Updated: Canon Speedlite 470EX-AI Additional Images & Information

[Canon Rumors Guy]

The Canon Speedlite 470EX-AI will be announced alongside the Canon EOS M50 in a few days. This was a flash that was supposed to be announced last summer but was delayed for unknown reasons.</p>
<p><strong>Canon Speedlite 470EX-AI Specifications:</strong> (via <a href="http://digicame-info.com/2018/02/470ex-ai.html">Digicame-Info</a>)</p>
<ul>
<li>Auto intelligent (AI) bounce mode of “Full Auto” and “Semi Auto”. With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows.</li>
<li>The guide number is 47 (105 mm). It is driven by 4 AA batteries.</li>
<li>It covers the field angle of 24-105mm. Wide panel (14 mm adapter) allows light to spread wide. Clip-on bounce adapter that makes light more soft.</li>
<li>Equipped with a dot matrix display that can display more detailed information.</li>
<li>The setting can be quickly changed with the shortcut button and the controller wheel.</li>
<li>Size 74.6 x 130.4 x 105.1 mm. Weight 385 g (excluding batteries).</li>
<li><strong>Price: $399 USD</strong></li>
</ul>
<p>This new flash is scheduled to begin shipping in April of 2018.</p>

		<style type='text/css'>
			#gallery-1 {
				margin: auto;
			}
			#gallery-1 .gallery-item {
				float: left;
				margin-top: 10px;
				text-align: center;
				width: 33%;
			}
			#gallery-1 img {
				border: 2px solid #cfcfcf;
			}
			#gallery-1 .gallery-caption {
				margin-left: 0;
			}
			/* see gallery_shortcode() in wp-includes/media.php */
		</style>
		<div id='gallery-1' class='gallery galleryid-33909 gallery-columns-3 gallery-size-thumbnail'><dl class='gallery-item'>
			<dt class='gallery-icon landscape'>
				<a href='http://www.canonrumors.com/wp-content/uploads/2018/02/ex470nok.jpg'><img width="168" height="168" src="http://www.canonrumors.com/wp-content/uploads/2018/02/ex470nok-168x168.jpg" class="attachment-thumbnail size-thumbnail" alt="" srcset="http://www.canonrumors.com/wp-content/uploads/2018/02/ex470nok-168x168.jpg 168w, http://www.canonrumors.com/wp-content/uploads/2018/02/ex470nok-144x144.jpg 144w" sizes="(max-width: 168px) 100vw, 168px" /></a>
			</dt></dl><dl class='gallery-item'>
			<dt class='gallery-icon portrait'>
				<a href='http://www.canonrumors.com/wp-content/uploads/2018/02/canon-4.jpg'><img width="168" height="168" src="http://www.canonrumors.com/wp-content/uploads/2018/02/canon-4-168x168.jpg" class="attachment-thumbnail size-thumbnail" alt="" srcset="http://www.canonrumors.com/wp-content/uploads/2018/02/canon-4-168x168.jpg 168w, http://www.canonrumors.com/wp-content/uploads/2018/02/canon-4-144x144.jpg 144w" sizes="(max-width: 168px) 100vw, 168px" /></a>
			</dt></dl><dl class='gallery-item'>
			<dt class='gallery-icon portrait'>
				<a href='http://www.canonrumors.com/wp-content/uploads/2018/02/canon_1-3.jpg'><img width="168" height="168" src="http://www.canonrumors.com/wp-content/uploads/2018/02/canon_1-3-168x168.jpg" class="attachment-thumbnail size-thumbnail" alt="" srcset="http://www.canonrumors.com/wp-content/uploads/2018/02/canon_1-3-168x168.jpg 168w, http://www.canonrumors.com/wp-content/uploads/2018/02/canon_1-3-144x144.jpg 144w" sizes="(max-width: 168px) 100vw, 168px" /></a>
			</dt></dl><br style="clear: both" />
		</div>

<span id="pty_trigger"></span>

 

[Sharlin]

Re: Canon Speedlite 470EX-AI Additional Images & Information

Damn. Looks like there’s actually automatic adjustment of the bounce angle.

 

[tron]

Re: Canon Speedlite 470EX-AI Additional Images & Information

This was a flash that was supposed to be announced last summer but was delayed for unknown reasons

Maybe it wasn't ... Intelligent enough ;D

 

[magarity]

Re: Canon Speedlite 470EX-AI Additional Images & Information

It looks like the guesses about somehow sensing the distance and color of the bounce surface win. There's a little round sensor-looking thingy next to where the diffuser slides into but nowhere looks like an opening for a second path for light to come out.

