fredmiranda.com
Login

Moderated by: Fred Miranda
  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

1       2       3              10              12       13       end
  

Archive 2020 · R5 artificial time limit

  
 
ilkka_nissila
Offline
• • • •
Upload & Sell: Off
p.11 #1 · R5 artificial time limit


Copper might not produce a significant benefit to heat transfer over aluminium since the bottleneck is likely elsewhere in the system. If the air were being moved by forced flow (vents and fan), copper might help but with the air not being moved past the sink, it seems the metal plate is designed to transfer heat to the metal chassis. Likely the material was chosen because higher thermal conductivity in this component would not significantly benefit overall cooling. Note that there is another metal surface and what seems to be a thermally conductive pad below the underside of the processor PCB. Finally, the copper in the processor board itself (one or more layers) conducts heat to the chassis as well. So there seem to be three or more metal plates which transfer heat from the processor and memory to the metal chassis of the body in parallel. I can't see what more they could have done without making a larger, ventilated, non-weather-sealed body.


Aug 22, 2020 at 12:42 PM
Zenon Char
Offline
• • • • • •
Upload & Sell: Off
p.11 #2 · R5 artificial time limit


I've said before I'm not an engineer. How much more time would the camera record with these great suggestions. 10 minutes? 2 hours?

I would have no trouble being convinced Canon kept the R5/6 bodies small so as not to compete with their new RF cinema bodies. Nothing here is a surprise. Months before the release people were asking how they would manage heat issues with that size of body and no fans. I also don't have an issue with that. Any company has the right to do whatever they want to. Consumers have the right to not buy a product if it does not suit their needs.



Aug 22, 2020 at 01:08 PM
Ray Still
Offline
• • • • •
Upload & Sell: Off
p.11 #3 · R5 artificial time limit


A damn pretty good video showing evidence of a thermal timer




Aug 22, 2020 at 02:07 PM
TeamSpeed
Offline
• • • • •
Upload & Sell: Off
p.11 #4 · R5 artificial time limit


Or perhaps the tracking of temperature rate of change.... Much more logical. Removing that means the camera has to start determining rate of change all over again and if the camera has reached a high temperature, the rate of change is about nil.


Aug 22, 2020 at 03:18 PM
wolf33d
Offline

Upload & Sell: Off
p.11 #5 · R5 artificial time limit


Weeks ago I said here it was for sure a Canon software limitation and a lot of people doubted or called trolling. I guess not so much anymore...




Aug 23, 2020 at 07:36 PM
quantumloop
Offline
• •
Upload & Sell: Off
p.11 #6 · R5 artificial time limit


wolf33d wrote:
Weeks ago I said here it was for sure a Canon software limitation and a lot of people doubted or called trolling. I guess not so much anymore...



I will still say it: if Canon has figured out a way for this little camera body to safely record 8K over that long of a period of time, then their engineers deserve a reward. Because it would truly be a breakthrough.

I mean while some would be angry at the so called Cripple Hammer, the reality is that Canon would have pulled off one of the great engineering feats of the modern video era. And that's not hyperbole. To overlook that while being mad at Canon for a business decision is to overlook the far greater consequence. I'm surprised that Andrew Reid doesn't realize that. Tony Northrup did.

The key question is of course, if one defeats that timer often, does it cause any long term damage to the unit? That would be the other part in the equation. And if for some reason this long 8K recording is viable, then again, Canon has created a massive engineering achievement.



Aug 23, 2020 at 08:03 PM
TeamSpeed
Offline
• • • • •
Upload & Sell: Off
p.11 #7 · R5 artificial time limit


Exactly, I think the wiping of nvram just removes the rate of change, so after a reset, there isn't much of a rate change since it is already hot, so the recording time is restored. Long term issues might actually result of one uses that method over and over.


Aug 23, 2020 at 09:10 PM
Zenon Char
Offline
• • • • • •
Upload & Sell: Off
p.11 #8 · R5 artificial time limit


I was just going to say the same thing. How do we know this was not intentional for a reason? What about when the ambient temperature is hot? Maybe the thermal limit method would have resulted in damaged cameras so they decided to limit by time.

If you are going to make this worth it you need at least one hour. Documentaries, shows, etc are typically an hour. Who will volunteer to subject their $4,000 camera to popping the battery out 4 times in a row get 1 hour of 8K video?

Maybe Canon did cripple it and maybe they didn't. I don't know but this is still not enough proof for me. I'm not an engineer but people months ago asked how is Canon going to manage heat filming 8K in that size of body with no fans. So we figure EOSHD and others are correct saying the R5 can go an hour + at 8K consistently without eventually getting fried? Weeks ago Canon Rumours said there will probably be a FW update but said don't expect much.

