The Canon EOS R3 will eliminate “Lag” and “Blackout” for stills shooters
- Thread starter Canon Rumors Guy
- Start date
The card won’t get too hot if you remove it after writing the files. If you leave the card reader connected after you’re done with it it continues to heat up. Don’t ask me why. I unplug it right after and it’s still cool.
Upvote
0
Until you try to lock focus in the corners of the frames lolSo maybe using a mirror and prism to see through the lens wasn’t such a bad idea after all……
Upvote
0
It's not physically possible for a mirrorless system to actually fully eliminate lag. Even SLRs can't fully, 100% eliminate lag, and they have a much faster and simpler pipeline to work with.
With any mirrorless system the light has to come in through the lens and hit the sensor, go through the ADC, be configured into RGB channels as dictated by the currently selected martix, encoded into a reference image using the currently selected colour profile, that compilation is then used to evaluate white balance, exposure and focus, those exposure and focus systems then have to adjust in whatever way the current shooting settings dictate, now the whole cycle so far is repeated again and compared to the first cycle, then assuming there are no more exposure or focus adjustments to be made (if there are the cycle will repeat), the currently-formed image is duplicated, with the buffer storing one copy for reference while another is resampled to the resolution of the viewfinder, which holds the image until the next screen refresh and then, finally, displays it in the viewfinder. Then as with any other design, your eye has to actually see the image, your finger has to press the shutter, the CPU has to recieve that capture command, the aperture has ot be stopped down to whatever the chosen setting is, then the shutter has to activate, the captured frame goes through that whole pipeline again, and the image is sent into the buffer to be saved to a card.
Okay, deep breath.
Did you get all that? Yeah, it's a lot. At least with an SLR the image doesn't have to be duplicated, rescaled and held for a screen refresh, since it just bounces up through the OVF at literally the speed of light. Even then, the fastest SLRs (1D & D6) have lag of around 35ms, only with the mirror locked up and when activated automatically (i.e. no human reflexes to slow them down). The least-laggy mirrorless camera right now is the Sony a1 at around 90ms with a mechanical shutter or 55ms (+/-) with the electronic shutter. (Canon claim the R5 can do 50ms electronic, but that's only with manual focus, manual exposure, the aperture left fully open and shooting a perfectly-lit medium grey card; I don't count that because it's simply such a totally unrealistic scenario.)
Physically, if we're charitable and allow the minimal electronic travel lag of 5ms, the least laggy an EVF can ever be is a delay of about 9ms, which is right on the edge of what the healthiest 20-something human can respond to anyway. That's if it has a refresh rate of 240Hz, which currently no OLED screen manufacturered for viewfinders achieves. 120hz is the current maximum manufactured, so the physical minimum lag currently is 13ms, which is a delay that most of the world's population between age 12 and 50 can perceive. This also assumes there's only a single frame refresh sync and no delay for anti-flicker systems, etc. No camera manufacturer wants to risk screen tearing, so the actual refresh sync delay is always two or three frames, bumping us up to a bare physical minimum of 21ms. And then, again, there is the shutter activation on top.
In camera shutter speed terms that's 1/48th. So if you think of the period of time your camera captures when set to a shuter of 1/50th, that's the period of time that is the physical minimum a camera can cut its total lag down to.
And, again, that's being extremely unrealistically charitable and minimising all that processing down to purely the electronic travel time, when in reality all that processing comes up as more like 25ms or so. Add the shutter activation, which is again a bare physical minimum of 5ms and more often closer to 20ms alone, and we're easily up to around the 50-60ms ballpark that the a1 currently maxes out at.
Further more, eliminating blackout is done by delaying the viewfinder by one or more frames, increasing visual lag. You can't get rid of blackout without increasing overall system lag; it's not physically possible.
So if we're really, really nice to Canon and pretend they've made some miracle breakthrough which allows them to either skip or simulataniously process multiple steps and they've somehow come up with a brand new EVF which refreshes at 240Hz even though they don't manufacturer EVFs and nobody that does makes one of that spec... even if we're that charitable to Canon, it's still not possible to eliminate lag in a mirrorless camera. Not lag that most people between 12 and 50 can perceive. There are simply too many parts which have to sync up, too much processing, and physically too far for the initial converted light signal to travel, to actually eliminate lag. At best—at best—they might be able to get the lag down to match the 1D X mk III. Beyond that, physics simply does not allow much improvement and the sheer nature of using an EVF and the camera having top process and display that video feed automatically enforces an absolute bare minimum of lag.
Now let's add in auto ISO or WB evalutation, the aperture stopping down, tracking focus adjustments...
Their marketing may claim otherwise, but you can't cheat physics.
With any mirrorless system the light has to come in through the lens and hit the sensor, go through the ADC, be configured into RGB channels as dictated by the currently selected martix, encoded into a reference image using the currently selected colour profile, that compilation is then used to evaluate white balance, exposure and focus, those exposure and focus systems then have to adjust in whatever way the current shooting settings dictate, now the whole cycle so far is repeated again and compared to the first cycle, then assuming there are no more exposure or focus adjustments to be made (if there are the cycle will repeat), the currently-formed image is duplicated, with the buffer storing one copy for reference while another is resampled to the resolution of the viewfinder, which holds the image until the next screen refresh and then, finally, displays it in the viewfinder. Then as with any other design, your eye has to actually see the image, your finger has to press the shutter, the CPU has to recieve that capture command, the aperture has ot be stopped down to whatever the chosen setting is, then the shutter has to activate, the captured frame goes through that whole pipeline again, and the image is sent into the buffer to be saved to a card.