 

[daleg]

maybe Canon held it back for the release of the "EOS 8D Mark II"

btw, what's is the "EOS 8D Mark II"?

see description in announcement !?!?

 

[wickedac]

Who is this even for? I don't get it. Does it spin around measuring distances, etc first? That would add a lot of time to the shot.I feel like if you're savvy enough to understand the usefulness of bouncing a speedlight off a wall or ceiling you'd rather choose for yourself how the light is shaped and just do it quickly.

 

[CanonGrunt]

maybe Canon held it back for the release of the "EOS 8D Mark II"

btw, what's is the "EOS 8D Mark II"?

see description in announcement !?!?

Oh boy!! I’ve been waiting forever for a replacement for my 8D!!!

 

[IglooEater]

What is “auto bounce”?

360 degree sensors evaluate the room and the subject, then a motorized head automatically rotates the flash to likely best exposure, based on your selection of 5 steps between "dramatic" and "symmetrical" 8)

Ok, well, here's wishin'

Okay Talys. Your idea was much closer than I expected - I have to give you that. Hat’s off.

 

[Ozarker]

What is “auto bounce”?

360 degree sensors evaluate the room and the subject, then a motorized head automatically rotates the flash to likely best exposure, based on your selection of 5 steps between "dramatic" and "symmetrical" 8)

Ok, well, here's wishin'

Okay Talys. Your idea was much closer than I expected - I have to give you that. Hat’s off.

I guess what I thought couldn't be done is now done. Wait for the lawsuit. Canon could not have invented this. Everyone knows Canon doesn't innovate. :

 

[neuroanatomist]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

 

[ajfotofilmagem]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

More difficult than geometry, is knowing the distance, the reflectivity and the color of the ceiling in the real world. If the flash is able to calculate all that and adjust properly, even with the camera in portrait orientation, it would be a great technological leap.

 

[neuroanatomist]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

More difficult than geometry, is knowing the distance, the reflectivity and the color of the ceiling in the real world. If the flash is able to calculate all that and adjust properly, even with the camera in portrait orientation, it would be a great technological leap.

If...

Where in, "With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows," do you interpret an ability to determine reflectivity and color? Incidentally, E-TTL will handle the reflectivity, there's no need for the flash to determine that a priori.

 

[rrcphoto]

maybe Canon held it back for the release of the "EOS 8D Mark II"

btw, what's is the "EOS 8D Mark II"?

see description in announcement !?!?

that's a machine translation error.

 

[ajfotofilmagem]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

More difficult than geometry, is knowing the distance, the reflectivity and the color of the ceiling in the real world. If the flash is able to calculate all that and adjust properly, even with the camera in portrait orientation, it would be a great technological leap.

If...

Where in, "With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows," do you interpret an ability to determine reflectivity and color? Incidentally, E-TTL will handle the reflectivity, there's no need for the flash to determine that a priori.

I really do not like the result of bouncing the flash on ceilings of any color, other than white. If this 470EX can measure and tell the camera the correct color of the reflected light, then it would be something beyond the beautiful words of the marketing staff.

 

[mb66energy]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

More difficult than geometry, is knowing the distance, the reflectivity and the color of the ceiling in the real world. If the flash is able to calculate all that and adjust properly, even with the camera in portrait orientation, it would be a great technological leap.

If...

Where in, "With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows," do you interpret an ability to determine reflectivity and color? Incidentally, E-TTL will handle the reflectivity, there's no need for the flash to determine that a priori.

But E-TTL cannot measure the reflectivity of the bounce surface so this has to be handled by the flash to mix the direct and bounced fractions of light correctly.

 

[neuroanatomist]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

More difficult than geometry, is knowing the distance, the reflectivity and the color of the ceiling in the real world. If the flash is able to calculate all that and adjust properly, even with the camera in portrait orientation, it would be a great technological leap.

If...

Where in, "With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows," do you interpret an ability to determine reflectivity and color? Incidentally, E-TTL will handle the reflectivity, there's no need for the flash to determine that a priori.

But E-TTL cannot measure the reflectivity of the bounce surface so this has to be handled by the flash to mix the direct and bounced fractions of light correctly.

I didn't mean to imply that E-TTL measures the reflectivity. But the reflectivity of the bounce surface determines how much bounced light reaches the subject, and E-TTL determines exposure based on light reflected from the subject. In that way, E-TTL 'handles' the reflectivity of the bounce surface.

What 'direct fraction' are you talking about mixing? If the head is elevated, it's all bounced (well, except for spill, but again, E-TTL handles that).