If I'm wrong - I'm wrong. We will see.




Aug 23, 2020 at 09:35 PM
MayaTlab
Offline
• • •
Upload & Sell: Off
p.11 #9 · R5 artificial time limit


quantumloop wrote:
I will still say it: if Canon has figured out a way for this little camera body to safely record 8K over that long of a period of time, then their engineers deserve a reward. Because it would truly be a breakthrough.


I don't think so really. Handling of 8K @30fps is already part of the specifications of Qualcomm smartphone chips. I'm pretty certain that, if designed for that purpose, Apple's chips would have zero trouble doing so as well given how much they outperform Qualcomm's latest SOCs. Both in designs that you'd think would be quite constrained thermally speaking.
Canon's SOC being mostly solely dedicated to imaging, you'd figure that they could allocate most of their transistor resources to image / video processing, in a packaging that could be a lot less thermally constrained.
Questions would legitimately arise regarding the actual image quality and characteristics of these 8K streams though. These two types of devices can't really be directly compared anyway as their purposes are quite fundamentally different.
Let's remember that Samsung's NX1 was doing H265 4K 6 years ago. I think it's just that as Camera manufacturers have habituated us to being quite a few years behind the tech curve, particularly in terms of processing power, that when they do things more on the edge it looks surprisingly modern.



Aug 24, 2020 at 05:06 AM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.11 #10 · R5 artificial time limit


MayaTlab wrote:
I don't think so really. Handling of 8K @30fps is already part of the specifications of Qualcomm smartphone chips. I'm pretty certain that, if designed for that purpose, Apple's chips would have zero trouble doing so as well given how much they outperform Qualcomm's latest SOCs. Both in designs that you'd think would be quite constrained thermally speaking.
Canon's SOC being mostly solely dedicated to imaging, you'd figure that they could allocate most of their transistor resources to image / video processing, in a packaging that could be a lot less thermally constrained.
Questions would legitimately arise regarding the actual
...Show more

What happens when you record 20 minutes of 8K video on a smartphone? (Temperature, storage, battery, etc)



Aug 24, 2020 at 05:10 AM
ilkka_nissila
Offline
• • • •
Upload & Sell: Off
p.11 #11 · R5 artificial time limit


It's not quite the same 8K. The S20 records 8K at 80 megabits per second whereas the R5's bit rate is 2600 Mbps, so the Canon records 30 times as much data per unit time as the smartphone. The quality is likely to be very different.

The S20 has a 5-min record limit.



Aug 24, 2020 at 05:11 AM
sturgis
Offline

Upload & Sell: Off
p.11 #12 · R5 artificial time limit


Tom_W wrote:
Having had video-capable cameras since at least the 5D3, I can say that I really haven't had much urge to do short videos after, or before stills.

I'm only now playing around with video, but I've learned a few things already - the cost of going to 8K on the computer side would be enormous right now, even 4K is a space hog, and there's a pretty stiff learning curve into turning a few video clips into something presentable.

So to answer your question, I have never had that urge in the past and I don't think I'll likely start doing
...Show more

Thanks for answering, and fair enough.

The 8K for me just is flexibility. I'm expecting proxies or something equivalent, but that for me is similar to letting whatever photo editing software generate previews.

Storage wise, yep - especially as I would not remove the original movie even once I have edited and then downsized / downsampled it to 4K, so I do agree it eats storage - but I just like to be able to flip to video record a few mins, and then flip back to stills. 8K gives me more flexibility in post.

In 2 years, once the imminent next gen Gfx cards come down in price, then we will have hw support for 4:2:2 which will help immensely. Until then, I might have to downsample to 4K, and "archive" the 8k for 2 years. I build my own PCs, and this time having not really upgraded in quite a few years, I built a 2nd one just so I can offload processing and both cost me less than the R6....

Horses for courses, I still re-process images I took with my 10D so I can be patient



Aug 24, 2020 at 05:54 AM
MayaTlab
Offline
• • •
Upload & Sell: Off
p.11 #13 · R5 artificial time limit


ilkka_nissila wrote:
The quality is likely to be very different.


You're absolutely right of course .

But that chip is a lot less dedicated to video processing than Canon's and I believe quite a bit smaller in surface area. Which it probably compensates to a degree by using a smaller node process. I'd wager that it might have to do a lot more work as well to clean up the output.
Let's imagine a situation where Canon uses the latest node process, the latest IPs, on a larger chip, of which most transistors would be dedicated to imaging, in a device that is (theoretically) a lot less thermally constrained. I'm not sure why 8K @ 30 fps would be, at least on the chip side, a huge challenge. That would simply amount to me as doing things at the forefront of what's possible today instead of being a few years behind. After all this is what you'd expect out of an imaging company.

