July 29, 2014, 02:59:59 AM

Author Topic: New DSLR on Tuesday, October 18, 2011. [CR3]  (Read 44951 times)

PeterJ

  • 7D
  • *****
  • Posts: 342
    • View Profile
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #165 on: October 15, 2011, 03:12:38 AM »
But the Frames Per Second issue is not the motor in the cameras , it's the speed in which the sensor can dispurse its electrical charge , and prepare for the next shot - the sensor must drain , and refill with each image shot , and the speed of cicuitry is slowed the larger the sensor is , and the more power the CPU uses - so although they are comming up with new faster ways to charge and discharge the sensor , the MP always going up slowly balances it out from giving huge FPS increases. All we can hope for is a new rapid current charge and discharge that does not affect battery , or more seriously , sensor life span.
This probably needs a "citation needed" tag :P. Unless I'm mistaken I thought CMOS sensors were effectively a photodiode with an active amplifier, so the "refill" is caused the photoelectric effect and the discharge would consist of removing the small charge held by the inherent capacitance of the photodiode through a transistor. You need a certain number of photons of light to hit the sensor but if you're taking a 1/100s shot there's no reason each inidividual cell can't approach that in terms of fps.

General point is valid though, but it's not so much how quickly you can discharge the sensor it's how long it takes to do something useful with the output of each cell, namely taking an analog to digital conversion. That's largely a function of cost / size / power and then it has to be processed by a CPU and stored to memory fast enough which again is a function of cost / size / power.

canon rumors FORUM

Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #165 on: October 15, 2011, 03:12:38 AM »

Joseph

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #166 on: October 15, 2011, 03:23:09 AM »
Quick wikipedia-ing: the EOS 1V did 10fps on a moving mirror, the EOS1nRS did 12fps on a pellicle mirror. And the current 1Dmk4 does 10fps on a moving mirror. (and the best I can find out about the F1 is 9fps, although there probably was a motor drive for 14fps that i didn't find).
Surely it's gotta be possible to go above thiese numbers, we've had 11 years of tech since the 1V, there's got to be a more efficient motor and materials capable of the G-forces involved by now. If the mirror can do it, the rest is just software.

I definitely like the idea of the "2 frames per mirror drop" (or 3).
- Firstly, it'll have to be via a custom function, only for those who want it, because it will almost definitely reduce hit-rate (at least in the first incarnation, coding and algorithms will get better over time).
- Second, they drop the mirror in between each shot for AF-tracking. See my post elsewhere quoting canon talking about their new "dream AF in the near future". If this AF really is that good, maybe it can predictively track moving targets with only dropping the mirror after every 2nd or 3rd shot? Or it can process the images it takes quickly and has improved contrast-based AF to help it keep up without dropping the mirror?
Speculation now, of course, and it may not be in this release, but it will be in *something* within a few years, i'd bet. It'll take processing power and good coding, but it's not impossible.

As for the 18MP FF sensor, I just don't believe it one bit. There is absolutely 0% chance they are going to drop the flagship MP-count DOWN, even if it does 1 million ISO as well as my 7D does iso400, it ain't gonna happen.
If it's 18MP, whether it's FF or APS-H, it's got a 1D-label (or 3D, or 6D, just NOT 1Ds or 5D).
18MP could be the "APS-H crop" mode (i'm too lazy to calculate how big an FF sensor would be for that). If they release an 18mp camera now, then they pretty much have to follow nikon's route with the D3/D3x and release a 30+mp 1Ds sooner rather than later, or too many actual pros will jump ship (and i'll be on ebay to get all their 1Ds3s when they do).

Alternatively, it might be an 18MP FF sensor, iff (which reads 'if and only if' to you non-maths nerds), it's a Foveon-type. (but will they brand it as a 54MP sensor a-la Sigma?).

