Canon EOS R5 Mark II to arrive before EOS R1? [CR2]

[Jethro]

Ultimately, CIPA is a data point. A test like all others comes with caveats and virtually all reviewers agree that real life usage gives substantially better battery performance than CIPA.

Assuming (and we don't know for sure) that Canon are using the exact same testing methodology for the R6 and R6ii, then there are substantial improvements in the new version.

Canon have always been a conservative company from an engineering perspective but the R5/R6 let their engineering team off the chain.
If Canon really cared about CIPA results on comparison spec sheets and they could change their testing methodology, I would guess that they would have done it by now.

That was my point - if companies are going to be dishonest about these things, they have just as much motivation to do so regarding their own products ('Look at the improvement from the R6 to R6 II!') as competitors'.

Post-Volkswagen, we all have a duty to question the limits of these types of data-fed metrics, but at some level they are all we have to go by.

 

[EOS 4 Life]

The mention of Gerald Undone makes me think this is about DR in video. And in that case, he is correct. The R5C allows shooting in CLOG2, which has a gamma curve that captures more dynamic range than CLOG3 (or regular non-log). The regular R5 doesn't allow CLOG2, only CLOG3.

You are correct that both share the same sensor and in stills mode, shooting RAW, you will get identical DR. But the R5-non-C captures and stores video in such a way that it limits the video DR. On top of that, video DR is measured in a different way than still DR, which complicates discussions even more.

Also, Gerald is very clear in his review that the difference in video DR isn't due to the sensor, but caused by the available log profiles.

Gerald compared CLOG2 on both cameras.
Canon did not go into detail on how they were able to get an extra stop of dynamic range but I suppose that has something to do with the active cooling.
The codecs are not quite the same either.
Also, the way Gerald measures DR is different than the way Canon measures it.

 

[Michael Clark]

So your view is that manufacturers will manipulate data etc to gain some advantage (in CIPA metrics) as against other manufacturers' products, but they will be scrupulously consistent in applying it to comparing their own products? Or just that basically all standards such as this are useless, and we are left with anecdotal evidence in making any comparisons at all?

It seems to me the manufacturers that set "defaults" to the best performing parameters for the tests would do the same with all of their cameras, so cameras within that manufacturer's catalog could all be compared when the settings are all set to the least power consuming options. On the other hand, Canon seems to use the actual default settings included when the camera actually is released to the market. Due to the way the CIPA tests are conducted, this results in battery ratings well below what most users experience in the real world, at least for users with DSLRs that shoot mostly using the OVF instead of shooting all of the time in LV with the screen on 100% of the time.

For example, the Canon EOS 7D Mark II's CIPA rating is 670 frames at 23°C tested in Live View and the built-in flash used on half the exposures (both per the CIPA guidelines). I routinely got around 2,200-2,400 frames shooting sports in high speed burst mode for several hours using the OVF, no flash, with very little chimping and still ended with 20-30% still left in the two brand new Canon LP-E6N batteries in the grip. That's 1,100-1,200 per battery plus a 20-30% reserve.

 

[Michael Clark]

I am not following you...
Kit said
"Do you know when the firmware 1.0 branch had its feature freeze and the firmware 1.1 started to be developed? Could as well be before the R5 was even publicly announced.
The presence of the temperature sensors alone tells us that this functionality was planned since the beginning, just wasn't ready for the 1.0 release."
and I replied to Kit about the firmware freeze.
Neuoro showed that the processor does downgrade from 14 to 12 bit from 12fps etc based on processor limitations and I subsequently calculated the bandwidth limits for R3/R5 in different modes showing a common limit of ~11gb/s

Kit also said in the comment to which I already linked, not the previous comment you quoted: "Nearly twofold increase in pixel rate with one-bit decrease in pixel bit depth points at the ramp-compare ADC as the most likely limitation."

 

[Michael Clark]

Ultimately, CIPA is a data point. A test like all others comes with caveats and virtually all reviewers agree that real life usage gives substantially better battery performance than CIPA.

