Are two cameras going to replace the Canon EOS R5? [CR]

[amorse]

Well, even for landscapes, I doubt many non professionals will NEED the 80 MP.
But, for heavy cropping, a tempting option.

I don't know - I see "need" as a pretty blurry line no-matter how we look at it. I don't disagree with you, but to play devil's advocate, is 40 fps bursts needed in a world where we have 20 fps bursts? I'd assume having 40 fps increases the likelihood of getting the perfect moment in a burst over a 20fps camera. I don't see that as being incredibly different to how 80mp (or insert whatever resolution here) may give some more cropping options than at 45, or allow detail that wouldn't otherwise be available for that unplanned "once in a lifetime" photo.

Need is a pretty hard bar to pass for a lot of advanced features given that people have been making a living shooting everything under the sun without those features for decades. At the end of the day, I think it just boils down to Canon deciding if enough people are willing to pay for x feature to make it worth their time to implement - whether a feature is really needed seems secondary to that.

 

[entoman]

I don't know - I see "need" as a pretty blurry line no-matter how we look at it. I don't disagree with you, but to play devil's advocate, is 40 fps bursts needed in a world where we have 20 fps bursts? I'd assume having 40 fps increases the likelihood of getting the perfect moment in a burst over a 20fps camera. I don't see that as being incredibly different to how 80mp (or insert whatever resolution here) may give some more cropping options than at 45, or allow detail that wouldn't otherwise be available for that unplanned "once in a lifetime" photo.

Need is a pretty hard bar to pass for a lot of advanced features given that people have been making a living shooting everything under the sun without those features for decades. At the end of the day, I think it just boils down to Canon deciding if enough people are willing to pay for x feature to make it worth their time to implement - whether a feature is really needed seems secondary to that.

Yep. Professionals have needs, hobbyists have desires. Most of us only use a fraction of the features that modern cameras provide, but we like to have them anyway! It's also nice to know that if at some stage in the future we want to try something different (e.g. focus stacking, pixel-shift, time-lapse, 8K etc), that those features are already in the camera, even if we don't need or want them right now.

As for MP, higher and higher resolutions require better and better stabilisation for handheld work, and better technique in general to get the most out of them. Also, high MP needs more storage space and slows down post processing. Lots to be taken into account when considering a high MP camera.

 

[entoman]

I know it's off-topic, but it's interesting to talk about AI and photography.

It won't be off-topic for long, it will very soon become an integral part of all photography. As long as we get the choice of whether to use it or not, it's fine with me. I've been so impressed with AI denoising, that I'm re-editing my entire photo collection!

 

[amorse]

Yep. Professionals have needs, hobbyists have desires. Most of us only use a fraction of the features that modern cameras provide, but we like to have them anyway! It's also nice to know that if at some stage in the future we want to try something different (e.g. focus stacking, pixel-shift, time-lapse, 8K etc), that those features are already in the camera, even if we don't need or want them right now.

As for MP, higher and higher resolutions require better and better stabilisation for handheld work, and better technique in general to get the most out of them. Also, high MP needs more storage space and slows down post processing. Lots to be taken into account when considering a high MP camera.

There’s certainly trade offs in using higher resolution bodies, but depending on the use case they’re manageable. For instance - I’m after a higher resolution option than the R5 and none of those limitations give me anxiety. Most of my images are on a tripod without the stabilizer on (probably 95% without autofocus on frankly), I shoot very few images so the over all storage load is low, I’m already using adobe’s AI upscale occasionally on the 5Div giving 120mp images and my computer system manages it fine. With that said, I’ve noticed that for images shot with lens stabilization on (at least for my ef 70-200), adobe’s upscaling occasionally creates some strange results.

I really want a higher resolution option but I’m not going to die or lose my job without it - I definitely don’t need it. With that said, it will be a big element in my decision on which body I buy next as the only feature improvement out there (over my trusty 5Div) which will have a meaningful impact on what/how I shoot is resolution.

The question is are there enough people like me out there for it to be worth it to Canon to produce something over 45mp. Only Canon knows that, but I’m hoping this CR1 comes true -even if it’s far fetched.

 

[David - Sydney]

There’s certainly trade offs in using higher resolution bodies, but depending on the use case they’re manageable. For instance - I’m after a higher resolution option than the R5 and none of those limitations give me anxiety. Most of my images are on a tripod without the stabilizer on (probably 95% without autofocus on frankly), I shoot very few images so the over all storage load is low, I’m already using adobe’s AI upscale occasionally on the 5Div giving 120mp images and my computer system manages it fine. With that said, I’ve noticed that for images shot with lens stabilization on (at least for my ef 70-200), adobe’s upscaling occasionally creates some strange results.

