Auto focus doesn't work on bird surrounded by foliage

[rfermat]

I have a Canon EOS 5D Mark III, purchased in late 2016.  I've taken at least 10000 photos with it.  I use either the EF 100-400 IS II or the 70-300 image stabilizer lens.  In either case, when I am photographing a bird that is perched on a branch and has background foliage, AF usually doesn't work.  I have lost at least 70% of the oppotunities for such photos, as eventually the bird flies off.  The %$^&#$ camera insists on focusing on the background or foreground.   These are situations where I am reasonably close to the bird, so that it opccupies maybe 1/12 or 1/16th of the image.

This is beyond frustrating.  I have experimented with many different AF settings and nothing seems to do any good.  

Anyone else have this problem?

[ebiggs1]

Keep in mind Single-point SPOT AF is not an option if you are in Auto mode.  You shouldn't be any way but just to be clear.

Av is the best choice.

[rfermat]

Thank you.  I did not know there is a difference betwen Single-point AF and Single-point SPOT AF.  I don't recall seeing thay fact anywhere before.

Also, I am astonished to see your blue and red rectangles.  I assumed single spot meant -- single spot.  Duh.  Evidently not.

[ebiggs1]

"Evidently not."

No it does mean just a single more tiny single point.  You will not see anything additional in the viewfinder just a smaller square. The colored example just indicates coverage you can expect.

I would not use it as a first option.  Let's get the basic settings a good try first.  When things don't go as expected it's always best to start off at the beginning.   The walk before you run thingy.

[TCampbell]

The 5D III (and IV, and 1D X, 1D X II, 7D II, etc.) are all very "technical" cameras.  They are very powerful... but they have loads of options that other cameras don't have. 

You really want to take the time to go through the manual and learn the features. 

Also, download and read Canon's 47 page PDF guide on the 5D III auto-focus system (this entire 47 page document is dedicated to JUST the focus system on your camera ... the number of pages alone should give you some idea that this camera has a LOT of options when it comes to focus.)

http://cpn.canon-europe.com/files/product/cameras/eos_5d_mark_iii/AF_guide_EOS5D_MarkIII_eng_January...

"Spot AF" is not the same as "Single Point AF".  Spot AF reduces the AF point to the smallest size possible to avoid possibility that another object will interfere with your subject and confuse focus.  But also... when you are in AI Servo mode, all the "case" settings (you'll read about those in the document) come into play.  You can tune the behavior of the camera as to whether it should ignore obstacles that momentarily block your subject ... or if should immediately snap the focus to that new subject.

A downside of "Spot AF" is that since the point is much smaller, it needs to be something wtih good contrast (to help the camera lock focus).  "Single Point AF" has a more generous area so that's more opportunity to find contrast to lock focus.  There's also "AF Point Expansion" that uses four additional AF points (above/below/left & right of your chosen AF ... these points "extend" the range of the single AF point.  There's also an AF Point Expansion that uses all 8 surrounding points (again... same behavior.  It wants to use the center but borrows the other points to "expand" the area in an effort to increase the area to search for good contrast.)  The risk is... as you expand the area it can search for good contrast... it might find good contrast on a tree branch or leaf and NOT on your subject.)

If you have a fast moving subject free of distracting obstacles, the camera can often work better (focus faster) if you give it a larger area.  Basically use more when you can.  When you realize there are competing obstacles... reduce the area so the camera HAS to use the tiny little spot you picked out to lock focus.

[rfermat]

Thanks for all the replies.

A perched bird is basically an ellipse for the body, an ellipse for the head, and a rectangular shape for a tail.  There may be a branch interesecting one of these.  Nonetheless, I don't see why software can't be written to search for a bird near the center of the field of view.  I say this as a mathematician and computer programmer.

This would establish a "perched bird mode".  It's more complicated for a flying bird, but I've not had trouble photographing them.  Usually there is no competing foliage.

[ebiggs1]

Remember camera sensors only see luminosity.  They can not distinguish a limb from a bird since it is just lighter or darker.

[kvbarkley]

That might work in live view, but through the view finder you only have at most 60 widely separated pixels.

And you might want a perched-bird focus mode, but I want a bee focus mode.

How many modes should canon add?

[coachboz68]

---

@rfermat wrote:

Nonetheless, I don't see why software can't be written to search for a bird near the center of the field of view.  I say this as a mathematician and computer programmer.

