The Canon EOS-1D X Mark II Specifications?

[Alastair Norcross]

Here's a brief explanation of what I mean

http://www.imaging-resource.com/PRODS/canon-7d-mark-ii/Z-7d2-65ptAF.jpg

That's the 7d mark 2 AF system. Each pair of lines you see on there opposite each other is a PDAF sensor - there's 30 of them total. Those sensors can be working on their own (in the case of single line direction sensors) or in concert with ones going the opposite direction (in the case of cross sensors). In this system, the UI breaks down the 30 PDAF sensors into 65 user selectable "points" , all of which are cross type, and the center is a double cross (that's the diagonal sensors on there)

The distance between the two pieces of each PDAF sensor pair is how sensitive they are - the horizontal and vertical sensors are all f/5.6 sensitive, and the high precision diagonal ones are f/2.8 sensitive (hence why they are spaced so much farther apart)

But looking at the layout, you can see that the AF 'points' you select in the UI actually have contiguous coverage between them, both horizontally and vertically, inside the 3 large zones (left, right and center) because it's not actually 65 discrete sensors on there, it's 30 large ones, which the UI breaks up into points to let you use just a portion of any one of them instead of the whole thing. (or an even smaller portion of them if you use the spot focus mode)

So all that contiguous area between the selectable points, is essentially "assist" points - when you are handing off point to point during tracking, the subject is still covered by the AF sensor and it can still see where it is and how it's moving before it reaches another user selectable point. Which is why I said if Canon counted the same way , they could tout that as a 173 point AF system, because there is over a hundred areas between the user selectable points that are still covered by af sensor. 102 actually, so I guess 167 rather than 173.

This is what the actual AF coverage area looks like
http://www.gauravmittal.com/wp-content/uploads/2014/12/Mittal_141217_27-Edit-2.jpg

So that's why a lot of us are suspect of what Nikon is claiming here - it's a very different way of counting AF points.

Any idea why Canon left the gaps between the three zones? When I'm using single point plus assist, it annoys me that I don't get a complete surround of assist point with the points at the outer edge of the middle zone, or the inner edges of the outer zones. I've sometimes wondered whether tracking birds, for example, I'd be better off using just one of the three large zones, rather than all 65 points in case the camera loses focus when a bird goes from one zone to the next. One of these days, I should test it. I'm not complaining about the 7DII's AF. In my experience, it's amazingly good. But I've always wondered why Canon left the gaps (albeit small gaps) between the zones.

 

[arbitrage]

Here's a brief explanation of what I mean

http://www.imaging-resource.com/PRODS/canon-7d-mark-ii/Z-7d2-65ptAF.jpg

That's the 7d mark 2 AF system. Each pair of lines you see on there opposite each other is a PDAF sensor - there's 30 of them total. Those sensors can be working on their own (in the case of single line direction sensors) or in concert with ones going the opposite direction (in the case of cross sensors). In this system, the UI breaks down the 30 PDAF sensors into 65 user selectable "points" , all of which are cross type, and the center is a double cross (that's the diagonal sensors on there)

The distance between the two pieces of each PDAF sensor pair is how sensitive they are - the horizontal and vertical sensors are all f/5.6 sensitive, and the high precision diagonal ones are f/2.8 sensitive (hence why they are spaced so much farther apart)

But looking at the layout, you can see that the AF 'points' you select in the UI actually have contiguous coverage between them, both horizontally and vertically, inside the 3 large zones (left, right and center) because it's not actually 65 discrete sensors on there, it's 30 large ones, which the UI breaks up into points to let you use just a portion of any one of them instead of the whole thing. (or an even smaller portion of them if you use the spot focus mode)

So all that contiguous area between the selectable points, is essentially "assist" points - when you are handing off point to point during tracking, the subject is still covered by the AF sensor and it can still see where it is and how it's moving before it reaches another user selectable point. Which is why I said if Canon counted the same way , they could tout that as a 173 point AF system, because there is over a hundred areas between the user selectable points that are still covered by af sensor. 102 actually, so I guess 167 rather than 173.

This is what the actual AF coverage area looks like
http://www.gauravmittal.com/wp-content/uploads/2014/12/Mittal_141217_27-Edit-2.jpg

So that's why a lot of us are suspect of what Nikon is claiming here - it's a very different way of counting AF points.

