Sign in with
Sign up | Sign in
Your question

Exposure problem with 420EX and 20D - please help :-(

Last response: in Digital Camera
Share
September 13, 2005 12:58:47 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Hi All,

I'm hoping someone can help me sort this out because it's kind of
stressing me out.

I just purchased a new 20D having owned a PowerShot G3 for 2 and a half
years, and mostly I'm happy with it. However I am extremely disappointed
by the flash exposure metering. It's notoriously inconsistent and
regularly underexposes the photos. I'm not talking about difficult
lighting situations either. I've been testing it just pointing at the
uniform colour flat objects and the exposures are all over the place. If
the flash is pointing straight ahead the exposures are generally OK,
however as soon as I start trying to bounce it I get really strange
results. I'm guessing it has something to do with the E-TTL II metering.

Having done some searches on the Internet I haven't found any useful
solutions. The suggestions I've seen for exposure problems is to just
use FEC. I know there will be situations where the automatic metering
needs a bit of help, but surely this should be the exception rather than
the norm. After all if I'm busy shooting at some kind of function I
don't want to have to be constantly second guessing the cameras
metering. Also often you can't ask your subject to pose again while you
correct the cameras metering bungle.

Anyway the way I did my testing was as follows...
1. Aim the camera down at a patch of beige carpet on the floor about a
meter away from me and take a photo with the flash pointing straight ahead.
2. Repeat this procedure with the flash angled in various other
directions, directly up, sideways, and backwards.

In all cases the green LED flashed on to indicate that the exposure was
sufficient, however the histogram indicated otherwise.

Basically I figured that as long as the flash had enough power, the
resulting exposure should be the same no matter where the flash head was
pointing. The pre-flash metering should surely tell the flash how to
compensate for the different indirect lighting angles.

I did the same test with the 420 connected to my G3 and the resulting
exposures were all spot on. It didn't matter what direction the flash
was pointing, it always managed to get the exposures consistent.

Another thing that bothers me is that on the 20D the flash doesn't seem
to compensate properly for diffusion. I held a couple of sheets of paper
in front of it and I noticed that the exposure started to fall off.
Again, shouldn't the flash power simply be boosted to compensate? As I
said in all situations the flash did not fire at full power and the LED
lit up to indicate correct exposure.

What could be going wrong here?

From my understanding the main difference between the G3 and the 20D as
far as flash metering is concerned is that the G3 is E-TTL rather than
E-TTL II. Could it be that the camera is basically ignoring the
pre-flash and using the lens focus distance instead? I know that E-TTL
II means (among other things) that the camera uses the subject distance
to help calculate flash power.

More often than not I have the flash pointing some other direction. I
almost never have it pointing directly at my subjects. I'll bounce it
off the walls or roof or whatever to get the kind of lighting that I
want. This is going to seriously screw me up if the 20D can't handle
this properly.

Most of the tests I did were with both cameras set to full automatic
mode. However I did a few tests in different creative modes and the
results didn't improve. I've also tried changing custom function #14 and
that didn't seem to make a lot of difference either.

One odd quirk with the 420 is that when the head is rotated (but not
tilted up) the zoom still operates. This seems a bit silly because
clearly if the flash is pointing at right angles to the lens there's no
benefit in it matching focal lengths. I'm guessing it's just a bit of a
design glitch or something. It never bothered the G3, but could it be
confusing the E-TTL II metering on the 20D? Like perhaps it still thinks
that the flash is pointing straight ahead maybe?

Anyway if anyone can suggest what might be wrong I would really
appreciate it. I'm kind of hoping that it's a fault in my camera rather
than a design fault. Of course if I've botched something up that would
be better still, although I've been searching for the "meter properly:
(yes|no)" function and can't find it ;-)

Thanks,

Eugene
September 13, 2005 1:20:00 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Just another really quick test that maybe someone could try to see if
their 20D+420 does the same thing.

1. In a darkened room stand a few meters away and take a photo of a
light coloured wall with the flash pointing forwards. The exposure
should be very even with a narrow histogram curve right in the center.
2. Now tilt the flash directly up and take another photo. The result
should be pretty much the same as #1.
3. Now try the same thing in portrait orientation again with the flash
poiting straight up (it'll be rotated 90 degrees this time rather than
tilted). In this situation I end up with a noticeable exposure fall-off
of about 1 stop.

This to me does not seem in any way reasonable. It just seems like it
must be a fault.

The results get even worse if the roof of wall that's being used to
bounce the light is a darker colour, because the flash power output
doesn't seem to compensate.
Anonymous
September 13, 2005 11:45:03 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Eugene, I share your frustrations.

In long using Metz on TTL with my Medium Format SLR to shoot weddings,
and getting perfect flash and fill flash all the time, flash TTL is
well appreciated by me. When I bought my 20D, I first discovered that
Metz does not advise using my 45CL in TTL mode with my 20D. Then I
discovered that the internal flash on my 20D underexposes like many
post on a variety for forums. Fortunately I can use my Metz on Auto,
but I lose the benefit of TTL. And if I mount of small softbox on my
Metz, its photosensor is blocked from the subject, so I cannot use Auto
either. So then I hope that the 580EX is my salvation, but then I see
a sample photo by someone on a forum, where he has to dial in FEC +2/3
in order to get it to expose properly. I thought the point of FEC was
to dial in - 1/2 EV or -1 EV when you wanted fill flash to be at that
level under ambient, not have to ask for more power simply to reach
ambient exposure!
I have been running tests, and I find that the 20D exposes an average
scene via ambient-only light perfectly on target, but it underexposes
with its internal flash. I suspect the Canon flash exposure software
has some inherent flaw making the underexposure a common complaint.
Perhaps if we all pressure Canon to get their camera's flash exposure
to be consistent with their ambient exposure for the same scene, they
might get their act together and issue a firmware update!

Meanwhile I absolutely will not order a Canon flash unless I first have
an opportunity to rent it (or borrow one) to run my tests.
Related resources
Anonymous
September 13, 2005 1:19:11 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Eugene wrote:
> Just another really quick test that maybe someone could try to see if
> their 20D+420 does the same thing.
>
> 1. In a darkened room stand a few meters away and take a photo of a
> light coloured wall with the flash pointing forwards. The exposure
> should be very even with a narrow histogram curve right in the center.
> 2. Now tilt the flash directly up and take another photo. The result
> should be pretty much the same as #1.
> 3. Now try the same thing in portrait orientation again with the flash
> poiting straight up (it'll be rotated 90 degrees this time rather than
> tilted). In this situation I end up with a noticeable exposure fall-off
> of about 1 stop.
>
> This to me does not seem in any way reasonable. It just seems like it
> must be a fault.
>
> The results get even worse if the roof of wall that's being used to
> bounce the light is a darker colour, because the flash power output
> doesn't seem to compensate.


Canon Speedlights on DSLRs are notorious for underexposing 1.5 to 2
stops. You can do one of two things. Either take it back for a refund
and buy a more appropriate Speedlight or live with the bugger and
compensate for the fault.

I use a 580EX on my 20D and it is always set to over compensate by 1.5
stops. It works OK that way. Canon (as usual) say there is nothing wrong
with the thing. How you can pay $800 AUD for a flash and it doesn't work
properly is beyond me. My Nikon Speedlight (on a Nikon camera) has no
such problems.

The other alternative is to shoot RAW and use RAWShooter (free) to
develop the images. this program is much better at developing under lit
(as opposed to under exposed) Canon 20D images than either Adobe ACR or
Canon's own developer.

There is always at least 1.5 stops of tolerance in a 20D image shot in
RAW mode. The only exception is when the scene has 60% or more of white
(bride's dress) and then you really need to under expose.

They work OK in full auto mode incidentally!

--
Douglas,
My name is but a handle on the doorway to my life.
September 13, 2005 7:23:37 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Thanks for the suggestions. Unfortunately I can't take the flash back as
it's nearly 2 years old. I've been using it on my G3 ever since I bought
it without the slightest trouble. It'd be good if the 20D had the
ability to downgrade to E-TTL instead of always using E-TTL II. I'm
assuming it must be the E-TTL II that's causing problems.