Let's just say that I'm skeptical that the R5's version of 8K 30fps or 4K 120 fps would result in 2020 in a chip that gets so hot that it can't be theoretically passively cooled in an enclosure such as the R5's... provided the camera's resources were allocated to that effect. It's unknown how effective the R5's cooling system is but it certainly looks very peculiar and counter-intuitive. It's entirely possible that it's just a compromise to the benefits of other areas of the camera and that Canon thought that 8K 30 fps or 4K 120 fps would be sufficiently niche applications that most users would use an external recorder instead and that they didn't warrant spending more of the available resources on a better passive cooling solution. That would be to me the nice, non-conspiracy theory explanation.




Aug 24, 2020 at 05:54 AM
sturgis
Offline

Upload & Sell: Off
p.11 #14 · R5 artificial time limit


MayaTlab wrote:
You're absolutely right of course .

But that chip is a lot less dedicated to video processing than Canon's and I believe quite a bit smaller in surface area. Which it probably compensates to a degree by using a smaller node process. I'd wager that it might have to do a lot more work as well to clean up the output.
Let's imagine a situation where Canon uses the latest node process, the latest IPs, on a larger chip, of which most transistors would be dedicated to imaging, in a device that is (theoretically) a lot less thermally constrained. I'm
...Show more
As you say, different chip fabrication - phones tend to use the latest the fabs can provide, I dont know who fabs Canon's procesor chips, but I would guess they are on an noticeably older size as it will be cheaper (and be less energy efficient)

Phones are on a massive scale in comparison to cameras as we all know. Look at the money in smartphones, hence the amount of effort gone into transforming them into what you see and use today.

80Mbps can use cheaper, slower storage. 2600Mbps as we know, takes CFExpress which further adds to the heat envelope.

But I still think they're doinig similar processing - x265 in HW (else it wouldnt be real time on either device), some lens correction, some noise processing, some sharpening and downsampling. I dont know how much phones do in HW vs SW, but I bet they do a lot. They release new CPU & GPUs almost annnually, and they iterate a lot quicker. Sure they do lots in SW, but anything they can offload into HW they will because of the reduction in processing times and heat.

Alas, no external recorder can support those high bitrate modes (yet), which is another indication for me about why 8K is bleeding edge, and even Canon's use of 4:2:2 rather than 4:2:0 means editing is rather difficult.

On the CR site, someone rightly pointed out that Sigma fp can do some of the modes that the R5 can do, without the same restrictions (and without some of the features that the R5 has).

I hope for Canon, that they are being protective of the Camera internals. Whether that is the Sensor, IBIS, the Image processor or something else, I doubt we will know - even if Canon did give some insight into the trade-offs.

No company is perfect, no product is perfect. If enough people don't buy the R5 then Canon will have to understand why. If that's because the wrong expectations were perhaps set in the pre-launch literature well then hopefully lesson learnt.

If I could sell my other Canon bodies back in the UK (here in HK pricing is much lower), I would buy a 2nd R5, even though the cool down times and stills impact on video times disappoint me, the rest of the camera is enough of an improvement for me, and different camera lines start to annoy me after a while - ha ha.



Aug 24, 2020 at 07:10 AM
Ziffl3
Offline
• • • • •
Upload & Sell: Off
p.11 #15 · R5 artificial time limit


MayaTlab wrote:
You're absolutely right of course .

But that chip is a lot less dedicated to video processing than Canon's and I believe quite a bit smaller in surface area. Which it probably compensates to a degree by using a smaller node process. I'd wager that it might have to do a lot more work as well to clean up the output.
Let's imagine a situation where Canon uses the latest node process, the latest IPs, on a larger chip, of which most transistors would be dedicated to imaging, in a device that is (theoretically) a lot less thermally constrained. I'm
...Show more

You are missing other aspects of this picture.....
Like the heat build up in the transfer to internal storage medium.
heat build up from the electricity used to move the data.
also the amount of electricity needed.

I don't remember the what node level canon is using in their main processing chips?
14nm, 10nm.... what? this would also influence heat requirements and specifications or use models.

another camera to look at as a possible cousin is the blackmagic pocket cinema camera 6K.
it has active cooling and runs a super 35 sensor processing 6K internal/external.
it is a relatively small design and has an canon EF mount.


as for niche applications.... i would love if 4k@120 could use external recording.
fyi... this is not a niche option. many event shooters and content creator love shooting in high frame rate.

beside the heat issue.... why did canon not have 1080P@120 or 240? that is very weird.




