Approve the Cookies
This website uses cookies to improve your user experience. By using this site, you agree to our use of cookies and our Privacy Policy.
OK
Index  •   • New posts  •   • RTAT  •   • 'Best of'  •   • Gallery  •   • Gear  •   • Reviews
Guest
New posts  •   • RTAT  •   • 'Best of'  •   • Gallery  •   • Gear  •   • Reviews
Register to forums    Log in

 
FORUMS News & Rumors Photography Industry News 
Thread started 14 Jun 2010 (Monday) 13:23
Search threadPrev/next
sponsored links
(this ad will go away when you log in as a registered member)

Canon's new HDR patent

 
Er1kksen
Member
175 posts
Joined Aug 2009
Location: Fredonia, NY
     
Jun 25, 2010 14:54 |  #16

phigment wrote in post #10385072 (external link)
Did I miss something? I didn't see them saying anything about exposure length. I thought they would modify the sensitivity of each photoreceptor on an individual bases.

Quote from the patent: "According to an exemplary embodiment of the present invention, preliminary image capturing is performed using the image capturing unit, and the exposure time is assigned to each pixel based on the result of the preliminary image capturing. Image capturing can thus be performed with a wide dynamic range without a loss of highlight detail and a loss of shadow detail."

Note the phrase "exposure time."

If they were simply altering sensitivity (gain) on a per-pixel basis you could do the same thing in post by exposing to preserve the highlights at the lowest ISO practical and then raising the exposure (increasing the gain) on selected areas of the image (even on a per-pixel basis if you like). That's already a viable technique for taming wide exposure ranges with the latest sensors and their high SNR.


40D, 85mm f1.8, DC 952AF flash

  
  LOG IN TO REPLY
sponsored links
(this ad will go away when you log in as a registered member)
HappySnapper90
Cream of the Crop
5,145 posts
Likes: 3
Joined Aug 2008
Location: Cleveland, Ohio
     
Jun 25, 2010 16:21 |  #17

Canon makes their own image sensors. They would be leaps and bounds ahead of sony and panasonic/samsung if they're able to map out different exposure times for different pixel groups acrsos their image sensor after an initial preview of the incoming light. This would be everything getting exposed correctly, not image blending.

Pixels in brightest areas receiving light for only 1/1000 while shadow areas might get the longest exposure at 1/30 let's say - a 5 stop difference. And that is exposure times not just leaving the dynamic range of the image sensor to try to provide for detail in the shadows.

Controlling how long each pixel, or pixel groups, are "turned on" for would be huge.




  
  LOG IN TO REPLY
Er1kksen
Member
175 posts
Joined Aug 2009
Location: Fredonia, NY
     
Jun 26, 2010 15:46 |  #18

HappySnapper90 wrote in post #10426868 (external link)
Canon makes their own image sensors. They would be leaps and bounds ahead of sony and panasonic/samsung if they're able to map out different exposure times for different pixel groups acrsos their image sensor after an initial preview of the incoming light.

Controlling how long each pixel, or pixel groups, are "turned on" for would be huge.

All of that is true, but keep in mind that the state of the art is still several steps away from that capability. The next step is global-shutter sensors, which Panasonic, Sony and Samsung have all confirmed that they are working on. Canon is quiet on that front thus far, but you can be sure they're also working on it. You can't simply jump to the technology required for the patent to work from the current technology; global-shutter sensors are a necessary stepping stone to reach that goal.

Someday Canon will be able to use this patent. It just won't be particularly soon. In the meantime, this patent allows them to lay first claim to the technology once they get there, and leave the rest of the sensor manufacturers to come up with their own means to the same end to work around the patent.


40D, 85mm f1.8, DC 952AF flash

  
  LOG IN TO REPLY
IanW
Goldmember
Avatar
1,276 posts
Gallery: 10 photos
Likes: 125
Joined Nov 2005
Location: Wiltshire, UK
     
Jun 26, 2010 15:53 |  #19

I like the idea and would certainly like it to feature in my next camera. Given that this is likely to be about 4 years away (based on owning my 20D for 4+ years), should be plenty of time for them to perfect it! :)

Ian.


