Canon EOS 5D Mark IV Talk [CR1]

[romanr74]

...
If you need an example of Dilbert's progress, review some of the recent silliness over DPReview. He's been much more rational and fair during the DPReview Witch Trials than the ubersensitive individuals who cannot seem to comprehend the whole concept of what a "review" actually means.
...

Perhaps, just perhaps, because I'm not a "fan boi", I don't behave irrationally when it comes to reviews. And maybe, just maybe, not being a "fan boi" means that in general, I'm less irrational in general regarding Canon but because everyone else here is caught up in emotional ways with Canon, the irrational appears rational and the rational appears irrational.

Just saying.

How is life at Shutter Island?

Just asking.

 

[slclick]

Wow, the sheer power of the CR1 has certainly raised some hackles with the 'Canon ala carte personal specification sensor and feature program' participants.

 

[pierlux]

Everyone here appears to be focused on cost, as concerns the possibility for the 5D IV to have dual Digic to allow more fps.

I've already mentioned at least a couple of times in this thread that I *suspect* cost is not the main determiner for the choice of limiting the speed in a 5D body, but battery power it is. It seems nobody is interested in discussing this possibility which I suggest might be the main reason for limiting speed. Let's use unfocused's way to analyze the facts.

fact: the 5D series is an ungripped body. Its specs are defined around the power which a single battery can deliver.

fact: the 5D III suffers from a fall in speed as soon as the battery charge falls below a certain percentage.

fact: the 7D II reaches 10 fps, but it's APS-C. The energy required to move its tiny mirror and re-penis the smaller shutter probably amounts to less than 1/3 of that required to achieve the same task in a full frame equivalent system (yet, but this may be a mere coincidence, a new, slightly more powerful battery was introduced along with it, the LP-E6n).

fact: there are cameras which need to be gripped, and with 2 batteries inserted, to achieve full speed.

We had a CR2 rumor claiming a new battery grip (BG-E20) is coming, then another CR1 rumor, reporting some specs, indicating 7 fps and a new battery (LP-E20). It is entirely possible that Canon had to develop a more powerful battery just to allow that single fps improvement in the 5D IV.

Any speculation based only on cost, disregarding the REAL limitations engineers must confront with when deciding which compromises to make in designing a camera, is flawed.

 

[nvsravank]

Isn't 7d ii a dual chip machine? In a similar size enclosure? With the same battery? With 20mp to 28 mp adding some difference but nothing to worry about.
Maybe that is why nobody is taking on that train of thought.

Everyone here appears to be focused on cost, as concerns the possibility for the 5D IV to have dual Digic to allow more fps.

I've already mentioned at least a couple of times in this thread that I *suspect* cost is not the main determiner for the choice of limiting the speed in a 5D body, but battery power it is. It seems nobody is interested in discussing this possibility which I suggest might be the main reason for limiting speed. Let's use unfocused's way to analyze the facts.

fact: the 5D series is an ungripped body. Its specs are defined around the power which a single battery can deliver.

fact: the 5D III suffers from a fall in speed as soon as the battery charge falls below a certain percentage.

fact: the 7D II reaches 10 fps, but it's APS-C. The energy required to move its tiny mirror and re-penis the smaller shutter probably amounts to less than 1/3 of that required to achieve the same task in a full frame equivalent system (yet, but this may be a mere coincidence, a new, slightly more powerful battery was introduced along with it, the LP-E6n).

fact: there are cameras which need to be gripped, and with 2 batteries inserted, to achieve full speed.

We had a CR2 rumor claiming a new battery grip (BG-E20) is coming, then another CR1 rumor, reporting some specs, indicating 7 fps and a new battery (LP-E20). It is entirely possible that Canon had to develop a more powerful battery just to allow that single fps improvement in the 5D IV.

Any speculation based only on cost, disregarding the REAL limitations engineers must confront with when deciding which compromises to make in designing a camera, is flawed.

 

[Loswr]

Isn't 7d ii a dual chip machine? In a similar size enclosure? With the same battery? With 20mp to 28 mp adding some difference but nothing to worry about.
Maybe that is why nobody is taking on that train of thought.

Everyone here appears to be focused on cost, as concerns the possibility for the 5D IV to have dual Digic to allow more fps.