For Canon, real life almost always gives much better performance than the CIPA rating. Not so for some of the others.

In real life the Canon 1D X Mark II and the contemporary Nikon D5 got about the same number of shots per battery according to my friend who shot with both before/after he changed jobs. The Nikon had a self-tested, self-reported, unaudited CIPA rating of 3780, the Canon had a self-tested, self-reported, unaudited CIPA rating of 1210.

 

[Michael Clark]

If Canon really cared about CIPA results on comparison spec sheets and they could change their testing methodology, I would guess that they would have done it by now.

Which is kind of my whole point. Canon knows what CIPA battery ratings really are.

As you have said, they are conservative and would rather promise less and deliver more than vice-versa. This goes far beyond battery life ratings. Canon's fps and buffer depth estimates are routinely lower than what Bryan Carnathan and others routinely find when they test Canon cameras under real world conditions.

 

[Michael Clark]

That was my point - if companies are going to be dishonest about these things, they have just as much motivation to do so regarding their own products ('Look at the improvement from the R6 to R6 II!') as competitors'.

Post-Volkswagen, we all have a duty to question the limits of these types of data-fed metrics, but at some level they are all we have to go by.

If a company seems to always hedge their bet and promise less based on a worst case scenario and then deliver more in real world application, as Canon has done with a wide variety of specs for decades, they'll probably continue doing so.

Canon has always seemed to be more concerned with exceeding their customers' expectations, based on what Canon has promised on spec sheets, than about winning spec sheet pissing contests with other camera makers who promise more than they deliver (Cough, cough, Sony 1st & 2nd generation α7 models, cough, cough).

Maybe that culture changed a bit after the fiasco regarding R5 8K overheating, maybe not. But institutions the size of Canon are like aircraft carriers when it comes time to change headings. It takes quite a while.

 

[David - Sydney]

Maybe that culture changed a bit after the fiasco regarding R5 8K overheating, maybe not. But institutions the size of Canon are like aircraft carriers when it comes time to change headings. It takes quite a while.

Oh dear... when are people going to recognise that it wasn't a fiasco. It was a beat up by reviewers and especially Sony owners that Canon out-engineered them at the time and Sony still can't provide 8k raw files even in a camera 50% more expensive.
Let's be very clear that thermal limits were only in 3 video modes: 8k raw, 4k120 and 4KHQ. All of which now have substantially better record times (and back-to-back times) after swift firmware updates.

Maybe and just maybe, Canon promised and delivered but still wasn't enough for complainers.
At the time (and to plagiarise an excellent historical commentary).

Somewhere in a galaxy not far from here, within 6 months of announcements…
Internet: Canon could never put 8K full frame in to a mirrorless camera. They don't have the technical capabilities to do that. The last time they innovated was when 5d mkii was born.
Canon: Our next 5 series will have 8K and demonstrate what we are capable of.
Internet: (spits coffee on screen) Yeah right. Maybe 8K timelapse mode
Canon: No, 8K video
Internet: Er no, 8K like 15 second movie burst mode then, because you know overheating would happen in such a small body. Do Canon engineers not know anything about physics? That's why refrigerators exist....
Canon: No full 8K video
Internet: Yeah right.... hahaha. It's not possible. So no AF or IBIS then. You won't be able to use the full technologies of the camera/
Canon: No. 8K Full frame, AF and IBIS enabled.
Internet: It's not possible, because you know.... overheating is a thing with technology in such a small package. If it was possible then Sony would have done it.
Canon: Hold our beer.....

Post launch
Internet: I demand a full unconditional apology from Canon. I demand immediately that they give us unlimited 8K full frame video for $4.99 and stop hiding behind this "overheating" scenario to protect their other cameras. I’ve never felt so personally insulted by a camera company as I have by Canon. To me honest behaviour is very important. I recognize that the only person that is perfect is me and people make mistakes. The important thing is to admit that and not to sell us totally unusable technology like this…

 

[David - Sydney]

Post-Volkswagen, we all have a duty to question the limits of these types of data-fed metrics, but at some level they are all we have to go by.