Canon 5DIII | BG-E11 | Sigma 50mm f1.4 Art | 24-105mm f4L | 70-200mm f2.8L IS II | 100-400mm f4.5-5.6L IS II | x1.4 & x2 III | 430EX
Yongnuo YN-622C | Kenko Tubes | LEE Filters | Manfrotto 055CXPRO3+498RC2, 694CX

flickr (external link)

  
  LOG IN TO REPLY
HappySnapper90
Cream of the Crop
5,145 posts
Likes: 3
Joined Aug 2008
Location: Cleveland, Ohio
     
Jun 26, 2010 16:16 |  #20

Er1kksen wrote in post #10431345 (external link)
All of that is true, but keep in mind that the state of the art is still several steps away from that capability.

How do you know what Canon is capable of and is doing in the R&D department? ;) Are you a worker in that lab?




  
  LOG IN TO REPLY
Pekka
El General Moderator
Avatar
17,756 posts
Gallery: 35 photos
Best ofs: 7
Likes: 1522
Joined Mar 2001
Location: Hellsinki, Finland
     
Jun 26, 2010 16:31 |  #21

Er1kksen wrote in post #10426468 (external link)
Quote from the patent: "According to an exemplary embodiment of the present invention, preliminary image capturing is performed using the image capturing unit, and the exposure time is assigned to each pixel based on the result of the preliminary image capturing. Image capturing can thus be performed with a wide dynamic range without a loss of highlight detail and a loss of shadow detail."

Note the phrase "exposure time."

If they were simply altering sensitivity (gain) on a per-pixel basis you could do the same thing in post by exposing to preserve the highlights at the lowest ISO practical and then raising the exposure (increasing the gain) on selected areas of the image (even on a per-pixel basis if you like). That's already a viable technique for taming wide exposure ranges with the latest sensors and their high SNR.

They are not claiming to alter per-pixel gain. They are claiming to alter per-pixel exposure. To preserve per-pixel exposure for highlights you only need to keep the pixel active shorter time.

How I read it: scene is metered so that shadow detail is preserved perfectly, i.e. exposed to the right and somewhat over. You have selected certain ISO for the shot. What the patent suggests is that there is a preliminary exposure which determines how much light each pixel receives in ISO X and time X. The the actual shot is taken, and inside exposure time individual pixels which receive highlights are are SHUT down when their "bucket is full". So this is not changing individual pixel gain, it is stopping pixels individually within exposure time.

So, for example a sunset shot: you set exposure time 1/4s to get perfect exposure of foreground, the sky would be overexposed X stops. With above patented feature you would take that 1/4s shot and the camera would stop pixels for the sky at 1/50 - 1/500 (smooth gradient) and run the rest for 1/4s.


The Forum Boss, El General Moderator
AMASS 2.1 Changelog (installed here now)

  
  LOG IN TO REPLY
Er1kksen
Member
175 posts
Joined Aug 2009
Location: Fredonia, NY
     
Jun 28, 2010 22:27 |  #22

HappySnapper90 wrote in post #10431440 (external link)
How do you know what Canon is capable of and is doing in the R&D department? ;) Are you a worker in that lab?

No, I am not. However, the industry is at where the industry is at. For Canon to make the leap that would be required to apply this technology now would be huge. Ridiculously huge. And again, it would require for them to have already developed global-shutter tech to step off from, and if they had already done so (it would put them well ahead of the competition), why wouldn't they have exploited it in their latest offerings? It would have given them a major competitive advantage- IF they had it.

The likeliest case is that Canon is hard at work on global shutter technology just like everyone else and that it'll start showing up around photokina.

Olympus has had a patent for very similar techniques for quite a while now as well and they don't have the technology to apply it yet either. The patents held by both companies are just textbook preemptory patenting: the design isn't feasible yet, but when it finally is, they have the patents to get a head start.

