A high megapixel camera is coming [CR2]

tron · Feb 23, 2022

unfocused said:
I can only speak to my personal experience. R5 with Raw in slot 1 and jpg in slot 2 as backup. If I'm shooting a basketball player running down the court and doing a layup, the R5 buffer will often fill before he or she reaches the basket (shooting mechanical shutter on the R5 High Speed +). Then, the camera is frozen while waiting for the buffer to clear (you can shoot once the buffer gets to a certain point, but it will immediately fill up again if it isn't completely cleared.

With the R3, you can usually complete the play before the buffer fills, but I've still had the occasional problem where it doesn't clear in time for the next play. Using the electronic shutter High Speed +.

I don't know how Brian conducted his tests, so I can't speak to them. I suppose it isn't the size of the buffer, but the speed at which it clears that is the problem.

As I've said many times before, it's never been a problem with birds in flight, but in sports, where one play can quickly follow another, it can present a challenge to capture that second play.

hello, regarding R5 what CFExpress card are you using exactly? I ask because for some brands the lower capacity CFExrpress cards are slower (for example 64,128,256 vs 512gb Sandisk). Of course camera's controller must be equally fast (I remember how slow 5DMk3 SD controller was!)

The compressed raw selection is a nice trick. I remembered I had set it so in my 90D where I had similar issues (and a UHS-II does not offer a big advantage). I had set it but it was set it and forget it ... literally and I remembered it thanks to your reference. Also it is not being used that much these days.

unfocused · Feb 23, 2022

David - Sydney said:
How many shots are you taking in a basketball game?
It is great that Canon offers a cRaw option at full res with very little shadow loss (as far as I can see in the raw files and read online tests).
I never understood the calls for small/medium/large raw option from the current 5Ds/r users vs cRaw

Too many.

Of course it is easy to get hooked on the high frame rate and I probably should shoot at a lower frame rate, but it's hard to do that when you know you might be missing a shot. (Volleyballs, Soccer Balls, Basketballs, Baseballs, Golf Balls all go really fast and if the ball isn't in the frame, you got nothing most times)

Honestly, I cannot see any difference between CRaw and Full Raw.

unfocused · Feb 23, 2022

tron said:
hello, regarding R5 what CFExpress card are you using exactly? I ask because for some brands the lower capacity CFExrpress cards are slower (for example 64,128,256 vs 512gb Sandisk). Of course camera's controller must be equally fast (I remember how slow 5DMk3 SD controller was!)

The compressed raw selection is a nice trick. I remembered I had set it so in my 90D where I had similar issues (and a UHS-II does not offer a big advantage). I had set it but it was set it and forget it ... literally and I remembered it thanks to your reference. Also it is not being used that much these days.

That's interesting. I never noticed that. Mostly I am using 256gb Sandisk, with some 128 Sandisk and ProGrade Gold thrown in.

I have one 512gb card, but I don't really like to use it because a weird thing happens -- with CRaw the number of files that the card can hold from the R3 exceeds the R3 default file numbering and Canon does a weird thing, where when it starts over and adds a "-1" to the file name. The result is that when you download the files, they are no longer in the shot order and even more frustrating, they alternate from one to another so that you might have 1,000 files where, say, a sequence has a single men's basketball game shot, then a single women's basketball game shot, then a men's, then a woman's. Royal pain in the butt to get them back into the proper order.

adrian_bacon · Feb 23, 2022

I'd love to see them take the pixel density of the 90D and go full frame with it with stacked BSI tech. That'd net ~11124x7416 px, or ~82.5 MP. Of course, just for video scaling ease and heat management, they likely go something more like 10560x7040 (~75MP) with a very slight crop down to 9600x5400 for UHD video and scale that down to either 8K or 4K/2K, which would literally be a straight 5x reduction for 1080p video. And for super 35 sized crop would give a solid 4K readout at 1:1. For DCI, they could either crop just a little more for native 1:1 pixel read out for DCI 8K, or have almost no crop at 10240x5760 and scale that down to the smaller DCI standards.