Any idea why Canon left the gaps between the three zones? When I'm using single point plus assist, it annoys me that I don't get a complete surround of assist point with the points at the outer edge of the middle zone, or the inner edges of the outer zones. I've sometimes wondered whether tracking birds, for example, I'd be better off using just one of the three large zones, rather than all 65 points in case the camera loses focus when a bird goes from one zone to the next. One of these days, I should test it. I'm not complaining about the 7DII's AF. In my experience, it's amazingly good. But I've always wondered why Canon left the gaps (albeit small gaps) between the zones.

It was an odd thing on the 7D2, on the 1DX and 5D3 there is a gap but the expansion points cross it so you always have your full square around your selected point.

 

[tr573]

Any idea why Canon left the gaps between the three zones? When I'm using single point plus assist, it annoys me that I don't get a complete surround of assist point with the points at the outer edge of the middle zone, or the inner edges of the outer zones. I've sometimes wondered whether tracking birds, for example, I'd be better off using just one of the three large zones, rather than all 65 points in case the camera loses focus when a bird goes from one zone to the next. One of these days, I should test it. I'm not complaining about the 7DII's AF. In my experience, it's amazingly good. But I've always wondered why Canon left the gaps (albeit small gaps) between the zones.

It was an odd thing on the 7D2, on the 1DX and 5D3 there is a gap but the expansion points cross it so you always have your full square around your selected point.

Software wise yeah, they let it include assist from the next zone, but there's still a coverage gap between the zones on that sensor also. As to why, I assume either manufacturing difficulty or expense or both limits how densely you can place the sensors / how long they can be continuously.

 

[Loswr]

tr573, that was enlightening, thanks.

Perhaps not so much.

Here's a brief explanation of what I mean

http://www.imaging-resource.com/PRODS/canon-7d-mark-ii/Z-7d2-65ptAF.jpg

That's the 7d mark 2 AF system.

But looking at the layout, you can see that the AF 'points' you select in the UI actually have contiguous coverage between them, both horizontally and vertically, inside the 3 large zones (left, right and center) because it's not actually 65 discrete sensors on there, it's 30 large ones, which the UI breaks up into points to let you use just a portion of any one of them instead of the whole thing. (or an even smaller portion of them if you use the spot focus mode)

So all that contiguous area between the selectable points, is essentially "assist" points...

Actually, no. That's an image of the 7DII's AF sensor, not the AF system. Another very important component of a phase detect AF system is the array of microlenses overlying the sensor, and those microlenses divide and focus the incoming light onto discrete sections of the AF sensor lines. So, in fact, AF points are actually AF points, not the continuous lines you seem to think exist.

Apologies for raining on the parade.

 

[tr573]

tr573, that was enlightening, thanks.

Perhaps not so much.

Here's a brief explanation of what I mean

http://www.imaging-resource.com/PRODS/canon-7d-mark-ii/Z-7d2-65ptAF.jpg

That's the 7d mark 2 AF system.

But looking at the layout, you can see that the AF 'points' you select in the UI actually have contiguous coverage between them, both horizontally and vertically, inside the 3 large zones (left, right and center) because it's not actually 65 discrete sensors on there, it's 30 large ones, which the UI breaks up into points to let you use just a portion of any one of them instead of the whole thing. (or an even smaller portion of them if you use the spot focus mode)

So all that contiguous area between the selectable points, is essentially "assist" points...

Actually, no. That's an image of the 7DII's AF sensor, not the AF system. Another very important component of a phase detect AF system is the array of microlenses overlying the sensor, and those microlenses divide and focus the incoming light onto discrete sections of the AF sensor lines. So, in fact, AF points are actually AF points, not the continuous lines you seem to think exist.

Apologies for raining on the parade.

glad you are here, because I know you have a ton of knowledge on this. (Your writeup on TDP, and Michael Clark's writeup about the 7D/70D sensor is what made me less confused about this topic, although not less enough since I left something out of this explanation!)

I found both articles when I was confused about why my 70D would focus on stuff WAY outside the marked "point" in the viewfinder if it had greater contrast. They helped me learn to think about the actual coverage area of the sensor that was assigned to that point by the software (and the coverage of the microlenses) so I would make better choices about where I was placing it.

It also made me think that taking the spot AF mode out of the 70D is responsible for a good 75% of the "help I upgraded to a 70D and the AF doesn't work" threads I see online.

So my question to you - since the coverage area projected by the microlens array effectively makes the AF point coverage area touch each other on the actual sensor, giving you that continuous coverage I explained poorly, do you see any difference between that and what Nikon is now advertising as 100+ extra 'assist points'? I'm curious to hear your opinion on this.