I really want a higher resolution option but I’m not going to die or lose my job without it - I definitely don’t need it. With that said, it will be a big element in my decision on which body I buy next as the only feature improvement out there (over my trusty 5Div) which will have a meaningful impact on what/how I shoot is resolution.

The question is are there enough people like me out there for it to be worth it to Canon to produce something over 45mp. Only Canon knows that, but I’m hoping this CR1 comes true -even if it’s far fetched.

There is the inherent optical issue of diffraction limits for high mp (smaller pixel size) sensors which limits the total resolution or loss of sharpness in the final image. Landscapers may need to stack multiple images at a wider aperture rather than a single image with - say - f13 or above.

Pixel size is one of the advantages of medium format besides the 16 bit depth. Whether these are key issues for your use cases is a different story.

 

[David - Sydney]

Pentax style star tracking using IBIS + GPS would be nice to have in all IBIS equipped bodies, but judging from previous behaviour, Canon will only include that in the astro bodies.

I can understand that GPS has been offloaded to paired phones (which is probably more accurate) and saves power/circuitry but keeping a good connection is problematic over longer periods and especially in areas without good mobile phone coverage - like astro photography locations.

 

[David - Sydney]

Me too, but I'm afraid it was only a rumor. The current 24-105 F4 is, in my opinion, a bit disappointing, mostly at 105mm.
But if the F 2,8 ever hits the market, my hopes are it could be optically as good as the 24-70 F2,8...

Agreed that it would be close to the poorest L lens but I guess that it is made to a price and be part of kit/bundles for the higher end bodies (and then sold separately by retailers to make more profit).
It is rare for a wider zoom to be as good as a smaller focal range. It can get close but generally costs a lot more. Not sure of the sales for a 24-105/2.8 vs Canon selling both a 24-70/2.8 + 70-200/2.8

 

[amorse]

There is the inherent optical issue of diffraction limits for high mp (smaller pixel size) sensors which limits the total resolution or loss of sharpness in the final image. Landscapers may need to stack multiple images at a wider aperture rather than a single image with - say - f13 or above.

Pixel size is one of the advantages of medium format besides the 16 bit depth. Whether these are key issues for your use cases is a different story.

Absolutely - and while diffraction is an issue, it’s again not insurmountable. I’m already focus stacking pretty often when on a tripod - for me, I’d likely just need more frames per stack if I have to bring down the aperture to address diffraction. Alternatively, when I’m shooting out of an airplane I’m pretty much always at infinity and f/2.8. Seems weird to shoot landscapes on burst at f/2.8 with a faster shutter speed, but I’ve found that to be most successful, and diffraction should be well managed at 2.8. With all that said, I’d be satisfied with the pixel density of an R7 in a full frame sensor.

I’ve debated medium format and the GFX system over and over but my issue is their lenses will just never be as long as a 100-400 (or equivalent) option on full frame. I really like the opportunity for more intimates on a 100-400 and I wouldn’t want to write that off - hence why a higher resolution full frame body is what I think fits my need best. Also I’ve seen some anecdotal notes that the GFX100s doesn’t have equivalent weather sealing to something like a 5D which would be a dealbreaker for me.

My point in all this is I do think there is a use case for such a body, but it’s up to canon to determine if there’s enough market to justify it. Time will tell I guess!

 

[EOS 4 Life]

I don't have evidence either way, just the observation that cRAW is turned off by default, which means a significant portion of users will ignore it.

I would like to think that is true but in the current Canon lineup, every camera from the R10 up has sports and wildlife features and would benefit from CRAW.
The R50 seems more like a JPEG camera for those activities.

 

[David - Sydney]

I think most people here at CR who shoot bursts and don't use JPEG (an JPEG is what R3 is for) use CRAW.

Now, if you use R5 as a landscape camera, you don't need bursts. But if you use it for something like birds in flight....

Clearly a generalisation but in my case (data set of n=1), I have no issues to shoot bursts in full raw with my M1 MBP.

cRAW would be faster to clear to buffer but Bryan's testing at
https://www.the-digital-picture.com...-R6-Buffer-Capacity-During-High-Speed-Capture
shows dual raw = 175 images for 14 bit/mechanical shutter and 104 images for 12 bit/electronic shutter
>14 seconds of burst for mechanical and >5 seconds @ 20fps is a long burst IMHO
without using cRAW.

