A speculative thought on Canon test bodies

[scottburgess]

Some rumors have indicated that Canon bodies in test eat batteries, which got me to thinking...

Suppose the new bodies are testing a triple-sensor design. That is, light is pushed through a trichromic prism to three CMOS sensors, one each for RGB. It is essentially a very old idea, but eliminating the Bayer sensor would improve both low light capability and effective resolution. Such a design would be bulky and would eat batteries, but it is a potential pathway to improved capabilities for high end DSLRs.

I suspect such a sensor array at around 40Mp (x 3) would be competitive with current MF cameras.

Just a thought. You may now return to your regularly scheduled rumors. ;D

 

[Lightmaster]

but it would not only be bulky, it would be at least 2x times more expensive i guess.

maybe an option for an "MF" competitor. the absolut top model in canons line.
but a 5D MK4 with this technology for 8000 euro... i guess not many will buy it.

i for one would be happy if any rumor would come true.
i thought 2014 would be a "canon" year.. but it is nearly may and nothing...

by the way, foveon sensors are reported to eat through batteries pretty fast too.

 

[dgatwood]

Laptops and cell phones eat batteries during testing, too. Whenever you're working with software, there's a good chance you're going to have bugs that cause excessive CPU utilization. Any time the CPU is doing work, it is consuming a lot more power than when it is idle. It only takes a tiny bit of activity every few milliseconds to seriously impact power consumption by preventing the CPU from ever reaching an idle state.

Hopefully, those bugs get fixed before the thing ships, but it isn't at all uncommon to have them during development. I'd be really surprised if anything other than the CPU were responsible for the high battery drain.

 

[mackguyver]

I think you're referring to what they call triple-CCD (or three-chip) cameras in the video world, right? I've wondered about that for a long time, too, but don't know about how practical it is in terms of size for an SLR. I know the pro video cameras had them for many years (and still do?) as they were supposed to produce better color and be better in low light.

 

[Drizzt321]

I think you're referring to what they call triple-CCD (or three-chip) cameras in the video world, right? I've wondered about that for a long time, too, but don't know about how practical it is in terms of size for an SLR. I know the pro video cameras had them for many years (and still do?) as they were supposed to produce better color and be better in low light.

Yea, but they're also quite a bit bulkier than a DSLR, even a 1D size body. At a guess, you'd need quite a bit more back-length at a minimum, because the prisms have to cover 100% (or slightly more) of a FF 35mm sensor. I'm pretty sure those 3-ccd video cameras had/have a much smaller sensor size than even APS-C, which requires a much smaller set of prisms.

 

[traingineer]

Laptops and cell phones eat batteries during testing, too. Whenever you're working with software, there's a good chance you're going to have bugs that cause excessive CPU utilization. Any time the CPU is doing work, it is consuming a lot more power than when it is idle. It only takes a tiny bit of activity every few milliseconds to seriously impact power consumption by preventing the CPU from ever reaching an idle state.

Hopefully, those bugs get fixed before the thing ships, but it isn't at all uncommon to have them during development. I'd be really surprised if anything other than the CPU were responsible for the high battery drain.

Well, Intel's CPUs generally use less power than AMD's CPUs.

 

[Drizzt321]

Laptops and cell phones eat batteries during testing, too. Whenever you're working with software, there's a good chance you're going to have bugs that cause excessive CPU utilization. Any time the CPU is doing work, it is consuming a lot more power than when it is idle. It only takes a tiny bit of activity every few milliseconds to seriously impact power consumption by preventing the CPU from ever reaching an idle state.

Hopefully, those bugs get fixed before the thing ships, but it isn't at all uncommon to have them during development. I'd be really surprised if anything other than the CPU were responsible for the high battery drain.

Well, Intel's CPUs generally use less power than AMD's CPUs.

Intel has greater IPC (instructions per clock) than similar level AMD CPUs at the same clock speed. More complicated than that, but basically Intel CPUs of similar generation as an AMD CPU at the same clock speed will do a good bit more work than the AMD CPU does. This lets the Intel CPU do a "race to sleep", which is to say get the CPU back to it's lowest power usage state. If a bit of software keeps the CPU active, even at low levels, it ends up using quite a bit more power than it does if it can if it is in it's lowest power state (at which point it's not really doing much at all).

Like dgatwood said, a tiny bit of activity every few (or even more often) milliseconds can result in the CPU staying at a higher power state longer than it otherwise should, and using much more power than it would otherwise. Also, don't forget that when it uses more power, it's also using part of that to generate waste heat. I don't know the formula, but I believe, in general, as the power usage goes up, for the same conductor efficiency goes down and more of the power is released as waste heat, rather than getting to where the work needs to get done.

 

[traingineer]

Laptops and cell phones eat batteries during testing, too. Whenever you're working with software, there's a good chance you're going to have bugs that cause excessive CPU utilization. Any time the CPU is doing work, it is consuming a lot more power than when it is idle. It only takes a tiny bit of activity every few milliseconds to seriously impact power consumption by preventing the CPU from ever reaching an idle state.

Hopefully, those bugs get fixed before the thing ships, but it isn't at all uncommon to have them during development. I'd be really surprised if anything other than the CPU were responsible for the high battery drain.

Well, Intel's CPUs generally use less power than AMD's CPUs.

Intel has greater IPC (instructions per clock) than similar level AMD CPUs at the same clock speed. More complicated than that, but basically Intel CPUs of similar generation as an AMD CPU at the same clock speed will do a good bit more work than the AMD CPU does. This lets the Intel CPU do a "race to sleep", which is to say get the CPU back to it's lowest power usage state. If a bit of software keeps the CPU active, even at low levels, it ends up using quite a bit more power than it does if it can if it is in it's lowest power state (at which point it's not really doing much at all).

Like dgatwood said, a tiny bit of activity every few (or even more often) milliseconds can result in the CPU staying at a higher power state longer than it otherwise should, and using much more power than it would otherwise. Also, don't forget that when it uses more power, it's also using part of that to generate waste heat. I don't know the formula, but I believe, in general, as the power usage goes up, for the same conductor efficiency goes down and more of the power is released as waste heat, rather than getting to where the work needs to get done.

Exactly.