I've already mentioned at least a couple of times in this thread that I *suspect* cost is not the main determiner for the choice of limiting the speed in a 5D body, but battery power it is. It seems nobody is interested in discussing this possibility which I suggest might be the main reason for limiting speed. Let's use unfocused's way to analyze the facts.

fact: the 5D series is an ungripped body. Its specs are defined around the power which a single battery can deliver.

fact: the 5D III suffers from a fall in speed as soon as the battery charge falls below a certain percentage.

fact: the 7D II reaches 10 fps, but it's APS-C. The energy required to move its tiny mirror and re-penis the smaller shutter probably amounts to less than 1/3 of that required to achieve the same task in a full frame equivalent system (yet, but this may be a mere coincidence, a new, slightly more powerful battery was introduced along with it, the LP-E6n).

fact: there are cameras which need to be gripped, and with 2 batteries inserted, to achieve full speed.

We had a CR2 rumor claiming a new battery grip (BG-E20) is coming, then another CR1 rumor, reporting some specs, indicating 7 fps and a new battery (LP-E20). It is entirely possible that Canon had to develop a more powerful battery just to allow that single fps improvement in the 5D IV.

Any speculation based only on cost, disregarding the REAL limitations engineers must confront with when deciding which compromises to make in designing a camera, is flawed.

Seems your train has left the thought station without you on it. Might want to re-read his post, I highlighted a rather important but that you left out.

 

[pierlux]

Isn't 7d ii a dual chip machine? In a similar size enclosure? With the same battery? With 20mp to 28 mp adding some difference but nothing to worry about.
Maybe that is why nobody is taking on that train of thought.

It is. In fact, we already extensively covered that matter. 20 MP@10 fps and advanced AF algorithms require the processing power of dual Digic. In the 7DII, a single battery is enough to power all that stuff. In a full frame equivalent of the 7DII, that wouldn't be enough, even disregarding the higher PM count.

What I'm saying is we must consider things it their entirety. The picture is not complete when focusing on single aspects, i.e. the cost of introducing a second chip in the 5D IV. I don't believe in conspiracy theories and I don't believe in "intentional crippling" of a product to "protect" sales of other products. And, actually, I don't think an "evil mind" at Canon decides not to allow 9-10 fps in the 5D IV because a second chip would rise the cost of it. It would, but I think not much. A few $ at worst.

The 5D IV will most probably run on a single Digic. And, on a side note, all the Magic Lantern users will rejoyce.

 

[Don Haines]

Isn't 7d ii a dual chip machine? In a similar size enclosure? With the same battery? With 20mp to 28 mp adding some difference but nothing to worry about.
Maybe that is why nobody is taking on that train of thought.

It is. In fact, we already extensively covered that matter. 20 MP@10 fps and advanced AF algorithms require the processing power of dual Digic. In the 7DII, a single battery is enough to power all that stuff. In a full frame equivalent of the 7DII, that wouldn't be enough, even disregarding the higher PM count.

What I'm saying is we must consider things it their entirety. The picture is not complete when focusing on single aspects, i.e. the cost of introducing a second chip in the 5D IV. I don't believe in conspiracy theories and I don't believe in "intentional crippling" of a product to "protect" sales of other products. And, actually, I don't think an "evil mind" at Canon decides not to allow 9-10 fps in the 5D IV because a second chip would rise the cost of it. It would, but I think not much. A few $ at worst.

The 5D IV will most probably run on a single Digic. And, on a side note, all the Magic Lantern users will rejoyce.

and to further confuse the issue, there is the possibility of a Digic 7, which would presumably have more computing power than the dual processors of the 7D2.....

The thing is, we don't know the code of the AF algorithm and if it works better on a dedicated processor, or if it can share a faster processor, or what the bandwidth limitations are of a new processor with A/D on chip, or what buffer sizes they will use, or a host of other questions. All we can do is speculate without firm data to back up our opinions and the result is a never ending argument....

Perhaps it has a digic6 for AF and a Digic 7 for data

 

[Don Haines]

Everyone here appears to be focused on cost, as concerns the possibility for the 5D IV to have dual Digic to allow more fps.