Pekka wrote in post #10431503 (external link)
They are not claiming to alter per-pixel gain. They are claiming to alter per-pixel exposure. To preserve per-pixel exposure for highlights you only need to keep the pixel active shorter time.

How I read it: scene is metered so that shadow detail is preserved perfectly, i.e. exposed to the right and somewhat over. You have selected certain ISO for the shot. What the patent suggests is that there is a preliminary exposure which determines how much light each pixel receives in ISO X and time X. The the actual shot is taken, and inside exposure time individual pixels which receive highlights are are SHUT down when their "bucket is full". So this is not changing individual pixel gain, it is stopping pixels individually within exposure time.

So, for example a sunset shot: you set exposure time 1/4s to get perfect exposure of foreground, the sky would be overexposed X stops. With above patented feature you would take that 1/4s shot and the camera would stop pixels for the sky at 1/50 - 1/500 (smooth gradient) and run the rest for 1/4s.

That is exactly the point I was making, which is why I emphasized the use of the phrase "exposure time" in the patent. The post you quoted was my explanation to a previous poster on exactly why exposure time was the important factor rather than gain (since, as I pointed out, they can't do anything by altering gain that we can't already do in post).

Were you intending to reply to the post I was correcting, or did you misread my post?


40D, 85mm f1.8, DC 952AF flash

  
  LOG IN TO REPLY
phigment
Member
230 posts
Likes: 2
Joined Jan 2009
Location: Waterloo, Ontario, Canada
     
Jun 29, 2010 10:50 |  #23

Er1kksen wrote in post #10444448 (external link)
The post you quoted was my explanation to a previous poster on exactly why exposure time was the important factor rather than gain (since, as I pointed out, they can't do anything by altering gain that we can't already do in post).

I admit I missed the part in the article about exposure time, but I beg to differ on this point.

If you have a sensor that can modify the gain on a per sensel level, you can have a greater dynamic range. You are basically saying there is no point to modifying the ISO setting of the sensor at all.

My point is that if you can detect that a certain sensel is going to saturate (or not even register anything) you can knock down (or up) the ISO level for that particular element.

In this way you wouldn't necessarily have to modify the exposure time on an individual level.


A7R, Firin 20mm, CV 21/4, CV 40/1.4, 100ZE, 300 4LIS, Tamron 150-600, Jupiter 8, RX100 IV

  
  LOG IN TO REPLY
Er1kksen
Member
175 posts
Joined Aug 2009
Location: Fredonia, NY
     
Jun 29, 2010 18:57 as a reply to  @ phigment's post |  #24

I see little advantage over exposing an image to preserve the highlights at ISO 200, pushing the midtones a stop and then pushing the shadows another stop or two as opposed to using a sensor that exposes the highlights at ISO 200, the midtones at ISO 400 and the shadows at 800 or 1600, at least with today's sensors. The amount of image noise in the end result would be about the same, and most scenes that present themselves will, in fact, fit within the dynamic range of a base-ISO exposure. Only scenes with extreme contrast outside the range of the sensor's DR would see an advantage in this situation, and it would mean some portions of the image would be extremely noisy while others would be very clean.

By altering exposure time per pixel, gain (and by extension image noise) is kept the same across the sensor (through the highlights and shadows alike), yielding a much superior result that modern post methods (except multiple-exposure-combining HDR) can't match by any means.


40D, 85mm f1.8, DC 952AF flash

  
  LOG IN TO REPLY
phigment
Member
230 posts
Likes: 2
Joined Jan 2009
Location: Waterloo, Ontario, Canada
     
Jun 29, 2010 19:08 |  #25

Er1kksen wrote in post #10450042 (external link)
I see little advantage over exposing an image to preserve the highlights at ISO 200, pushing the midtones a stop and then pushing the shadows another stop or two as opposed to using a sensor that exposes the highlights at ISO 200, the midtones at ISO 400 and the shadows at 800 or 1600, at least with today's sensors. The amount of image noise in the end result would be about the same, and most scenes that present themselves will, in fact, fit within the dynamic range of a base-ISO exposure. Only scenes with extreme contrast outside the range of the sensor's DR would see an advantage in this situation, and it would mean some portions of the image would be extremely noisy while others would be very clean.