---

Sorry if this is a bit of a diversion from topic... but I am in tech as well, so your comment caught my eye.  Google is doing quite a lot with its AI platform wrt photography.  In this [1] case, they worked on using AI to sift through "street view" images and basically work out which clips would make "good" photographs.  They've also been doing well in competitions against human photographers where people "vote" on images.  And here [2] is an example where they are doing more with short video clips that are determined worthy of capture with their AI.  I suspect that phones will be the faster adopters of AI technology for photography and professional-grade cameras will lag a bit due to the use cases being somewhat different.  

[1] https://petapixel.com/2017/07/14/google-uses-ai-create-professional-photos-street-view-shots/

[2] https://ai.googleblog.com/2018/05/automatic-photography-with-google-clips.html 

[TCampbell]

---

@coachboz68 wrote:

---

@rfermat wrote:

Nonetheless, I don't see why software can't be written to search for a bird near the center of the field of view.  I say this as a mathematician and computer programmer.

---

Sorry if this is a bit of a diversion from topic... but I am in tech as well, so your comment caught my eye.  Google is doing quite a lot with its AI platform wrt photography.  In this [1] case, they worked on using AI to sift through "street view" images and basically work out which clips would make "good" photographs.  They've also been doing well in competitions against human photographers where people "vote" on images.  And here [2] is an example where they are doing more with short video clips that are determined worthy of capture with their AI.  I suspect that phones will be the faster adopters of AI technology for photography and professional-grade cameras will lag a bit due to the use cases being somewhat different.  

[1] https://petapixel.com/2017/07/14/google-uses-ai-create-professional-photos-street-view-shots/

[2] https://ai.googleblog.com/2018/05/automatic-photography-with-google-clips.html 

---

The unfortunate thing about AI is that it has to be "trained" ... and the training requires loads and loads of data (which, fortunately, Google has in abundance.)  

There are quite a few cloud-based products that can use AI / machine-learning to enhance product functionality in ways just not possible (or certainly not practical) for individual users or running on non-cloud computers simply because of the impracticality of it (individual users don't have nearly enough data).

Machine learning isn't "new" ... the algorithms being used are things developed back in the 1980s.  But you're reading so much more about now because today, with the internet and cloud computing, it's possible to feed these systems enough data with processors that can churn through it fast enough to be practical.  But there are "data lakes" and "server farms" involved ... not individual computers or individual storage devices.

When you use a voice-assistant (Siri, Alexa, Google, etc.), you'll notice these devices are all cloud-connected and that's because your phone or smart speaker doesn't have the data or capacity to learn... but the servers in the cloud do.

So the question is:  Could your camera recognize the bird?

The reality of how this would have to work (and why it's not a viable strategy ... at least not yet) is that the camera would need to capture a sample image (or a few sample images ... for example 1 second worth of video) and send it up to a cloud server.  AI could churn through it and recognize that there's a bird in the photo and also recognize where the bird is located ... directing the camera's focus attention to the correct spot.

But think about the problems...

(a) the camera needs to be not just wi-fi enabled, but internet-connected (currently you can have your phone connect to your camera, but the camera only has a private connection to the phone... it can't use the Internet)

(b) this network needs to be fast and reliable.  It needs to work everywhere you could possibly imagine taking a photo (and you can imagine taking photos in a lot of places that are beyond the reach of wireless networks.)

(c) the servers need to be able to respond very quickly.  Basically you want to half-press the shutter and, all in the time span of a few fractions of a second, have the camera record a short clip, upload it, get it analyzed, download the results, focus, and shoot... and it all needs to happen fast enough that the camera owner doesn't complain about the horrible lag time.

So there's two answers:  

From a "could we have an algorithm that does this?" perspective, the answer is... sure can!

From a "technically can we implement an architecture that provides an experience that consumers would like?" perspective, the answer is... sorry, not yet.

[rfermat]

TCampbell:  I don't think AI, massive sifting through data, or an internet connection is necessary.  Software to analyze an image looking for circles and parallelograms has been around for quite a while.  It's really not that complicated.  I don't think it should be hard to add ellipses.  I've worked on somewhat similar image analysis problems.  I'm guessing that a fast chip and a few gigabytes of RAM would be enough.