I've already mentioned at least a couple of times in this thread that I *suspect* cost is not the main determiner for the choice of limiting the speed in a 5D body, but battery power it is. It seems nobody is interested in discussing this possibility which I suggest might be the main reason for limiting speed. Let's use unfocused's way to analyze the facts.

fact: the 5D series is an ungripped body. Its specs are defined around the power which a single battery can deliver.

fact: the 5D III suffers from a fall in speed as soon as the battery charge falls below a certain percentage.

fact: the 7D II reaches 10 fps, but it's APS-C. The energy required to move its tiny mirror and re-penis the smaller shutter probably amounts to less than 1/3 of that required to achieve the same task in a full frame equivalent system (yet, but this may be a mere coincidence, a new, slightly more powerful battery was introduced along with it, the LP-E6n).

fact: there are cameras which need to be gripped, and with 2 batteries inserted, to achieve full speed.

We had a CR2 rumor claiming a new battery grip (BG-E20) is coming, then another CR1 rumor, reporting some specs, indicating 7 fps and a new battery (LP-E20). It is entirely possible that Canon had to develop a more powerful battery just to allow that single fps improvement in the 5D IV.

Any speculation based only on cost, disregarding the REAL limitations engineers must confront with when deciding which compromises to make in designing a camera, is flawed.

Quite correct, power is a limitation that either requires drastically improved battery technology to overcome, bigger/more batteries, or much more efficient electronics.

Power also brings its friend heat into the equation....yet another variable that keeps cameras from running flat-out, particularly for those who like video and live-view.... When you run down the battery you can pop in a fresh one, but heat is a lot harder to deal with...

Heat and power are why the 1DX2 is such a large heavy beast.... Unless Canon wants to make the 5D4 similarly large, they can't go very far down that path, and as a result, power and heat are one of the main design hurdles.

 

[mb66energy]

I see only roughly 500mW average power for each the shutter and the mirror at 10 fps for a FF system (@10% efficiency). Resulting in 10 images per Joule electrical energy per second (1W). LP-E6 keeps 45 000 Joules energy (equivalent of 1.5 ml of gasoline).

CPU power should be in the region of 10 Watts (or 10 Joules per second).

A better argument is maybe the force on the mirror components which urges the switch from "simple materials" (some carbon fiber reinforced plastics) to very special materials (carbon fibre hand woven in moonlight) between ~ 6 fps and 8 fps.

The price of the cpu is IMO negligible and in the region of some dollars - maybe Canon selects them by operating voltage/max frequency and sells the slower ones in Powershots etc. so nothing is wasted.

 

[pierlux]

... All we can do is speculate without firm data to back up our opinions and the result is a never ending argument....

That's the fun of it! Speculating, guessing, being corrected by others - there's always someone who knows something more - it's a way to get educated, I've learned a lot of things on CR, thanks to all of you!

Perhaps it has a digic6 for AF and a Digic 7 for data

Similarly to the 1DX, which has a dedicated Digic 4 for AF? Doubt it, but, if there is, Digic 6 would be overkill. BTW, does anyone know if the 1DX II has a dedicated Digic for AF, too? I'm too lazy to go and check now...

 

[pierlux]

Power also brings its friend heat into the equation....yet another variable that keeps cameras from running flat-out, particularly for those who like video and live-view.... When you run down the battery you can pop in a fresh one, but heat is a lot harder to deal with...

Good point. Actually, engineers have to deal with so many constraints and variables the vast majority of which we common mortals have absolutely no idea they even exist. This awareness further reinforces my convinction that no intentional crippling, no conspiracy exists. Of course, the marketing department has always the last word, but a Company which wants to stay healthy in the market just can't overlook profit.

 

[kaihp]

I see only roughly 500mW average power for each the shutter and the mirror at 10 fps for a FF system (@10% efficiency). Resulting in 10 images per Joule electrical energy per second (1W). LP-E6 keeps 45 000 Joules energy (equivalent of 1.5 ml of gasoline).

CPU power should be in the region of 10 Watts (or 10 Joules per second).

Be careful not to compare average and peak current capacities.

A good analogy for a battery is a barrel of water with a tap (or hole) at the bottom. The barrel size is equivalent to the battery capacity (mAh). The tap size is equivalent to the battery's peak current delivery.