And just because nikon may or may not have a patent on hardware-cropping, doesn't mean canon can't work around it. P&S cameras have had "digital zoom" for years. the 60/600D has digitalzoom/hardwarecropping for video, so some are already getting around it in some way or another.
Does nikon have the patent just for APS-C cropping? canon can get around that by using APS-H cropping, or even their own APS-C cropping (seeing as canon's APS-C is a different size than nikon's APS-C). Depends on how the patent (if it exists) is written as to how easily they can get around it.

More things to speculate on that noone's mentioned yet.
- Modularity? We can already change focussing screens, how about the whole prism (ie, to a sports-finder?). Will than affect Weather-Sealing too much?
- Flip-Screen? Is pretty much out on a pro-body, for ruggedness and sealing. How about a separete plug-in screen? Even a battery-powered screen via the HDMI plug? Pretty much the same as tethered to a laptop, just a smaller screen and more portable.
- Square sensor? You're going to lose a lot of the image circle with a 36mm square sensor. And using one to crop 36x25 and 25x36 images wastes a lot of real-estate around the edges. Very unlikely.


I don't know if someone already mentioned this , since I am commenting on this as I read it.

But the Frames Per Second issue is not the motor in the cameras , it's the speed in which the sensor can dispurse its electrical charge , and prepare for the next shot - the sensor must drain , and refill with each image shot , and the speed of cicuitry is slowed the larger the sensor is , and the more power the CPU uses - so although they are comming up with new faster ways to charge and discharge the sensor , the MP always going up slowly balances it out from giving huge FPS increases. All we can hope for is a new rapid current charge and discharge that does not affect battery , or more seriously , sensor life span.


I did not know that. Very interesting!

I'm curious.. So the expansion pack on the faster film cameras merely provided more power?


I can't be certain with the film camera's - I am all digital myself - in search of the best digital technologies , and theories I can find within each of the manufacturers.

It sounds likely though - although not for the same reason , since film cameras have no sensor -

But I can see the film cameras needing more power to flip up the mirror , advance the film , and actuate the shutter.

In SLR's I'm sure they had to make the mirror retract mechanism stronger , needing a more powerful motor to flip it up each time , and the same with the shutter curtain - it would require a quicker , more accurate actuation as well. Then the film has to advance quick enough to keep up with everything else , while still allowing enough exposure time per frame. Man , people had to be all over high ASA (800+) films back then to keep the camera exposure on par with the shutter speeds lol , I mean if they wanted to keep increasing FPS at that point.

Joseph

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #167 on: October 15, 2011, 03:37:11 AM »
But the Frames Per Second issue is not the motor in the cameras , it's the speed in which the sensor can dispurse its electrical charge , and prepare for the next shot - the sensor must drain , and refill with each image shot , and the speed of cicuitry is slowed the larger the sensor is , and the more power the CPU uses - so although they are comming up with new faster ways to charge and discharge the sensor , the MP always going up slowly balances it out from giving huge FPS increases. All we can hope for is a new rapid current charge and discharge that does not affect battery , or more seriously , sensor life span.
This probably needs a "citation needed" tag :P. Unless I'm mistaken I thought CMOS sensors were effectively a photodiode with an active amplifier, so the "refill" is caused the photoelectric effect and the discharge would consist of removing the small charge held by the inherent capacitance of the photodiode through a transistor. You need a certain number of photons of light to hit the sensor but if you're taking a 1/100s shot there's no reason each inidividual cell can't approach that in terms of fps.

General point is valid though, but it's not so much how quickly you can discharge the sensor it's how long it takes to do something useful with the output of each cell, namely taking an analog to digital conversion. That's largely a function of cost / size / power and then it has to be processed by a CPU and stored to memory fast enough which again is a function of cost / size / power.