 

[David - Sydney]

Absolutely - and while diffraction is an issue, it’s again not insurmountable. I’m already focus stacking pretty often when on a tripod - for me, I’d likely just need more frames per stack if I have to bring down the aperture to address diffraction. Alternatively, when I’m shooting out of an airplane I’m pretty much always at infinity and f/2.8. Seems weird to shoot landscapes on burst at f/2.8 with a faster shutter speed, but I’ve found that to be most successful, and diffraction should be well managed at 2.8. With all that said, I’d be satisfied with the pixel density of an R7 in a full frame sensor.

I’ve debated medium format and the GFX system over and over but my issue is their lenses will just never be as long as a 100-400 (or equivalent) option on full frame. I really like the opportunity for more intimates on a 100-400 and I wouldn’t want to write that off - hence why a higher resolution full frame body is what I think fits my need best. Also I’ve seen some anecdotal notes that the GFX100s doesn’t have equivalent weather sealing to something like a 5D which would be a dealbreaker for me.

My point in all this is I do think there is a use case for such a body, but it’s up to canon to determine if there’s enough market to justify it. Time will tell I guess!

I've never used burst from an airplane but am taking many shots to look for different perspectives. I am looking for abstracts more that traditional landscapes though so sharp banking is common (if stomach churning) Sharpness is not critical for abstracts and that helps as I am using the 24-105mm/4 in any case.

My main issue with the GFX system is the lack of UWA lenses. It would seem like going backwards to need to shoot panoramas (as well as stacking) for an image. Having one FF system and lenses from 8-500mm covers everything so far

 

[neuroanatomist]

I don't have evidence either way, just the observation that cRAW is turned off by default, which means a significant portion of users will ignore it.

I’m not sure your logic holds up. RAW is also turned off by default. The default option on a camera fresh from the box is Large JPG. If one decides to shoot RAW, it’s one click of the wheel for RAW and two clicks for cRAW.

 

[EOS 4 Life]

I read, a few years ago, that some major agencies no longer accept "raw", for the reason they are too easy to manipulate, unlike Jpegs. I think it was either Reuters or AFP.

The problem with RAW is that it requires manipulation.
They want an image unaltered.
It is pretty easy to manipulate a JPEG or manipulate a RAW image into a JPEG.
It is just against the rules.

 

[David - Sydney]

24-105 MM f/2.8 makes even more sense with the release of the 100-300 f/2.8.
There are many people would be set with two cameras with one camera having the v24-105 and the other having the 100-300.

That is a fair point but what would be the % of users in that situation? Generally sports will benefit from the 100-300mm zoom vs 70-200/2.8 + 300/2.8. Would they need to also cover the shorter focal lengths?

 

[EOS 4 Life]

Would they need to also cover the shorter focal lengths?

It depends on the sport.
Combining a 70-200 with the EF 200-400 f/4 x 1.4 or the rumored RF 200-500 f/4 x 1.4 would probably suit a lot more sports.

 

[David - Sydney]

As I'm sure you know, most photo competitions use the "traditional darkroom technique" or similar standard. The digital equivalents of dodging, burning, spotting, etc. are generally considered acceptable.

From what I have seen of competitions, they are creating different categories specifically for different amounts of processing. Specifying global or small amounts of local dodge/burn is okay but composites aren't. In any case, all original images used have to be taken by the person submitting them. There was one case for Australia's top awards (no longer in place now) where composites were allowed and won but the photographer used some stock images in her work = kerfuffle.

To try to keep it simple, what I'm getting at is that, for me, I want to be able to capture what was really there and what I experienced.

For me, it is the memory of me being there expressed in an image... but also my post processing to produce an artwork. Critters underwater are great at camouflage and there are lots of distractions but the beauty is exposing the subject to a wider audience vs what a tourist might take or someone who is newer to the craft. It is hard to do it well and that is the challenge for me. AI won't do that for me but there may be aspects eg noise reduction or upscaling where I would take advantage of it if I felt it was needed.

I just don't think AI is going to be the death of photography as we know it, because I really believe most people are after the experience of doing it themselves.

I believe that there will be different groups of users as "photography" evolves. Everything is digital now so the "painting with light" definition is somewhat quaint and seems akin to the "is photography art?" question by artists.

AI will be used to replace stock or graphic designer/website images for sure!
Midjourney CEO's admission to using a “hundred million” images without consent to train his AI image generator in an interview with Forbes last year will be a massive justification to a class action lawsuit making a whole category of workers redundant. Uber needing to compensate taxi owners for decreasing their state enforced taxi license value is another example.

 

[EOS 4 Life]

It depends on the sport.
Combining a 70-200 with the EF 200-400 f/4 x 1.4 or the rumored RF 200-500 f/4 x 1.4 would probably suit a lot more sports.