You cannot get more water out of the barrel on a sustained basis than what is set by the size of the tap. If you need more water (current) than what the tap can deliver, but only at intervals, you can 'stage' the water into a smaller bucket, which has a larger tap (which can be turned on/off). Close the tap of the staging bucket, fill it up, open the staging tap, get a big splash. Repeat.


Even though the driver for the mirror is active for a fraction of the time*, maybe 5-8 of the time at max FPS, the current needed while it's driving the mirror that can be crucial in terms of the battery ability to deliver that current.
Let's for a moment assume that your 500mW average is correct, and that the SlomoGuys video is reliable (I measured a 15msec mirror-up time). At 7.4V and an on-time of 15msec and off-time of 110ms (the 7D does 8fps), that would be 500mW/7.4*(125ms/15ms) = 563mA. Yikes!

A CPU power consumption of 10Watts is way out of line for two reasons. First, the camera wouldn't be able to get dissipate that without getting very hot, very quickly. My guess would be less than 500mW. Secondly, the battery won't be able to sustain such current draw.

To get a bound for the power consumption, I looked at a day when I took 2500 shots with a 5D3, with relatively little chimping. The battery was full in the morning, and had ~27% power left at the end. That's 1800mAh*(1-27%)/2500 shutters = 0.53mAh/shutter actuation, or 0.146mAs. If everything is consumed by the shutter in 15msec/167msec** that's a peak of 5.85mA. Power for all the electronics, the memory card, screen, lens AF and IS needs to be deducted from this.

Now, the actual numbers can definitely be debated, as the are plenty of ass-u-me's in here. If we had real data for the power used by the various modules it would be much easier to say something more precise, but I do believe that the above values should be in the right ballpark.

*) The mirror-up time is 15msec for a 7D, if the video from The Slo Mo Guys is reliable and I measured it correctly.

**) 6fps for the 5D3.

 

[jrista]

Dilbert, why do you persist? I don't think a single person on these forums has any respect for your arguments. I myself find them...bewildering most of the time to say the least, and often just...ignorant. You really gain NOTHING by sticking around these forums and continuing your crusade... It certainly isn't going to gain you any respect, and only raises the ire of Neuro most of the time. You should seriously cut your losses and find an environment of like minded individuals...or just move on to better things in life. It isn't worth wasting away your life being the one ultra-contrary voice on a forum full of Canon fans.

Hey, Dilbert has made a lot of progress in the past year. Of course, he started from a very low bar, but I detect an effort on his part to at least try to be slightly less uninformed and irrational. Yes, at times it seems like he takes two steps backward after taking one step forward, but that's not always the case. (although in this case, he is certainly dead wrong. To digress briefly, the entire debate over processors has been silly and back*sswords. To repeat, the processors do not determine the feature set. The desired feature set is determined first and that decides the processing power needed.)

If you need an example of Dilbert's progress, review some of the recent silliness over DPReview. He's been much more rational and fair during the DPReview Witch Trials than the ubersensitive individuals who cannot seem to comprehend the whole concept of what a "review" actually means.

I'm not comfortable trying to drive anyone off of this site. I've seen too many talented, professional photographers give up in frustration because a handful of ultra-prolific commenters want to be self-appointed experts on everything.

Not trying to "drive" him off the site. Just trying to free him from an endeavor that will never gain him anything. It isn't worth it. He could be enjoying the time he usually spends here doing something else, rather than always having everyone look down on him, even if it is usually for good reason.

There is more to life than...this forum. Than always fighting a losing battle.

 

[jrista]

...
If you need an example of Dilbert's progress, review some of the recent silliness over DPReview. He's been much more rational and fair during the DPReview Witch Trials than the ubersensitive individuals who cannot seem to comprehend the whole concept of what a "review" actually means.
...

Perhaps, just perhaps, because I'm not a "fan boi", I don't behave irrationally when it comes to reviews. And maybe, just maybe, not being a "fan boi" means that in general, I'm less irrational in general regarding Canon but because everyone else here is caught up in emotional ways with Canon, the irrational appears rational and the rational appears irrational.

Just saying.

*SMH* Wow. He came to your defense, man.

