Canon EOS R8 specifications

[Exploreshootshare]

I justify the 45MP of the R5 because I photograph nervous animals, birds and insects that can be difficult to approach, so I need to be able to crop fairly heavily sometimes. I could get around that by using a longer lens, but that means more weight and much higher cost.

That'd be a great advantage for me as well, because I do heavily crop. I have images with only 4 MP left from 30mp The reason is I "only" shoot with the RF100-500mm which is limited at times. I used both extenders but I was very unhappy with them only working between 300-500mm. Therefore, I choose to crop.

30MP is a sweet spot and ideal for most subjects.

It really is. File-size is very manageable, it's great for landscape, portrait, people, sports...just not for cropping wild-life.

and this is where in-camera merging will become a major feature for eliminating noise, merging focus-brackets, and hand-held pixel-shift high resolution. You won't have to review 100 shots and pick the best one yourself - the camera will automatically pick the sharpest shots and merge them into a single final image that will be ultra sharp and noise free. But that will require very fast readouts, powerful processors and smaller sensors. M43 is the future.

Sounds like a great new world and finally a way in which MILC might have bigger advantages compared to SP. I actually planed on one more "last" camera because I don't see anything tempting in terms of MP, FPS, DR, ergonomics, but the features you mentioned will probably drive camera sales in 5-10 years and I'll be in line to get the successor of my next camera Or I'll just shoot with my EOS R until then

 

[AlanF]

Some people shooting sports or very fast-moving birds will probably benefit from 20fps or maybe a bit more, but I only need about 8-10fps at most, and usually shoot single frames. Things will change in a few years though - if a camera could shoot half a dozen frames at 100fps there would be virtually no subject or camera movement between shots, and this is where in-camera merging will become a major feature for eliminating noise, merging focus-brackets, and hand-held pixel-shift high resolution. You won't have to review 100 shots and pick the best one yourself - the camera will automatically pick the sharpest shots and merge them into a single final image that will be ultra sharp and noise free. But that will require very fast readouts, powerful processors and smaller sensors. M43 is the future.

If a camera shoots 6 frames in 100fps at say 1/600s each, then merging them will give the same s/n as 1 frame at 1/100s at 1/6 the iso since the same number of photons hit the sensor in both cases. Repeating an exposure 6x increases s/n over a single one by sqrt6 = 2.44x. You won't eliminate noise let alone increase s/n over a single longer shot by that procedure.

 

[entoman]

If a camera shoots 6 frames in 100fps at say 1/600s each, then merging them will give the same s/n as 1 frame at 1/100s at 1/6 the iso since the same number of photons hit the sensor in both cases. Repeating an exposure 6x increases s/n over a single one by sqrt6 = 2.44x. You won't eliminate noise let alone increase s/n over a single longer shot by that procedure.

That's odd, because I've read several times that merging frames does reduce noise. I can't recall the source unfortunately. I get the mathematics, but is there another explanation why merging frames can reduce noise?

Edit: Here is one statement found by a quick google search:

"The stacking technique makes use of the random nature of noise. By making a series of images without moving the camera, every single image will have a slightly different noise pattern. When combining these images in a smart way, it is possible to cancel out the noise, without loss of detail."

Source: fstoppers

 

[nunataks]

That'd be a great advantage for me as well, because I do heavily crop. I have images with only 4 MP left from 30mp The reason is I "only" shoot with the RF100-500mm which is limited at times. I used both extenders but I was very unhappy with them only working between 300-500mm. Therefore, I choose to crop.

It really is. File-size is very manageable, it's great for landscape, portrait, people, sports...just not for cropping wild-life.

