Patent: Microadjustment Automated

[Canon Rumors]

<div name="googleone_share_1" style="position:relative;z-index:5;float: right; /*margin: 70px 0 0 0;*/ top:70px; right:120px; width:0;"><g:plusone size="tall" count="1" href="http://www.canonrumors.com/?p=15494"></g:plusone></div><div style="float: right; margin:0 0 70px 70px;"><a href="https://twitter.com/share" class="twitter-share-button" data-count="vertical" data-url="http://www.canonrumors.com/?p=15494">Tweet</a></div>
<p>A patent has been published outlining automatic microadjustment for Canon DSLRs and lenses. This would definitely be a handy feature going forward. Especially if you rent lenses.</p>
<ul>
<li><span>Patent Publication No. 2013-254194</span>
<ul>
<li><span>Publication date 2013.12.19</span></li>
<li><span>Filing date 2012.5.11</span></li>
</ul>
</li>
<li><span>Issues surrounding the AF adjustment</span>
<ul>
<li><span>The looseness of the phase difference AF sensor, because the AF accuracy is worse, need AF adjustment</span></li>
<li><span>It is not done correctly the AF adjustment in your environment</span></li>
<li><span>Necessary facilities such as service center</span></li>
</ul>
</li>
<li><a href="http://egami.blog.so-net.ne.jp/tag/Canon"><span>Canon</span></a><span> patents</span>
<ul>
<li><span>Automation of the AF adjustment</span></li>
<li><span>Procedure</span>
<ul>
<li><span>Let me be the difference adjusted value of the phase difference AF and contrast AF</span></li>
<li><span>To determine the reliability</span></li>
</ul>
</li>
</ul>
</li>
</ul>
<p>Source: [<a href="http://egami.blog.so-net.ne.jp/2014-01-13" target="_blank">EG</a>]</p>
<p><strong><span style="color: #ff0000;">c</span>r</strong></p>

 

[Loswr]

Let me be the difference adjusted value of the phase difference AF and contrast AF
To determine the reliability

Great...except that we know contrast detect AF isn't necessarily the gold standard for precision. Given the specified and observed precision of both phase and contrast AF, I really think multiple tests are required for an accurate result. It they implement this method, and it does just one contrast detect AF, I certainly won't be using it!

 

[Jamesy]

And to be clear, this wouldn't make it into any existing camera bodies in the line-up, right?

 

[mb66energy]

Let me be the difference adjusted value of the phase difference AF and contrast AF
To determine the reliability

Great...except that we know contrast detect AF isn't necessarily the gold standard for precision. Given the specified and observed precision of both phase and contrast AF, I really think multiple tests are required for an accurate result. It they implement this method, and it does just one contrast detect AF, I certainly won't be using it!

One part of the patent seems to be the reliability check of a contrast AF measurement:
"Canon patent to determine the reliability , seems to validate whether can be adjusted correctly . [...] The play of the phase difference AF if that is the cause , because there should be aging , It is what you want also a mechanism to be calibrated on a regular basis ." (via google translate)

There might be an algorithm which excludes a blue sky as welll suited CDAF object.

Additionally I would expect a calibration of all AF points - if they differ to much this might invoke some cross checks or a new measurement procedure.

 

[ScottyP]

Is this a computer translation from Japanese? The text is kind of garbled.

I.E.

◦It is not done correctly the AF adjustment in your environment ??????
Let me be the difference adjusted value of the phase difference AF and contrast AF ??????

 

[East Wind Photography]

No doubt this will be a premium feature offered only on high end models. Current model contrast AF is prett poor. However I bet the new dual pixel sensors may have improved in this area. I think it's a good move provided they fix some of the current systemic issues. No reason this can't be done in Camera.

 

[unfocused]

I figured this was coming sooner or later. Manual microadjustment seems like a very primitive approach and prone to errors.

 

[Don Haines]

I figured this was coming sooner or later. Manual microadjustment seems like a very primitive approach and prone to errors.

+1

This approach allows the camera to keep a running average of dozens or hundreds of shots and allows a matrix of calibrations using focal lengths and distances. Any inaccuracy of an individual measurement gets lost in a long-term average...

 

[Mt Spokane Photography]

I pulled up the translation of the patent to English from the Japanese patent site. As usual, the Translation is poor, and difficult to read.

Its also written by one of those guys who, when asked what time it is, starts back with the big bang and tells of the formation of the galaxy, the movement of stars and planets, he can't just say how it works, but includes all the formulas for contrast detection, there is so much extraneous data that the real information is hard to find.

It does sound like they have all the bases covered, and understand the limitations of contrast detection.

BTW, it appears that while contrast detection on some models is unreliable, the 6D and subsequent models have been quite accurate and consistent. It makes me assume that they have improved the accuracy or discovered a flaw in the older method. If that's the case though, why haven't we seen a firmware update?

