Canon R5 Announced with 8k

[Bassman2003]

I would trend towards the 4k 12 or 14bit too but then there is the belief that 8k on a 4k timeline improves everything. With these mirrorless RAW implementations, is RAW always 4:4:4? I do not trust specification expectations in the small camera world when it comes to video.

 

[ahalpert]

I think that 8K on a 4K timeline would improve things vs 4K with the same color subsampling... but maybe I would prefer 4K 444 over 8K 422 although I'd want to see a side by side comparison...

David Mullen says this:

"raw is an image off of the sensor pre-conversion to RGB, un-bayered, whereas 4:4:4, 4:2:2, etc. are all forms of color subsampling, which implies working with an RGB signal in the first place in order to then subsample it."

https://cinematography.com/index.php...e-same-as-444/

So, I think that every form of raw will give you full color bandwidth in every channel. You have a color reading from each photosite

I get confused by the different forms of "raw" out there and can't remember all the discussions about which ones aren't "truly" raw. I know that braw does a partial debayer before encoding...can't remember the specifics of that...

But basically I don't think you get to call it RAW if it's been debayered. And if it hasn't been debayered, then you've basically got 444

Bit depth is probably just as important as color subsampling, although you shouldn't be able to see a difference once you surpass a certain level. (12-bit log or 16-bit linear seems to be the consensus.) But what irks me about color subsampling is that you're literally skipping over color values and effectively smudging the colors. That being said, I only own 422 cameras (although I could get ProRes raw from my A7SIII with a ninja)

EDIT: I just worked out the feasibility of recording ProRes raw for my longform shooting... nope, still not realistic. But definitely if I'm going out to record for less than an hour, I could probably do it. Certainly for my fine art b-roll shoots, that could be cool...

 

[NorBro]

I would trend towards the 4k 12 or 14bit too but then there is the belief that 8k on a 4k timeline improves everything. With these mirrorless RAW implementations, is RAW always 4:4:4? I do not trust specification expectations in the small camera world when it comes to video.

Depends on who you're asking, but any internal "RAW" video technically isn't supposed to be real RAW video because of patents.

So most of them are likely extremely high-quality proprietary formats engineered in their own ways.

CRL is 10-bit and 12-bit (4K/60p is limited to 10-bit).

External RAW is open to more options.

 

[ahalpert]

I'm looking at the ProRes raw white paper right now and it does say that it records the data without debayering it. You can pull a 444 file later

PS the ninja v has really held its value over the years, which is unfortunate because I don't own one

 

[NorBro]

That's probably what some of the lawsuit was about with RED...Apple likely wanted it inside cameras - maybe even their phones - and couldn't (lost), so it's still limited to external.

 

[Bassman2003]

Thanks for your replies. In the end, it's all good to some extent compared to just a few years ago.

 

[ahalpert]

Has anyone tested the new R5 firmware? Are overheating shutdowns a thing of the past?

I find it hard to believe that the R5 would never or should never shut down for such a reason when even my a7siii and fx3, which are never thermally taxed as much nor were ever suspected of a cripple conspiracy, will shut down to preserve themselves under high-stress conditions.

Nevertheless! If the R5 performed nearly as well as those then I probably would have bought it instead way back when, if only because I already owned Canon af lenses

 

[NorBro]

If the setting works how it does in a few of their other cameras then the recording times will drastically improve (and a few tests on YouTube from the last few days already seem to confirm that notion).

 

[AtticusLake]

Has anyone tested the new R5 firmware? Are overheating shutdowns a thing of the past?

I'm currently working my way through a battery of tests -- each test takes half a day, basically, because of needing to let the camera cool completely between tests. So I'm not done yet. But so far:

• Overheat=High, 4k, 23.98 FPS, HQ: 3 hours continuous record, no warnings, no shutdown
• Overheat=High, 8k, 23.98 FPS, IPB: 4 hours plus continuous record, intermittent warnings, no shutdown
• Overheat=High, 4k, 59.94 FPS, LQ: 3 hours continuous record, no warnings, no shutdown

In Overheat=standard, times look pretty normal: like 40 minutes shooting in 4k HQ before shutdown. So I have to say, the update is looking good.

After running 3-4 hours continuous, the camera didn't even seem especially hot -- warm, but not painful. The memory card was too hot to hold, though.

 

[Clermond]

I'm currently working my way through a battery of tests -- each test takes half a day, basically, because of needing to let the camera cool completely between tests. So I'm not done yet. But so far:

• Overheat=High, 4k, 23.98 FPS, HQ: 3 hours continuous record, no warnings, no shutdown
• Overheat=High, 8k, 23.98 FPS, IPB: 4 hours plus continuous record, intermittent warnings, no shutdown
• Overheat=High, 4k, 59.94 FPS, LQ: 3 hours continuous record, no warnings, no shutdown

In Overheat=standard, times look pretty normal: like 40 minutes shooting in 4k HQ before shutdown. So I have to say, the update is looking good.