Aug 24, 2020 at 08:43 AM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.11 #16 · R5 artificial time limit


Ziffl3 wrote:
beside the heat issue.... why did canon not have 1080P@120 or 240? that is very weird.



An uneducated guess:

The full sensor can only be read 60 times per second. 120 Times per second with line skipping. Reading 240 fps would mean skipping 3 out of 4 lines, and the result would not look good.



Aug 24, 2020 at 09:01 AM
sturgis
Offline

Upload & Sell: Off
p.11 #17 · R5 artificial time limit


alundeb wrote:
An uneducated guess:

The full sensor can only be read 60 times per second. 120 Times per second with line skipping. Reading 240 fps would mean skipping 3 out of 4 lines, and the result would not look good.


I think they are coming. Canon USA said originally HD120 was there. I guess it didnt make the cut, but the rumours suggest that may come with a future firmware (which seems to be the norm with Canon - viz a vie 24fps on prev models)...



Aug 24, 2020 at 08:39 PM
sturgis
Offline

Upload & Sell: Off
p.11 #18 · R5 artificial time limit


Ziffl3 wrote:
You are missing other aspects of this picture.....
Like the heat build up in the transfer to internal storage medium.
heat build up from the electricity used to move the data.
also the amount of electricity needed.

I don't remember the what node level canon is using in their main processing chips?
14nm, 10nm.... what? this would also influence heat requirements and specifications or use models.

another camera to look at as a possible cousin is the blackmagic pocket cinema camera 6K.
it has active cooling and runs a super 35 sensor processing 6K internal/external.
it is a relatively small design and has an canon EF mount.

as for
...Show more

One of the Sony cameras can do 4k120 to the Atomos, but does so by somehow sending it as 4k60 and doubling the recording length. Presumably that would require Canon to change the firmware to do this, and might also need them to change from 4:2:2 to 4:2:0 ie unlikely on both counts.



Aug 24, 2020 at 08:41 PM
davedv
Offline
• •
Upload & Sell: Off
p.11 #19 · R5 artificial time limit


One of the Sony cameras can do 4k120 to the Atomos, but does so by somehow sending it as 4k60 and doubling the recording length. Presumably that would require Canon to change the firmware to do this, and might also need them to change from 4:2:2 to 4:2:0 ie unlikely on both counts.

The only Sony cameras that could do 4K 120fps output to the Atomos recorders were the Sony FS700 and FS5. They did this by recording a 4K 120 fps burst (up to 4 seconds in length) to an internal buffer in the camera, and then playing it out over and SDI port at 4K 60 fps when the burst was complete.

So the recorder only had to handle processing a lower 4K 60fps video bandwidth on input. Incidentally, the format of this video was actually a proprietary compressed Sony RAW format sent over a 3G SDI port. This kind of approach isn't really suitable for continuous video recording and required a special high speed memory buffer within the camera for it to work.

Technically speaking, uncompressed 4K 120fps video at 8-bit 4:2:0 subsampling can be sent over an HDMI 2.0b port, although I am not aware of any portable video recorders that accept this input format for video recording yet. Nor am I aware of any of video cameras that output this signal format over HDMI.

4K 120fps video with 10-bit color depth or 4:2:2 or higher subsampling requires the higher bandwidth of an HDMI 2.1 port: https://www.hdmi.org/spec21Sub/EightK60_FourK120



Aug 24, 2020 at 09:19 PM
Howie4life
Offline
• •
Upload & Sell: On
p.11 #20 · R5 artificial time limit




sturgis wrote:
Thanks for answering, and fair enough.

The 8K for me just is flexibility. I'm expecting proxies or something equivalent, but that for me is similar to letting whatever photo editing software generate previews.

Storage wise, yep - especially as I would not remove the original movie even once I have edited and then downsized / downsampled it to 4K, so I do agree it eats storage - but I just like to be able to flip to video record a few mins, and then flip back to stills. 8K gives me more flexibility in post.

In 2 years, once the imminent next
...Show more

What's your video workflow for the R5 if you down mind me asking. With my old pocket 6k I use to trim the Raws with resolve media player without any quality loss to save on space. I would then transcode the clips the clips I didn't use in my movie to h265. Sony also provided a tool to quickly trim my A9 videos. I'm struggling to find a good workload on my R5 to limit storage. I love shooting 8k raw and my laptop can easily play it back in resolve in real time, but those files are huge. People keep saying current computers can edit the h265 R5 files, but mine plays everything back in resolve except 8k h265.



Aug 24, 2020 at 10:19 PM
1       2       3              10              12       13       end




FM Forums | Canon Forum | Join Upload & Sell

1       2       3              10              12       13       end
    
 

Welcome back
Log in to your account