclusively. I just try to learn the strengths and limitations of each, and select the best one for the situation.

I like to use Average E-TTL for a lot of indoor situations, but only because I find it's easier for me to predict its behavior and adjust FEC accordingly. That doesn't mean it's best for everyone.

For those who feel the Nikon system is superior, well, they sell some nice cameras too. There's not much point in using a system you think is inferior and then relentlessly complaining about it.

[BestVisuals]

I beg to differ. The focusing points make all the difference in the WORLD when using E-TTL II. If they didn't, changing the algorithm to AVERAGE would make no difference. I have test shots taken with central point focusing that show lots of difference between the central focusing dot being on a light subject and being on a dark subject only inches apart. Using a single, center point of focus like I do makes the situation all that much worse.

The reason the focusing points DO make a difference is that Canon was *trying* to do you a favor and meter on the subject instead of a dark background. Since the camera can select the subject for you (it's the brightest and closest thing the camera sees), Canon was trying to meter the exposure based upon the reflected light from the focusing points. It's good in theory but failed miserably in practice as it turned the metering system into multiple spot meters. I've got wedding photo after wedding photo showing the focusing spots on a white wedding dress (with severe under exposure) even though th dress is not the majority of the scene. It's like spot metering your flash shots to be 18% grey. Once I switched to the AVERAGE mode (which turns OFF the evaluative algorithm), my flash exposures went to 95% accurate.

I quit using multi-point focusing with the Canon camera because about 15% of the time it selects the wrong thing to focus on. Even when locked on a tripod, if I push the shutter release 1/2 way down to establish focus, the &$^!! camera never picks the same focus points twice when nothing has changed. I therefore use a single, central focus point for everything.

Let's all admit that severe underexposure with flash is well documented with the Canon cameras. Setting the algorithm to AVERAGE fixes the problem completely for me. Our modern cameras shouldn't have to be constantly biased to get good flash exposures. I remember great exposures with old Vivitar automatic flashes that were better than with E-TTL, because auto flashes averaged the scene, not one spot.

With a college degree in photography (and PPA Certified), I *do* understand how metering systems work. I never said ANY system KNOWS if the subject is white or black and adjusts for it. No system can know this, I don't know where you get that from. An 18% calibrated meter renders light subjects darker (to make them grey), and darker subjects lighter (to make them grey). The reason the Canon flash algorithm consistently underexposes its photos with flash is that the focusing system selects the brightest, closest subject areas to focus on and meter on. Disconnect the focusing selection from the flash evaluation and you fix the problem.

Canon is no better than Nikon. I give a head nod to Canon for slightly better glass, but that's it. I would be indifferent between either brand (now that I know how to get around Canon's flash problems).

Focusing points make no difference with ETTL-II



Let's say we can agree to disagree. The problem with setting flash metering to "averaging" is that it averages the entire scene without taking into account (or trying to) what the subject is and getting that properly exposed. It does not try to differentiate if the thing it is metering is 3 feet away or 30 feet.

There is a very good tutorial with examples of testing all modes here: http://super.nova.org/DPR/Canon/TTL/

If find evaluative to be the smartest and most consistent. You just need to understand how it attempts to make it's "decisions" and know when it needs to be tweaked.



You seem to think that a metering system which is set to meter to 18% grey can tell if the thing it is metering is actually white or black and automatically adjust for it. Also, by your statement it does not appear you understand what ETTL or ETTL-II do and their differences.

Frankly, if I were that much "in love" with Nikon, I would be shooting them.

[jrsforums]

I beg to differ. The focusing points make all the difference in the WORLD when using E-TTL II. If they didn't, changing the algorithm to AVERAGE would make no difference. I have test shots taken with central point focusing that show lots of difference between the central focusing dot being on a light subject and being on a dark subject only inches apart. Using a single, center point of focus like I do makes the situation all that much worse.