How can they claim that E-TTL II is more advanced and more accurate etc.
Afterall the proof is in the pictures and to me the old E-TTL pre-flash
metering just seems to work much better. I wish I could just tell it to
ignore the subject focus distance and look just at the light that's
bouncing back.

I guess perhaps I could try taking the 20D back and see if I can get a
refund. I'm kind of peeved that after spending that much money it can't
do the exposures as well as my 2 and half year old PowerShot.

I was hoping that perhaps the problem might be limited to the 420,
however if you also notice it with the 580 I guess I'm out of luck there.

I actually have been trying it in full-auto mode and that's when I
noticed the problems. For most situations I use the creative modes, but
sometimes I prefer to just let the camera do it's thing. Now if only the
silly thing had a bit more of a clue I'd be fine. I can see clearly that
it's underexposing why can't the 20D's processor!? It's not exactly
rocket science.

> Eugene wrote:
>
>> Just another really quick test that maybe someone could try to see if
>> their 20D+420 does the same thing.
>>
>> 1. In a darkened room stand a few meters away and take a photo of a
>> light coloured wall with the flash pointing forwards. The exposure
>> should be very even with a narrow histogram curve right in the center.
>> 2. Now tilt the flash directly up and take another photo. The result
>> should be pretty much the same as #1.
>> 3. Now try the same thing in portrait orientation again with the flash
>> poiting straight up (it'll be rotated 90 degrees this time rather than
>> tilted). In this situation I end up with a noticeable exposure
>> fall-off of about 1 stop.
>>
>> This to me does not seem in any way reasonable. It just seems like it
>> must be a fault.
>>
>> The results get even worse if the roof of wall that's being used to
>> bounce the light is a darker colour, because the flash power output
>> doesn't seem to compensate.
>
>
>
> Canon Speedlights on DSLRs are notorious for underexposing 1.5 to 2
> stops. You can do one of two things. Either take it back for a refund
> and buy a more appropriate Speedlight or live with the bugger and
> compensate for the fault.
>
> I use a 580EX on my 20D and it is always set to over compensate by 1.5
> stops. It works OK that way. Canon (as usual) say there is nothing wrong
> with the thing. How you can pay $800 AUD for a flash and it doesn't work
> properly is beyond me. My Nikon Speedlight (on a Nikon camera) has no
> such problems.
>
> The other alternative is to shoot RAW and use RAWShooter (free) to
> develop the images. this program is much better at developing under lit
> (as opposed to under exposed) Canon 20D images than either Adobe ACR or
> Canon's own developer.
>
> There is always at least 1.5 stops of tolerance in a 20D image shot in
> RAW mode. The only exception is when the scene has 60% or more of white
> (bride's dress) and then you really need to under expose.
>
> They work OK in full auto mode incidentally!
>
Anonymous
September 13, 2005 7:23:38 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

I am replying to this thread possibly late in the game w/o having read all
of the replies. However, if not already posted, the OP may find the
writings on EOS flash at the following site helpful:
http://photonotes.org/articles/
Anonymous
September 13, 2005 7:23:38 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

I use a 420EX with a 350D, and the flash exposures are right on the money.
Same E-TTLII. Perhaps there is a problem with your 20D.

-- Martin

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g5noo$s78$1@otis.netspace.net.au...
> Thanks for the suggestions. Unfortunately I can't take the flash back as
> it's nearly 2 years old. I've been using it on my G3 ever since I bought
> it without the slightest trouble. It'd be good if the 20D had the ability
> to downgrade to E-TTL instead of always using E-TTL II. I'm assuming it
> must be the E-TTL II that's causing problems.
>
> How can they claim that E-TTL II is more advanced and more accurate etc.
> Afterall the proof is in the pictures and to me the old E-TTL pre-flash
> metering just seems to work much better. I wish I could just tell it to
> ignore the subject focus distance and look just at the light that's
> bouncing back.
>
> I guess perhaps I could try taking the 20D back and see if I can get a
> refund. I'm kind of peeved that after spending that much money it can't do
> the exposures as well as my 2 and half year old PowerShot.
>
> I was hoping that perhaps the problem might be limited to the 420, however
> if you also notice it with the 580 I guess I'm out of luck there.
>
> I actually have been trying it in full-auto mode and that's when I noticed
> the problems. For most situations I use the creative modes, but sometimes
> I prefer to just let the camera do it's thing. Now if only the silly thing
> had a bit more of a clue I'd be fine. I can see clearly that it's
> underexposing why can't the 20D's processor!? It's not exactly rocket
> science.
>
>> Eugene wrote:
>>
>>> Just another really quick test that maybe someone could try to see if
>>> their 20D+420 does the same thing.
>>>
>>> 1. In a darkened room stand a few meters away and take a photo of a
>>> light coloured wall with the flash pointing forwards. The exposure
>>> should be very even with a narrow histogram curve right in the center.
>>> 2. Now tilt the flash directly up and take another photo. The result
>>> should be pretty much the same as #1.
>>> 3. Now try the same thing in portrait orientation again with the flash
>>> poiting straight up (it'll be rotated 90 degrees this time rather than
>>> tilted). In this situation I end up with a noticeable exposure fall-off
>>> of about 1 stop.
>>>
>>> This to me does not seem in any way reasonable. It just seems like it
>>> must be a fault.
>>>
>>> The results get even worse if the roof of wall that's being used to
>>> bounce the light is a darker colour, because the flash power output
>>> doesn't seem to compensate.
>>
>>
>>
>> Canon Speedlights on DSLRs are notorious for underexposing 1.5 to 2
>> stops. You can do one of two things. Either take it back for a refund and
>> buy a more appropriate Speedlight or live with the bugger and compensate
>> for the fault.
>>
>> I use a 580EX on my 20D and it is always set to over compensate by 1.5
>> stops. It works OK that way. Canon (as usual) say there is nothing wrong
>> with the thing. How you can pay $800 AUD for a flash and it doesn't work
>> properly is beyond me. My Nikon Speedlight (on a Nikon camera) has no
>> such problems.
>>
>> The other alternative is to shoot RAW and use RAWShooter (free) to
>> develop the images. this program is much better at developing under lit
>> (as opposed to under exposed) Canon 20D images than either Adobe ACR or
>> Canon's own developer.
>>
>> There is always at least 1.5 stops of tolerance in a 20D image shot in
>> RAW mode. The only exception is when the scene has 60% or more of white
>> (bride's dress) and then you really need to under expose.
>>
>> They work OK in full auto mode incidentally!
>>
Anonymous
September 13, 2005 7:23:38 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g5noo$s78$1@otis.netspace.net.au...
> Thanks for the suggestions. Unfortunately I can't take the flash back as
> it's nearly 2 years old. I've been using it on my G3 ever since I bought
> it without the slightest trouble. It'd be good if the 20D had the ability
> to downgrade to E-TTL instead of always using E-TTL II. I'm assuming it
> must be the E-TTL II that's causing problems.
>
> How can they claim that E-TTL II is more advanced and more accurate etc.
> Afterall the proof is in the pictures and to me the old E-TTL pre-flash
> metering just seems to work much better. I wish I could just tell it to
> ignore the subject focus distance and look just at the light that's
> bouncing back.
>
> I guess perhaps I could try taking the 20D back and see if I can get a
> refund. I'm kind of peeved that after spending that much money it can't do
> the exposures as well as my 2 and half year old PowerShot.
>
> I was hoping that perhaps the problem might be limited to the 420, however
> if you also notice it with the 580 I guess I'm out of luck there.
>
> I actually have been trying it in full-auto mode and that's when I noticed
> the problems. For most situations I use the creative modes, but sometimes
> I prefer to just let the camera do it's thing. Now if only the silly thing
> had a bit more of a clue I'd be fine. I can see clearly that it's
> underexposing why can't the 20D's processor!? It's not exactly rocket
> science.
>
Anonymous
September 13, 2005 7:23:39 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

> I use a 420EX with a 350D, and the flash exposures are right on the money.
> Same E-TTLII. Perhaps there is a problem with your 20D.

