Could Someone Else do this Test with 70D Please?

Started 4 months ago | Discussions
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Could Someone Else do this Test with 70D Please?
4 months ago

I’m finding big exposure differences when using View Finder compared with Live View when using a non-EF lens on my 70D.

For example, I’m using an FD 1:3.5 50 mm Macro lens connected via an FD to EF adaptor - one with no optics, so no degradation, but no infinity focussing. Set on a tripod looking at a close subject and with a shield over the eyepiece. I set the camera on Aperture Priority.

Here are the results:

Lens wide open at f3.5:

Using VF the speed is 400, using LV the speed is 250 – ½ stop difference

Lens set at f8:

Using VF the speed is 60, using LV the speed is 30 – 1 stop difference

Lens set at f11:

Using VF the speed is 30, using LV the speed is 8 – 2 stops difference

In all cases, the Live View exposures are about right and the VF ones are under-exposed, progressively more so as the lens is stopped down. Why should this be?

Could someone try this test with any non-EF lens on their 70D, ie where the electric contacts are not used to stop down the lens. If your results produce good exposure in both VF and LV, there’s a problem with my camera. If you replicate my results, Canon seems to have a design problem as this just shouldn’t be.

Many thanks

Peter

PS - I wrote to Canon about this.  They phoned back and said they "don't support use with non-EF lenses", but suggested I get the camera checked and recalibrated at a Canon centre.  Their Maunal says nothing about problems using non-EF lenses, or about different results with VF vx LV.

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Canon EOS 70D
If you believe there are incorrect tags, please send us this post using our feedback form.
TTMartin
Senior MemberPosts: 4,728Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

I suspect this is a similar problem to the one that is occurring using bounce flash.

It would seem the Dual Pixel sensor may have a different sensitivities for different apertures than anticipated, and Canon is going to have to tweak the firmware.

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon EOS 400D Canon EOS 7D Canon EOS M +23 more
Reply   Reply with quote   Complain
Aaron Corey
New MemberPosts: 6Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

I'm not exactly and expert on stop-down metering with manual lenses, but here's my theory on what you're seeing based on some research I've done into the topic:

When you use an EF lens, the camera does it's metering with the lens wide open and calculates what the proper exposure would be when the lens is stopped down.  An EF lens is not actually stopped down to the selected aperture until the instant the photo is taken.

With stop down metering, you manually stop the lens down to the desired aperture with the aperture ring and allow the camera to take a reading.  In this case the camera will just base the exposure on what it's measuring without doing any calculation (the camera knows that it can't detect what aperture is being used).  Because the lens is stopped down, less light is getting into the camera.

When you use the optical viewfinder, the camera is using its dedicated exposure sensor in the viewfinder, but this sensor needs lots of light to function at it's best.  In live view, the camera's imaging sensor is being used to determine the exposure.  From personal experience, I've seen that live view can produce more accurate exposures in low light.  For example, when I use a 6-stop ND filter (to shoot waterfalls), I have to use live view to get good, consistent exposures. I think it's because the "exposure simulation" in live view is amplifying the brightness of the image.

So I suspect that what's happening is that you have your manual lens stopped down far enough that the dedicated exposure sensor isn't getting enough light to work at it's best.  Live view seems to handle low light conditions better, which might explain why it's giving you better results.

 Aaron Corey's gear list:Aaron Corey's gear list
Canon EOS 60D Canon EOS 1100D Canon EF-S 15-85mm f/3.5-5.6 IS USM Sigma 70-300mm F4-5.6 DG Macro Canon EF 28mm f/2.8 +2 more
Reply   Reply with quote   Complain
borno
Contributing MemberPosts: 857Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

My guess is the VF metering is taken through the focus screen and calibrated to take the amount of darkening due to the lens max aperture into account using an ef lens. When using a manual lens the camera meter can't be trusted as the focus screen darkening may not be linear and it doesn't know the aperture. Just a guess as I had similar problems using a pentax with old lenses. A workaround was to meter wide open and recalculate. With the pentax a fix was to use an older model focus screen, it was a little darker though.