The more that I think about it, Canon should probably replace the RF 400 f/2.8 with a 300-400 f/2.8 zoom lens.
Same thing with the 600 f/4.
That would leave 100-300 f/2.8 and 300-400 f/2.8 with 24-105 f/2.8 completing the trinity.
The F/4 trinity would be 70-200 f/4*, 200-500 f/4, and 500-600 f/4.

*I think people in this price range would just opt for the RF 70-200 f/2.8 and stop it down to f/4 when they needed to.

 

[David - Sydney]

I know it's off-topic, but it's interesting to talk about AI and photography. I enjoy taking photos, and now rely on the Canon AF to get exact focus. But it's just a tool that I control to tell it what to focus on, and removing that complexity from me allows me to focus on composition of the image which I find the most important thing.

Just to help with the definitions... Generative AI (midjourney, Chat GPT etc) will generate new images based on their training set. This can now be downloaded and it can take 500 of your processed images and create something as well. Things are moving fast!

Machine learning is different (although a part of AI) and it related to better pattern matching eg adding train or car AF tracking as a different category to animals and people options

On the way out, the art director (whom I know as I've displayed my art with him) told me they were all generated and painted by an AI algorithm that a single person had instructed. That person was not an artist, but someone who could pull the levers of a new artificial intelligence that would eventually know of no limits on what it could create, and of all the livelihoods of artists it could eventually replace. Why would anyone not want to buy these lovely works of art, in just the style they wanted to see, at such a low price that they could not resist buying? And how quickly will AI, and the machines that run it, replace us?

The big question for me is around copyright and I guess the legal framework around it.
Who owns the copyright to an AI generated image? The person who entered the commands or the AI tool or the thousands of images whose copyright has been "borrowed"?

The PETA question of whether the macaque owned the image since they pushed the button or now was a great test... but ultimately the result was based on that only a person can hold a copyright (at this time).

On a separate note, the people who buy Ikea framed photos for their walls staggers me as the quality is mediocre at best and I could do better (IMHO).
The convenience to buy there outweighs quality like MP3 quality and number of downloaded songs on small devices vs portable CD players that had quality but skipping/size etc seems to be a similar nexus.

 

[David - Sydney]

The more that I think about it, Canon should probably replace the RF 400 f/2.8 with a 300-400 f/2.8 zoom lens.
Same thing with the 600 f/4.
That would leave 100-300 f/2.8 and 300-400 f/2.8 with 24-105 f/2.8 completing the trinity.
The F/4 trinity would be 70-200 f/4*, 200-500 f/4, and 500-600 f/4.

*I think people in this price range would just opt for the RF 70-200 f/2.8 and stop it down to f/4 when they needed to.

That's a lot of "trinity's"!
Previously it was only 16-35/2.8 + 24-70/2.8 + 70-200/2.8 + 300/2.8 + 400/2.8
Now we have 15-35/2.8 + 28-70/2 + 70-200/2.8 + 100-300/2.8 + 400/2.8
And 14-35/4 + 24-105/4 + 100-400/4 + 200-400/4 + 600/4
Too many choices

 

[Ph0t0]

In theory I would agree with you, but I've been using Topaz software for some time now, and the AI generated detail is extremely convincing, e.g. it reveals the detail of the microscopic scales in butterfly wings very accurately. It's amazingly good at generating natural looking detail.

As for generating detail from blurred car number plates, there was recently a test conducted by dpreview, where commercial AI software produced accurate and legible alphanumerics from blurred and illegible origin images.

If it isillegible it is illegible. So obviously if it produced accurate alphanumerics there was enough data to sharpen and interperet the image correctly. And extremely convincing is different than real.
Even if the subject is an everyday object -> if it has an abnormality that is not recorded by the sensor, then it wont be shown on an AI image either.
In some cases it might even be remowed, even if it is captured by the sensor.

And don't get me wrong I tottaly understand that for most cases AI generated detail is good enough. But there are also plenty of cases where you might want the real detail.
I shoot a lot of commercial stuff for ads and usualy I don't care if those images are AI enhanced or not.
But I'm also working on some editorial stuff and some personal projects. Last week I photographed a 92 year old farmer/hunter in front of his house with his fathers rifle in his hand and some tools and trophies up on the wooden walls of the house. There were plenty of details that are interesting. Some are visible some not. But in this case I would rather choose a higher MP camera to show the hand made engravings on his rifle, the details on his face, engravings on trophies, details on old tools, then have AI make up "extremely convincing" ones.

And about MP needs.... even when the AI generated detail is convincing enough or when you just need to sharpen or to clean the image with the help of AI, there are cases when you might have a need for a high MP camera. Otherwise you might just want to tell the AI to create everything from scratch.