It's certainly possible, as I've mentioned many (perhaps TOO many) times,
I received a 300D which would grossly underexpose with any flash, even
though all of the flashes worked fine on other 300D cameras.

steve
Anonymous
September 13, 2005 7:37:51 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Martin Schiff" <martin@stevegoldman.comDA-DE-DANospam> wrote in message
news:XNydnYeWgKFvm7reRVn-hg@fdn.com...
>I use a 420EX with a 350D, and the flash exposures are right on the money.
>Same E-TTLII. Perhaps there is a problem with your 20D.
>
> -- Martin
>

It's too consistent of a problem with the 20D for the camera to be faulty.
Almost everyone I've talked to with that camera and Canon flashes says the
same thing. Except the ones who shoot with fast glass. That seems to be
the rub, but lenses won't open up enough to expose the sensor correctly,
with all of the folderol of sensing the distance to subject, etc. We had
that problem consistently until we started using f2.8 zooms...

--
Skip Middleton
http://www.shadowcatcherimagery.com
September 14, 2005 10:24:48 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Skip

does this mean the flash works well with a F2.8 lens when its set at 2.8 or
when its set at other apertures as well?

regards

Don

"Skip M" <shadowcatcher@cox.net> wrote in message
news:%wIVe.33$2o.1@fed1read05...
> "Martin Schiff" <martin@stevegoldman.comDA-DE-DANospam> wrote in message
> news:XNydnYeWgKFvm7reRVn-hg@fdn.com...
>>I use a 420EX with a 350D, and the flash exposures are right on the money.
>>Same E-TTLII. Perhaps there is a problem with your 20D.
>>
>> -- Martin
>>
>
> It's too consistent of a problem with the 20D for the camera to be faulty.
> Almost everyone I've talked to with that camera and Canon flashes says the
> same thing. Except the ones who shoot with fast glass. That seems to be
> the rub, but lenses won't open up enough to expose the sensor correctly,
> with all of the folderol of sensing the distance to subject, etc. We had
> that problem consistently until we started using f2.8 zooms...
>
> --
> Skip Middleton
> http://www.shadowcatcherimagery.com
>
September 14, 2005 10:46:48 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Hi Martin,

What about when you bounce the flash of something like a wall? That's
when I encounter problems. With the flash aiming directly at the subject
the results are fine.

It appears to be mainly when I rotate the flash head rather than tilt it
that the exposures get mucked up.

I think I might take the camera back and try to demonstrate the problem
and at least see what they say.

Thanks for the reply,

Eugene

> I use a 420EX with a 350D, and the flash exposures are right on the money.
> Same E-TTLII. Perhaps there is a problem with your 20D.
>
> -- Martin
>
September 14, 2005 10:56:24 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Hi Wilt,

Well I'm kind of glad that I'm not alone in thinking that this kind of
thing really isn't acceptible.

It's very frustrating when that little green LED lights up every time to
indicate that the exposure is successful when it's very obviously not.

Canon obviously know how to get flash exposures right, because as I said
the G3 flash exposures are perfect.

Maybe what I'd like to see is a firmware update to add a new custom
function (if that's possible?) to switch between E-TTL and E-TTL II.
There are just too many situations for me where the lens focus distance
isn't useful at all. This information would surely only be useful when
the flash is pointing directly at the subject without any kind of
diffuser. It seems the 20D is putting too much faith in non-relevant data.

It seems crazy that we should have to go back to auto-flash or manually
setting FEC simply to correct the cameras fumblings. Not good enough
Canon!!!

Anyway thanks very much for your response.

Eugene

> Eugene, I share your frustrations.
>
> In long using Metz on TTL with my Medium Format SLR to shoot weddings,
> and getting perfect flash and fill flash all the time, flash TTL is
> well appreciated by me. When I bought my 20D, I first discovered that
> Metz does not advise using my 45CL in TTL mode with my 20D. Then I
> discovered that the internal flash on my 20D underexposes like many
> post on a variety for forums. Fortunately I can use my Metz on Auto,
> but I lose the benefit of TTL. And if I mount of small softbox on my
> Metz, its photosensor is blocked from the subject, so I cannot use Auto
> either. So then I hope that the 580EX is my salvation, but then I see
> a sample photo by someone on a forum, where he has to dial in FEC +2/3
> in order to get it to expose properly. I thought the point of FEC was
> to dial in - 1/2 EV or -1 EV when you wanted fill flash to be at that
> level under ambient, not have to ask for more power simply to reach
> ambient exposure!
> I have been running tests, and I find that the 20D exposes an average
> scene via ambient-only light perfectly on target, but it underexposes
> with its internal flash. I suspect the Canon flash exposure software
> has some inherent flaw making the underexposure a common complaint.
> Perhaps if we all pressure Canon to get their camera's flash exposure
> to be consistent with their ambient exposure for the same scene, they
> might get their act together and issue a firmware update!
>
> Meanwhile I absolutely will not order a Canon flash unless I first have
> an opportunity to rent it (or borrow one) to run my tests.
>
September 14, 2005 11:46:42 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Thanks for that link. It does have some useful information. I think what
may be happening is that when the flash head is rotated rather than
tilted it may not be indicating this to the camera, so that camera still
thinks that the unit is pointing straight at the subject and meters
accordingly based on subject distance. This would certainly account for
the problems I've had bouncing the flash off the roof when using
portrait orientation.

When the flash head is rotated rather than tilted the zoom feature still
functions, even if the head is rotated a full 180 degrees. This suggests
to me that the flash itself doesn't even know that it's no longer
pointing at the subject.

This would explain also why it works properly on the G3. The G3 being
regular E-TTL doesn't use focus distance to meter the flash, so it'd
just be using the pre-flash, which will give the correct results no
matter where the flash is pointing.

Based on this I think what would be useful from Canon would be a new
custom function to prevent the lens focus distance information from
influencing the flash exposure.

> I am replying to this thread possibly late in the game w/o having read all
> of the replies. However, if not already posted, the OP may find the
> writings on EOS flash at the following site helpful:
> http://photonotes.org/articles/
>
>
Anonymous
September 14, 2005 11:46:58 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

>>Metz were past masters at "auto" flash exposure. So much so that they
became to only portable speedlight worth having. Their latest offering
for Canon (digital) cameras is pretty much disappointing. <<

Are you commenting on Metz offering for Canon cameras (like the SCA3000
series)? Or is this comment about the 580EX? (My English teacher
would indicate that 'they' is a reference to Metz...and I would be
interested to hear why any thoughts about the 54MZ or 70MZ should be
given greater scrutiny as a replacement for my 45CL on Auto when used
with the 20D.


>>I prefer the CT series with a EOS digital adapter and let the flash look
after itself and make it's own decisions. Since going that way, I have
never had a failed exposure due to flash. Before that I had 10 ~30%
underexposed shots using Canon's latest and greatest 580EX... So much
for progress! <<

I just wish I could continue to put a mini softbox on my Metz for
softened lighting, with the 45CL and 20D. Problem is, the mini softbox
blocks the photosensor in the 45CL. When I use the 45CL with my ETRSi,
the TTL flash control gives me perfect exposures (or perfect fill
flash) and the on-board photosensor is not used. But since the 20D
cannot tell the 45CL when to cut off its output, I have to resort to
Auto on the Metz and do without the minibox. That's the reason for my
strong interest in the topic of Canon flash with Canon dSLR (and all
the shortcomings of said design!).
Anonymous
September 14, 2005 11:57:40 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

>>Would a Metz be capable of working TTL on the 20D? Anyone had experience
with this? <<

I ran a test in the past week. Now I understand why Metz' web site
says the combination of the 45CL with Metz 3000 adapter and 3102 module
is 'not recommended'! I thought it was because you did not have so
many of the ETTL features, but I decided that I would use the Metz 45
on TTL mode with the 20D and still have TTL functionality similar to
with my Bronica ETRSi or Olympus OM-4...Wrong! Metz on TTL control
with the 20D is a total failure. So I leave the Metz in Auto mode, and
the Metz adapter and module simply provide synch and viewfinder flash
status.
Anonymous
September 14, 2005 3:01:24 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Eugene,

No problems with bounced flash either.

