A new Canon ILC has hit certification

usern4cr

R5
CR Pro
Sep 2, 2018
1,376
2,308
Kentucky, USA
I hope the quad AF does not mean quad Bayer as well.
*EDIT* - This post is about claiming 16 Quad Bayer pixels vs the normal 4 Bayer pixels. It is *not* about QP vs DP design.

I still don't see what benefit a Quad Pixel AF "Quad-Bayer" (16 Bayer pixels) could offer. For pixels in focus (where you would want extra resolution), phase detect DP or QP gives the same dual or quad values for all values (by design) so there's no benefit there. With out of focus pixels (where the QP values differ) , who cares about increasing resolution as it's just a smear which could be done in post.

So I only see "QP AF, Quad-Bayer" as marketing BS overclaiming of resolution which I hope Canon has the integrity to avoid.

Also, to be clear, I see QP AF (by itself) as a brilliant AF method and the #1 new feature which I *really* want to have in my next camera! I have fully given over AF (with my use of "spot focus" and "eye focus" back buttons) to the camera and QP would be a massive improvement over the already brilliant DP design in the R5.
 
Last edited:
  • Like
Reactions: 1 user
Upvote 0
Apr 1, 2016
348
321
About quad pixel AF, here's a copy from an answer of this, that explains why it could be beneficial:

"Right now with dual pixel AF (DPAF) sensors, you can focus reliably while the camera is in the horizontal position and your edge of contrast that you are locking on are vertical. If the edge is horizontal (or parallel to the camera orientation) then it has extreme difficulty in locking on. This is because all the pixels are arranged in one direction for dual pixel AF. What Canon needs is a quad pixel, where the pixel is split up, not once but twice, allowing for different phase different arrangements."
 
  • Like
Reactions: 1 users
Upvote 0
Mar 10, 2021
73
85
I still don't see what benefit a Quad Pixel AF "Quad-Bayer" could offer. For pixels in focus (where you would want extra resolution), phase detect DP or QP gives the same dual or quad values for all values (by design) so there's no benefit there. With out of focus pixels (where the QP values differ) , who cares about increasing resolution as it's just a smear which could be done in post.

So I only see "QP AF, Quad-Bayer" as marketing BS overclaiming of resolution which I hope Canon has the integrity to avoid.

Also, to be clear, I see QP AF (by itself) as a brilliant AF method and the #1 new feature which I *really* want to have in my next camera!
dual pixel AF (left vs right phase detect) have issues focusing a pattern with thin perfectly horizontal lines being the only contrasting elements in the AF area. Quad pixel AF would resolve this challenge.
 
  • Like
Reactions: 1 user
Upvote 0

usern4cr

R5
CR Pro
Sep 2, 2018
1,376
2,308
Kentucky, USA
About quad pixel AF, here's a copy from an answer of this, that explains why it could be beneficial:

"Right now with dual pixel AF (DPAF) sensors, you can focus reliably while the camera is in the horizontal position and your edge of contrast that you are locking on are vertical. If the edge is horizontal (or parallel to the camera orientation) then it has extreme difficulty in locking on. This is because all the pixels are arranged in one direction for dual pixel AF. What Canon needs is a quad pixel, where the pixel is split up, not once but twice, allowing for different phase different arrangements."
I completely agree. It's my #1 "want to have" in my next body. (just don't call it "quad Bayer" and claim 4x more pixel resolution than it really is)
 
  • Like
Reactions: 1 users
Upvote 0
Jan 29, 2011
10,675
6,121
I still don't see what benefit a Quad Pixel AF "Quad-Bayer" could offer. For pixels in focus (where you would want extra resolution), phase detect DP or QP gives the same dual or quad values for all values (by design) so there's no benefit there. With out of focus pixels (where the QP values differ) , who cares about increasing resolution as it's just a smear which could be done in post.

So I only see "QP AF, Quad-Bayer" as marketing BS overclaiming of resolution which I hope Canon has the integrity to avoid.

Also, to be clear, I see QP AF (by itself) as a brilliant AF method and the #1 new feature which I *really* want to have in my next camera! I have fully given over AF (with my use of "spot focus" and "eye focus" back buttons) to the camera and QP would be a massive improvement over the already brilliant DP design in the R5.

Surely the value is the same, only more so, that DP. You have the option to use the sum of those photodiodes to get more detail in highlights. With QP you have the potential to increase DR by two stops over a standard array.
 