Reply   Reply with quote   Complain
tr573
Regular MemberPosts: 166
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to borno, 4 months ago

Adding an AF confirm chip will help in this case, because you can program in the proper max aperture of the lens, and meter wide open. Then stop down for the actual exposure (to the meters suggested settings)

Reply   Reply with quote   Complain
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Aaron Corey, 4 months ago

Aaron Corey wrote:

I'm not exactly and expert on stop-down metering with manual lenses, but here's my theory on what you're seeing based on some research I've done into the topic:

Many thanks Aaron for your input.  I think much of what you say explains the problem, BUT:

When you use an EF lens, the camera does it's metering with the lens wide open and calculates what the proper exposure would be when the lens is stopped down. An EF lens is not actually stopped down to the selected aperture until the instant the photo is taken.

With stop down metering, you manually stop the lens down to the desired aperture with the aperture ring and allow the camera to take a reading. In this case the camera will just base the exposure on what it's measuring without doing any calculation (the camera knows that it can't detect what aperture is being used). Because the lens is stopped down, less light is getting into the camera.

Yes, that's precisely as I understand how an SLR works.

When you use the optical viewfinder, the camera is using its dedicated exposure sensor in the viewfinder, but this sensor needs lots of light to function at it's best.

Not sure I follow this.  My 70D is my first camera with Live View.  I've never experienced any problems with Canon cameras having metering difficulties in low light.  Apart from an EOS 300D, my cameras have all been film cameras.  Their metering is always done in the pentaprism via the mirror.  Metering is always pretty accurate, whether the lens is manually stopped down (as all FD ones have to be in Apperture Priority) or in Auto stop down.

In live view, the camera's imaging sensor is being used to determine the exposure. From personal experience, I've seen that live view can produce more accurate exposures in low light. For example, when I use a 6-stop ND filter (to shoot waterfalls), I have to use live view to get good, consistent exposures. I think it's because the "exposure simulation" in live view is amplifying the brightness of the image.

My experience with LV is non-existant, but I would have thought that "normal" metering (ie using mirror and VF) SHOULD BE more accurate than Live View.  Clearly with digital cameras that offer LV, the accuracy of "normal" metering has been sacrificed.

So I suspect that what's happening is that you have your manual lens stopped down far enough that the dedicated exposure sensor isn't getting enough light to work at it's best. Live view seems to handle low light conditions better, which might explain why it's giving you better results.

You are probably right, although I still get this 2-stop inaccuracy when there's plenty of light eg when using f8 and the correct shutter speed is 1/250 or faster - surely plenty of light for the VF meter.  No previous (VF only) camera would have offered a shutter speed 2 stops from accurate.

Is there a way of tricking the camera into ignoring its misguided expectation that the lens will be stopped down when the shutter is released?  In other words, to calculate the shutter speed solely on the light that's actually entering the camera.

Peter

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Reply   Reply with quote   Complain
asad137
Contributing MemberPosts: 669
Like?
Metering mode?
In reply to Peter Marchant, 4 months ago

What metering mode were you set in?

In Live View, there's only one -- a form of Evaluative Metering that uses the image sensor as the metering sensor. In normal operation, of course, you can pick from multi (evaluative), center-weighted, partial, or spot. And even then the evaluative metering for through-the-viewfinder shooting is different than Live View.

The problem is that the presence of a subject in the frame can affect the metering in the different modes (as well it should).

My suggestion: Try taking a shot of a featureless, evenly-illuminated frame. If you have an 18% grey card that you can use to fill the frame, that would be ideal. Make sure it is illuminated uniformly so that variations in the brightness of the target don't confuse the results.

In all metering modes (including Live View), this should give the same exposure. If it doesn't, there's a problem.

Reply   Reply with quote   Complain
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Re: Metering mode?
In reply to asad137, 4 months ago

asad137 wrote:

What metering mode were you set in?

In Live View, there's only one --

Are you sure about that?  There's nothing I can find in the Manual that says Live View limits your metering options to a single, take-it-or-leave-it choice of one!

a form of Evaluative Metering that uses the image sensor as the metering sensor. In normal operation, of course, you can pick from multi (evaluative), center-weighted, partial, or spot. And even then the evaluative metering for through-the-viewfinder shooting is different than Live View.

My tests were with Centre Weighted as the chosen mode. but similar results were found with Evaluative.

The problem is that the presence of a subject in the frame can affect the metering in the different modes (as well it should).

Of course, but why should the same subject produce such different results just by selecting different lens apartures?  With View Finder metering, the image at f3.5 is only about 1/2 stop out, whereas it's 2 stops out when the lens is stopped down to f11 - same subject, same meter method, only difference is the manually set aperture.