Try shooting in manual mode. The flash will still work automatically.

-- Martin

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g7drl$1hts$1@otis.netspace.net.au...
> Hi Martin,
>
> What about when you bounce the flash of something like a wall? That's when
> I encounter problems. With the flash aiming directly at the subject the
> results are fine.
>
> It appears to be mainly when I rotate the flash head rather than tilt it
> that the exposures get mucked up.
>
> I think I might take the camera back and try to demonstrate the problem
> and at least see what they say.
>
> Thanks for the reply,
>
> Eugene
>
>> I use a 420EX with a 350D, and the flash exposures are right on the
>> money. Same E-TTLII. Perhaps there is a problem with your 20D.
>>
>> -- Martin
>>
September 14, 2005 7:39:53 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Sorry to sound like a broken record here, but can someone tell me why
the PowerShot G3 is able to get the flash exposures correct, but the 20D
can't? I've read the information at
http://photonotes.org/articles/eos-flash/#ettlii and I now have a better
idea of why it's doing what it's doing, but I'm still confused as to why
they can't solve the problem is they're aware of it. Could they not
simply get the DSLR's to meter the flash the same way that the G3 does?

I just tried going into a dark room and taking photos of a wall with the
flash pointed in various direction. The exposure was perfect every time
with the G3. However with the 20D even +2 FEC sometimes wasn't able to
correct it.

Is it something to do with the G3 using the CCD to read the pre-flash?
Is that how it works?

I'm not an engineer, but I'm sure it shouldn't be that hard. The
processor logic should just be something like...

2) Fire pre-flash
3) Measure bounced light at the active focus point sensors
4) Adjust flash power to suit results from pre-flash. If the pre-flash
metering measured a 2 stop drop in illumination then simply give the
flash 2 stops more power.

Is this hard??? It doesn't seem like it should be. All the extra E-TTL
II stuff about it comparing the ambient and flash results to dissregard
shiny objects etc. is all well and good, but it's really just candy. The
bottom line is that it should get the basic exposure right.

I've done some more tests and I'm almost certain that it's ignoring the
pre-flash. If I take a photo of a wall and then take another with a
piece of paper held in front of the flash then it comes out about 1 stop
under-exposed.

I understand why digitals can't use regular TTL, because the sensor is
too shiny and wouldn't give useful OTF (or OTS as the case may be)
results, but measuring the pre-flash through the lens should still work.

How does Nikon get around the shiny CCD TTL issue?
September 14, 2005 7:39:54 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

In article <dg8d35$1t3m$1@otis.netspace.net.au>, nospamthanks@nospam.com
says...
> Sorry to sound like a broken record here, but can someone tell me why
> the PowerShot G3 is able to get the flash exposures correct, but the 20D
> can't? I've read the information at
> http://photonotes.org/articles/eos-flash/#ettlii and I now have a better
> idea of why it's doing what it's doing, but I'm still confused as to why
> they can't solve the problem is they're aware of it. Could they not
> simply get the DSLR's to meter the flash the same way that the G3 does?
>
> I just tried going into a dark room and taking photos of a wall with the
> flash pointed in various direction. The exposure was perfect every time
> with the G3. However with the 20D even +2 FEC sometimes wasn't able to
> correct it.
>
> Is it something to do with the G3 using the CCD to read the pre-flash?
> Is that how it works?
>
> I'm not an engineer, but I'm sure it shouldn't be that hard. The
> processor logic should just be something like...
>
> 2) Fire pre-flash
> 3) Measure bounced light at the active focus point sensors
> 4) Adjust flash power to suit results from pre-flash. If the pre-flash
> metering measured a 2 stop drop in illumination then simply give the
> flash 2 stops more power.
>
> Is this hard??? It doesn't seem like it should be. All the extra E-TTL
> II stuff about it comparing the ambient and flash results to dissregard
> shiny objects etc. is all well and good, but it's really just candy. The
> bottom line is that it should get the basic exposure right.
>
> I've done some more tests and I'm almost certain that it's ignoring the
> pre-flash. If I take a photo of a wall and then take another with a
> piece of paper held in front of the flash then it comes out about 1 stop
> under-exposed.
>
> I understand why digitals can't use regular TTL, because the sensor is
> too shiny and wouldn't give useful OTF (or OTS as the case may be)
> results, but measuring the pre-flash through the lens should still work.
>
> How does Nikon get around the shiny CCD TTL issue?

He-he. By not putting out "beta" products for the general public to test
for them!

As always I have to blow Nikons trumpet. I have been using the new SB-
800 unit on all my Nikon cameras in its various modes. I have used it on
an FE2, F4s, F2 Photomic, F100 and also the D70. It's never failed me.

The Canon 550EX is a lot better than the 420EX for an abundance of
reasons. There seems to be something inherently "wrong" with the 420EX,
not only in its exposure, but also in its colouring. I really battled
with the white balance using that flash (back when I was using Canon).

--
Look. See. Click. Share.
www.leica.co.za
www.dallasdahms.com
Anonymous
September 14, 2005 7:39:55 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"DD (Rox)" <roxy@empirerods.com> wrote in message
news:MPG.1d91fc8a15dbdec8989717@news.mweb.co.za...
> In article <dg8d35$1t3m$1@otis.netspace.net.au>, nospamthanks@nospam.com
> says...

> As always I have to blow Nikons trumpet. I have been using the new SB-
> 800 unit on all my Nikon cameras in its various modes. I have used it on
> an FE2, F4s, F2 Photomic, F100 and also the D70. It's never failed me.
>
> The Canon 550EX is a lot better than the 420EX for an abundance of
> reasons. There seems to be something inherently "wrong" with the 420EX,
> not only in its exposure, but also in its colouring. I really battled
> with the white balance using that flash (back when I was using Canon).
>


I'm sorry but I just don't see so many problems with my 420EX and 20D. And
as a mater of fact, I have MORE problems with my older 550EX on both my 20D
and 300D over the 420EX. I've used both flashes in all orientations both
fill and sole light and I just don't see these problems all that often.
Occasionally yes but all the time no and a quick adjustment of the FEC
corrects any issues. As to WB, I shoot mostly in RAW so it can be fixed
easily but even there I rarely have to adjust WB.

I'm not saying that people haven't had problems, but if you read through
this thread you're lead to believe that the flash just doesn't work at all
and that's just not true.

--

Rob
September 14, 2005 9:09:48 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Presumably it'd be all the time. Until the exposure starts the lens is
always wide open anyway.

> Skip
>
> does this mean the flash works well with a F2.8 lens when its set at 2.8 or
> when its set at other apertures as well?
>
> regards
>
> Don
>
>
September 14, 2005 9:17:17 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Update:

I went back to the store and demonstrated the problem. It was pretty
easy to show that something was wrong when I also showed them the
results from my G3. While I was there I also tried it on a 350D and same
problem. Someone from Canon is going to call me back about it. I think
if this is a fundamental problem with the DSLR's I'm going to ask for a
refund (Canon that is, not my local camera store). If they object to
that I'll just ask them to point out where this limitation is documented
in the manual.

It's just really not good enough as far as I'm concerned. I should be
able to bounce the flash however I please and the camera should still be
able to calculate the correct exposure. I can't afford to get F2.8 glass
and I wasn't aware that I would need to. I certainly don't blame the
people at the store, I think this might be the first DSLR they've sold,
certainly the first 20D at least.