By altering exposure time per pixel, gain (and by extension image noise) is kept the same across the sensor (through the highlights and shadows alike), yielding a much superior result that modern post methods (except multiple-exposure-combining HDR) can't match by any means.

So I take it that means you only ever shoot your 40D at ISO 100 then if everything can be done in post via pushing the exposure.


A7R, Firin 20mm, CV 21/4, CV 40/1.4, 100ZE, 300 4LIS, Tamron 150-600, Jupiter 8, RX100 IV

  
  LOG IN TO REPLY
Er1kksen
Member
175 posts
Joined Aug 2009
Location: Fredonia, NY
     
Jun 29, 2010 21:07 as a reply to  @ phigment's post |  #26

No. I said TODAY'S sensors. ;-)a

I no longer have the 40D (should update my sig) but the sensors on the market today blow it out of the water.

However, shooting my 40D at ISO 100, underexposed to push in post, and shooting an ISO or two higher at "proper" exposure never showed much of a difference in noise. Pushing several stops showed a small but visible difference, but it's all moot with my K-x (or a 7D, or a D90, or an NEX-5, or any modern FF...).

Bottom line was that I shot almost exclusively at ISO 100 with the 40D because I hated the files it put out above ISO 200. :p


40D, 85mm f1.8, DC 952AF flash

  
  LOG IN TO REPLY
PaintedSaint
Member
97 posts
Likes: 2
Joined Apr 2010
     
Jun 29, 2010 21:46 |  #27

I feel like an automode to hdr wouldnt yeild the best results but im sure they said that about autofocusing at first too...




  
  LOG IN TO REPLY
Er1kksen
Member
175 posts
Joined Aug 2009
Location: Fredonia, NY
     
Jun 30, 2010 11:25 |  #28

PaintedSaint wrote in post #10450964 (external link)
I feel like an automode to hdr wouldnt yeild the best results but im sure they said that about autofocusing at first too...

...it's not exactly an "auto-HDR" mode, per se, but a sensor feature that allows more of the scene's dynamic range to be captured within the SNR "sweet spot" of the sensor. Whether the end result looked like a straight capture or an overcooked HDR (or a natural-looking HDR) would depend entirely on how the user set the camera or interpreted the data in post.


40D, 85mm f1.8, DC 952AF flash

  
  LOG IN TO REPLY
sponsored links
(this ad will go away when you log in as a registered member)

7,312 views & 0 likes for this thread
Canon's new HDR patent
FORUMS News & Rumors Photography Industry News 
AAA
x 1600
y 1600

Jump to forum...   •  Rules   •  Index   •  New posts   •  RTAT   •  'Best of'   •  Gallery   •  Gear   •  Reviews   •  Member list   •  Polls   •  Image rules   •  Search   •  Password reset

Not a member yet?
Register to forums
Registered members may log in to forums and access all the features: full search, image upload, follow forums, own gear list and ratings, likes, more forums, private messaging, thread follow, notifications, own gallery, all settings, view hosted photos, own reviews, see more and do more... and all is free. Don't be a stranger - register now and start posting!


COOKIES DISCLAIMER: This website uses cookies to improve your user experience. By using this site, you agree to our use of cookies and to our privacy policy.
Privacy policy and cookie usage info.


POWERED BY AMASS forum software 2.1forum software
version 2.1 /
code and design
by Pekka Saarinen ©
for photography-on-the.net

Latest registered member is Constantin
917 guests, 218 members online
Simultaneous users record so far is 15144, that happened on Nov 22, 2018

Photography-on-the.net Digital Photography Forums is the website for photographers and all who love great photos, camera and post processing techniques, gear talk, discussion and sharing. Professionals, hobbyists, newbies and those who don't even own a camera -- all are welcome regardless of skill, favourite brand, gear, gender or age. Registering and usage is free.