Canon Dual Pixel CMOS AF

how many fucking retarded shill threads are you going to make in one day

holy fuck literally kill yourself

>>3114560
>no 4k
think of how many eos m5 and 6 they could have sold.
people bought sony and panasonic instead.

>>3114567
But, that would require Canon actually do R&D on sensors.

>>3114568
>4k
>sensor r&d
how would they be related at all

>>3114604
You know how high-end video cameras have fewer megapixels, but capture better videos than DSLRs with more megapixels?

Every time you capture from a sensor, it generates heat.

Photo CMOS are optimized for at most 15 captures per second in burst while video CMOS are optimized for sustained 60+ captures per second.

DSLR image processors get around this by activating only every fourth or third pixel for video capture and even moving them around the sensor to keep the sensor from getting too hot.

Sony got the photo CMOS sensor getting too hot for capture problem down pat with photo sensors that generate less heat and image processors that are better at moving the active parts around to spread out the heat.

Canon hasn't.

The reason Canon can't push the video to 4K in the new DSLRs is because the number of pixels that have to be active on the CMOS increases 4x going from 1080p to 4K, and their photo CMOS will overheat when 4 times as many pixels are active in sustained basis and damage the camera (unlike Sony's).

>>3114635
thanks for that actual explanation
didn't expect that on /p/ these days, good on you

>>3114568
wow. what a retard.
they just need a processor that doesn't shit out.

>>3114635
what is pixel bin?
what is line skipping?
what is cropping?

protip: canon is shooting 4k on 1dxii

File: 1490127871498.gif (893KB, 400x554px) Image search: [Google]

893KB, 400x554px

>>3114635
Nice explanation, anon. Thanks for sharing.

>>3114635
and yet sony is most the brand most famous for overheating and dying.
Fuck off shill retard.

>>3114604
I don't believe they needed to do r&d. look at the 1Dc. they had a chance to not gimp themselves but look at their cinema line.

File: 1483601366503.jpg (121KB, 1024x768px) Image search: [Google]

121KB, 1024x768px

>>3114560
Since no one has answered you about the dual pixel autofocus it is actually as good as canon states. Possibly better.

>>3114560
Just go to a store and try it out. It's pretty good.

>>3114933
They think if they gimp their DSLRs, people will run out to buy $28,000 Cinema EOS cameras.

Whenever you're baffled by Canon's retarded business decisions, remember the CEO of Canon is 81 years old, and the average age of Canon's executive team is 62.

>>3114677
>canon is shooting 4k on 1dxii
with a huge crop and bloated format
>>3114635
thanks for explaining pixel binning in an extremely long-winded comment. /s

File: .png (339KB, 815x611px) Image search: [Google]

339KB, 815x611px

>>3114560
https://www.youtube.com/watch?v=ubYqXApI1qc
Canon marketed DP-PDAF really well. In reality it's only somewhat faster than their previous gen Hybrid AF. Still far better than CDAF but also slower than dedicated PDAF.

>>3114635
wow anon, thank you

>>3115100
Just keep in mind, DP-PDAF is limited due to the image-sensors small photo-cites, where dedicated PDAF allows optimization for af.
The dedicated PDAFs advantage is speed and reliability in difficult situations. The DP-PDAFs advantage is accuracy, since measurement takes place on exactly the image plane, no misalignment possible

File: .png (68KB, 524x950px) Image search: [Google]

68KB, 524x950px

>>3116495
Well seems like even Sony managed to do it.

>>3116497
Which model and where are the af photo-cites? I tried to found docs about sonys af tech

>>3116920
a9. earlier sony stuff was shit.
http://www.imaging-resource.com/PRODS/sony-a9/sony-a9A6.HTM

>>3116497
Sony is narrowing the gap with the A9, but dedicated AF sensors are still better.

>>3116929
Actually scratch this >>3116497 and >>3116922. It's only shutter lag and not focus acquisition time. There isn't really a standard to compare focus acquisition since different subjects and lighting can greatly change AF speed.

>>3116929
>>3116935
Only way to get comparable AF speed tests is side by side. But then again, lenses are also a factor in AF speed among other things. Here's a test. Not sure how accurate.
https://www.youtube.com/watch?v=KX1sfy__7A4

>>3116922
That link doesn't explain the tech behind DPAF on Sony cameras. Is it stacking or also dualpixels? Googling always leads me to Canon DualPixel related stuff and the Samsung Galaxy S7

File: ZTECH_PDAF_PIXELS.gif (35KB, 250x250px) Image search: [Google]

35KB, 250x250px

>>3116945
Don't think Sony uses DP-PDAF in anything outside of smartphone camera sensors. They've always just used a standard on sensor PDAF design and just upped the AF sensor count each gen. Seems like it might just be a combo of brute force (more AF points) and software optimization/algorithms.
Also see >>3116935

>>3116957
I see!, thanks! Of course averaging more data gives better results for every sensor. I envy the large coverage. Sadly, CaNikon is more like, let's improve the center coverage.