After running 3-4 hours continuous, the camera didn't even seem especially hot -- warm, but not painful. The memory card was too hot to hold, though.

thanks for doing these tests and sharing the results.
it’s literally cool!

 

[ahalpert]

.

After running 3-4 hours continuous, the camera didn't even seem especially hot -- warm, but not painful. The memory card was too hot to hold, though.

Sounds like a data corruption risk. Somewhat challenging to mitigate if you need cfexpress card speeds since it only has one cfexpress slot, eliminating the possibility of dual recording

Also, are you conducting tests at room temperature?

Still, sounds like the new update pretty much solves the issue for the type of intermittent b-roll shooting I would have primarily used the camera for

Thanks for sharing your results!

 

[NorBro]

You can also use SDXC but just not for everything. The cards always get super hot in their cameras, it's fine.

 

[AtticusLake]

Also, are you conducting tests at room temperature?

See the write-up on my web site: https://moonblink.info/MudLake/gear/r516

 

[ahalpert]

See the write-up on my web site: https://moonblink.info/MudLake/gear/r516

Thanks. Looks like the camera still has reliability concerns but I suppose that's to be expected. Probably good enough for anyone who's interested in saving money going this route anyway

 

[NorBro]

P.S. AtticusLake...if you ever have a chance, check out the R3. It is such a good camera, it's the best Canon camera I've personally ever used (never tried the C300 Mark III or C500 Mark II).

Per usual, Canon is a simple company who does not entertain bargains; pay them the most money in their hybrid world and you'll get their best. The AF, the IS, the low-light, even the colors somehow seem nicer.

 

[ahalpert]

, even the colors somehow seem nicer.

I've spoken to photographers who say that in their experience, the 1D series cameras have nicer color than the 5D series cameras. But they couldn't put their finger on why exactly, or explain the cause. I suspect there is a difference in the quality or fine-tuning/quality control of components, perhaps in the sensor circuitry, that does not get reported in a spec sheet

 

[NorBro]

I've spoken to photographers who say that in their experience, the 1D series cameras have nicer color than the 5D series cameras. But they couldn't put their finger on why exactly, or explain the cause. I suspect there is a difference in the quality or fine-tuning/quality control of components, perhaps in the sensor circuitry, that does not get reported in a spec sheet

I would guess it's software-based. Subtle tweaks in color management that make it just a bit more noticeably better...maybe a saturation of 8 instead of 5 (poor analogy but you get it).

Blackmagic has improved their color science in the same camera over the years (and even highlight roll-off) so it's at least possible. RED too, including for cameras like 5+ years later with IPP2. Suddenly colors pop more in sensors that were shipping years before.

 

[AtticusLake]

P.S. AtticusLake...if you ever have a chance, check out the R3. It is such a good camera, it's the best Canon camera I've personally ever used

Well, I'm shooting video, and the R3 strikes me as more of a stills-oriented camera -- like the successor to the 1D. I know well that it has great video abilities, but it's huge, and yet no actve cooling or ND filters?

When you say "it's the best Canon camera I've personally ever used", have you compared it to the R5C or the C70? I have both of those (I know, way too many cameras) and theyre great.

 

[NorBro]

Yeah, but the R5 you're testing above is and always will be a stills camera first as well...

It's big but they cut down the size from the original bigger bodies. It doesn't overheat so does it need active cooling? (Def. no ND filters.)

And I haven't compared them to the aforementioned because I don't do any lab-like tests like that (for me they are useless because real life run-and-gun productions have different circumstances), but I would like some more time with the R5 C. I definitely prefer it over the C70 (and it's smaller than it in 2 out of 3 ways).

This is it compared to the 1DX Mark III (source: Lensrentals).

 

[ahalpert]

I would guess it's software-based. Subtle tweaks in color management that make it just a bit more noticeably better...maybe a saturation of 8 instead of 5 (poor analogy but you get it).

Blackmagic has improved their color science in the same camera over the years (and even highlight roll-off) so it's at least possible. RED too, including for cameras like 5+ years later with IPP2. Suddenly colors pop more in sensors that were shipping years before.

I'm not sure I buy that explanation. For one thing, new color science can be rolled out to all BM and RED cameras, low-end and high-end alike. And if that hypothesis were correct, you certainly wouldn't expect a newer 5D to have worse color than an older 1D, as is the case.

My guess is that it's more like with the stitched-sensor ARRIs yielding more dynamic range. John Brawley told me that only cream of the crop ALEV sensors were selected to create larger-format cameras, and that they performed slightly better than average with regard to DR, but that otherwise there was no reason why there would be a DR advantage over the S35 Alexa. Obviously, that's not what goes on with the 1D vs 5D but there could be some other unreported quality control or component difference.