 

[Don Haines]

<snip>
I'm not trying to argue that processors have anything to do with features (in this thread), rather that they don't come for free. A Canon 5DIII with multiple DIGIC chips is going to be more expensive than a Canon 5DIII with a single DIGIC
<snip>

Why is anyone arguing this? Of course it costs more to put in extra circuitry...... Is there anyone out there who thinks that it doesn't?

 

[pierlux]

I see only roughly 500mW average power for each the shutter and the mirror at 10 fps for a FF system (@10% efficiency). Resulting in 10 images per Joule electrical energy per second (1W).

I respectfully think you underestimate this aspect, even if not by much. This is not a case of constant, linear motion. The inertial mass is low, but its movement implies sudden, extremely fast repeated and frequent acceleration.

I'm not sure what you mean by @10% efficiency: do you mean that 90% energy is wasted as heat? Or rather you estimate a 90% efficiency for the motor? If the latter, though true that the latest electrical engines reach such efficiency, again it's not true for small motors and linear EM actuators, especially when engaged with gearwheel connected mechanisms which dissipate energy, and undergoing repeated frequent cycles of opening/closing of an electrical circuit. Anyway, I agree the power required is in the order of a few W units. But, lacking any evidence, this is pure guessing, we may both be wrong.

CPU power should be in the region of 10 Watts (or 10 Joules per second).

Based on the images of the CPU and the fact it doesn't require the presence of a heatsink, I'd say less than 10 W, more approx 2-4 W.

The price of the cpu is IMO negligible and in the region of some dollars - maybe Canon selects them by operating voltage/max frequency and sells the slower ones in Powershots etc. so nothing is wasted.

Totally agree. I'm sure they do exactly this, like Intel do. Nothing is wasted.

A better argument is maybe the force on the mirror components which urges the switch from "simple materials" (some carbon fiber reinforced plastics) to very special materials (carbon fibre hand woven in moonlight) between ~ 6 fps and 8 fps.

Concerning both the effect on overall cost and the possibility of fps limitation, that's another valid point.
_ _ _

And, at the end, there's the significant power required by the attached lens for focusing and IS which has to be taken into account.

 

[davidhfe]

We can (and no doubt will!) discuss battery capacity and cpu throughput until August, but the reality is a 24mp, 6.5FPS FF camera already exists at the $2000 price point. While there are no doubt challenges bumping that up to a round 7fps that doesn’t drop when a battery is at 50%, it'd be legitimate to question a $1500 premium for it.

For the extra money, it seems reasonable that the 5D4 would need to exceed expectations in at least one of the "core" feature areas (MP, AF+Meter system, FPS, sensor quality). Thankfully, I'm not the one at Canon who gets to figure out what that is.

 

[Loswr]

<snip>
I'm not trying to argue that processors have anything to do with features (in this thread), rather that they don't come for free. A Canon 5DIII with multiple DIGIC chips is going to be more expensive than a Canon 5DIII with a single DIGIC
<snip>

Why is anyone arguing this? Of course it costs more to put in extra circuitry...... Is there anyone out there who thinks that it doesn't?

Well, you know dilbert. He likes to argue, even if only with himself.

While no one with any sense would argue that an additional chip would be 'free' to Canon, it likely would not add to the cost to consumers. Maybe in dilbertland Canon would charge $3301.27 for a camera, but in the real world it wouldn't lead to a $100 increase, but rather a slight decrease in Canon's per-unit profit...which is ample reason not to include anything that's not deemed necessary.

 

[PFerrara]

I have a 70D. I upgraded to it after 5 years with a T1i. I love it for the 7fps I get when shooting my son's swim meets & soccer games. Video is good too, but I rarely use it. But it is my camera for everything and when used at indoor family events with poor lighting it struggles.

I would rather have clean images than more megapixels and the ability to crop.

I want to upgrade to a FF for the better low light performance, but want the decent fps so I can still use it for the kids sports.

So here's hoping for 7+ fps, some good low light performance and built in GPS would be nice for all the national park trips we like to take.

Reading through these comments, many seem to think that for Canon, this is a zero sum game. That is, one camera's sales will/may take away from another. I'm choosing to believe that the $$ guys at Canon are going to launch this camera with an eye toward taking market share from the Nikon and Sony camps. If they do, then I think we are all going to be very happy.