Sounds like a great new world and finally a way in which MILC might have bigger advantages compared to SP. I actually planed on one more "last" camera because I don't see anything tempting in terms of MP, FPS, DR, ergonomics, but the features you mentioned will probably drive camera sales in 5-10 years and I'll be in line to get the successor of my next camera Or I'll just shoot with my EOS R until then

I remember when 6 megapixels was “the perfect resolution” hahaha

 

[Exploreshootshare]

I remember when 6 megapixels was “the perfect resolution” hahaha

Heheh that's really a crazy asumption
At the beginning of the century, MP wars started to gain speed and every little bump had a bit of an improvement. But as technology matures, there is a point where "standards" are set. Canon pretty much thinks 24 MP is enough for most use cases. A more extreme example is the car industry where HP/ PS in regular cars (city cars e.g.) has been pretty much the same for nearly 20 years or more now.

 

[AlanF]

That's odd, because I've read several times that merging frames does reduce noise. I can't recall the source unfortunately. I get the mathematics, but is there another explanation why merging frames can reduce noise?

Edit: Here is one statement found by a quick google search:

"The stacking technique makes use of the random nature of noise. By making a series of images without moving the camera, every single image will have a slightly different noise pattern. When combining these images in a smart way, it is possible to cancel out the noise, without loss of detail."

Source: fstoppers

What I wrote is basic 101 statistical physics - it's my day job analysing experimental data. S/n, experimental error etc scale as the square root as the number of times you repeat the measurement - eg to increase accuracy or s/n 10x you need to do 100 measurements. Supposing you do the stacking technique stated by fstoppers of say stacking 100 images of exposure 1 µs each, you would indeed improve s/n tenfold, but you would also get the same s/n in 1 exposure of 100µs.

 

[Del Paso]

I justify the 45MP of the R5 because I photograph nervous animals, birds and insects that can be difficult to approach, so I need to be able to crop fairly heavily sometimes. I could get around that by using a longer lens, but that means more weight and much higher cost. Also it can be difficult to frame fast-moving subjects accurately, so I like to have some space around the subject as a safety margin, and it gives me lots of options to rotate and crop in post. But for my landscapes, fungi and other static subjects, 20MP would be enough. 30MP is a sweet spot and ideal for most subjects.

Some people shooting sports or very fast-moving birds will probably benefit from 20fps or maybe a bit more, but I only need about 8-10fps at most, and usually shoot single frames. Things will change in a few years though - if a camera could shoot half a dozen frames at 100fps there would be virtually no subject or camera movement between shots, and this is where in-camera merging will become a major feature for eliminating noise, merging focus-brackets, and hand-held pixel-shift high resolution. You won't have to review 100 shots and pick the best one yourself - the camera will automatically pick the sharpest shots and merge them into a single final image that will be ultra sharp and noise free. But that will require very fast readouts, powerful processors and smaller sensors. M43 is the future.

Cropping bis indeed the reason why my next camera will be a higher MP one (R5 or R5II).
Yet, for landscapes, I'm pretty happy with my 5 DIV, R or M 240 (30 and 24 MP).
But for macros, I do some cropping (greater distance, more DOF,adequate aperture before difraction sets in).

 

[Del Paso]

That'd be a great advantage for me as well, because I do heavily crop. I have images with only 4 MP left from 30mp The reason is I "only" shoot with the RF100-500mm which is limited at times. I used both extenders but I was very unhappy with them only working between 300-500mm. Therefore, I choose to crop.

It really is. File-size is very manageable, it's great for landscape, portrait, people, sports...just not for cropping wild-life.

Sounds like a great new world and finally a way in which MILC might have bigger advantages compared to SP. I actually planed on one more "last" camera because I don't see anything tempting in terms of MP, FPS, DR, ergonomics, but the features you mentioned will probably drive camera sales in 5-10 years and I'll be in line to get the successor of my next camera Or I'll just shoot with my EOS R until then

There's never a "last" camera.
The last camera is the camera before the next one...

 

[Dragon]

What I wrote is basic 101 statistical physics - it's my day job analysing experimental data. S/n, experimental error etc scale as the square root as the number of times you repeat the measurement - eg to increase accuracy or s/n 10x you need to do 100 measurements. Supposing you do the stacking technique stated by fstoppers of say stacking 100 images of exposure 1 µs each, you would indeed improve s/n tenfold, but you would also get the same s/n in 1 exposure of 100µs.