 

[Loswr]

I see your point about 'continuous' (contiguous might be a better descriptor). I suspect the D5 simply has smaller microlenses resulting in smaller 'points'. When not in tracking mode, I wonder if it will integrate across several points to achieve focus.

Nikon can also pack in more points since they all use a less accurate f/5.6 baseline vs. the longer f/4 baseline for many of the 1D X points.

 

[R1-7D]

I see your point about 'continuous' (contiguous might be a better descriptor). I suspect the D5 simply has smaller microlenses resulting in smaller 'points'. When not in tracking mode, I wonder if it will integrate across several points to achieve focus.

Nikon can also pack in more points since they all use a less accurate f/5.6 baseline vs. the longer f/4 baseline for many of the 1D X points.

Sorry to change topics, but you wrote an article on TDP (as mentioned above by tr573)? I'd love it read it. Can I search for it under the same screen name you use here?

 

[tr573]

I see your point about 'continuous' (contiguous might be a better descriptor). I suspect the D5 simply has smaller microlenses resulting in smaller 'points'. When not in tracking mode, I wonder if it will integrate across several points to achieve focus.

Nikon can also pack in more points since they all use a less accurate f/5.6 baseline vs. the longer f/4 baseline for many of the 1D X points.

Sorry to change topics, but you wrote an article on TDP (as mentioned above by tr573)? I'd love it read it. Can I search for it under the same screen name you use here?

http://www.the-digital-picture.com/Photography-Tips/canon-eos-dslr-autofocus-explained.aspx

Assumed it's the same neuroanatomist (not exactly a common handle)

 

[R1-7D]

I see your point about 'continuous' (contiguous might be a better descriptor). I suspect the D5 simply has smaller microlenses resulting in smaller 'points'. When not in tracking mode, I wonder if it will integrate across several points to achieve focus.

Nikon can also pack in more points since they all use a less accurate f/5.6 baseline vs. the longer f/4 baseline for many of the 1D X points.

Sorry to change topics, but you wrote an article on TDP (as mentioned above by tr573)? I'd love it read it. Can I search for it under the same screen name you use here?

http://www.the-digital-picture.com/Photography-Tips/canon-eos-dslr-autofocus-explained.aspx

Assumed it's the same neuroanatomist (not exactly a common handle)

Cheers! Neuro sent it to me too. Excellent read.

 

[gwilson]

Thanks Neuro. I was quite happy with my Canon AE-1's split prism focus system for 34 years after purchasing it in 1977. After 34 years of use I didn't think it owed me anything so I decided to get into digital and to learn about it on the fly. It must have been hard to photograph birds in flight with the split prism focus system.
Thanks for the great explanation of autofocus in Canon EOS DSLRs. It will help in selecting my next DSLR body, perhaps a 1Dx or II.

 

[Jack Douglas]

Wow, that was more than worth the read, and free. And now I know where the phrase "the life of Reilly" comes from. This is shaping up to be a good week.

Jack

 

[d]

Just received information from a source of mine deep within Canon HQ...he's been reading this thread with great interest, and thought it might be helpful if he shared with us from where Canon have taken their design cues for this revised version of the 1DX. Also, the designation 'Mark II' was simply a working title for the development period; the final camera will officially be the 1DX Mark T...


d.

 

[Gnocchi]

Just received information from a source of mine deep within Canon HQ...he's been reading this thread with great interest, and thought it might be helpful if he shared with us from where Canon have taken their design cues for this revised version of the 1DX. Also, the designation 'Mark II' was simply a working title for the development period; the final camera will officially be the 1DX Mark T...


d.

Pity the fool who picks on this camera..

 

[expatinasia]

CR member SereneSpeed found the leaked 1DX Mark II in this thread:

http://www.canonrumors.com/forum/index.php?topic=28966.0

http://procam.com/canon-eos-1d-x-mark-ii-dslr-camera-body-only

At least that is what it seems to be.

 

[Kwwund]

To the extent that there is new technology in the 1DX2, how much of it do we expect to trickle down into the next series of EOS camera introductions due this year (5d4, 80D)?

 

[KeithBreazeal]

Listen to the FPS music ;D

https://www.youtube.com/watch?feature=player_detailpage&v=I6JXCsVipcA

 

[Dick]

Anti climax. :-\