VW was a different beast in that their SW changed the engine profile based on whether it detected that a 3rd party was doing an emissions test or not.
If the cars were keeping within the limits then the performance was less but VW advertised their car's performance to the wrong emissions level.

That is heavy duty fraud and the financial and reputational cost was high.
The emissions standards were at a country level and independently tested so a little different to CIPA.
CIPA is voluntary and self tested and has less impact on society than emissions

Now, datasets feeding AI is a whole different kettle of fish as biased inputs give biased outputs and it isn't clear how AI (as a black box) gives the final answer. Using non-consensual data ie scraping is also ethically dubious. AI will change people's lives from education to white collar jobs and the speed of change will hit people hard. Having regulatory guardrails is pretty important for that but legislation is always slower than innovation.

 

[scyrene]

The ghost of my eighth grade English teacher still haunts me to this day. If I was less pedantic she would torture me with that blasted ruler for all eternity.

Sometimes pedantry is warranted. But essentially claiming the technical definition of a word in a specialist field trumps a general usage doesn't demonstrate your superiority; I would argue, quite the opposite.

 

[scyrene]

Getting cheaper on this side of the pond.

View attachment 207764

Possibly beside the point but I was at a retailer event yesterday with representatives from all the major camera manufacturers and they were offering the R5 new for £3469.10 (but an even bigger % discount was offered on the R62, strangely). I'm not sure who takes the hit there but presumably Canon doesn't mind them undercutting the RRP if their guys are actually there.

 

[unfocused]

Oh dear... when are people going to recognise that it wasn't a fiasco. It was a beat up by reviewers and especially Sony owners that Canon out-engineered them at the time and Sony still can't provide 8k raw files even in a camera 50% more expensive.
Let's be very clear that thermal limits were only in 3 video modes: 8k raw, 4k120 and 4KHQ. All of which now have substantially better record times (and back-to-back times) after swift firmware updates.

Maybe and just maybe, Canon promised and delivered but still wasn't enough for complainers...

I'm sorry, but I feel like we fanboys (and I happily include myself in that category) just have to take the loss on this one. And, I say this as a stills photographer who was never impacted by the issue.

Canon overpromised and underdelivered. And, including some asterisk disclaimers really didn't excuse them. Canon realized their mistake and took measures to correct it as best they could. I give them credit for ultimately owning the problem and addressing it. But it doesn't do anyone any good to act like it was never a problem or that it was okay to release a product that had a feature that wasn't ready for prime time.

That said, it is time for the complainers to move on as well and quit acting like this one-time screw up was somehow indicative of an endemic problem with Canon.

 

[neuroanatomist]

That said, it is time for the complainers to move on as well and quit acting like this one-time screw up was somehow indicative of an endemic problem with Canon.

One time? Oh, puuuleeez. Remember the debacle with the 1DIII AF system? And what about the cave? Remember their failure at the cave.

 

[padam]

I'm sorry, but I feel like we fanboys (and I happily include myself in that category) just have to take the loss on this one. And, I say this as a stills photographer who was never impacted by the issue.

Canon overpromised and underdelivered. And, including some asterisk disclaimers really didn't excuse them. Canon realized their mistake and took measures to correct it as best they could. I give them credit for ultimately owning the problem and addressing it. But it doesn't do anyone any good to act like it was never a problem or that it was okay to release a product that had a feature that wasn't ready for prime time.

The first paragraph explains everything. The R6 and R5 stock will be around for a few years for those who don't care about overheating for less money than an R6 II or the upcoming R5 II.
It wasn't really a 'mistake', from the get go they've published all the overheating times (so they didn't overpromise) which in case of the R5 with the High Temperature mode they've improved it up to a point where it's not really an issue in many use cases.
Maybe that was enough to delay the R5 II a little bit.
This 'fiasco' is just some ways to mimic what Sony is doing, keeping around older models who don't need all the bells and whistles of the newer ones or incentivize upgrading to the new model within 3 years.
It is still only the revised processor that is making most of the difference. They can still add better heat dissipation in future models - (if they think it is more profitable than selling another one with a built-in fan and different video interface, etc.)