  • Like
Reactions: 1 user
Upvote 0

DBounce

Canon Eos R3
May 3, 2016
500
544
There was a rumour (credibility unknown) that the R1 will be closer to $8500 as it may be well above the A1 and Z9. The R5 also pushed up market in price and specs.
If Canon is going to ask that kind of money, this thing has better be near perfect... if it is, I’ll buy it happily. But if it falls short they will lose my business. This isn’t a luxury camera... it has to perform.
 
Upvote 0

usern4cr

R5
CR Pro
Sep 2, 2018
1,376
2,308
Kentucky, USA
dual pixel AF (left vs right phase detect) have issues focusing a pattern with thin perfectly horizontal lines being the only contrasting elements in the AF area. Quad pixel AF would resolve this challenge.
Yes, but my comment was about claiming 16 Bayer pixels instead of 4 Bayer pixels. I've already mentioned QP AF (by itself) as my #1 "want" in the next camera.
 
Upvote 0

usern4cr

R5
CR Pro
Sep 2, 2018
1,376
2,308
Kentucky, USA
Surely the value is the same, only more so, that DP. You have the option to use the sum of those photodiodes to get more detail in highlights. With QP you have the potential to increase DR by two stops over a standard array.
I don't see how existing technology DP or QP increases dynamic range. You haven't increased the total well count between 1 normal pixel, 2 half pixels or 4 quarter pixels (in fact with the added complexity you probably have reduced the total well count). And if the sub-pixels are all the same size (and sensitivity) then you haven't given different dual/quad sensitivity ranges between the sub-pixels to allow for oversaturation of the overall image in order to expand overall dynamic range in the entire image. If you do have different sensitivity of the sub-pixels then it will complicate the AF design substantially and possibly cripple it as you now have blown out sub-pixels being compared to non-blown out sub-pixels which will cause AF errors.

In summary, I see DP or QP having a massive benefit to AF (it's intended purpose) as well as offering the creation of depth masks to allow for computational algorithms to increase blur in OOF areas (both or which are truly great!). But I don't see DP or QP having an effect on dynamic range or signal-to-noise. If you want to increase overall image dynamic range by correcting blown-out exposure, then use different sensitivity sub-pixels which are not used for phase detect AF.
 
Last edited:
Upvote 0
Jan 29, 2011
10,675
6,121
I don't see how existing technology DP or QP increases dynamic range. You haven't increased the total well count between 1 normal pixel, 2 half pixels or 4 quarter pixels (in fact with the added complexity you probably have reduced the total well count). And if the sub-pixels are all the same size (and sensitivity) then you haven't given different dual/quad sensitivity ranges between the sub-pixels to allow for oversaturation of the overall image in order to expand overall dynamic range in the entire image. If you do have different sensitivity of the sub-pixels then it will complicate the AF design substantially and possibly cripple it as you now have blown out sub-pixels being compared to non-blown out sub-pixels which will cause AF errors.

Maths.


 
  • Like
Reactions: 4 users
Upvote 0
D

Deleted member 381342

Guest
If Canon is going to ask that kind of money, this thing has better be near perfect... if it is, I’ll buy it happily. But if it falls short they will lose my business. This isn’t a luxury camera... it has to perform.

If it is markably more expensive than the Z9 and A1(Which becomes very expensive when you add the grip compared to the D6 and 1DxIII) I would expect it has some killer technology like a global shutter that sets it apart. Until then it is just speculation, but no more nor less credible than any other rumour.
 
Upvote 0

DBounce

Canon Eos R3
May 3, 2016
500
544
If it is markably more expensive than the Z9 and A1(Which becomes very expensive when you add the grip compared to the D6 and 1DxIII) I would expect it has some killer technology like a global shutter that sets it apart. Until then it is just speculation, but no more nor less credible than any other rumour.
Honestly, until specs are released it’s hard to say if it’sa buy... but I’m more excited about the R1 than any camera in the last 3 years.
 
Upvote 0
Mar 19, 2021
117
148
Finally a compact full frame ? Like the Sony A7C ? Please ...
Its possibe ,there isnt any small gripped R camera yet for asian small hands.
But they may still postpone it and make just one last M camera too.
Or making cheap mechanical ibis for M might be too unreliable and they jump straight to global shutter digital ibis R
 
Upvote 0