I know, wishful thinking on my part, but still... now that it seems Canon has woken up to designing their sensors to make it easier to do oversampled and scaled down video, it'd behoove them to set the native pixel dimensions to even multiples of video resolutions just so scaling or binning is easier and less computationally expensive. Part of heat management is not having to burn CPU cycles doing unnecessary math.

EOS 4 Life · Feb 23, 2022

landscaper said:
Sony will probably leap- Frog Canon with their 100 megapixel sensor that has Already been developed in 2019

Sony Made a 100MP and 6K Full-Frame Sensor for Consumer Cameras

It seems the race to launch a 100-megapixel full-frame sensor is on. Sony has reportedly developed a new 35mm full-frame CMOS sensor that shoots

petapixel.com

Unlike Canon, Sony does not only make sensors for themselves.
Binning 100 MP down to 6K would be like the a1 binning 50 MP down to 6K for oversampled 4K.

EOS 4 Life · Feb 23, 2022

dtaylor said:
We need a high MP FAQ at the head of every rumor about a high MP body

Lenses/sensors do not "out resolve" each other. Improving either will improve the final result. There are of course diminishing returns. But as demonstrated by Bob Atkins in his 5Ds review, even the worst lens will produce a better result on a higher MP sensor.

High ISO noise performance is dominated by sensor size, not pixel size. A high MP RF body is very likely to have the same high ISO performance as the R5.

Pixel size has not been strongly correlated with base ISO DR for years. It's likely a high MP RF body will have at least as much DR as the R5.

There is a bit of truth to motion blur claims. A high MP sensor can sometimes pickup blur that a lower MP sensor does not. But we're talking about 0.5-1 stop faster shutter speeds to compensate. No, a high MP RF body will not be stuck on a tripod.

There is more too it than just megapixels but modern sensors can pick up aberrations in lenses that older sensors were not able to.

koenkooi · Feb 23, 2022

unfocused said:
That's interesting. I never noticed that. Mostly I am using 256gb Sandisk, with some 128 Sandisk and ProGrade Gold thrown in.

I have one 512gb card, but I don't really like to use it because a weird thing happens -- with CRaw the number of files that the card can hold from the R3 exceeds the R3 default file numbering and Canon does a weird thing, where when it starts over and adds a "-1" to the file name. The result is that when you download the files, they are no longer in the shot order and even more frustrating, they alternate from one to another so that you might have 1,000 files where, say, a sequence has a single men's basketball game shot, then a single women's basketball game shot, then a men's, then a woman's. Royal pain in the butt to get them back into the proper order.

After having 2 cameras that hit the same numbers at the same time, I have exiftool rename files on import to date-camera-lens-originalname. And in the viewer I sort them based on capture date, DPP4 is really bad at it, LR and DXO pretty good.

Depending on how you download the files, the filesystem can keep the modification date when copying. That only works if you’re using a card reader, though.

NKD · Feb 23, 2022

Sure a few have allready posted.
My wishlist hahahaa
- R3 body
- What ever dynamic range - id probably be happy with same as R5 15stops (coming form 5dsR)
- Cancel AA filter for extreme sharpening in post!
- Flawless firmware on release
- Good battery life (LPN6's dual even better!)
- Illuminated rear buttons
- Good spot for pyical tether cord for vertical and horiz mounting via arca swiss (RRS plates)
- Wireless tether
- GPS
- Hybrid EVF ith electronic overlay
- 3fps is fine - Although 10 would be nice (even with crop)
- 4k 60fps or 8k 30fps

Wishfull thinking.
- God knows even a new filtype? CR4...The sensor can just capture light and in editing not overepose

- Dont release another susseor for another 8 years - This gunna be a huge investment with new RF glass

ColinJR · Feb 23, 2022

keithcooper said:
I appreciate it's not something many want, but I've never stepped up in resolution with a new camera and thought 'that's too much'

Exactly. People say they don't need it but I'd bet those same people would not be willing to go down in resolution from what they currently use.

adrian_bacon · Feb 23, 2022

SwissFrank said:
GENERATING 1/4 the pixels by averaging four neighboring pixels might be hard, though, I don't know.