My suggestion: Try taking a shot of a featureless, evenly-illuminated frame. If you have an 18% grey card that you can use to fill the frame, that would be ideal. Make sure it is illuminated uniformly so that variations in the brightness of the target don't confuse the results.

In all metering modes (including Live View), this should give the same exposure. If it doesn't, there's a problem.

That's why I'm asking someone to do the same test - to verify if my camera is faulty or whether the 70D in unable to accurately set the correct speed on VF when fitted with a manually-set non-EF lens.

Thanks

Peter

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Reply   Reply with quote   Complain
TTMartin
Senior MemberPosts: 4,728Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

To clarify, when shooting LiveView metering is done by the imaging sensor. When using the optical viewfinder metering is done by a separate metering sensor.

There seems to be some differences in the light sensitivity of the dual pixel sensor that can hasn't fully accounted for. This is resulting in some exposure differences when using bounce flash and LiveView with manual lenses.

Whether Canon is fully aware of it yet, and if they are going to come out with a firmware update to correct it, I don't know.

I doubt they would do a firmware update for the manual lenses in LiveView alone, as they told you, they don't support that. But, you might get lucky and if they fix the metering for bounced flash it might fix the LiveView with manual lenses too.

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon EOS 400D Canon EOS 7D Canon EOS M +23 more
Reply   Reply with quote   Complain
asad137
Contributing MemberPosts: 669
Like?
Re: Metering mode?
In reply to Peter Marchant, 4 months ago

Peter Marchant wrote:

asad137 wrote:

What metering mode were you set in?

In Live View, there's only one --

Are you sure about that? There's nothing I can find in the Manual that says Live View limits your metering options to a single, take-it-or-leave-it choice of one!

Apologies -- on the 60D there's only one live-view metering mode so I assumed it was the same on the 70D. Bad assumption -- looking at the 70D manual you're right, it looks like you can change the metering mode.

Carry on!

Reply   Reply with quote   Complain
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to TTMartin, 4 months ago

TTMartin wrote:

To clarify, when shooting LiveView metering is done by the imaging sensor. When using the optical viewfinder metering is done by a separate metering sensor.

There seems to be some differences in the light sensitivity of the dual pixel sensor that can hasn't fully accounted for. This is resulting in some exposure differences when using bounce flash and LiveView with manual lenses.

Whether Canon is fully aware of it yet, and if they are going to come out with a firmware update to correct it, I don't know.

I doubt they would do a firmware update for the manual lenses in LiveView alone, as they told you, they don't support that. But, you might get lucky and if they fix the metering for bounced flash it might fix the LiveView with manual lenses too.

Thanks, but use of flash is a different matter.  The method of exposure calculation VF vs LV may be differnent, but it shouldn't make an appreciable difference with the end result - both should be properly exposed.

I'm using an FD macro lens, so the lens has to be manually stopped down - a practice that surely all cameras should be able to cope with in Aperture Priority, whether using VF or LV.  The camera knows it has no electric contact with the iris (the Properties show f0), so it must surely measure the light it actually sees through the lens, as it knows that pressing the release won't change that.

Can someone else do this test on their 70D please?  All you need is an FD or any other manual lens and do a few exposures with VF and LV.  I need to know whether my camera is faulty, rather than theories about why it might be happening.

Thanks again

Peter

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Reply   Reply with quote   Complain
TTMartin
Senior MemberPosts: 4,728Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

Peter Marchant wrote:

TTMartin wrote:

To clarify, when shooting LiveView metering is done by the imaging sensor. When using the optical viewfinder metering is done by a separate metering sensor.

There seems to be some differences in the light sensitivity of the dual pixel sensor that can hasn't fully accounted for. This is resulting in some exposure differences when using bounce flash and LiveView with manual lenses.

Whether Canon is fully aware of it yet, and if they are going to come out with a firmware update to correct it, I don't know.

I doubt they would do a firmware update for the manual lenses in LiveView alone, as they told you, they don't support that. But, you might get lucky and if they fix the metering for bounced flash it might fix the LiveView with manual lenses too.

Thanks, but use of flash is a different matter. The method of exposure calculation VF vs LV may be differnent, but it shouldn't make an appreciable difference with the end result - both should be properly exposed.