I'm not sure why people think that this 'AI auto bounce' feature is addressing flash exposure...that's what E-TTL does. The point is, E-TTL cares about how much light reflects back from the subject, not how that light gets to the subject. Whether direct or bounced, coming from one flash or 12 flashes, it doesn't matter. The point of this feature seems exactly as described: determining the optimal bounce position to eliminate unnecessary shadows.

Also, consider this: the exposure calculations are done based on a pre-flash, which occurs with almost no temporal separation from the actual exposure. The electric head rotation motor isn't going to move the flash head instantaneously. In other words, the AI auto bounce head movement will happen before the pre-flash/exposure sequence.

 

[mb66energy]

'Auto Intelligence' seems like an awfully grandiose moniker for simple geometry. :

More difficult than geometry, is knowing the distance, the reflectivity and the color of the ceiling in the real world. If the flash is able to calculate all that and adjust properly, even with the camera in portrait orientation, it would be a great technological leap.

If...

Where in, "With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows," do you interpret an ability to determine reflectivity and color? Incidentally, E-TTL will handle the reflectivity, there's no need for the flash to determine that a priori.

But E-TTL cannot measure the reflectivity of the bounce surface so this has to be handled by the flash to mix the direct and bounced fractions of light correctly.

I didn't mean to imply that E-TTL measures the reflectivity. But the reflectivity of the bounce surface determines how much bounced light reaches the subject, and E-TTL determines exposure based on light reflected from the subject. In that way, E-TTL 'handles' the reflectivity of the bounce surface.

What 'direct fraction' are you talking about mixing? If the head is elevated, it's all bounced (well, except for spill, but again, E-TTL handles that).

I'm not sure why people think that this 'AI auto bounce' feature is addressing flash exposure...that's what E-TTL does. The point is, E-TTL cares about how much light reflects back from the subject, not how that light gets to the subject. Whether direct or bounced, coming from one flash or 12 flashes, it doesn't matter. The point of this feature seems exactly as described: determining the optimal bounce position to eliminate unnecessary shadows.

Also, consider this: the exposure calculations are done based on a pre-flash, which occurs with almost no temporal separation from the actual exposure. The electric head rotation motor isn't going to move the flash head instantaneously. In other words, the AI auto bounce head movement will happen before the pre-flash/exposure sequence.

I think I know what E-TTL does. But how can the camera "decide" if there is a good angle of the flash without knowing what is in front of the flash reflector? IMO there is a need for some device in the flash that helps to measure the reflectivity of the bounce surface. This has to be used for flash positioning. And at least: If there is no bounce surface the flash has to detect this and switch to direct light only instead of wasting light into the sky.
And a 30 degree angle - measured against the optical axis of the lens - should emit a substantial fraction of direct light onto the object but a substantial fraction of bounced light (if there is a bounce surface).
The round thing (upper right corner of reflector section) is maybe a sensor to do this: E.g. a small LIDAR which (1) measures distance, (2) can also obtain reflectivity data and (3) scan the room geometry by scanning the distance during a sweep of the head.

 

[brad-man]

So what's the deal? The motor sucks so much juice from the batteries that they didn't want to include RT functionality? So this thing has less than half the power of the 430-RT and costs 60% more. I wonder who would choose this flash over the 430?

 

[tron]

I do not believe that the flash will be very AI at the end. Probably more like N.S. ;D

P.S Feel free to ask for clarifications about the acronym
Very helpful Hint: It is the opposite of AI ;D ;D ;D

 

[neuroanatomist]

I think I know what E-TTL does. But how can the camera "decide" if there is a good angle of the flash without knowing what is in front of the flash reflector? IMO there is a need for some device in the flash that helps to measure the reflectivity of the bounce surface. This has to be used for flash positioning. And at least: If there is no bounce surface the flash has to detect this and switch to direct light only instead of wasting light into the sky.

I thkn you're sounding like those who saw the 'dual-sensing IS' feature on the M50 and concluded that it has IBIS: reading too much into this.

With an electric head rotation device, the 470EX-AI automatically determines the optimal bounce position eliminating unnecessary shadows.

The goal is to eliminate unnecessary shadows, which is a geometry problem. The solution requires knowing the distance to the subject (which the lens provides, that's E-TTL II), and the distance to the bounce surface (which, presumably, is what the sensor on the front of the flash head measures).

Sure, it's possible that the flash measures the reflectivity, color, angle (not all ceilings are flat), temperature, roughness, and/or Rayleigh scattering properties of the bounce surface. But I'll bet it measures only distance. That would cover the case where someone tries to use it under the open sky, or under a 40' ceiling.