Don't average them, just lower the ADC bit depth and sum them. Doing 12 bit ADC and summing the two green channels (RGGB to RGB) would net 12 bit red and blue and 13 bit green. Taking a 2x2 RGB and halving it again (for quarter resolution) would net 14 bits red and blue, and 15 bits green. 12 bit ADC gives faster readout too. Half size would give slightly smaller files, but quarter size would give much smaller files, though either would no longer be bayer CFA but rather just full raw RGB.

EOS 4 Life · Feb 23, 2022

NKD said:
- Flawless firmware on release

This is the least likely part of your list

EOS 4 Life · Feb 23, 2022

ColinJR said:
Exactly. People say they don't need it but I'd bet those same people would not be willing to go down in resolution from what they currently use.

Plenty of people went from the EOS R to the R6

ColinJR · Feb 23, 2022

EOS 4 Life said:
Plenty of people went for the EOS R to the R6

I wonder what the stat is on that. As much of a "beta" camera the R is compared to the R6, I went for the R5 (from the R) because I didn't want to take a resolution hit... I find too much value in cropping and enjoy the look of the higher res sensor.

adrian_bacon · Feb 23, 2022

EOS 4 Life said:
Plenty of people went for the EOS R to the R6

I actually went even more drastic. I went from an R5 and R6 (primary and backup) to two R6s about 3-4 months ago. I still have the R5, but it's used in a dedicated copy stand setup for something else and not for general purpose shooting like my two R6s. I know a lot of wedding/event photographers that went from 5DM4 or the original R to the R6. Do I wish the R6 had more resolution? Heck yeah, but at the same time, I primarily shoot portraits and headshots and weddings and events and most of my work is either printed at book size or used directly online as profile photos, and frankly, if your output is 240-300 pixels per inch on paper (or even smaller for screen), the R6 is more than enough resolution. I have grips on both my R6s and would love to do 2 R3s, but they're just too expensive, so two R6s is where it's at for me for at least the next couple of years. If I absolutely have to have the resolution, I can take the R5 off the copy stand and use it, but in all reality, the R6 is plenty of resolution for my type of work, and for stills photography is basically a mirrorless 1DXMIII with two fast SD cards. I used to have an original R (before the R5) and still use an RP for my personal camera and in all honesty, from a resolution standpoint, I'm pretty hard pressed to see sharpness differences between them without looking at the metadata to see which is which. I'm sure if I take exactly the same shot with the same lens but different body and compare them I'd see the differences, if I pixel peeked, but outside of that. Nope. They all look sharp and I'm rarely cropping in so far that my final output isn't still scaled down.

David_D · Feb 23, 2022

AlanF said:
If you are using the same focal length lens and viewing at the same metric size, then: minor focus errors; subject movement; camera-shake; high ISO noise; and diffraction-related softness should all be near enough identical. Fine detail should be better for the high resolution sensor, but perceived edge detail may artefactually look lower because smooth transitions may lose their texture and look sharper on the lower Mpx sensor. If there are lens aberrations, high noise, excessive diffraction, movement, bad focus etc the difference in fine detail will decrease, but the high resolution sensor will never give poorer images than the lower.

I completely agree, if looking at the RAW files, but given there is 4 times the information (although potentially a little noisier) that will give the image processing algorithms more to work with, so that the end result should appear much better with the higher Mpx sensor?

koenkooi · Feb 23, 2022

David_D said:
I completely agree, if looking at the RAW files, but given there is 4 times the information (although potentially a little noisier) that will give the image processing algorithms more to work with, so that the end result should appear much better with the higher Mpx sensor?

It depends on the size you are using to view it. On an 8x10 print at arms length you would see less difference, if any, compared to viewing it full screen on a 27" monitor.

Generally more resolution means better results, which might not show up in the end product. People here like to use the 800x800px Instagram post as an example for that

Exploreshootshare · Feb 23, 2022

Interesting rumor.
77mp sounds like a good compromise between high pixel, noise handling etc. and a great resolution. It seems like Canon once again chose to produce files which can be handled with ease (like 24MP for R3) rather than follow the call for extreme pixel peeping.