I can understand that , but thats why they have the BUFFER - to store the first set of fast paced shots , I dont know if the buffers only 256mb or what , but it can be easily increased to 1GB , and can store the RAW unprocessed data while the CPU catches up - I mean the 1D series have dual processors , and memory (BUFFER) speeds are at rediculous high levels now , exceeding over 2000mhz and processing data anywhere from 20 to 160 gigabytes a second + , NOT saying this is in our cameras , but it could be if they wanted too , especially if this made the FPS count higher , I don't think they are scared to charge more for superior stats - it has to be more than just that ya know , because theres not much stopping them from increasing CPU speeds , lowering memory latencies , and adding more memory in general - it seems these are too simple of problems to be the reason FPS arent at 20fps + right now. SO WHAT IS IT CANON :)

Mark D5 TEAM II

  • 1D X
  • *******
  • Posts: 1177
  • Proud N0ink 0wnz0r / crApple iFruitcake H4t3r
    • View Profile
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #168 on: October 15, 2011, 07:19:06 AM »
Something is really going on...

Today i received an email from CPS Mexico with the subject "Wait for the best for profesional photographers" and this image:



Nice find, thanks for sharing. For numerology/symbology nuts, the above logo would be implying, 3 or 8 ( EOS-3/8?) or 3-layer Foveon-type full-color RGB sensel array sensor (?). :o - edited to satisfy the anal-retentive types...  :P
« Last Edit: October 15, 2011, 08:05:55 AM by Mark D5 TEAM II »
Nikon NSF16 5-Leaf Dual-Blade Industrial Fan ¦ Nikon NTMFI-H1 Dry Heavy Flat Iron ¦ Nikon Tough Mama NTMRC1-2S Rice Cooker “Limited Edition” ¦ Nikon Tough Mama NTMJK18-S Stainless Electric Kettle Plate Type ¦ Nikon NHT 2-in-1 Curling Iron & Straightener in One ¦ Nikon N4004s Decision Master System™

Gothmoth

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #169 on: October 15, 2011, 07:40:20 AM »
[quote author=dr croubie link=topic=1875.msg33677#msg33677 date=But the Frames Per Second issue is not the motor in the cameras , it's the speed in which the sensor can dispurse its electrical charge , and prepare for the next shot - the sensor must drain , and refill with each image shot , and the speed of cicuitry is slowed the larger the sensor is , and the more power the CPU uses - so although they are comming up with new faster ways to charge and discharge the sensor , the MP always going up slowly balances it out from giving huge FPS increases. All we can hope for is a new rapid current charge and discharge that does not affect battery , or more seriously , sensor life span.

i don´t think that´s a problem.  your on the wrong way here.
you see cmos cameras that can do much more then DSLR cameras when it comes to FPS.

12-14 FPS are no problem at all from a charge/discharge viewpoint.


Quote
Nice find, thanks for sharing. For numerology/symbology nuts, the above logo would be implying, 3 or 8 ( EOS-3/8?) or 3-layer Foveon(?).

please stop calling it a foveon sensor.
it´s not a foveon sensor.

foveon is a company who produces a special cmos sensor (foveon x3).
i have already customers who ask if new canon cameras will come with a sigma foveon sensor.   :o

Quote
edited to satisfy the anal-retentive types

well it´s simply wrong calling it a foveon sensor... has nothing to do with counting beans.



« Last Edit: October 15, 2011, 08:45:47 AM by Gothmoth »

neuroanatomist

  • CR GEEK
  • ********
  • Posts: 13594
    • View Profile
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #170 on: October 15, 2011, 11:34:26 AM »
I'm curious.. So the expansion pack on the faster film cameras merely provided more power?
But I can see the film cameras needing more power to flip up the mirror , advance the film , and actuate the shutter.

In SLR's I'm sure they had to make the mirror retract mechanism stronger , needing a more powerful motor to flip it up each time , and the same with the shutter curtain - it would require a quicker , more accurate actuation as well. Then the film has to advance quick enough to keep up with everything else , while still allowing enough exposure time per frame. Man , people had to be all over high ASA (800+) films back then to keep the camera exposure on par with the shutter speeds lol , I mean if they wanted to keep increasing FPS at that point.