The patent does appear to all be in firmware, no physical changes to a camera, so it could be done on existent bodies, but will likely appear on new bodies first. The timing of the patent release makes it possible that we will see it on a 7D MK II and whatever new FF bodies appear.

 

[takesome1]

It was just a matter of time.

Sux to be Reikan.

 

[Marauder]

Its also written by one of those guys who, when asked what time it is, starts back with the big bang and tells of the formation of the galaxy, the movement of stars and planets, he can't just say how it works, but includes all the formulas for contrast detection, there is so much extraneous data that the real information is hard to find.

LOL With an all-but-infinite number of "LOL's" behind it!

 

[BL]

It was just a matter of time.

Sux to be Reikan.

but very profitable while it lasted

 

[helpful]

If this would be done properly, it would be the best advancement ever (that I can imagine) in the current professional workflow. I personally would use it at just about every shoot, setting up my camera and calibrating the focus adjustment to middle of the exact expected distance to my subjects, in just the same way that I set custom white balance to fit exactly the environment I am shooting. I have noticed, for example, that perfect calibration of my 400mm at 80 m does not result in perfect calibration at 25 m, and vice versa, so hence the usefulness to a professional of AF microadjustment in the field under working conditions.

Hopefully there will be an option to base the AF micro adjustment on not only the results of contrast AF comparison, but on several real automated test photos. There could be a "faster" and a "full precision" option. It would surely be possible to AF calibrate for a certain target using contrast AF comparison without taking any real photos in a matter of 1-2 seconds using Newton's method (possibly with least squares fitting) to refine the results. And with the option to use several actual photographs for exact data, it would hopefully be possible to complete the entire process in less than 5 seconds and obtain results as absolutely perfect as the maximum theoretical potential precision of the AF phase detection sensors, by centering their accuracy on dead center.

 

[TWI by Dustin Abbott]

This is actually a (potentially) very exciting development. This could be a very real benefit, and I'm excited about it even though I invested in FoCal pro. Canon takes a lot of heat around here for a (perceived) lack of innovation, but this development along with DualPixel AF are actually probably some of the most innovative new features in recent years.

 

[Woody]

Improvement on MagicLantern's implementation of the dot tune method? Sounds cool. Anything is welcome... as long as it's free...

 

[Ruined]

Improvement on MagicLantern's implementation of the dot tune method? Sounds cool. Anything is welcome... as long as it's free...

Sure, if you count buying a new camera as free

 

[verysimplejason]

I just hope all of these are purely software enhancements. My 6D's doing fine most of the time...

 

[Rienzphotoz]

<div name=\"googleone_share_1\" style=\"position:relative;z-index:5;float: right; /*margin: 70px 0 0 0;*/ top:70px; right:120px; width:0;\"><glusone size=\"tall\" count=\"1\" href=\"http://www.canonrumors.com/?p=15494\"></glusone></div><div style=\"float: right; margin:0 0 70px 70px;\"><a href=\"https://twitter.com/share\" class=\"twitter-share-button\" data-count=\"vertical\" data-url=\"http://www.canonrumors.com/?p=15494\">Tweet</a></div>
<p>A patent has been published outlining automatic microadjustment for Canon DSLRs and lenses. This would definitely be a handy feature going forward. Especially if you rent lenses.</p>
<ul>
<li><span>Patent Publication No. 2013-254194</span>
<ul>
<li><span>Publication date 2013.12.19</span></li>
<li><span>Filing date 2012.5.11</span></li>
</ul>
</li>
<li><span>Issues surrounding the AF adjustment</span>
<ul>
<li><span>The looseness of the phase difference AF sensor, because the AF accuracy is worse, need AF adjustment</span></li>
<li><span>It is not done correctly the AF adjustment in your environment</span></li>
<li><span>Necessary facilities such as service center</span></li>
</ul>
</li>
<li><a href=\"http://egami.blog.so-net.ne.jp/tag/Canon\"><span>Canon</span></a><span> patents</span>
<ul>
<li><span>Automation of the AF adjustment</span></li>
<li><span>Procedure</span>
<ul>
<li><span>Let me be the difference adjusted value of the phase difference AF and contrast AF</span></li>
<li><span>To determine the reliability</span></li>
</ul>
</li>
</ul>
</li>
</ul>
<p>Source: [<a href=\"http://egami.blog.so-net.ne.jp/2014-01-13\" target=\"_blank\">EG</a>]</p>
<p><strong><span style=\"color: #ff0000;\">c</span>r</strong></p>

Sounds good ... any improvement in technology, that makes life a little more convenient, is always welcome ... it might not be "perfect" but its a step in the right direction and in time it will only improve.

 

[sanj]

God be praised!

 

[horshack]

I'd have to read through the patent to see if there are any unique elements to it but generally speaking the idea of using CDAF to auto-tune PDAF has lots of prior art online.