Eugene
Anonymous
September 14, 2005 9:57:20 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Skip M <shadowcatcher@cox.net> wrote:
<skip wrote nothing>

No offense ... but I am sort of ornry. The only thing I hate more than
top posting is top posting without a message body ;-)

--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
Spammers please contact me at renegade@veldy.net.
Anonymous
September 15, 2005 12:01:45 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Thomas T. Veldhouse" <veldy71@yahoo.com> wrote in message
news:43286480$0$63621$8046368a@newsreader.iphouse.net...
> Skip M <shadowcatcher@cox.net> wrote:
> <skip wrote nothing>
>
> No offense ... but I am sort of ornry. The only thing I hate more than
> top posting is top posting without a message body ;-)
>
> --
> Thomas T. Veldhouse
> Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
> Spammers please contact me at renegade@veldy.net.
>

I know, I'm the same way. I'm also still trying to figure out a) what I
meant to say, there, and b) what the devil happened to it! ;-)

--
Skip Middleton
http://www.shadowcatcherimagery.com
Anonymous
September 15, 2005 12:03:15 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Don" <mackie.don@bigpond.com> wrote in message
news:QmPVe.45150$FA3.23634@news-server.bigpond.net.au...
> Skip
>
> does this mean the flash works well with a F2.8 lens when its set at 2.8
> or when its set at other apertures as well?
>
> regards
>
> Don
>

It works best at f2.8, I've come to terms working in either AV or Manual
mode, though it's a royal pain in the ass with the variable lighting
conditions at a wedding reception...

--
Skip Middleton
http://www.shadowcatcherimagery.com
Anonymous
September 15, 2005 12:04:19 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g8ibo$1v0b$1@otis.netspace.net.au...
> Presumably it'd be all the time. Until the exposure starts the lens is
> always wide open anyway.
>

That just allows it to find focus faster and better. It really doesn't work
any better at f4.5 than it did with my 28-135 IS

--
Skip Middleton
http://www.shadowcatcherimagery.com
Anonymous
September 15, 2005 12:11:00 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g8d35$1t3m$1@otis.netspace.net.au...
> Sorry to sound like a broken record here, but can someone tell me why the
> PowerShot G3 is able to get the flash exposures correct, but the 20D
> can't? I've read the information at
> http://photonotes.org/articles/eos-flash/#ettlii and I now have a better
> idea of why it's doing what it's doing, but I'm still confused as to why
> they can't solve the problem is they're aware of it. Could they not simply
> get the DSLR's to meter the flash the same way that the G3 does?
>
> I just tried going into a dark room and taking photos of a wall with the
> flash pointed in various direction. The exposure was perfect every time
> with the G3. However with the 20D even +2 FEC sometimes wasn't able to
> correct it.
>
> Is it something to do with the G3 using the CCD to read the pre-flash? Is
> that how it works?
>
> I'm not an engineer, but I'm sure it shouldn't be that hard. The processor
> logic should just be something like...
>
> 2) Fire pre-flash
> 3) Measure bounced light at the active focus point sensors
> 4) Adjust flash power to suit results from pre-flash. If the pre-flash
> metering measured a 2 stop drop in illumination then simply give the flash
> 2 stops more power.
>
> Is this hard??? It doesn't seem like it should be. All the extra E-TTL II
> stuff about it comparing the ambient and flash results to dissregard shiny
> objects etc. is all well and good, but it's really just candy. The bottom
> line is that it should get the basic exposure right.
>
Because the G3 is ETTL, not ETTL-II. The "II" measures distance to subject,
and reduces flash power to avoid overexposure. Since Canon's flashes are
biased toward "fill", this will result, if there is any level of ambient
light, in an underexposure. The G3 won't do this, thus the exposures are
better. I get consistently great results with my 420EX and my old D30, on
the 20D, the 420EX has been a disaster, until, like I said earlier, I
started to use f2.8 L glass. That reduced, but did not eliminate, the
problem.
--
Skip Middleton
http://www.shadowcatcherimagery.com
September 15, 2005 2:53:07 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Hi Rob,

Thanks for the response. Since you're using the same gear as me I was
wondering could you possibly do a quick test to see if it's doing the
same thing.

Could you try the following...

1. Set the camera to full-auto mode and take a photo of a blank light
coloured wall with the flash head pointing directly ahead.
2. Take another photo with the flash pointing directly up. Obviously
you'll need to be somewhere with a ceiling ;-)
3. Then another with the camera in portrait orientation and the flash
head swivelled 90 degrees to point straight up again.

When I do this with the 20D the first one is exposed correctly, the
second might be about .5 to 1 stop under, the third is approximately 1
to 2 stops under. Is yours doing something similar?

When I do this with my G3 all the exposures are exactly the same, with
the histogram curve right in the centre.

The 350D at the shop did the same thing as my 20D, however I'm not able
to test on another 20D because they don't have another in stock.

I don't think it could be a fault with my flash unit, since it works
fine on the G3, however I haven't been able to verify this because I
don't have access to another 420.

Thanks,

Eugene

>
> I'm sorry but I just don't see so many problems with my 420EX and 20D. And
> as a mater of fact, I have MORE problems with my older 550EX on both my 20D
> and 300D over the 420EX. I've used both flashes in all orientations both
> fill and sole light and I just don't see these problems all that often.
> Occasionally yes but all the time no and a quick adjustment of the FEC
> corrects any issues. As to WB, I shoot mostly in RAW so it can be fixed
> easily but even there I rarely have to adjust WB.
>
> I'm not saying that people haven't had problems, but if you read through
> this thread you're lead to believe that the flash just doesn't work at all
> and that's just not true.
>
> --
>
> Rob
>
>
>
Anonymous
September 15, 2005 2:53:08 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g96fh$278l$1@otis.netspace.net.au...
> Hi Rob,
>
> Thanks for the response. Since you're using the same gear as me I was
> wondering could you possibly do a quick test to see if it's doing the same
> thing.
>
> Could you try the following...
>
> 1. Set the camera to full-auto mode and take a photo of a blank light
> coloured wall with the flash head pointing directly ahead.
> 2. Take another photo with the flash pointing directly up. Obviously
> you'll need to be somewhere with a ceiling ;-)
> 3. Then another with the camera in portrait orientation and the flash head
> swivelled 90 degrees to point straight up again.
>
> When I do this with the 20D the first one is exposed correctly, the second
> might be about .5 to 1 stop under, the third is approximately 1 to 2 stops
> under. Is yours doing something similar?
>
> When I do this with my G3 all the exposures are exactly the same, with the
> histogram curve right in the centre.
>
> The 350D at the shop did the same thing as my 20D, however I'm not able to
> test on another 20D because they don't have another in stock.
>
> I don't think it could be a fault with my flash unit, since it works fine
> on the G3, however I haven't been able to verify this because I don't have
> access to another 420.


If I get a chance I'll check however, I must tell you I never shoot in Auto
mode. I'm not that failure with the problem as I noted but might the
results you're seeing have to do with auto mode? I'm used to doing things
manually anyway so maybe I'm compensating for the problems that you are
seeing in auto mode.

--

Rob
September 15, 2005 11:47:00 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Yep, tried that also, still get problems. Hopefully when the Canon tech
calls me back I'll be able to get an idea what's going wrong.

> Eugene,
>
> No problems with bounced flash either.
>
> Try shooting in manual mode. The flash will still work automatically.
>
> -- Martin
September 15, 2005 12:09:36 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

I don't often use auto-mode either, but I think it should still work
correctly.

There may be situations where I want to give the camera to a (trusted)
sibling or friend at a function and get them to grab a few shots. I
don't want to have to sit down with them to explain how to do manual
flash compensation.

It's nice to be able to put the camera in auto and know that it'll be
able to get good (correctly exposed) shots.

Incidentally are you using fast glass? Someone else mentioned that the
problem occured when he wasn't using F2.8 lenses.