I don't see a reason for more power to more the mirror or shutter just because the medium is film vs. digital 9assuming we're talking about a FF digital sensor).  Current cameras may need less power because advances in materials used to construct the mirror and shutter have reduced the weight of those components.  I suspect the main effect of the 'power booster' grips was to drive the film advance motors faster.  But then, Nikon dSLRs get higher frame rates with the battery grip - however, that may not be for technical reasons, but rather for marketing reasons (i.e. to boost sales of the grips). 
EOS 1D X, EOS M, and lots of lenses
______________________________
Flickr | TDP Profile/Gear List


canon rumors FORUM

Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #171 on: October 15, 2011, 12:11:51 PM »

Gothmoth

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #172 on: October 15, 2011, 12:58:39 PM »
can we stop that off topic discussion here?  thanks!

Mt Spokane Photography

  • Canon EF 600mm f/4L IS II
  • ********
  • Posts: 8272
    • View Profile
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #173 on: October 15, 2011, 01:50:38 PM »
Moved to a new topic.

bornshooter

  • 5D Mark III
  • ******
  • Posts: 702
  • Love L series glass
    • View Profile
    • my flickr
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #174 on: October 15, 2011, 02:57:18 PM »
can we stop that off topic discussion here?  thanks!
+ 1

Daveco

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #175 on: October 15, 2011, 03:48:36 PM »
As I recall, the 1nRS had 10 fps speed, and AIservo was not capable at high speed.  I have to manual somewhere, and still own the camera, but I haven't used it in a lot of years.

trailgoer

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #176 on: October 15, 2011, 09:56:04 PM »
Anyone know what time the announcement is scheduled?

I have not seen that but I may have missed it.

mediobarco

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #177 on: October 16, 2011, 12:54:17 AM »
18 mpx, 3 Digit V, 1 of them only for AF. Max fps (12/14) only with the new serie II lenses.

canon rumors FORUM

Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #177 on: October 16, 2011, 12:54:17 AM »

djw

  • Power Shot G16
  • **
  • Posts: 34
    • View Profile
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #178 on: October 16, 2011, 05:56:15 PM »
Amazing--less than 36 hours until the 18th and no further rumour leaks. Either Canon runs a water tight ship or ...

If there is something coming, I would have thought Canon marketing would try to spice things up a bit--c'mon, Canon, that's what makes rumour following a sport!

Picsfor

  • Guest
Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #179 on: October 16, 2011, 06:38:08 PM »
Quick wikipedia-ing: the EOS 1V did 10fps on a moving mirror, the EOS1nRS did 12fps on a pellicle mirror. And the current 1Dmk4 does 10fps on a moving mirror. (and the best I can find out about the F1 is 9fps, although there probably was a motor drive for 14fps that i didn't find).
Surely it's gotta be possible to go above thiese numbers, we've had 11 years of tech since the 1V, there's got to be a more efficient motor and materials capable of the G-forces involved by now. If the mirror can do it, the rest is just software.

I definitely like the idea of the "2 frames per mirror drop" (or 3).
- Firstly, it'll have to be via a custom function, only for those who want it, because it will almost definitely reduce hit-rate (at least in the first incarnation, coding and algorithms will get better over time).
- Second, they drop the mirror in between each shot for AF-tracking. See my post elsewhere quoting canon talking about their new "dream AF in the near future". If this AF really is that good, maybe it can predictively track moving targets with only dropping the mirror after every 2nd or 3rd shot? Or it can process the images it takes quickly and has improved contrast-based AF to help it keep up without dropping the mirror?
Speculation now, of course, and it may not be in this release, but it will be in *something* within a few years, i'd bet. It'll take processing power and good coding, but it's not impossible.

As for the 18MP FF sensor, I just don't believe it one bit. There is absolutely 0% chance they are going to drop the flagship MP-count DOWN, even if it does 1 million ISO as well as my 7D does iso400, it ain't gonna happen.
If it's 18MP, whether it's FF or APS-H, it's got a 1D-label (or 3D, or 6D, just NOT 1Ds or 5D).
18MP could be the "APS-H crop" mode (i'm too lazy to calculate how big an FF sensor would be for that). If they release an 18mp camera now, then they pretty much have to follow nikon's route with the D3/D3x and release a 30+mp 1Ds sooner rather than later, or too many actual pros will jump ship (and i'll be on ebay to get all their 1Ds3s when they do).