Okay, deep breath.
Did you get all that? Yeah, it's a lot. At least with an SLR the image doesn't have to be duplicated, rescaled and held for a screen refresh, since it just bounces up through the OVF at literally the speed of light. Even then, the fastest SLRs (1D & D6) have lag of around 35ms, only with the mirror locked up and when activated automatically (i.e. no human reflexes to slow them down). The least-laggy mirrorless camera right now is the Sony a1 at around 90ms with a mechanical shutter or 55ms (+/-) with the electronic shutter. (Canon claim the R5 can do 50ms electronic, but that's only with manual focus, manual exposure, the aperture left fully open and shooting a perfectly-lit medium grey card; I don't count that because it's simply such a totally unrealistic scenario.)
Physically, if we're charitable and allow the minimal electronic travel lag of 5ms, the least laggy an EVF can ever be is a delay of about 9ms, which is right on the edge of what the healthiest 20-something human can respond to anyway. That's if it has a refresh rate of 240Hz, which currently no OLED screen manufacturered for viewfinders achieves. 120hz is the current maximum manufactured, so the physical minimum lag currently is 13ms, which is a delay that most of the world's population between age 12 and 50 can perceive. This also assumes there's only a single frame refresh sync and no delay for anti-flicker systems, etc. No camera manufacturer wants to risk screen tearing, so the actual refresh sync delay is always two or three frames, bumping us up to a bare physical minimum of 21ms. And then, again, there is the shutter activation on top.
In camera shutter speed terms that's 1/48th. So if you think of the period of time your camera captures when set to a shuter of 1/50th, that's the period of time that is the physical minimum a camera can cut its total lag down to.
And, again, that's being extremely unrealistically charitable and minimising all that processing down to purely the electronic travel time, when in reality all that processing comes up as more like 25ms or so. Add the shutter activation, which is again a bare physical minimum of 5ms and more often closer to 20ms alone, and we're easily up to around the 50-60ms ballpark that the a1 currently maxes out at.
Further more, eliminating blackout is done by delaying the viewfinder by one or more frames, increasing visual lag. You can't get rid of blackout without increasing overall system lag; it's not physically possible.
So if we're really, really nice to Canon and pretend they've made some miracle breakthrough which allows them to either skip or simulataniously process multiple steps and they've somehow come up with a brand new EVF which refreshes at 240Hz even though they don't manufacturer EVFs and nobody that does makes one of that spec... even if we're that charitable to Canon, it's still not possible to eliminate lag in a mirrorless camera. Not lag that most people between 12 and 50 can perceive. There are simply too many parts which have to sync up, too much processing, and physically too far for the initial converted light signal to travel, to actually eliminate lag. At best—at best—they might be able to get the lag down to match the 1D X mk III. Beyond that, physics simply does not allow much improvement and the sheer nature of using an EVF and the camera having top process and display that video feed automatically enforces an absolute bare minimum of lag.
Now let's add in auto ISO or WB evalutation, the aperture stopping down, tracking focus adjustments...
Their marketing may claim otherwise, but you can't cheat physics.
Upvote
0
There's no actual physical reason why an SLR couldn't have the same focus point coverage and sensitivity that mirrorless does. It's just a bit harder, more expensive, and all-round more effort to produce. Physically it's possible; manufacturers just couldn't be bothered to do it. (See also: IBIS, f/11 primes, f/2 zooms, etc.)
Upvote
0
If you have a Canon DSLR with a transmissive LCD (7D, 5DIII, 70D or newer), try an experiment: look through the VF with the camera in sleep mode, or even powered off. Then, remove the battery and look through the VF again. (Spoiler: there’s a difference, and the only reason there would be a difference is if power is being drawn from the battery even in sleep mode or with the power switch set to off.)
Upvote
0
What makes you think the R1 will be high mp ? The R1 is going to be aimed at Pro sports shooters just like the 1DXiii and they have no need for very high resolution just high speed , dependability and efficient workflow.
Canon have stated that the R3 is not their replacement for the 1DXiii so the R1 probably is.
There's a rumoured high mp R5 variant which seems much more likely
Last edited:
Upvote
0
Upvote
0
No mention of GPS, somewhere the eye-focus is not mentioned anymore (although those who do have the camera talk about it), and no mnetion of how fast the AF tracking is (especially for fast moving objects like little birds).
Upvote
0
Last week, I shot an outdoor theatre performance on an RP, with 200mm and 85mm primes, and lemme tell you . . . I have never worked harder, and I have never been more irritated to not have a 100-400 or 70-200 zoom + a good DSLR with which to capture the action --- especially having no lag, a la an OVF.
The RP was incredibly slow and hard to use, and in that moment, I would have given my left nut for a professional camera and a great zoom. I still got enough good shots that my client was happy, but I was undeniably stressed while doing it.
The RP was incredibly slow and hard to use, and in that moment, I would have given my left nut for a professional camera and a great zoom. I still got enough good shots that my client was happy, but I was undeniably stressed while doing it.
Upvote
0
I have definitely noticed the battery life shortcomings of the R5. Whenever I used to read about the battery life on the 5D or the original R, I couldn't understand why the reviewers or even Canon's stated battery life was so far below what I was getting. I could shoot all day with one battery, over a 1000 frames no problem. Now, I shoot for an hour and the battery is already half drained. It's definitely the worst battery life of any Canon I've owned, and this is the 6th DSLR I've had that used this kind of battery.
Upvote
0
Similar threads
- Contains 1 staff post(s)