 

[pvalpha]

It took me until just five seconds ago to realize something: a plethora of people who speculate here don't actually have fun - they speak every word as though its truth or falsehood might be some slight against the purest core of their existence. Either that, or they just don't use enough emoticons.

Most of a processor's cost is in its development. And even then, that cost is likely minimal in a Digic - if only because its a lightweight processor - on a par with an ARM driven package. If they've designed it correctly and are using an appropriate process, the cost per chip is negligible. Each package equates to about $100-$200 of the consumer price, assuming a top of the line brand new design. Thus a dual CPU package is about $200-$400. Which is piddily compared to the other components in a camera in the X series.

Power - again, arguments can be made, but the draw is probably somewhere in the 1500-3000 milliwatt range at max draw. So double would be 3w to a maximum of 6w - and that's a peak number. For most processes, the device is going to sip power at somewhere in the 300-700 mW range per package. Its probably the fourth most power hungry system in the camera, following the sensor, lens, and shutter/mirror assembly.

So if the device required it, I see no reason why canon would not use a dual processor system if it gave them the ability to give the 5DmIV an edge over its direct competitors and to make it attractive to those who own its predecessor.

As for Megapixels - this is the one thing Canon is more likely to focus on in this camera. Its been their end game for the last few generations with it and its the one thing they are okay with the 5 series outperforming their 1 series cameras on. And they don't really have to do much to achieve a decent jump from iteration to iteration. Moores law dictates that computational capacity doubles roughly every 18-24 months. The Digics, unlike most high end CPUs are likely nowhere near that threshold and have probably another decade before their designs start pushing the limits of physics. Assuming a spherical cow... You're probably going to see a practical gain of 20%-30%. So taking a 22.4 megapixel sensor and taking it up to 29.12 megapixels or thereabouts should be easily done without any drastic changes in design or capabilities while imparting the same image performance. Canon could also achieve a similar gain by switching to back side illumination on the sensor without changing the performance of the processor. Or they could put in two processors for another 20%-30%.

The latest Digics support C-Fast. You could pump 32 megapixel raws to the c-fast card at 10fps. So if you have a moderate buffer (4gb), you could easily sustain a 10fps burst for 12 seconds before experiencing a mild falloff to about 9.6 fps (due to the transfer delay from file to file) until the card was filled. For UHS-II, the maximum fps rate to card would be about 7fps at 32mp, and about 8fps at 29.12mp but there is no sign the digic system supports UHS-II SpeedClass 3. ISO 100-400. More noise = more megabytes after lossless compression so... *shrug*

Canon is unlikely to put 10fps on the 5DmIV. 8fps might happen, but odds are we will see 7fps so that either CFast or UHS-II cards can offer near "limitless" buffer given a fast enough card and keeping the high speed buffer storage down to a minimum. Its this that will take $$$ out of the cost of the electronics as memory is the most expensive component in the system due to how fast it must be in order to transfer data from sensor through to the processor(s), then back to the card. 7fps will require a small push to the AF system, but they can just take what the 7DmII already has and adapt it. They would likely also need to dampen the shutter a bit better.

If they wanted to make a sensor capable of being used for 8k DCI video, they'd need something in the 42-50mp range. Which would be limited to about 3-6fps at buffer exhaustion depending on storage types. They'd also need an 8gb buffer to store a significant amount of raws. And if they wanted to improve noise performance, they'd need to switch to BSI, as there's a law of diminishing processor returns once you start taking the pixel density to where the 7DmII and 5DS/r are unless you have appropriately increased processing capacity. I'd say that's a 2020 or so thing.

Again, that's assuming that canon isn't in its usual conservative "Thou shall not cannabilze thy sales" mode. But I really have a hard time not seeing canon bump to 29-32 mp, as it would cost them little to do so. And because of the 4k video (assuming its there and uses mjpeg) they really have nothing to lose. You're going to need CFast media for that so you're already going to have to make the investment in the cards. They wouldn't lose sales or performance with the megapixel bump. If they forego the 4k video... or use a highly compressed format... then UHS-II and CF would work just as well... and you'd only need a 24mp sensor. Although I think such a iterative upgrade would lose more new customers than bumping up would lose existing customers, and would make them even farther behind the field when it comes to video capability.

My two cents... for what its worth.