Agreed if you are light limited, but if there is enough light that both the single shot and and multi-shot are at base ISO and the same shutter speed you will see improvement with the merge. There is also the case where the camera is unsteady and a short shutter is required in either case. In that situation, the merge can register the multilpe images and provide noise improvement. That could be useful for shots from moving platforms such as cars or ships. Phones aleready do this kind of stuff.

 

[entoman]

What I wrote is basic 101 statistical physics - it's my day job analysing experimental data. S/n, experimental error etc scale as the square root as the number of times you repeat the measurement - eg to increase accuracy or s/n 10x you need to do 100 measurements. Supposing you do the stacking technique stated by fstoppers of say stacking 100 images of exposure 1 µs each, you would indeed improve s/n tenfold, but you would also get the same s/n in 1 exposure of 100µs.

Sorry but I'm still confused by the apparent contradictions.

Suppose a scene needs an exposure of 1/100 @ F10 @ ISO 1000. If I'm following you correctly, you are talking about reducing the ISO to 100 and taking 10 shots at 1/100 @ F10, so that the overall amount of light falling on the sensor is the same as in the single shot?

That makes perfect sense to me if I was making an accumulative exposure, but my (possibly incorrect) interpretation of the statements made by fstoppers, dpreview etc is that each of the 10 shots would be taken at 1/100 @ F10 @ ISO 1000, and that software would enable these to be merged and end up with a correctly exposed image that had far less noise than a single shot at 1/100 @ F10 @ ISO 1000.

 

[entoman]

There's never a "last" camera.
The last camera is the camera before the next one...

.... until the money dries up!

 

[neuroanatomist]

@AlanF and @entoman, I suspect you (and DPR) are talking about different methods of ‘merging’ images.

Say you have a scene where you need a 1 s exposure for your chosen aperture and ISO settings.

Merging ten exposures of 1/10 s into a single exposure of 1 s will not yield lower noise than a single 1 s exposure. Mathematically, that’s adding image data and thus you’re adding noise as well.

Merging ten exposures of 1 s into a single exposure of 1 s will yield less noise than a single exposure of 1 s. Mathematically, that’s averaging image data and random image noise will be reduced.

 

[entoman]

@AlanF and @entoman, I suspect you (and DPR) are talking about different methods of ‘merging’ images.

Say you have a scene where you need a 1 s exposure for your chosen aperture and ISO settings.

Merging ten exposures of 1/10 s into a single exposure of 1 s will not yield lower noise than a single 1 s exposure. Mathematically, that’s adding image data and thus you’re adding noise as well.

Merging ten exposures of 1 s into a single exposure of 1 s will yield less noise than a single exposure of 1 s. Mathematically, that’s averaging image data and random image noise will be reduced.

Thanks for simplifying it neuro. I'm pretty sure that what fstoppers, dpreview etc were talking about is what you refer to in your last sentence, and that is the type of merging that I anticipate being used in future cameras. As @Dragon points out, it already happens with phones. With conventional cameras such tech will almost certainly appear first in M43 and APS-C.

 

[Alastair Norcross]

8 FPS but with full AF adjustment just about 3fps… it’s very slow indeed

Actually, the R shoots at 5fps with full AF adjustment. They misleadingly call the 3fps speed "AF priority" and the 5fps speed "shooting priority" (or something like that), but in fact, there's no difference in servo AF accuracy between the 3fps and the 5fps rate. This is from extensive experience shooting at both speeds. The 6D shot at 4.5fps, so the R isn't slower than that. It's slightly faster. I find that 5fps is actually fast enough for everything except fast sports, and even then I get plenty of great shots at 5fps. I remember when I upgraded from the 300D (2.5fps for a full 4 frame buffer) to the the 20D (5fps for a buffer somewhere in the teens). The 20D felt insanely fast!

 

[AlanF]

@AlanF and @entoman, I suspect you (and DPR) are talking about different methods of ‘merging’ images.

Say you have a scene where you need a 1 s exposure for your chosen aperture and ISO settings.

