Youngnuo w/Canon Flash question

lodious

lodious

Messages
181
Edit My Images
Yes
I just got a YN-568 to use with a 5D. I'm pretty impressed with the build quality...the plastics and buttons are better than expected. I do have a question though.....

With my EX-420, if i'm using the flash indoors, I put the camera in Manual, set the shutter, ISO and aperture to something sensible, and let the flashgun do the work to give the correct exposure. With the Yonguno, it doesn't appear to work. I have the flash in ETTL mode, but it doesn't appear to expose the picture correctly. Am I doing something wrong? This is a pretty major bummer for me, especially as the auto adjustment of ISO to keep above minimum shutter speed does not work with a flash (Canon or otherwise) attached.

Any idea's would be gratefully received.
 
I've just done some more testing, it is varying the flash power, but just not very well!. If I dial in +2 FEC when bouncing the flash, it's spot on every time, but if I then point the flash directly at the target, it's overexposed by two stops.

It's pretty much the same in Av and Manual....it exposes OK when firing directly at target, but it's under by 2.5stops when bouncing. How strange, I thought the metering would take into account the fact the light was bounced?
 
The plot thickens...it exposes correctly at wide angle zoom settings (on the lens). It get's progressively worse as I zoom in. I must be doing something wrong....although with a 420EX it works fine. :thinking:
 
Strange. Have you anything on the flash, like a Stofen-type diffuser cap? That can upset things if the flash is fired directly at the subject, when the camera tends to use distance information that overrides everything else, and doesn't know the Stofen is attached.

Maybe switch from evaluative to centre-weighted metering (for flash only). It's more consistent if the situation is 'difficult' for the camera to read, when evaluative's smart interpretation is sometimes a bit clever for its own good and can be prone to unpredictable shifts.
 
Last edited:
It's a good thought, but i'm not using any modifiers with it. I've been manually setting the flash zoom to 24 when bouncing. It's fine with the lens at 24mm....it's unusable at 105mm. I've just tried the 420ex again, and it's spot on, every time. I've reset the flash to default settings...i've even read the manual! I had a horrible thought that i'd bought the Nikon version, but it says Canon on the box ;-)
 
I've just tried all the metering modes....it seems to be performing exactly the same regardless of metering. How strange. It's as if it's applying some (too much) compensation for either focal length or focus distance...it's very consistent in how it behaves, but it's not getting it right.
 
Here are some examples to illustrate what appears to be happening.....

Not bad @ 24mm (bounce flash) ....


2J8A3703.jpg by Lodious, on Flickr

Not good at 105mm (bounce flash)....


2J8A3704.jpg by Lodious, on Flickr

OK with flash directly aiming at target @ 105......


2J8A3706.jpg by Lodious, on Flickr

I've tried the flash in lots of different situations. Canon 420EX is spot on every time. I can't see what I can be doing wrong?
 
lodious said:
Any idea's before I order something that actually works?
Click to expand...

Well if you've tried everything, it looks like it must a faulty unit.
 
There's defo something funny going on....if I use the same settings, hold the flash above the hotshoe (i.e. same position), and trigger the flash with a pair of 622's, it works fine :thinking:
 
Just to close the thread out, I can get reliable exposures everytime with YN-568 and 622's. It just doesn't work when mounted on the hotshoe! There are a few threads from people who have similar underexposure problems w/5D and 580EX, so there is a possibility it's a compatibility problem which is not all Youngnuo's fault.

Thanks Hoppy for your suggestions...it's been an interesting problem. I think i'll be keeping the YN-568, as I like the 622 triggers. I can use the lighter 420EX on the camera and remote trigger the YN-568.
 
lodious,

have you been on to the 622C thread and asked Clive Bolton about this. He seems to be the expert with his 'Other YN-622C Guide' Although the 420EX is not listed as compatible, neither is the 270EX, but a friend of mine uses that one without difficulty with the 622's
 
Last edited:
Cheers for the suggestion Alex, but the problem isn't with the triggers, it works fine with the triggers....the flash doesn't work correctly when mounted on the hotshoe! God knows why...there is something funny going on.:wacky:

This was taken with the flash via the triggers.....


2J8A3814.jpg by Lodious, on Flickr

Which looks pretty well exposed. Everything with the flash on the hotshoe is 3 stops underexposed.

