EDP Manuals

EOS Documentation Project

  EDP Home
  Who
  Manuals
  Legal
  Links
  Join
  Credits

Flash Negative Evaluative Exposure Compensation (NEVEC)

by Tom Musselman and Julian Loke

Contents

Introduction

E-TTL is Canon's latest, greatest flash system, which promised EOS toting photographers a better balanced, smarter, and more versatile flash system. But once all the hype had settled down, several reports pointed to the existence of undocumented "smart" features that took control away from the photographer.

Negative Evaluative Exposure Compensation (NEVEC)

The phenomenon of NEVEC was discovered when users of the EOS 3 and EOS 1V reported that exposure settings in Av mode would change as soon as the flash was switched on. By careful measurement with a 1V, I found that automatic ambient reduction was not a function of EV, but of light levels. In other words, it is a combination of EV and ISO. Here is a brief table that demonstrates this:

EOS 1V

Adjustment to ambient exposure, in stops, with flash on.

Av or Tv mode
EV ISO 100 ISO 400 ISO 1600
7 -1 -1 -1
8 -1/2 -1 -1
9 0 -1 -1
10 0 -1/2 -1
11 0 0 -1
12 0 0 -1/2
13 0 0 0

My camera was set to 1/2 stop intervals, which is what I use in this example. Set your camera to ISO 100, in either Av or Tv mode, and point it at a scene that meters to EV 8 (for example f/8 and 1/4 second). When you turn on the flash, the camera will change the shutter speed from 1/4 to 1/6 if the camera is in Av mode. If you change the ISO setting of the camera and meter the same scene, the meter returns a different EV, but the results are the same. For example, change the ISO to 1600 and the meter will return EV 12 (f/8 and 1/60), but the reduction will still be 1/2 stop. When the flash is turned on the shutter speed will change from 1/60 to 1/90.

If you are photographing a scene that is much brighter than the transition zone of the automatic ambient reduction (EV 8 at ISO 100), then there is no danger of the camera acting on your behalf and reducing the ambient exposure. Likewise, if the scene is much darker than the transition zone, you can dial in +1 stop exposure compensation and totally defeat the automatic function.

Problems arise when the scene is close to the transition zone. For example, you go to photograph an event in a large room. You are using ISO 400 film. You set the camera to ISO 400, point it to various parts of the room and see that it meters between EV 9 and 11. Depending on the scene you are photographing, the background could be 0, 1/2 or 1 stop underexposed. The camera does not indicate when it is helping your photos. And, you cannot escape the transition zone by changing film speeds! You can either add or remove light, or keep your flash turned off, or change cameras.

Why does it happen?

The response from Canon technical support simply stated this behaviour was normal for the EOS 1V, and it could not be turned off by the user. But E-TTL is an evolving system, with older cameras like the EOS Elan IIE (EOS 50E) being unable to take advantage some of the more advanced system features, such as modelling flash or wireless flash ratio control. Could there be differences between different EOS cameras?

The answer is yes. "Type B" camera bodies, such as the EOS 5 and EOS 1000FN, predate the release of E-TTL, and do not exhibit flash NEVEC behaviour.

Flash NEVEC appears to be programmed into the camera body, rather than being a feature of any specific flash unit. Even the type of flash did not matter. On the Elan 7E (which has a popup flash), flash NEVEC is seen as soon as ANY flash is switched on, whether the flash is a TTL-only popup flash, an A-TTL EZ series Speedlites, or an E-TTL EX series Speedlite.

One day I was using centre-weighted metering on my Elan 7E, and was pleasantly surprised that flash NEVEC did not occur! Further testing showed that flash NEVEC returned, but only when evaluative (and not partial or centre-weighted) metering mode was selected. The camera operating mode (Av, Tv, or M) did not influence NEVEC.

So, it seems that on the Elan 7E, but not on the 1V, flash NEVEC can be disabled by simply changing to partial metering mode or center-weighted metering mode.

  EDP Home
  Who
  Manuals
  Legal
  Links
  Join
  Credits

No choice?

Going back to the EOS 1V, I set up a scene and put the camera on a tripod. The scene was designed to give different meter readings for each of the 4 metering modes of the 1V: spot, partial spot, center-weighted, and evaluative. With a setup like that, it's easy to see that when the flash is turned on in Manual exposure mode, the camera switches to evaluative and stays there. It cannot be switched to any other mode.

Because the amount of flash NEVEC changes with the ambient light, it is possible to apply positive exposure compensation, but the amount required will change as the ambient light changes.

Never say never

So, if you MUST use flash together with Av mode in a fast changing situation, you could always get an Elan 7E.

Another option is to fool the camera that there is no flash mounted in the hotshoe. One way is to use a PC cord to synchronize an automatic flash unit. This can also be simulated by using a carefully placed insulator that only exposes the large centre contact.