I'm using an FD macro lens, so the lens has to be manually stopped down - a practice that surely all cameras should be able to cope with in Aperture Priority, whether using VF or LV. The camera knows it has no electric contact with the iris (the Properties show f0), so it must surely measure the light it actually sees through the lens, as it knows that pressing the release won't change that.

Can someone else do this test on their 70D please? All you need is an FD or any other manual lens and do a few exposures with VF and LV. I need to know whether my camera is faulty, rather than theories about why it might be happening.

No you camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate.

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon EOS 400D Canon EOS 7D Canon EOS M +23 more
Reply   Reply with quote   Complain
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to TTMartin, 4 months ago

TTMartin wrote:

No you camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate.

Why should a Canon camera not be designed to work with manually operated aperture?  I've used Canon SLRs since the days of the AE-1 and the original F-1 and never had any metering problems whatever the lens is attached, or whether the lens is automatically stopped down or is set manually.

Explain why a 70D should not be able to accurately measure the light that falls on its sensor and why it measures that light so differently when the VF is used to compose compared with the LCD screen?  Or for that matter, why the difference between VF and LV varies (from accurate) so much by changing the lens aperture?

I don't want to "learn how to compensate" as this should not be necessary and I particularly don't want to learn multiple compensations depending on whether I choose to use LV or VF, or by how much the lens is manually stopped down.

But thanks anyway.

Peter

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Reply   Reply with quote   Complain
TTMartin
Senior MemberPosts: 4,728Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

Peter Marchant wrote:

TTMartin wrote:

No you camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate.

Why should a Canon camera not be designed to work with manually operated aperture? I've used Canon SLRs since the days of the AE-1 and the original F-1 and never had any metering problems whatever the lens is attached, or whether the lens is automatically stopped down or is set manually.

Explain why a 70D should not be able to accurately measure the light that falls on its sensor and why it measures that light so differently when the VF is used to compose compared with the LCD screen? Or for that matter, why the difference between VF and LV varies (from accurate) so much by changing the lens aperture?

I don't want to "learn how to compensate" as this should not be necessary and I particularly don't want to learn multiple compensations depending on whether I choose to use LV or VF, or by how much the lens is manually stopped down.

But thanks anyway.

The Dual Pixels of the 70D's sensor are directional.

As you decrease the aperture you change the angle of the light entering the pixel.

Image from http://electronics.howstuffworks.com used for educational purposes.

This would change the exposure, compared to a traditional non-directional pixel.

Evidently Canon worked this out for when there is a known lens attached, and normal lighting situations.

It seems for your manual lenses, and for calculating ETTL using a bounce flash that Canon didn't program/take into account/etc. this factor in the exposure calculations.

Or in other words, if Canon used normal ETTL calculations from a non-Dual Pixel calculations on a 70D they would result in an under exposure when using a bounce flash.

This would explain why you have issues with your older manual lenses, and other individuals are having an issue when using a bounced flash.

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon EOS 400D Canon EOS 7D Canon EOS M +23 more
Reply   Reply with quote   Complain
TTMartin
Senior MemberPosts: 4,728Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to TTMartin, 4 months ago

TTMartin wrote:

Peter Marchant wrote:

TTMartin wrote:

No you camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate.

Why should a Canon camera not be designed to work with manually operated aperture? I've used Canon SLRs since the days of the AE-1 and the original F-1 and never had any metering problems whatever the lens is attached, or whether the lens is automatically stopped down or is set manually.

Explain why a 70D should not be able to accurately measure the light that falls on its sensor and why it measures that light so differently when the VF is used to compose compared with the LCD screen? Or for that matter, why the difference between VF and LV varies (from accurate) so much by changing the lens aperture?

I don't want to "learn how to compensate" as this should not be necessary and I particularly don't want to learn multiple compensations depending on whether I choose to use LV or VF, or by how much the lens is manually stopped down.

But thanks anyway.

The Dual Pixels of the 70D's sensor are directional.

As you decrease the aperture you change the angle of the light entering the pixel.

Image from http://electronics.howstuffworks.com used for educational purposes.

This would change the exposure, compared to a traditional non-directional pixel.

Evidently Canon worked this out for when there is a known lens attached, and normal lighting situations.

It seems for your manual lenses, and for calculating ETTL using a bounce flash that Canon didn't program/take into account/etc. this factor in the exposure calculations.