Merging ten exposures of 1/10 s into a single exposure of 1 s will not yield lower noise than a single 1 s exposure. Mathematically, that’s adding image data and thus you’re adding noise as well.

Merging ten exposures of 1 s into a single exposure of 1 s will yield less noise than a single exposure of 1 s. Mathematically, that’s averaging image data and random image noise will be reduced.

What I commented on was:

- if a camera could shoot half a dozen frames at 100fps there would be virtually no subject or camera movement between shots, and this is where in-camera merging will become a major feature for eliminating noise, merging focus-brackets, and hand-held pixel-shift high resolution. You won't have to review 100 shots and pick the best one yourself - the camera will automatically pick the sharpest shots and merge them into a single final image that will be ultra sharp and noise free. But that will require very fast readouts, powerful processors and smaller sensors. M43 is the future.

a) half a dozen shots can reduce noise by no more than 60%, and not eliminate noise and be noise free.
b) half a dozen rapid shots will in any case have no less noise than one shot encompassing the same time period.

That example by entoman is not an example of the "the merging ten exposures of 1 s into a single exposure of 1 s", but refers to a series of 6 very rapid short time shots at 100 fps, when one could take one slower shot at lower fps.

 

[cerealito]

Man! I can live without IBIS but please give us a real battery not that shitty LP-e17. What's the point of a lighter body if we have to carry additional batteries and a charger?
I just love how I can take my 70d out for a whole week and leave the charger at home. I know a fully charged battery will be enough. Could I do the same with the RP? with the R8?

 

[AlanF]

Agreed if you are light limited, but if there is enough light that both the single shot and and multi-shot are at base ISO and the same shutter speed you will see improvement with the merge. There is also the case where the camera is unsteady and a short shutter is required in either case. In that situation, the merge can register the multilpe images and provide noise improvement. That could be useful for shots from moving platforms such as cars or ships. Phones aleready do this kind of stuff.

He wasn’t talking about the case of unsteady camera and had eliminated that situation by saying that the rapid acquisition of shots would mean that there would be virtually no movement between frames.

 

[Dragon]

cin

hat yo

Man! I can live without IBIS but please give us a real battery not that shitty LP-e17. What's the point of a lighter body if we have to carry additional batteries and a charger?
I just love how I can take my 70d out for a whole week and leave the charger at home. I know a fully charged battery will be enough. Could I do the same with the RP? with the R8?

Battery life is the price to be paid for being able to see your subject in the VF when you can't see it with your naked eye . Whether the manufacturer puts a huge battery in the cameara or obliges you to put a few extras in your pocket is an ergonomics choice that varies with the intended purpose of the camera. Just be thankful that camera manufacturers haven't gone down the phone road with non-replaceable lithium-polymer batteries (one of key reasons I still carry a flip phone). Maybe we will see an LP17NH . That would be a nice consolation prize for M owners.

 

[SHAMwow]

The 5DIV is a solid workhorse for my needs and is still producing images I'm very happy with. The only thing I really want to improve upon against the 5DIV is resolution, and I'm content to wait and see if Canon releases something with more resolution than the R5 with an R mount. Eventually I'll get impatient and maybe end up buying an R5 or an R5 II, but since I'm not exactly in a rush or being really heavily limited at the moment, I don't see an urgent need to make the move.

Oh okay. Your initial post in read as in you thought there wasn't a suitable replacement. I'd say the 45 is enough on the R5 especially if you're saying that your 5DIV is your workhorse. My R5 files already eat up space as is.

 

[adrian_bacon]

Looks like the R8 is potentially the replacement for the RP. Not directly, but in much the same way the RP was a low cost entry level full frame camera, it looks like the R8 is that. Should be interesting to see if the R8 really does have the same/similar sensor as the R6II, or maybe a revised R sensor. Who knows until it\'s actually announced.

I\'m pretty happy with the R6II sensor, and would be happy to see the R8 use the same sensor and just be cost reduced through the removal of IBIS, and other features to get the form factor and size down.