The reason the focusing points DO make a difference is that Canon was *trying* to do you a favor and meter on the subject instead of a dark background. Since the camera can select the subject for you (it's the brightest and closest thing the camera sees), Canon was trying to meter the exposure based upon the reflected light from the focusing points. It's good in theory but failed miserably in practice as it turned the metering system into multiple spot meters. I've got wedding photo after wedding photo showing the focusing spots on a white wedding dress (with severe under exposure) even though th dress is not the majority of the scene. It's like spot metering your flash shots to be 18% grey. Once I switched to the AVERAGE mode (which turns OFF the evaluative algorithm), my flash exposures went to 95% accurate.

Please read this: http://photography-on-the.net/forum/showthread.php?t=46311&amp;highlight=ettlii

It explains that while ETTL used focus points, ETTL-II does not.

I quit using multi-point focusing with the Canon camera because about 15% of the time it selects the wrong thing to focus on. Even when locked on a tripod, if I push the shutter release 1/2 way down to establish focus, the &$^!! camera never picks the same focus points twice when nothing has changed. I therefore use a single, central focus point for everything.

That is a whole different discussion and not germane to the present topic....but I do agree with you.

Let's all admit that severe underexposure with flash is well documented with the Canon cameras.

It does what it does. If you know about it, you adjust to compensate with FEC....which is why I suggested to the OP that -FEC was the wrong direction for indoors.
Setting the algorithm to AVERAGE fixes the problem completely for me. Our modern cameras shouldn't have to be constantly biased to get good flash exposures.

I glad it works for you. I was an advocate for 'averaging' indoors for some time, with 'evaluative' outdoors. However, I found that 'averaging' had it's set of problems also. One big one was if the scene was predominantly dark or the subject was much closer, the subject had a tendency to be significantly overexposed, as the metering was only going to look at the "overall" scene. "Evaluative" is pretty smart and consistent. owever, if you have something closer than the subject you want, it is going to meter on the close item and your "subject" is going to be underexposed. It can only assume that the close thing was the subject you wanted. You then can either recompose with zoom or feet or adjust FEC.
I remember great exposures with old Vivitar automatic flashes that were better than with E-TTL, because auto flashes averaged the scene, not one spot.

Thyristor flashes are not much different than what you are doing with ETTL-II set to 'averaging'.....except that with ETTL the scene being averaged will be what the camera sees, not what the flash sees, which "should" give better results.

With a college degree in photography (and PPA Certified), I *do* understand how metering systems work. I never said ANY system KNOWS if the subject is white or black and adjusts for it. No system can know this, I don't know where you get that from. An 18% calibrated meter renders light subjects darker (to make them grey), and darker subjects lighter (to make them grey). The reason the Canon flash algorithm consistently underexposes its photos with flash is that the focusing system selects the brightest, closest subject areas to focus on and meter on. Disconnect the focusing selection from the flash evaluation and you fix the problem.

As it says in my reference above, the 'evaluative' flash metering is dependent on determining the subject based on the difference in reflective light based on the preflash, not the focual points.

With either flash metering method (or with thyristor) there will be situations where the metering will be thrown off. You will then have to use your experience with the system to adjust.

Canon is no better than Nikon. I give a head nod to Canon for slightly better glass, but that's it. I would be indifferent between either brand (now that I know how to get around Canon's flash problems).

OK...a little different than your original statement.

[randy98mtu]

I experimented tonight indoors at my parents' place with my niece as the subject again. No white cats around. The first few shots I still had my FEC at -1. They were definitely underexposed. So I bumped it back to 0. They looked much better in camera. I haven't gotten them on the computer yet (I brought the card home, but forgot the card reader was with the camera!)

Thanks for the posts and information. It's all very useful.

[Bob_A]

Why do you have FEC at minus 1? Typically ETTL II underexposes, not overexposes so FEC need to be a positve number. My starting point is +2/3, and I adjust from there. Also, as already mentioned shiny surfaces in the frame can mess things up as can flash metering off of something white or black.

For tricky situations flash exposure lock is your friend (* button when using flash).