By doing so, the camera will only send an X-sync signal to the flash like using a PC-cord. The flash head will not zoom, nor will the flash quench for TTL automatic exposure. So, this mode will work, but it works best if you use fully manual flash or auto-thyristor flash, being careful to zoom the flash head to match your focal length.

Conclusion

I think, overall, the Canon flash system is pretty darn good. It seems clear to me that with these automatic features - the fill flash reduction in bright light, evaluative metering, and ambient reduction in dim light - it is designed for fast shooting in a variety of lighting conditions. I appreciate that because fast shooting is part of my style. The 1V was not meant to be a studio camera, but a super-fast action camera. No wonder Canon has a lock on the sports photography market!

What I can't understand is why two of these three automatic features are undocumented and cannot be disabled. I want to turn off the automation when doing careful work. What if I care about the background-to-subject flash ratio, and still want to shoot fast? Who knows what else is undocumented!

Revision History

v0.1b2003Dec29 jul: edits
v0.1 2003Jan06 Tm & jul: Original draft

Comments

Great work.

I contacted Canon tech support over 1 year ago with NEVEC on my EOS 3, and they wanted me to send the camera in for repairs.

It may be helpful to compile a list of bodies which do have NEVEC. It also could be useful to compile a chart which lists if the ambient meter pattern switches automatically to evaluative when a flash is attached.

For starters, the Elan II and D60 do not have NEVEC and the meter pattern does not automatically switch to evaluative when a flash is attached.
Ken Katz 1/9/2003 5:00:56 PM

I knew about this "NEVEC" phenomenon, or should I say feature but thanks for quantifying it. BTW, what does the acronym stand for? I rarely use Av or Tv mode with flash so I am not overly concerned. I recently took some pictures of some different colored peppers with and without flash fill in Av mode and the ones with flash fill were overexposed. I was not concerned about the background as it was a tight shot. As far as I'm concerned the camera could have reduced the flash exposure more.

Anyhow, since Canon programmed this feature into the EOS-1V, one can only assume they had a reason. They must believe it produces better pictures from the standpoint of exposure ratio between subject and background. So far, I have not heard from anyone who said their pictures were ruined by the effect. So, are the pictures better or worse?

I know some are concerned with maintaining control of all aspects of their photography and get all bent out of shape when the camera does the thinking for them. I have an open mind and if the camera is smarter than I am in a certain instance, I am willing to relinquish control. However, I am still on the fence about this subject because I don't know if it's a good thing or bad.

Anyhow, that's my two cents' worth.
Bill Goldman 24/01/2003 04:21:42

Boy, is my face red. I just noticed the title and I now know what NEVEC stands for.
Bill Goldman 24/01/2003 04:23:41

After some thought, it occurred to me that a better name for this phenomenon might be Automatic Background Exposure Compensation or ABEC. What do you guys think?
Bill Goldman 24/01/2003 06:21:13

I have a number of comments about this effect, and on the ETTL system.

One thing that runs through my mind as I read about your experiment is the fact that in ETTL flash mode, with no CFs set, the camera's exposure algorithm is designed so that the flash part of the exposure will primarily affect the area around the focus point, which is assumed to be the 'subject'. And _how much_ the flash affects the subject will be controlled by the measured ambient light level, varying from no flash at high levels to all flash at low levels.

Bear in mind, as the ETTL designers did, the geometry of a typical picture taking situation where the photographer turns on the flash: The subject at focus point is near, and the background far. This leads to the design decision that the flash will attempt to illuminate mainly the subject, and therefore will be metered there and not elsewhere.

The Canon engineers had to decide what kind of exposure produces aesthetically pleasing photos in this common case, and they knew that they had to try to prevent the background from appearing very dark in the final photo. And because of the aforementioned geometry, they had to design a system that would go easy on the flash so as to allow the background to come up to a reasonable level of exposure with mainly ambient illumination. Even at that, many picture taking situations involve dark backgrounds well behind the subject that cannot be brought up to a reasonable level, due to constraints of film speed, max aperture, and max handheld exposure time.

I have one more guess to make about the aesthetic judgements made by the designers on our behalf: I conjecture that they found the most pleasing flash photos to be the ones in which the background was not real dark, but was not quite as bright as the subject either. Perhaps one stop down looked best.

Now all of this implies that when you switch on the flash, the weighting of the ambient metering segments changes, biasing its ambient calculation _away_ from the subject and _toward_ the background, and vice versa for the flash.

Interestingly, it does this without knowing yet what will be the effect of the flash; it won't know that until you press the shutter release all the way down, at which time it does the pre-exposure evaluation flash.*

In calculating the desired flash energy, the system uses data from the same exposure sensor array, but the data come from the moment the preflash is emitted. And, the weighting is the _inverse_ of the weighting chosen for the ambient, so that the combination of ambient and flash illumination on the subject (focus point) adds up to the right value (i.e., one that will produce a medium density on the film).

And, as stated before, the system proportions illumination at the subject between ambient and flash depending on ambient levels: more flash when the ambient is low, and less flash when the ambient is high.