 

[ata bora]

Considering that the expected Canon Eos R5 Mark II will be announced much later than the R6 Mark II, I think a newly designed processor will come with more features for 8K. We will see together what the heating, freezing and time limit will be with the R5.

 

[padam]

Considering that the expected Canon Eos R5 Mark II will be announced much later than the R6 Mark II, I think a newly designed processor will come with more features for 8K. We will see together what the heating, freezing and time limit will be with the R5.

The R5C already does everything except power draw limits for 8k60p
They may leave out 8k60p from the R5 II altogether and keep it for the R5C and just have 4k60p Fine recording, who knows.

 

[Jethro]

...
It wasn't really a 'mistake', from the get go they've published all the overheating times (so they didn't overpromise) which in case of the R5 with the High Temperature mode they've improved it up to a point where it's not really an issue in many use cases.
Maybe that was enough to delay the R5 II a little bit.
This 'fiasco' is just some ways to mimic what Sony is doing, keeping around older models who don't need all the bells and whistles of the newer ones or incentivize upgrading to the new model within 3 years.
It is still only the revised processor that is making most of the difference. They can still add better heat dissipation in future models - (if they think it is more profitable than selling another one with a built-in fan and different video interface, etc.)

It was absolutely a fiasco in a marketing sense, but not design. You're right that they published time limits, but they were buried deep in the handbooks, while the marketing stuff prominently featured '8K'. And they deserved grief for doing that, because people (sometimes maybe even genuinely) thought they were getting a cutting edge pro-level cinematography camera for 1/3 what it would normally cost. If you stick your chin into a boxing ring, you can't much complain about bringing it out with bruises.

In reality though, at the time, the response to Canon announcing the R5 had 8K capacity was ... incredulity. It can't happen! It'll overheat immediately! The laws of thermodynamics! If it was possible Sony would have already done it!! In fact it worked fine, but for relatively short periods of time. The design issue was that there seemed to be overly-conservative software triggering 'overheating' warnings and shutdowns. And that was fixed with firmware updates. Which absolutely should not have been necessary.

 

[Jethro]

One time? Oh, puuuleeez. Remember the debacle with the 1DIII AF system? And what about the cave? Remember their failure at the cave.

What was the cave?

 

[neuroanatomist]

What was the cave?

 

[davidhfe]

It was absolutely a fiasco in a marketing sense, but not design. You're right that they published time limits, but they were buried deep in the handbooks, while the marketing stuff prominently featured '8K'. And they deserved grief for doing that, because people (sometimes maybe even genuinely) thought they were getting a cutting edge pro-level cinematography camera for 1/3 what it would normally cost. If you stick your chin into a boxing ring, you can't much complain about bringing it out with bruises.

In reality though, at the time, the response to Canon announcing the R5 had 8K capacity was ... incredulity. It can't happen! It'll overheat immediately! The laws of thermodynamics! If it was possible Sony would have already done it!! In fact it worked fine, but for relatively short periods of time. The design issue was that there seemed to be overly-conservative software triggering 'overheating' warnings and shutdowns. And that was fixed with firmware updates. Which absolutely should not have been necessary.

Exactly. What makes it a fiasco is that it was entirely avoidable, and that there were multiple ways Canon could have avoided it. It overshadowed the launch and damaged Canon's rep to some degree. I still hear people say "the R5 can't shoot 4K without overheating" which was never true at standard quality aka the quality of all other high mp cameras for at least a year post R5 launch?

I had a Sony guy at Best Buy talk about how A1 was better and I said "it's almost twice as much money it should be??" and I just stood there watching his brain try to overcome the cognitive dissonance.