Alternatively, it might be an 18MP FF sensor, iff (which reads 'if and only if' to you non-maths nerds), it's a Foveon-type. (but will they brand it as a 54MP sensor a-la Sigma?).

And just because nikon may or may not have a patent on hardware-cropping, doesn't mean canon can't work around it. P&S cameras have had "digital zoom" for years. the 60/600D has digitalzoom/hardwarecropping for video, so some are already getting around it in some way or another.
Does nikon have the patent just for APS-C cropping? canon can get around that by using APS-H cropping, or even their own APS-C cropping (seeing as canon's APS-C is a different size than nikon's APS-C). Depends on how the patent (if it exists) is written as to how easily they can get around it.

More things to speculate on that noone's mentioned yet.
- Modularity? We can already change focussing screens, how about the whole prism (ie, to a sports-finder?). Will than affect Weather-Sealing too much?
- Flip-Screen? Is pretty much out on a pro-body, for ruggedness and sealing. How about a separete plug-in screen? Even a battery-powered screen via the HDMI plug? Pretty much the same as tethered to a laptop, just a smaller screen and more portable.
- Square sensor? You're going to lose a lot of the image circle with a 36mm square sensor. And using one to crop 36x25 and 25x36 images wastes a lot of real-estate around the edges. Very unlikely.


I don't know if someone already mentioned this , since I am commenting on this as I read it.

But the Frames Per Second issue is not the motor in the cameras , it's the speed in which the sensor can dispurse its electrical charge , and prepare for the next shot - the sensor must drain , and refill with each image shot , and the speed of cicuitry is slowed the larger the sensor is , and the more power the CPU uses - so although they are comming up with new faster ways to charge and discharge the sensor , the MP always going up slowly balances it out from giving huge FPS increases. All we can hope for is a new rapid current charge and discharge that does not affect battery , or more seriously , sensor life span.


I did not know that. Very interesting!

I'm curious.. So the expansion pack on the faster film cameras merely provided more power?


I can't be certain with the film camera's - I am all digital myself - in search of the best digital technologies , and theories I can find within each of the manufacturers.

It sounds likely though - although not for the same reason , since film cameras have no sensor -

But I can see the film cameras needing more power to flip up the mirror , advance the film , and actuate the shutter.

In SLR's I'm sure they had to make the mirror retract mechanism stronger , needing a more powerful motor to flip it up each time , and the same with the shutter curtain - it would require a quicker , more accurate actuation as well. Then the film has to advance quick enough to keep up with everything else , while still allowing enough exposure time per frame. Man , people had to be all over high ASA (800+) films back then to keep the camera exposure on par with the shutter speeds lol , I mean if they wanted to keep increasing FPS at that point.


To clarify, the power grip that came with my A1 allowed to wind wind the film on at a rate of 5 frames a second. There was no buffer on film it just had to be set right.

No, asa800 film wasn't needed to use it at that rate either. It was like any other situation, sunny day and you could get away with asa100. Grey miserable day and use asa800...

The problem with 5fps on film was, only 7 seconds of shooting with a 36exp film!

It wasn't the ability to do 10fps, it was the why? 3.5 seconds worth of exposures, all requiring to be developed. You would have needed multiples of bodies with lenses etc to keep up with that rate, and some one to keep changing the film.

64gb cf card will hold nearly 2000 shots from a 5D2 at full RAW mode producing images in the 20-30mb size. No change of film, body or lens required. And if the light changes, just alter the iso!

Yep, digital is much easier than film. And that's not a pop, technology has made it easier...

canon rumors FORUM

Re: New DSLR on Tuesday, October 18, 2011. [CR3]
« Reply #179 on: October 16, 2011, 06:38:08 PM »