I'm going to try the flash on a couple of different bodys at the weekend to see how it works with different combinations of equipment.
 
alexam said:
'Other YN-622C Guide' Although the 420EX is not listed as compatible, neither is the 270EX, but a friend of mine uses that one without difficulty with the 622's
Click to expand...
The compatibility section is is being revised to make clear what is intended.

Both the 270EX and the 420EX will be listed in the class 3 flashes - ETTL but no manual levels.

When the 270EX was listed as "fully compatible", there were those that thought that by using the YN-622s, their no-manual-levels flashes would be magically imbued with such powers.
 
Just to follow up on this.....

I've found a workaround. Setting the External Speedlight Control > E-TTL II meter > Average (the default is Evaluative) sorts the problem with the flash mounted on the camera :)

That's a relief....really pleased with the flash now. I'm not sure if i'm loosing out by not being able to use Evaluative, but the results look correctly exposed to me, so I'm happy.
 
Lodious - Canon's E-TTL II does some remarkable stuff, and sometime it is not helpful. Depending on camera exposure settings, E-TTL will lower ambient exposure to make room for the flash contribution.

ETTL mode/Average throws that out, along with most adjustments. I think that your workaround indicates this as source of your problem.

Search for NEVEC for more info.
 
SeedyBee said:
Lodious - Canon's E-TTL II does some remarkable stuff, and sometime it is not helpful. Depending on camera exposure settings, E-TTL will lower ambient exposure to make room for the flash contribution.

ETTL mode/Average throws that out, along with most adjustments. I think that your workaround indicates this as source of your problem.

Search for NEVEC for more info.
Click to expand...

I thought most Canon shooters used average metering for flash because it's more reliable? That's certainly been my experience, and I'm sure it's been mentioned on this site before.
 
Thanks for the NEVEC search suggestion, i'd never heard of it! I think it's explained why some pictues I took this weekend turned out underexposed when my flash (triggered using the 622's) failed to fire (the batteries were running low) and the pic's were under exposed. I thought that in AV mode the exposure was unchanged when using a flash, but it's makes sense that the exposure is reduced a little by NEVEC.
 
Interesting that when triggering with the 622's, there is no difference in exposure between Evaluative and Average E-TTL metering....does this mean that the 622's don't support Evaluative, and that's why the flash exposure was OK when using the triggers, and not OK when the flash was on the hotshoe?
 
lodious said:
Interesting that when triggering with the 622's, there is no difference in exposure between Evaluative and Average E-TTL metering....does this mean that the 622's don't support Evaluative, and that's why the flash exposure was OK when using the triggers, and not OK when the flash was on the hotshoe?
Click to expand...

Evaluative or average metering is a camera function, but who knows. Evaluative and average will only give different readings in unusual situations, when areas of dark/light tone or other factors in evaluative mode tie in with its (not very) smart brain.

It then makes a choice to favour one tone over the other, sometimes it gets it wrong, and then again slight changes in composition can make it flip the other way - hence the inconsistency. You tend to get more of these ticky lighting situations with flash. I use evaluative for normal light.
 
Perhaps I'm putting two and two together and making 3, but my thinking is that the 568 doesn't correctly interprate some of the information the camera is sending when using evaluative metering. I'm guessing the 622's strip out some of the information, so the flash responds correctly using the triggers, but not when it receives all the information which evaluative metering provides (like perhaps focal length?... seems funny the flash exposes OK at shorter focal lengths...is the camera sending information which a Canon flash would use to adjust the exposure).

I dunno, it just seems strange that the flash works OK when the signal is passed over the triggers?
 
Last edited:
lodious said:
Perhaps I'm putting two and two together and making 3, but my thinking is that the 568 doesn't correctly interprate some of the information the camera is sending when using evaluative metering. I'm guessing the 622's strip out some of the information, so the flash responds correctly using the triggers, but not when it receives all the information which evaluative metering provides (like perhaps focal length?... seems funny the flash exposes OK at shorter focal lengths...is the camera sending information which a Canon flash would use to adjust the exposure).

I dunno, it just seems strange that the flash works OK when the signal is passed over the triggers?
Click to expand...

All of which assumes that your flash is actually working correctly. It may be.

I salute anyone that tries to get to the bottom of these things. Auto-TTL flash is fiendishly complicated, operations vary between cameras and guns and triggers, manufacturers change things without telling anyone, and even the third party manufacturers don't understand it fully or they wouldn't need to keep making firmware updates. Unless Canon just likes to throw a spanner in the works now and then (entirely possible IMHO).

As I understand it, the clever exposure stuff and firing sequence is done by the camera, which then issues simple instructions. The trigger does nothing other than convert those instructions to radio, transmit them, and convert them back at the other end. All the flash then has to do is fire, at the instructed output, when it's told.

Not saying any of that is easy to do, especially at incredible speed, but it's very easy in principle.
 
Yes, agreed. I'm going to try the flash on a couple of different bodys this week, so that should give a clear indication of if the flash is working OK. I'm thinking of getting another 568, so that would also be an interesting test.

For the level i'm operating at, I'm pretty much happy with how the setup is performing...I still have a lot to learn about lighting. I got two wireless flashes reliably triggering in two groups tonight, which is something I've been aiming for quite a while....little steps and all that ;-)
 
Keep us posted (y)
 
Tried the flash with a 550D body tonight....it worked fine in both evaluative and average ETTL metering ....so the YN-568 does not appear to like the 5D3's evaluative metering :-(

Anybody else got this combination who could confirm if they are playing happily together?
 
Phil V was right, but not necessarily helpful, when he said that Average E-TTL gave more consistent exposures than Evaluative E-TTL.

Average mimics film technology and is designed for the old hands who have already developed their art. It is far less precise. You could get even more consistent results by set the flash to M 1/1 for everything.

People who set out to learn how E-TTL works will discover that Evaluative is much more precise in a wider range of images, but this is more demanding of the photog.

The 5D III has many more exposure zones (49?) than the 550D, so has greater precision. Therefore it requires greater care, as the slightest shift in framing can produce a significant change in exposure. It's a skill that needs to be developed.

E-TTL at its heart is designed on the assumption that there is a subject (human) and a background, and the subject dominates in terms of a "correct" exposure. Trying to get E-TTL to work with only a plain wall, or no specific subject is not helpful - it's outside the parameters. (My usual test is with a tripod and Patsy, a blonde head, with a grey card below folded 90 degrees, and a white backdrop. The lighting on the head and card is evaluated, not the backdrop.) When doing comparisons, care must be taken to get a pixel for pixel match in the framing.

As HoppyUK says, E-TTL is done in the camera, and their programming varies between models. The YN-568 just does as it is told, but may have a bug with the advanced 5D3.

I think that your problems arise from a poorly-constructed test environment, and not recognising the importance of framing. E-TTL is that precise.
 
interesting reading for me as I have some YN flashes
however, when you bounce the flash I thought you had no option generally and the light focusing mechanism in the flash head defaults to wide when the flash is angled up at all
 
SeedyBee said:
Phil V was right, but not necessarily helpful, when he said that Average E-TTL gave more consistent exposures than Evaluative E-TTL.

Average mimics film technology and is designed for the old hands who have already developed their art. It is far less precise. You could get even more consistent results by set the flash to M 1/1 for everything.

People who set out to learn how E-TTL works will discover that Evaluative is much more precise in a wider range of images, but this is more demanding of the photog.

The 5D III has many more exposure zones (49?) than the 550D, so has greater precision. Therefore it requires greater care, as the slightest shift in framing can produce a significant change in exposure. It's a skill that needs to be developed.

E-TTL at its heart is designed on the assumption that there is a subject (human) and a background, and the subject dominates in terms of a "correct" exposure. Trying to get E-TTL to work with only a plain wall, or no specific subject is not helpful - it's outside the parameters. (My usual test is with a tripod and Patsy, a blonde head, with a grey card below folded 90 degrees, and a white backdrop. The lighting on the head and card is evaluated, not the backdrop.) When doing comparisons, care must be taken to get a pixel for pixel match in the framing.

As HoppyUK says, E-TTL is done in the camera, and their programming varies between models. The YN-568 just does as it is told, but may have a bug with the advanced 5D3.

I think that your problems arise from a poorly-constructed test environment, and not recognising the importance of framing. E-TTL is that precise.
Click to expand...

Evaluative is usually very good, I use it for normal metering, but it has some inherant problems with flash and I don't think 'recognising the importance of framing' is always practical. I assume you mean using FEL (flash exposure lock)?

Framing is framing, and the picture is what it is. You can't change that without changing the shot. We don't know how evaluative works in detail, because Canon doesn't say though we do know from experience that the alogorithms change as they tweak things with new cameras, as well as changing technical details that throw third-party manufacturers and force them to upgrade their firmware.

The inherant problem with evaluative is its 'smart' thinking when combined with unusual distribution of light and dark areas in the image. That's quite common with flash and evaluative tries to read that and compensate if necessary. But given the unpredictable nature of flash subjects, it sometimes gets it wrong, and then again if you do change the framing slightly, in some situations that will make evaluative's simple brain switch emphasis and change the exposure quite a lot. The result can be a set of pictures, all of the same subject in the same situation, but with exposures varying. With average metering, at least you know how it works, always, and can make an assessment/adjustment knowing that it will hold good.

Another thing that can upset the metering is using accessories and diffusers when the head is pointing directly forward. Fortunately, most accessories are used with the flash head tilted, but if not the system uses distance information as an overriding measure for calculating exposure (logically enough, inverse square law) and ignores the reading from its own pre-flash. Tilting or rotating the head disables this, as does just pressing the tilt release button on my Canon 580EX guns.

FEL is very useful to lock the exposure once you've got that right, but inconvenient when you want to take a series of pictures. That's when I sometimes switch to manual. One simple feature that could help a lot with this would be an indicator that shows the power output of the gun in E-TTL mode. You could then simply read that and apply it manually - quick, easy, reliable. Even better if combined with a button that applies that power setting for you and switches to manual automatically. E-TTL is not just about ease of use, but speed in fluid social situations.
 
Last edited:
Thanks for the insight Hoppy....I think your spot on.

I think we are getting a bit carried away here....my opinion, the YN-568 does not work correctly with a 5d3 when the camera is set to meter using evaluative E-TTL.

It's not to do with learning how the flash works, I think I know enough to know it does not work correctly. It's very unlikely (IMHO) the flash is faulty, It's very unlikely (IMHO) the camera is faulty. For some reason, they just don't work together.

I'm happy to live with using Average E-TTL metering, and overall I'm happy with the flash.....but it would be nice to get to the bottom of this for the benefit of me and other users who stumble on this thread, if for nothing else, than to prove that it's user error on my part so it does not deter anybody from buying a YN-568 to use with a 5d3.

If anybody has some test scenarios they want me to try, I'm more than happy to give them a go. I'm defiantly not after turning this thread into one of the DP Review style threads where someone slates a product to big themselves up.

I think the flash has a problem with the 5d3, why else would it work perfectly with a set of wireless triggers in between them?

The 5d3 works perfectly with a 420EX.

The YN-568 works perfectly with a 550D.


:thinking:
 
Last edited:
Lodious - It seems that you have identified a programming bug. Have you contacted YongNuo? They may be able to clear up the matter and have some useful advice. They moved quickly when people reported that the YN568 had a problem at x-sync with some bodies.

HoppyUK - I did not mean FEL. I rarely use it.

I was trying to suggest that the tests Lodious was doing needed a high level of consistency to rule out some of the variables, and a definable subject. (Refer to his posted images.) Like a slight change in framing can cause points of light to impinge on a different sector of the exposure meter, producing noticable shifts in lighting (as you say). To get a good comparison, you need matching framing. It is not an issue for real photography, just for tests.

Evaluative = E-TTL II largely suspended.

Distance info most certainly does not over-ride other considerations. It informs the exposure evaluation of the location of the subject within the metering zones.

One thing I appreciated while testing the YN-568. The zoom is still menu-settable even when the head is rotated or tilted. In OCF use, the zoom makes a useful light modifier, which Canon blocks with its forced irrelevant settings. Good one, YN.
 
SeedyBee said:
Lodious - It seems that you have identified a programming bug. Have you contacted YongNuo? They may be able to clear up the matter and have some useful advice. They moved quickly when people reported that the YN568 had a problem at x-sync with some bodies.

HoppyUK - I did not mean FEL. I rarely use it.

I was trying to suggest that the tests Lodious was doing needed a high level of consistency to rule out some of the variables, and a definable subject. (Refer to his posted images.) Like a slight change in framing can cause points of light to impinge on a different sector of the exposure meter, producing noticable shifts in lighting (as you say). To get a good comparison, you need matching framing. It is not an issue for real photography, just for tests.

Evaluative = E-TTL II largely suspended.

Distance info most certainly does not over-ride other considerations. It informs the exposure evaluation of the location of the subject within the metering zones.

One thing I appreciated while testing the YN-568. The zoom is still menu-settable even when the head is rotated or tilted. In OCF use, the zoom makes a useful light modifier, which Canon blocks with its forced irrelevant settings. Good one, YN.
Click to expand...

I agree that very careful tests are needed to check these things. And a good deal of understanding (and time!). Good luck to anyone willing to take it on, but my experience is that there are so many variables with cameras and guns and triggers, that even when you think you've found something, another person doesn't get the same result!

Your point about distance info is a good example. Maybe you're not getting this effect, but I first noticed it when using a Rayflash ring flash, where I was getting persistant but variable under-exposure when I could see no good reason for it (Canon 5D2, EX580 on camera). Then I noticed that just pressing the tilt-release button disabled the range/distance readout on the back of the gun, and good exposure was restored. Same when using a Stofen pointing straight forward (though you shouldn't use one like that) but it always under-exposed when I did.

This strongly suggests to me that overriding distance info was being used, and that makes perfect sense - it should be very reliable with the gun used in that way. Presumably when using a remote trigger, the distance info is disabled anyway - as it should be. I don't have any E-TTL triggers here ATM to check that.
 
HoppyUK said:
This strongly suggests to me that overriding distance info was being used, and that makes perfect sense - it should be very reliable with the gun used in that way. Presumably when using a remote trigger, the distance info is disabled anyway - as it should be. I don't have any E-TTL triggers here ATM to check that.
Click to expand...
I think that it was probably NEVEC coming into play, dropping exposure on Canon's assumptions.
 
SeedyBee said:
I think that it was probably NEVEC coming into play, dropping exposure on Canon's assumptions.
Click to expand...

Why do you think that? Why would distance data not be used when it's such a reliable exposure guide in that situation? Why with distance data disabled was exposure not the same, ie under-exposed? Why did it immediately restore accuracy?
 
HoppyUK said:
Why do you think that? Why would distance data not be used when it's such a reliable exposure guide in that situation? Why with distance data disabled was exposure not the same, ie under-exposed? Why did it immediately restore accuracy?
Click to expand...

Sorry, not interested in arguing or providing training, just trying to help Lodious. Your assumption that only thre distance data is affected by a tilt or rotate is wrong. It moves the assessment from "main light" to "fill light". Did you look up NEVEC?

Best wishes with your problem, Lodious.
 
SeedyBee said:
Sorry, not interested in arguing or providing training, just trying to help Lodious. Your assumption that only thre distance data is affected by a tilt or rotate is wrong. It moves the assessment from "main light" to "fill light". Did you look up NEVEC?

Best wishes with your problem, Lodious.
Click to expand...

Not interested in arguing either, just asking why you think distance data is not a factor. NEVEC and distance data enabled/disabled appear to be unrelated on my camera/flash.
 
Sorry Guys, I left my camera at work this weekend (agghhh), so I can't do the testing I'd like to have done. I plan to do some more considered testing, upload the results to flickr, then mail Youngnuo to see if they have any ideas.

As I said earlier, I'm still happy with the flash, but it would be nice to understand what's going on!
 
I bought two of these YongNuo YN568EX from DZONE HONG KONG LTD

When ordering, I gave the following note to them:
I am going to use this flash together with a Canon EOS 5D Mark III and a YN-622C Flash Trigger. If there is any compatibility problem please let me know before shipping.
Click to expand...

However the delivered equipment suffered from the symptoms described in this thread.

I asked DZONE2 a few days ago If they were aware of the incompatibility problem and what we can do in order to solve the problem.

DZONE2's short reply that I received today was:
Yes the problem was with the Oct made flashes only
Click to expand...

That's great news because then I'm sure it's a problem related to the flash and it's probably been solved after October 2012.

I had to write them back again though because they did not answer if they were willing to help me (warranty, goodwill or whatever...). Let's see what they say about it.

photo-george
 
Last edited:
You must log in or register to reply here.
Back
Top