>
>
> If I get a chance I'll check however, I must tell you I never shoot in Auto
> mode. I'm not that failure with the problem as I noted but might the
> results you're seeing have to do with auto mode? I'm used to doing things
> manually anyway so maybe I'm compensating for the problems that you are
> seeing in auto mode.
>
> --
>
> Rob
>
>
>
September 15, 2005 12:13:03 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

The problem isn't just limited to auto mode. It's just an easy way for
me to demonstrate it, because it shows that there's nothing else going
on to bias the results. For example if I'm using one of the creative
modes then it could be because of something I'm doing wrong, but in
auto-mode the camera is controlling everything.

>
> If I get a chance I'll check however, I must tell you I never shoot in Auto
> mode. I'm not that failure with the problem as I noted but might the
> results you're seeing have to do with auto mode? I'm used to doing things
> manually anyway so maybe I'm compensating for the problems that you are
> seeing in auto mode.
>
> --
>
> Rob
>
>
>
September 15, 2005 5:29:45 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Is it just me then or does E-TTL II seem like a backwards step? Is there
actually any kind of situation where it would result in a more accurate
exposure?

If it causes so many problems for bounce flash exposures then there
should at least be an option to turn it off and revert back to regular
E-TTL.

Incidentally some of the tests I've done have been in almost total
darkness and it still underexposes.

>
> Because the G3 is ETTL, not ETTL-II. The "II" measures distance to subject,
> and reduces flash power to avoid overexposure. Since Canon's flashes are
> biased toward "fill", this will result, if there is any level of ambient
> light, in an underexposure. The G3 won't do this, thus the exposures are
> better. I get consistently great results with my 420EX and my old D30, on
> the 20D, the 420EX has been a disaster, until, like I said earlier, I
> started to use f2.8 L glass. That reduced, but did not eliminate, the
> problem.
Anonymous
September 15, 2005 7:41:25 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D gapr3$2r3i$1@otis.netspace.net.au...
> Is it just me then or does E-TTL II seem like a backwards step? Is there
> actually any kind of situation where it would result in a more accurate
> exposure?
>
> If it causes so many problems for bounce flash exposures then there should
> at least be an option to turn it off and revert back to regular E-TTL.
>
> Incidentally some of the tests I've done have been in almost total
> darkness and it still underexposes.
>
>>
>> Because the G3 is ETTL, not ETTL-II. The "II" measures distance to
>> subject, and reduces flash power to avoid overexposure. Since Canon's
>> flashes are biased toward "fill", this will result, if there is any level
>> of ambient light, in an underexposure. The G3 won't do this, thus the
>> exposures are better. I get consistently great results with my 420EX and
>> my old D30, on the 20D, the 420EX has been a disaster, until, like I said
>> earlier, I started to use f2.8 L glass. That reduced, but did not
>> eliminate, the problem.

I remarked several months ago that E-TTL II seemed to be a step back, at
least compared to my old D30's E-TTL...

--
Skip Middleton
http://www.shadowcatcherimagery.com
Anonymous
September 15, 2005 8:50:08 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Eugene <nospamthanks@nospam.com> writes:
> Would a Metz be capable of working TTL on the 20D? Anyone had
> experience with this?

Plain TTL will not work on a Canon Digital, but Metz sell several
units that is supposed to work with E-TTL. I have a friend that uses
a Metz 54 MZ with the SCA-3102 M4 adapter on a 20D, and he is happy
with it.

For a quick survet E-TTL compatible flash guns, see this webpage:
http://folk.uio.no/gisle/photo/flash.html
--
- gisle hannemyr [ gisle{at}hannemyr.no - http://folk.uio.no/gisle/ ]
------------------------------------------------------------------------
Kodak DCS460, Canon Powershot G5, Olympus 2020Z
------------------------------------------------------------------------
Anonymous
September 16, 2005 1:06:00 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

In article <1126709860.315996.269880@g14g2000cwa.googlegroups.com>,
wiltw@aol.com (wilt) wrote:

> *From:* "wilt" <wiltw@aol.com>
> *Date:* 14 Sep 2005 07:57:40 -0700
>
> >>Would a Metz be capable of working TTL on the 20D? Anyone had
> experience
> with this? <<
>
> I ran a test in the past week. Now I understand why Metz' web site
> says the combination of the 45CL with Metz 3000 adapter and 3102 module
> is 'not recommended'! I thought it was because you did not have so
> many of the ETTL features, but I decided that I would use the Metz 45
> on TTL mode with the 20D and still have TTL functionality similar to
> with my Bronica ETRSi or Olympus OM-4...Wrong! Metz on TTL control
> with the 20D is a total failure. So I leave the Metz in Auto mode, and
> the Metz adapter and module simply provide synch and viewfinder flash
> status.
>
>
Canon DSLRs can't do TTL metering because they don't have any
off-the-sensor flash sensors.

When Nikon first brought out their own DSLRs they had to introduce a
new flashgun range at the same time, and no earlier flashguns would
work with the D1. This was because the reflectivity of the sensor was
much different to that of film or shutter blinds. So getting TTL flash
metering to work demanded flashguns which understood this.

Canon took the opposite tack and left TTL metering out of the cameras
altogether. This means that you must use some variant of ETTL which
uses preflash and reads the flash exposure with the ambient metering
sensor.

The other nasty little trick they built in was to helpfully read the
flash exposure from the sensor area corresponding to the active focus
point. While exactly what you'd think you needed, if you focus and
recompose before firing, the flash metering is being performed on where
that sensor is pointing at the moment of pre-exposure, not where you
think the subject is. This can produce wild variations in flash
exposure alone.

Finally there's the fact that all Canon SLRs use fill-in-flash metering
by default in all modes except M and Av. So if you think the flash is
the main light source, you have to use manual as Av will almost always
set a very long shutter speed!. The flash still autoexposes, but this
time to completely light the subject.

Hopefully!
Anonymous
September 16, 2005 4:00:52 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Skip M wrote:
>
>
> I remarked several months ago that E-TTL II seemed to be a step back, at
> least compared to my old D30's E-TTL...
>
Canon's answer?
"The 580 Speedlight is equipped with functionality to either over or
under expose, depending on how you manage it's controls. The 20D also
has custom functions to manage flash exposures with '+' or '-' exposure
compensation. If you took the time to properly understand how the custom
functions of your speedlight work in conjunction with the custom
functions in your camera, I'm sure you will be impressed with their
capabilities".

So for me, when I (rarely now) use the 20D and 580 Speedlight, I set the
dial on the speedlight to over expose 1.5 stops and it seems to do all
it's claimed to. When I use an omni 'mini soft box' I over expose 2.5
stops and it works like I expected it to in the first place.

My personal opinion is that the 580 Speedlight and ETTL2, along with a
20D, work pretty well together. It took me quite a while before it sank
in that the default settings for the flash are moderate in the extreme
and it has adjustments so you can adjust it... A bit like the 20D
camera, really. RTFM is highly relevant with all gear you do not
understand. Modern flash technology is verging on black art.

So now comes the punch line... Who can actually reprogram their
speedlight on the fly? For that matter, who can reprogram their 20D on
the fly? Oh, that's odd. No one!

So take your gear out and experiment. Find an area you need to
understand and read the manual. When you are competent with that, do it
again with another area until you understand one of the most complex SLR
cameras ever made. I have owned mine for almost a year and am only now
*retaining* what I learned about it. That's the trick, isn't it?
Retaining what you discover.

--
Douglas,
Have gun will travel... Said his card.
I didn't care, I shot him anyway.
1/125th @ f5.6. R.I.P. Mamiya.
September 16, 2005 5:07:44 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

The trouble is that although it's not difficult to dial in a bit of
ambient or flash exposure compensation, it still takes time. There also
has to be the oportunity to actually do another shot. I can just imagine
the happy couple walking down the isle and then me saying "can you stop
there for a minute, I just have to fuss around with my camera controls
for a bit". Wouldn't go down well I suspect.

I'm thinking that as someone else suggested it might be primarily
because of the speed of my lens. When I was testing with my G3 I didn't
think to try setting it to the same aperture as my EOS lens (F4.5 -
F5.6). I was comparing both cameras in auto-mode, so my G3 was
automatically setting a much wider aperture of F2 - F3.

I still maintain that the logic in the processor must be a bit screwy.
If the flash is indicating that it has enough power for the shot then
the image should be properly exposed. End of story! I think that the
only time it should be reasonable to get under-exposure is when the
flash just doesn't have enough power. It also shouldn't be assuming that
you only want fill flash. It should just measure the ambient light and
use the flash as required to boost it up to correct exposure.

Anyway I've probably ranted about this particular topic enough ;-) I'm
sure I'll get used to it, as I got used to the AF on my G3. I guess all
cameras have their limitations.

Just out of curiosity. When you set FEC on the flash unit itself, does
it then pass this information to the camera? Like is it basically the
same as setting it on the camera? Is it possible to get +4 by dialing in
+2 on the camera and +2 on the speedlight, or does one override the
other? My 420 doesn't have this facility built-in obviously, and my
local camera shop doesn't keep Canon speedlights in stock because
they're a little too expensive for regular happy snappers. I haven't
therefore been able to play with the high end flash units.


> Canon's answer?
> "The 580 Speedlight is equipped with functionality to either over or
> under expose, depending on how you manage it's controls. The 20D also
> has custom functions to manage flash exposures with '+' or '-' exposure
> compensation. If you took the time to properly understand how the custom
> functions of your speedlight work in conjunction with the custom
> functions in your camera, I'm sure you will be impressed with their
> capabilities".
>
> So for me, when I (rarely now) use the 20D and 580 Speedlight, I set the
> dial on the speedlight to over expose 1.5 stops and it seems to do all
> it's claimed to. When I use an omni 'mini soft box' I over expose 2.5
> stops and it works like I expected it to in the first place.
>
> My personal opinion is that the 580 Speedlight and ETTL2, along with a
> 20D, work pretty well together. It took me quite a while before it sank
> in that the default settings for the flash are moderate in the extreme
> and it has adjustments so you can adjust it... A bit like the 20D
> camera, really. RTFM is highly relevant with all gear you do not
> understand. Modern flash technology is verging on black art.
>
> So now comes the punch line... Who can actually reprogram their
> speedlight on the fly? For that matter, who can reprogram their 20D on
> the fly? Oh, that's odd. No one!
>
> So take your gear out and experiment. Find an area you need to
> understand and read the manual. When you are competent with that, do it
> again with another area until you understand one of the most complex SLR
> cameras ever made. I have owned mine for almost a year and am only now
> *retaining* what I learned about it. That's the trick, isn't it?
> Retaining what you discover.
>
Anonymous
September 16, 2005 5:07:45 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D gdctp$otm$1@otis.netspace.net.au...
> The trouble is that although it's not difficult to dial in a bit of
> ambient or flash exposure compensation, it still takes time. There also
> has to be the oportunity to actually do another shot. I can just imagine
> the happy couple walking down the isle and then me saying "can you stop
> there for a minute, I just have to fuss around with my camera controls for
> a bit". Wouldn't go down well I suspect.


Like anything, once you get to know your equipment you'll know how to set
things before you start shooting.

--

Rob
Anonymous
September 17, 2005 3:23:05 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

In article <43260d12$1@dnews.tpgi.com.au>, Pix on Canvas
<canvaspix@yahoo.com.au> writes
>
>Canon Speedlights on DSLRs are notorious for underexposing 1.5 to 2
>stops. You can do one of two things. Either take it back for a refund
>and buy a more appropriate Speedlight or live with the bugger and
>compensate for the fault.
>
>I use a 580EX on my 20D and it is always set to over compensate by 1.5
>stops. It works OK that way. Canon (as usual) say there is nothing
>wrong with the thing. How you can pay $800 AUD for a flash and it
>doesn't work properly is beyond me. My Nikon Speedlight (on a Nikon
>camera) has no such problems.
>
>The other alternative is to shoot RAW and use RAWShooter (free) to
>develop the images. this program is much better at developing under lit
>(as opposed to under exposed) Canon 20D images than either Adobe ACR or
>Canon's own developer.
>
>There is always at least 1.5 stops of tolerance in a 20D image shot in
>RAW mode. The only exception is when the scene has 60% or more of white
>(bride's dress) and then you really need to under expose.
>
>They work OK in full auto mode incidentally!
>
One thing puzzles me: what do you think is the difference between an
under-lit image and an under-exposed one?

David
--
David Littlewood
Anonymous
September 17, 2005 3:59:12 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

In message <432a2755$1@dnews.tpgi.com.au>,
Pix on Canvas <canvaspix@yahoo.com.au> wrote:

>My personal opinion is that the 580 Speedlight and ETTL2, along with a
>20D, work pretty well together. It took me quite a while before it sank
>in that the default settings for the flash are moderate in the extreme
>and it has adjustments so you can adjust it... A bit like the 20D
>camera, really. RTFM is highly relevant with all gear you do not
>understand. Modern flash technology is verging on black art.

For most situations, you can easily adapt with FEC, but this really
limits how positive you can get. I don't know about you, but if I am
shooting something mostly white, I want +2 to +2.6 EC. You can't do
that when the -2 to +2 range is shifted to -3.5 to +0.5. They should
just give you a big fat range, like -3 to +3, and have a custom function
to reduce for flash fill in bright light, like the 10D. Canon suffers
from too much praise from the magazines that they feed. They forget the
user in a lot of their decisions, and they need a rude awakening.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
Anonymous
September 17, 2005 4:01:30 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

In message <zsOdnaiB9PEY2rfenZ2dnUVZ_sudnZ2d@giganews.com>,
"Robert R Kircher, Jr." <rrkircher@hotmail.com> wrote:

>Like anything, once you get to know your equipment you'll know how to set
>things before you start shooting.

That works fine in some situations, but in others, conditions are
constantly changing, and are hard to keep up with. That's where
intelligent automation comes in.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
Anonymous
September 18, 2005 2:53:44 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"wilt" <wiltw@aol.com> wrote in message
news:1126622703.314783.193740@g49g2000cwa.googlegroups.com...
> Eugene, I share your frustrations.
>
> In long using Metz on TTL with my Medium Format SLR to shoot weddings,
> and getting perfect flash and fill flash all the time, flash TTL is
> well appreciated by me. When I bought my 20D, I first discovered that
> Metz does not advise using my 45CL in TTL mode with my 20D. Then I
> discovered that the internal flash on my 20D underexposes like many
> post on a variety for forums. Fortunately I can use my Metz on Auto,
> but I lose the benefit of TTL. And if I mount of small softbox on my
> Metz, its photosensor is blocked from the subject, so I cannot use Auto
> either. So then I hope that the 580EX is my salvation, but then I see
> a sample photo by someone on a forum, where he has to dial in FEC +2/3
> in order to get it to expose properly. I thought the point of FEC was
> to dial in - 1/2 EV or -1 EV when you wanted fill flash to be at that
> level under ambient, not have to ask for more power simply to reach
> ambient exposure!
> I have been running tests, and I find that the 20D exposes an average
> scene via ambient-only light perfectly on target, but it underexposes
> with its internal flash. I suspect the Canon flash exposure software
> has some inherent flaw making the underexposure a common complaint.
> Perhaps if we all pressure Canon to get their camera's flash exposure
> to be consistent with their ambient exposure for the same scene, they
> might get their act together and issue a firmware update!
>
> Meanwhile I absolutely will not order a Canon flash unless I first have
> an opportunity to rent it (or borrow one) to run my tests.
>

I wonder if there is unit to unit variation then. I did tests when I first
got the 20D and found the ambient and internal flash exposure to be good
under a range of lighting conditions. I also checked both exposures against
reflective and incident light ambient and flash meters.

There was some variation in flash exposure if the target reflectance was
high or low but this was less than +/- 0.5 stop. I also played about with
different active focus points to see if this made much difference with the
same sort of result.

I did note that although correctly exposed in terms of average tone the
typical flash exposure had less dynamic range that an ambient exposure due
to there being no light source included unless there were areas of
significant specula highlights. This can make flash shots "look" a bit
underexposed in terms of filling the histogram to the right, although they
are not in terms of average tonality.

Lester
Anonymous
September 19, 2005 5:10:43 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

Eugene, let us know when you hear something from Canon!

In exchanges I have had with a number of people (including pros who
understand the Canon approach) with the 580EX and the 20D, it seems
they all have 'gotten used to ' having to dial in about +2/3 EV or +1
EV of FEC in order for 'main flash' ( vs. 'fill flash') to work for
them. To me that is backward...If I want flash to equal ambient in
brightness, I should select FEC=0, and if I want fill to be -1 EV
darker than ambient, I select FEC = -1, according to the logic applied
in other systems which I am accustomed to using professionally.

To carry the scenario further, the way it ought to work, if I take an
ambient-only exposure with 20D (meter says 1/8 f/3.5) and get a certain
brightness level via the auto exposure, I should then also be able to
point the flash at the same subject with FEC=0, and have the flash
provide the ***same brightness*** in the resultant exposure as the
ambient-only. That is, if I chose Av f/3.5 the camera and flash is on,
it flashes the 580EX and uses shutter speed 1/250 (if Custom Function
is set to always use 1/250) so that the plane of focus (and other items
at that same plane -- ETTL-II !) are exposed similarly.

In actual testing, it does NOT behave this way. In a purely planar
setting (photographing a large painting) the ambient-only is lit
properly, but the flash-only is seriously underexposed.
Anonymous
September 20, 2005 10:04:06 AM

Archived from groups: rec.photo.digital.slr-systems (More info?)

This seems to be a common problem with the 20D and the flashes. The
built in flash does the same thing. When I first got mine, I noticed
it, and when I investigated, found out most people were having this
problem. A couple people sent the camera and flash in for calibration,
and when they got it backl, exposure was spot on. I have tried it with
a 420EX and a 580EX with the same results. Someday I'll send them all
to Canon for calibration. I just hate to part with a brand new camera
and flash right after I buy them!

Greg
September 20, 2005 12:22:36 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

I didn't get very far with the tech support guy from Canon. He basically
just explained to me that when the flash is bounced there's going to be
light loss. I was well aware of that though. I just don't understand why
the camera can't adequately compensate for this.

One interesting thing I've observed is that using FEL actually improved
the exposure, even if I'm not recomposing the shot. It gets it closer,
but it's still not right.

The issue with it underexposing more when the flash head is rotated
rather than tilted I think may be related to a design fault of the
flash. It seems as though when the flash head is rotated it doesn't
signal to the camera that it's no longer pointing at the subject. What
I'm mainly basing this assumption on is the fact that the auto-zoom on
the flash still operates even when it's rotated a full 180 degrees. This
doesn't really make any sense to me, but suggests that the flash itself
doesn't know that it's pointing in another direction.

I think as someone else suggested that the problem also could be related
to slow glass. I found that when I stopped my G3 down to the same
aperture as my 20D then I started observing similarly inconsistent
exposures. The green LED was always lighting up though and the flash
wasn't firing at full power, so I don't know why this is a problem.

Anyway I think I'm just going to have to get used to the limitation. The
camera obviously has a lot of other redeeming qualities.

> Eugene, let us know when you hear something from Canon!
>
> In exchanges I have had with a number of people (including pros who
> understand the Canon approach) with the 580EX and the 20D, it seems
> they all have 'gotten used to ' having to dial in about +2/3 EV or +1
> EV of FEC in order for 'main flash' ( vs. 'fill flash') to work for
> them. To me that is backward...If I want flash to equal ambient in
> brightness, I should select FEC=0, and if I want fill to be -1 EV
> darker than ambient, I select FEC = -1, according to the logic applied
> in other systems which I am accustomed to using professionally.
>
> To carry the scenario further, the way it ought to work, if I take an
> ambient-only exposure with 20D (meter says 1/8 f/3.5) and get a certain
> brightness level via the auto exposure, I should then also be able to
> point the flash at the same subject with FEC=0, and have the flash
> provide the ***same brightness*** in the resultant exposure as the
> ambient-only. That is, if I chose Av f/3.5 the camera and flash is on,
> it flashes the 580EX and uses shutter speed 1/250 (if Custom Function
> is set to always use 1/250) so that the plane of focus (and other items
> at that same plane -- ETTL-II !) are exposed similarly.
>
> In actual testing, it does NOT behave this way. In a purely planar
> setting (photographing a large painting) the ambient-only is lit
> properly, but the flash-only is seriously underexposed.
>
Anonymous
September 25, 2005 2:26:05 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g7drl$1hts$1@otis.netspace.net.au...
> Hi Martin,
>
> What about when you bounce the flash of something like a wall? That's when
> I encounter problems. With the flash aiming directly at the subject the
> results are fine.
>
> It appears to be mainly when I rotate the flash head rather than tilt it
> that the exposures get mucked up.
>
> I think I might take the camera back and try to demonstrate the problem
> and at least see what they say.
>
> Thanks for the reply,
>
> Eugene
>

Is your lens one that transmits distance info. When the head is turned to
bounce the flash exposure algorithm stops using the distance info so this
might account for the difference.
Anonymous
September 25, 2005 2:37:17 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Eugene" <nospamthanks@nospam.com> wrote in message
news:D g8d35$1t3m$1@otis.netspace.net.au...
> Sorry to sound like a broken record here, but can someone tell me why the
> PowerShot G3 is able to get the flash exposures correct, but the 20D
> can't? I've read the information at
> http://photonotes.org/articles/eos-flash/#ettlii and I now have a better
> idea of why it's doing what it's doing, but I'm still confused as to why
> they can't solve the problem is they're aware of it. Could they not simply
> get the DSLR's to meter the flash the same way that the G3 does?
>

Have you tried checking the flash exposure with an 18% grey card as a target
filling the frame.

If the exposure is off then it is clearly the camera.

Out of curiousity what metering mode and active focus point do you have set?
Anonymous
September 25, 2005 2:42:35 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Skip M" <shadowcatcher@cox.net> wrote in message
news:9D5We.251345$E95.70033@fed1read01...
<snip>
> Because the G3 is ETTL, not ETTL-II. The "II" measures distance to
> subject, and reduces flash power to avoid overexposure. Since Canon's
> flashes are biased toward "fill", this will result, if there is any level
> of ambient light, in an underexposure. The G3 won't do this, thus the
> exposures are better. I get consistently great results with my 420EX and
> my old D30, on the 20D, the 420EX has been a disaster, until, like I said
> earlier, I started to use f2.8 L glass. That reduced, but did not
> eliminate, the problem.
> --
> Skip Middleton
> http://www.shadowcatcherimagery.com
>
>

Probably also explains the difference when the head is rotated.

I am beging to think what is needed is a CF to turn-off the fill in flash
part of the algorithm so the exposure is not biased by high ambient EVs.
Anonymous
September 25, 2005 2:49:07 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

"Ender W." <greg54@operamail.com> wrote in message
news:qb9vi1lbuksafinqutm2h9fsv3j1lsfg3t@4ax.com...
> This seems to be a common problem with the 20D and the flashes. The
> built in flash does the same thing. When I first got mine, I noticed
> it, and when I investigated, found out most people were having this
> problem. A couple people sent the camera and flash in for calibration,
> and when they got it backl, exposure was spot on. I have tried it with
> a 420EX and a 580EX with the same results. Someday I'll send them all
> to Canon for calibration. I just hate to part with a brand new camera
> and flash right after I buy them!
>
> Greg

My internal flash is spot on with a grey card and internal meter and does
basic indoor flash exposure fine.
!