Or in other words, if Canon used normal ETTL calculations from a non-Dual Pixel camera on a 70D they would result in an underexposure when using a bounce flash.

This would explain why you have issues with your older manual lenses, and other individuals are having an issue when using a bounced flash.

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon EOS 400D Canon EOS 7D Canon EOS M +23 more
Reply   Reply with quote   Complain
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to TTMartin, 4 months ago

TTMartin wrote:

Peter Marchant wrote:

TTMartin wrote:

No you camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate.

Why should a Canon camera not be designed to work with manually operated aperture? I've used Canon SLRs since the days of the AE-1 and the original F-1 and never had any metering problems whatever the lens is attached, or whether the lens is automatically stopped down or is set manually.

Explain why a 70D should not be able to accurately measure the light that falls on its sensor and why it measures that light so differently when the VF is used to compose compared with the LCD screen? Or for that matter, why the difference between VF and LV varies (from accurate) so much by changing the lens aperture?

I don't want to "learn how to compensate" as this should not be necessary and I particularly don't want to learn multiple compensations depending on whether I choose to use LV or VF, or by how much the lens is manually stopped down.

But thanks anyway.

The Dual Pixels of the 70D's sensor are directional.

As you decrease the aperture you change the angle of the light entering the pixel.

Image from http://electronics.howstuffworks.com used for educational purposes.

This would change the exposure, compared to a traditional non-directional pixel.

Evidently Canon worked this out for when there is a known lens attached, and normal lighting situations.

It seems for your manual lenses, and for calculating ETTL using a bounce flash that Canon didn't program/take into account/etc. this factor in the exposure calculations.

Or in other words, if Canon used normal ETTL calculations from a non-Dual Pixel calculations on a 70D they would result in an under exposure when using a bounce flash.

This would explain why you have issues with your older manual lenses, and other individuals are having an issue when using a bounced flash.

All very interesting, but with all due respect, irrelevant. I am not using flash, so let's not complicate matters with that. The pixels on the sensor don’t measure the light and set the speed, as far as I’m aware.

What you haven't explained is why the meter is reasonably accurate at all manually-set apertures when the LCD is on and Live View is being used, whereas with the conventional (I call it this because "Live View" is a new feature of SLRs) View Finder is used to compose the picture; the exposure is reasonably accurate when the iris is wide open, but it gets progressively more inaccurate as the iris is closed down.

Whatever type of sensor pixels, the METER should be able to reasonably accurately measure the light and choose an appropriate shutter speed. For some reason (not explained in your thesis) my camera is failing when in Live View. I'm asking someone who has a 70D to do similar tests with a non-EF lens to see if my findings are peculiar to my particular camera.

Thanks

Peter

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Reply   Reply with quote   Complain
TTMartin
Senior MemberPosts: 4,728Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

Peter Marchant wrote:

TTMartin wrote:

Peter Marchant wrote:

TTMartin wrote:

No you camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate.

Why should a Canon camera not be designed to work with manually operated aperture? I've used Canon SLRs since the days of the AE-1 and the original F-1 and never had any metering problems whatever the lens is attached, or whether the lens is automatically stopped down or is set manually.

Explain why a 70D should not be able to accurately measure the light that falls on its sensor and why it measures that light so differently when the VF is used to compose compared with the LCD screen? Or for that matter, why the difference between VF and LV varies (from accurate) so much by changing the lens aperture?

I don't want to "learn how to compensate" as this should not be necessary and I particularly don't want to learn multiple compensations depending on whether I choose to use LV or VF, or by how much the lens is manually stopped down.

But thanks anyway.

The Dual Pixels of the 70D's sensor are directional.

As you decrease the aperture you change the angle of the light entering the pixel.

Image from http://electronics.howstuffworks.com used for educational purposes.

This would change the exposure, compared to a traditional non-directional pixel.

Evidently Canon worked this out for when there is a known lens attached, and normal lighting situations.

It seems for your manual lenses, and for calculating ETTL using a bounce flash that Canon didn't program/take into account/etc. this factor in the exposure calculations.

Or in other words, if Canon used normal ETTL calculations from a non-Dual Pixel camera on a 70D they would result in an under exposure when using a bounce flash.

This would explain why you have issues with your older manual lenses, and other individuals are having an issue when using a bounced flash.