Now to (try to) relate this to your experiments.

If the subject and background in your experiment were the same brightness under ambient lighting, e.g., a flat gray or white surface, then a 1/2 to 1 stop reduction in the EV reading could be the result of the aesthetic decision I conjecture the designers made (in the fourth paragraph, above).

If your ambient-illuminated subject brightness was different from the background, then the results would be hard to interpret.

That's my $/50.

* I wonder whether the algorithm ever modifies the s & a from what it displayed before the actual exposure command, while the shutter button was halfway down; there could be benefits if it had that degree of freedom, because it doesn't know the entire illumination situation until after the preflash, and the optimum s & a might differ from those calculated before. I guess this could be determined by figuring out the appropriate test case and doing a Flash Exposure Lock on it, and seeing if the a & s changed.



Don Weiss 2/23/2003 10:52:34 PM

Does the NEVEC scenario extend to EV levels below 7; more specifically, does the negative compensation diminish by more than the minus 1 compensation at light levels lower than EV 7? Thank you.
R. Moreno 6/11/2003 9:07:48 PM

Apparently not.
Julian Loke 6/30/2003 2:24:49 AM

I was pleased to have stumbled across this article - I was mystified when I found my EOS-1N was doing exactly this some of the time (obviously, only when light levels where low, which I didn't appreciate at the time). It's extremely annoying - I usually turn the fill-flash ratio right down to -1.7 or -2 with AFR switched off to get the most subtle efect possible (for nature close-ups). I don't want the background underexposed and 'compensating for the compensation' is not really a viable option. So the article hasn't solved my problem but it's nice to know I'm not paranoid! (It also worries me that AFR is ON rather than OFF by default, and that it even kicks in when the camera is in manual mode for total flash photography, but that's another story.)

Anyway, thanks for the useful information.

Chris
Chris Mattison 8/26/2003 8:59:14 PM

In my previous example where I stated that I experienced overexposure of the subject shot in E-TTL flash in Av mode, I have since determined that the flash was closer to the subject than minimum flash range and that was the reason for overexposure.

Since then I've done a test with my EOS-1v, shooting the same subject with flash in P mode and Av mode. These were shot indoors with fairly bright daylight thru large windows.

They were shot on Provia (ISO 100) slide film. They consisted of a vase on a bar and other wooden cabinetry and a light yellow painted wall. The first was shot in P mode at 1/60 and f/3.5 with a 50mm f/1.4 lens. The second shot was in Av mode at 1/15 and f/2.0 (reduced due to NEVEC).

Exposures were virtually identical but color rendition was better in Av mode. The P mode shot was too cool and the wall looks more white than yellow. As far as this test was concerned, I have no complaints. E-TTL seems to working great in all modes.
Bill Goldman 10/7/2003 1:58:56 AM

I have just found this article and have conflicting feelings. First feeling is good, Ì`m happy I am not insane. Second feeling is bad. I want my camera too adjust for the whole scene like all cameras always do, and set the flash off just to eliminate the harsh shadows when shooting outdoor subjects. The article also leaves me feeling, do I get a different camera or just accept it and adjust all my pics post processing. My backgrounds are underexposed, and I am angry, I work in creative modes and I love my flash outdoors, but give me the result your manual promises. I have the elan 7e but want to buy the rebel XT, now I dont know what to do. Great article, very informative.
Adam 2005Oct24 09:19:58 -0700

Users of the Elan 7E or recent Digital SLRs, such as the Digital REBEL XT / EOS 350D, can switch off NEVEC by choosing PARTIAL or CENTRE-WEIGHTED metering mode.
Julian Loke 2005Nov07 15:39:27 -0800

I gather that this NEVEC feature is a changeable situation and since your 7E has a somewhat more controlled approach to it, I might guess that E-TTLII had something to do with it. I noticed a situation that seems to be nearly the same when I shot in AV mode with flash, with my 30D. The pictures looked fine on the camera screen, but when I downloaded them to the computer, they were very much darker. I was able to adjust them back to "normal" appearance with the software. One person at the camera store said that it could be a function of ISO 100 and to boost ISO to 400. So far I've not checked that out. I am doing a lot of metering with the spot, so that could take care of it. Any way, I too like ETTL and it treats me quite kindly most of the time.
Gordon Kraeger 2006May03 14:06:26 -0700

[URL=http://rmzjatle.com]ucdjfgbu[/URL] ktnhinks http://qllyzttq.com xhthetmq thkxgkdy <a href="http://lfmeakui.com">eplekzwm</a>
phgeuixk 2008Mar30 12:18:20 -1000


© 2003 Dec 29 Tom Musselman and Julian Loke for EOS Documentation Project. All Rights Reserved.

Redistribution of any EDP document, text, image, or code, or any work derived therefrom, in any form, electronic or printed, is not allowed without prior written permission from EDP. Copyright notices, including this notice, must remain intact. Use of this document constitutes agreement to our terms and conditions.