On the other hand, I guess they'll lose the "high-mega pixel" to Sony shortly after the release of the high MP R5s (or similiar) because Sony will easily manage to install an 80+ MP sensor in the 7RV.

Exploreshootshare · Feb 23, 2022

Genuine question:
Would a 77MP be a good astro camera or not? I imagine there'll be a ton of crazy detail on a tracked Milky Way shot. But I fear micro shocks would easily make the image loose sharpness. And what about low-light capabilities? Would they be much worse than R6 level specs?

cayenne · Feb 23, 2022

kaihp said:
As I recall, the world-wide market for MF bodies are in the few thousands. 5-6K maybe.
@neuroanatomist will surely remember the number off the top of his head.

I sincerely doubt that Canon has any interest in going into such a small market when they would have to make both bodies and lenses for an MF format. That is, unless they use an existing bayonet system. Which would seem very unlike Canon to be.

With all these "this ain't Canon" statements, surely someone is going to point out that Canon did do something like XXX at some point in time

Like I said, I'm not holding my breath for Canon to do it.

However, with the popularity of the Fuji GFX system, especially with the release of the 100MP GFX100S, at the fairly low price of $6K, I would have to imagine that the uptake is well north of 5000-6000 units sold.

I believe they are also revamping the 50MP digital MF GFX cameras, which will be even lower in price.

No, they aren't high FPS shooters, but that's not expected from medium format....you do gain in other areas significantly.

But again, I was just kinda wish-listing for Canon to look into digital MF.

I'm just wondering if they won't soon hit the limit of trying to squeeze more MP into a FF sensor....at some point, you run into problems with pixel density, size of pixels, etc.

At some point, you would gain more by going with a larger sensor size.

Hell, while I'm wishing, I wish someone would come with a true panoramic sensor....something you could take native XPan type images with.
I'd be all over that in digital, or even if someone put out a fully mechanical new 35mm film camera version of it.

I'd hock a kidney or two for that.

I guess I might as well throw in a wish for a pony too.

C

Fischer · Feb 23, 2022

dtaylor said:
This isn't entirely true. A higher MP camera can pick up motion blur that a lower MP camera does not. It may be an optical phenomenon, but the motion is across a pixel grid. It is possible for a finer grid to detect a moving line across two pixels where on a coarser grid the moving line never leaves the bounds of one.

And if the higher MP sensor does pickup motion blur, you can see it at the same view size just like you can see increased sharpness and fine detail in a non-blurred shot. This depends of course on the view size and how well the medium preserves the full resolution. It's a narrow window between both cameras showing blur and only one showing blur. Or perhaps neither showing it at the intended view size. But that window does exist. I can't reliably hand hold a 5Ds to the same shutter speeds as a 5D mark III.

That said, people make way too much of this. I shoot my 5Ds at the same shutter speeds I used to shoot my 7D at. So the higher resolution costs me...a half stop? Maybe a 75mp camera would cost a full stop to be on the safe side? How much do we gain from IBIS? From improved OIS in lenses? When the 5Ds and 5DsR came out there were people who acted like they had to be tripod mounted at all times, which is complete nonsense even without IS lenses. A slight increase in the possibility of capturing some motion blur is no reason to avoid a higher MP body.

As to AF: Canon did take steps to improve accuracy on the 5Ds/5DsR. And it worked. But more to the point, this new body will use dual or quad pixel AF. Focusing will be at the sensor with millions of points analyzing the scene. I doubt AF accuracy will be an issue.

Oh please... I am not continuing this.

A high megapixel camera is coming [CR2]

EOS-1D X Mark III

Photos/Photo Book Reviews: www.thecuriouseye.com

Photos/Photo Book Reviews: www.thecuriouseye.com

EOS 90D

EOS 5D Mark IV

EOS 5D Mark IV

EOS 5D Mark IV

EOS M50

EOS R5

EOS 90D

EOS 5D Mark IV

EOS 5D Mark IV

EOS R5

EOS 90D

EOS M6 Mark II

EOS 5D Mark IV

EOS RP

EOS RP

EOS R6

EOS RP

Similar threads