All very interesting, but with all due respect, irrelevant. I am not using flash, so let's not complicate matters with that. The pixels on the sensor don’t measure the light and set the speed, as far as I’m aware.

What you haven't explained is why the meter is reasonably accurate at all manually-set apertures when the LCD is on and Live View is being used, whereas with the conventional (I call it this because "Live View" is a new feature of SLRs) View Finder is used to compose the picture; the exposure is reasonably accurate when the iris is wide open, but it gets progressively more inaccurate as the iris is closed down.

Yes, I have told you why you just don't want to listen.

When using LiveView the camera is metering off of the sensor itself.

When using the optical view finder the camera uses a separate metering sensor. With the lenses specified to be used with the camera, the camera then has to calculate what the exposure will be when the photo is taken, using not only the amount of light, but, the focal length of the lens, and the aperture to determine the angle the light will be hitting the sensor, and calculate the correct exposure.

The focal length and aperture of the lens both impact the angle the light comes from when it hits the sensor.

With a manual lens camera has no way of knowing what the focal length and aperture that the light the separate sensor is see is coming from. So the camera doesn't know what angle the light rays hitting the image sensor will be coming from, so it has no way of calculating the correct exposure.

When using LiveView the camera is metering right off the image sensor, so it is seeing the light same light, at the same angles that will being used when the picture is taken.

Your camera is not faulty. It is not designed to work with those lenses, and if you want to use those lenses you need to learn how to compensate, for both the varying focal length and aperture, as those will both impact the proper exposure from the Dual Pixel sensor.

Whatever type of sensor pixels, the METER should be able to reasonably accurately measure the light and choose an appropriate shutter speed. For some reason (not explained in your thesis) my camera is failing when in Live View. I'm asking someone who has a 70D to do similar tests with a non-EF lens to see if my findings are peculiar to my particular camera.

Thanks

Peter

 TTMartin's gear list:TTMartin's gear list
Canon PowerShot S100 (2000) Canon PowerShot SD1000 Canon EOS 400D Canon EOS 7D Canon EOS M +23 more
Reply   Reply with quote   Complain
WilbaW
Veteran MemberPosts: 8,213Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

Peter Marchant wrote:

Can someone else do this test on their 70D please? All you need is an FD or any other manual lens and do a few exposures with VF and LV.

You can do the testing with any EOS-compatible lens, like this - set the exposure aperture, hold down the DOF preview button, hold down the lens release button, and rotate the lens the few degrees required to get the aperture display to show "F00". You now have a manual stopped-down lens. Don't forget to turn the lens back to the locked position when you're done!

I need to know whether my camera is faulty, rather than theories about why it might be happening.

I agree, don't pay attention to fanciful ideas of why it might be happening, just work with evidence.

Reply   Reply with quote   Complain
yellodog
Senior MemberPosts: 1,042Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to Peter Marchant, 4 months ago

For what it's worth I repeated the test with an old Olympus 35 2.0

At F2.0  VF => 1/50 s

LV =>  1/40 s

At F8.0  VF => 0,6 s

LV => 1/4 s (0.25 s)

Reply   Reply with quote   Complain
Peter Marchant
Regular MemberPosts: 272Gear list
Like?
Re: Could Someone Else do this Test with 70D Please?
In reply to yellodog, 4 months ago

yellodog wrote:

For what it's worth I repeated the test with an old Olympus 35 2.0

At F2.0 VF => 1/50 s

LV => 1/40 s

At F8.0 VF => 0,6 s

LV => 1/4 s (0.25 s)

Thanks for that.  It seems to confirm my findings that LV will reasonably accurately measure a manually stopped-down lens and set the correct speed.  In VF mode the accuracy when the lens is fully open is OK (your result is under 1/4 stop out compared with LV) but is becomes less accurate as you stop the lens down.  Your exposure difference is far less than mine (just over 1/2 stop compared with my 2 stop difference) but it does show that VF mode is not the best choice with manual lenses.

What do the results look like?  My guess is that 3 look well exposed, but the f8 VF one will be noticably out.

Peter

 Peter Marchant's gear list:Peter Marchant's gear list
Canon EOS 70D Canon EF-S 18-135mm f/3.5-5.6 IS Canon EF 75-300mm f/4.0-5.6 III Canon EF 100mm f/2.8 Macro USM Canon EF 28mm f/2.8 +6 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads